Temperature dependence of spontaneous otoacoustic emissions in the edible frog (Rana esculenta)

ELSEVIER

HBEIrlG R[SSIRCH Hearing Research 98 (1996) 22 28

Temperature dependence of spontaneous otoacoustic emissions in the edible frog (Rana esculenta) Glenis R. Long a,., Pim Van Dijk u, Hero P. Wit b ~' Department (~/'Audiology and Speech Sciences, Purdue UnicersiO'. West lx~f~o'ette, IN 47907, USA b ENTDepartment/Audiology, Unicersio, Hospital, P.O. Box 30.001, 9700 RB Groningen, The Netherhmds

Received 27 July 1995;revised 29 February 1996;accepted 9 March 1996

Abstract

The change in frequency of individual emissions in the European edible frog ( R a n a esculenta) when the temperature of the fiog is modified, is part of a complex pattern of interaction between spontaneous otoacoustic emissions. At high temperatures (above 24°C) two emissions are always detected (e.g., one near 800 Hz and one near 1200 Hz). The higher-frequency emission is lower in level and has a wider bandwidth than the lower-frequency emission. It is also often asymmetric and sometimes breaks into two emissions when an external suppressor tone is applied. When the temperature is decreased, these emissions are reduced in frequency at a rate of 0.04 octave/°C. The higher-frequency emission becomes narrower and taller, and the lower-frequency emissions becomes broader and less intense. At approximately 18°C the lowest of these emissions (now between 600 and 700 Hz) disappears and is replaced by a new emission approximately 100 Hz lower in frequency. When the temperature is carefully controlled the two emissions can exist simultaneously. The lowest-frequency emission changes 0.015°C/octave suggesting that the mechanisms controlling the frequency of this emission may be different than those determining the frequencies of the other emissions. All but the lowest-frequency emissions are maximal in level and have minimal bandwidth when the frequency is close to 700 Hz, which is interpreted as evidence that these emissions are filtered by a temperature-independent process. Keywords: Frog; Spontaneousotoacoustic emission; Temperature

1. I n t r o d u c t i o n

The amphibian inner ear appears to be a result of divergent evolution (reviewed in Wever, 1985; Lewis and Lombard, 1988). In contrast to the auditory organs of mammals, the hair cells in the papillae do not sit on a flexible membrane, but are suspended from a limbic (cartilage-like) shelf in the frog's ear. A tectorial structure is attached to the hair cells and is suspended in the endolymph. There are two such papillae in the frog's ear. Each occupies its own recess with a flexible membrane separating the recess from the perilymphatic space. The basilar papilla appears to be a relatively simple transducer in which all nerves are tuned to the same frequency (Ronken, 1990). In contrast, the amphibian papilla appears to be a specialized organ with complex hair cell orientation (Lewis et al., 1982). The amphibian papilla is a complicated transducer exhibiting two-tone suppression and c o m -

Corresponding author. Tel.: (317) 494-3815: Fax: (317) 494 077 I.

bination tone generation (Lift et al., 1968; Capranica and Moffat, 1980; reviewed in Zakon and Wilczynski, 1988; Narins and Benedix, 1994). The differences in structure of the inner ear in frogs and mammals may lead one to anticipate that there would be many differences in function and it may be surprising that the frog's inner ear generates the narrow band sound from the inner ear known as spontaneous otoacoustic emissions (reviewed in KiSppl, 1995). The mechanisms responsible for spontaneous emissions are not well understood, but they are thought to be generated by the same mechanisms responsible for good frequency resolution and high sensitivity (see Van Dijk and Wit, 1995; McFadden and Mishra, 1993). In mammals this mechanism is thought to depend on outer-hair-cell motility (reviewed in Brownell, 1990; Probst et al., 1991). Hair cells in the non-mammalian species do not appear to be motile, but are electrically tuned (reviewed in Eatock et al., 1993). Spontaneous emissions in all species (with the possible exception of humans; Wilson, 1986) are dependent on

0378-5955/96/$15.00 Copyright ~5 1996 Elsevier Science B.V. All rights reserved. Pll S0378-5955(96)00057-3

G.R. Long et al. / Hearing Research 98 (1996) 22-28

body temperature (reviewed in Van Dijk and Wit, 1995). Changes in spontaneous emissions are not always accompanied by equivalent changes in the best frequency of nerve fibers (reviewed in Van Dijk and Wit, 1995). If we assume that the frequency of otoacoustic emission depends on electrical tuning of hair cells, the temperature dependence of emission frequency may result from the temperature dependence of calcium channel kinetics. Changes in mammalian emissions, which do not appear to depend on electrical tuning, are harder to interpret. Wiener kernel analysis of neural data lead Van Dijk and Wit (1995) to postulate that the temperature sensitivity of both neural fibers and otoacoustic emissions above 600 Hz can be explained by a sandwich model in which the frequency response is determined by a temperature-insensitive bandpass filter and a temperature-dependent low-pass filter. When there are several emissions in a single ear they do not appear to be independent. Stable emissions in human ears are separated by approximately 0.4 mm (reviewed in Talmadge et al., 1993). Most frogs investigated to date have been described as having two emissions, one above

23 i

~""m"'--~

1"

v

-

55

-

50

-

45

-

40

•

35

-

30

-

25

-

20

•

15

'E

v ~

• 10 Temperature (C° ) 15 2~

!

5

..-.L2

9O

i

250

i

o

10 20 30 500 750 1000 1250 1500 Temperature (°C) Frequency (Hz) I

]

I

-i

~

;8O Fig. 2. Waterfall plot o f the spontaneous emission spectra (64 averages)

from the right ear of Frog 2 (1 Nov. 1991). 7O

"6

6O

Q.

'50 F40

.30

2O

1000 Hz and one below (Van Dijk et al., 1994). These authors suggested that the higher-frequency emission may come from the basilar papillar and the lower-frequency emission from the amphibian papilla. The observed changes in frequency, level and bandwidth of these emissions with changes in body temperature were investigated. Although the changes in frequencies were systematic (decreasing temperature reduces the frequencies), the correlated changes in level and bandwidth were complex and not understood. In an attempt to better characterize the pattern of emissions in the edible frog (Rana esculenta) a detailed examination of the changes with temperature was undertaken.

10

2. Methods

400 600 800 1000 12001400 Frequency (Hz) Fig. I. Waterfall plot of the spontaneous emission spectra (32 averages each) from the right ear of Frog 1 on 25 Oct. 1991. The change in temperature with time is plotted to the right of the waterfall plot so that the peaks of the plots line up with the time and frequency of the frog at the time each spectrum was obtained.

The flogs were anesthetized via immersion in a 0.1% MS222 solution, then placed in a water bath and covered by a thin gauze soaked in the MS222 solution. The mouth of the frog was propped open by insertion of a small tube to reduce the acoustic coupling between the two ears and thus permit separate examination of the two ears of one flog. A Sennheiser 13-2272 microphone was connected to

G.R. Long et al./ Hearing Research 98 (1996) 22-28

24

a small tube that encircled the tympanic membrane. The microphone signals were amplified (1000 × ) and filtered (highpass 200 Hz, Krohn Hite 3550) before being digitized and stored on tape. On line inspection of the data and much of the analysis was done with an Advantest Spectrum Analyzer. Averaged spectra of the microphone signal were obtained by adding 32 or 64 FFT spectra of successive time windows (50% overlapped). The experiments were performed in Groningen. Some additional analysis was done at Purdue using the Wavetek 5820A spectrum analyzer. Frequency, bandwidth, peak height and area of the emissions was determined by fitting a Lorenzian curve to the data (see Van Dijk and Wit, 1995; Talmadge et al., 1993). Temperature was controlled by Peltier devices under the water bath. The frog's temperature was monitored by a small thermocouple placed in the mouth. The DC voltage from the thermocouple modulated the frequency of a Wavetek VCG oscillator. The oscillator signal was recorded on the second channel of the tape recorder. In the offline analysis, demodulation of this temperature signal provided the frog's body temperature with a precision of greater than 0.01°C during the experiment.

100

90

80

70

6O

g

!

50

rn

.

~2 ~ I..-

._

O

40

rn "1o

30

20

10

Temperature (°C) 15 25 60

~

"~

.

~ 0 10 15 20 25 400

600

800

1 0 0 0 1 2 0 0 1400

Co

Frequency (Hz)

50 Fig. 4. Waterfall plot of the spectra (64 averages) obtained from the left ear of Frog 1 (20 Nov. 1991).

40

The care and use of animals reported on in this study was approved by the Dutch government.

.i

30~

3. Results

~

i.=

,,.,n

~

....

- ~ ' ~ - J ~ - ~

400 600 800 1000 12001400 Frequency(Hz)

20

10

o

Fig. 3. Waterfall plot o f the spontaneous emission spectra (64 averages) from the right ear of Frog 3 (11 Nov. 1991 ).

Detailed examination of the spontaneous emissions in six ears of four frogs were undertaken. Representative waterfall plots of the changes seen with temperature are presented in waterfall plots (Fig. lFigs. 2-4). At the start of each experiment, when the frogs' temperature was near 20°C, each ear had two emissions: one above 1000 Hz and one between 700 and 800 Hz. Reducing the frogs' temperature reduced the frequency of the emissions. The bandwidth of the high-frequency emission reduced systematically with the reduction of the frequency of the emission and the peak height increased until the emission reaches around 750 Hz. The lower-frequency emissions changed frequency more slowly and, when they reached approximately 600 Hz, the emission disappeared into the noise floor and was replaced by another lower-frequency emission near 500 Hz. For a narrow temperature region both

25

G.R. Long et al./Hearing Research 98 (1996) 22-28 10

-1

Warm Frog (22.5 c)

External tone at 950 Hz

External Tone at 1250 Hz

Atn ~10

10

22

2

~3 0

500

1000 Frequency

1500

18

25

19

26

20

28

21

3O

2000

(Hz)

24

22

c

22

:.=

23

Fig. 5. Detailed spontaneous emission spectrum obtained from the left ear of Frog 3 on 6 Dec. 1991. The dark solid line is a least-squares fit to the data based on the sum of two Lorenzians.

24

36

~

4o

~

25

v

low-frequency emissions were simultaneously present (see Fig. 2). The exact pattern depended on the frog. In one ear (Fig. 3) the frequencies of the two lower-frequency emissions were so close that the peaks appear to merge. In other ears (Fig. 4) the 500 Hz emission is near the noise floor and would probably be missed in other circumstances. In all these waterfall plots most of the emissions reach peak amplitude between 700 and 800 Hz. It is in this region that the bandwidth o f the emissions are the narrowest (see Figs. 1-4). W e investigated this relationship more systematically by fitting each emission with a Lorenzian curve and obtaining estimates o f the frequency, peak, bandwidth and area of each emission. Many of the highfrequency emissions (see Fig. 5) were not easily fit by a single Lorenzian as they appeared to be asymmetric. A good fit was, however obtained when they were fit as two overlapping emissions. Independent confirmation that the high-frequency emission is indeed a combination of two emissions comes from a separate experiment in which external tones were played to the frog to characterize suppression of the emission by external tones. Fig. 6 displays some of the results from this experiment. As a 950 Hz external tone was increased in level, the frequency of the low-frequency emission is lowered. At the same time, the high-frequency emission splits into two peaks. Similar peak splitting is observed when a 1250 Hz tone is presented. This tone has no observable effect on the low-frequency emission. When the frequencies obtained by this fitting procedure are plotted as a function of the temperature of the frog (Fig. 7) there is a high correlation ( r > 0.96) between the frequency of each emission and the temperature of the frog. The slope for the emissions with frequencies above 600 Hz are close to 0.04 octaves per degree Celsius change in temperature (see solid lines in Fig. 7). The slope for the emissions below 600 Hz is close to 0.015 o c t a v e s / ° C ) . The temperature was measured in the mouth, and one would anticipate that the inner ear temperature would lag

26 27

42

29

45

3O 32 33 35 36 I end start

600

800

1000

1200

600

800

Frequency (Hz)

1000

F'requency

1200

(Hz)

Fig. 6. Spontaneous emission spectra obtained from Frog 3 on 16 Dec. 1991 when the temperature was kept constant at 22.5°C while external tones at 950 and 1250 Hz were played through a speaker placed above the frog. Note that the peak near 1100 Hz breaks into two peaks as the attenuation of the external tones is decreased.

1000 ~

900

=_.. o"

-

0.04 OCt/°C

Boo 700

°oo

500

400

'

10

,

t

12

.

,

~ . 14

•

~. 16

,

i 18

Temperature

,

,

~ , 20

,.,

. 22

.

t

,

24

(°C)

Fig. 7. Frequency as a function of temperature for data obtained in from the right ear of Frog 3 (18 Nov. 1991, circles), left ear of Frog 1 (20 Nov. 1991, squares) and the left ear of Frog 3 (12 Dec. 1991, triangles) each in a single experimental session. Solid lines indicates slopes of 0.04 oct a v e s / ° C and 0.015 o c t a v e s / ° C .

G.R. Long et a l . / Hearing Research 98 (1996) 22-28

26 500

bandwidth of all the emissions is plotted as a function of frequency (Fig. 9) we find that the emissions form a pattern with a minimal width near 700 Hz and a maximum width near 500 Hz (see for example the triangles). It appears that, independent of temperature, the emission nearest to this frequency will always be the tallest and narrowest. As either emission departs from this frequency its peak level is reduced and its bandwidth is increased.

200

IO0

80

-~ 50 m 20

4. D i s c u s s i o n

10 8 i

0.602

0.005

0.010

0.020

,

i

0.050

0.100

Peak (arbitary units) Fig. 8. Bandwidth of spontaneous otoacoustic emissions plotted as a function of the peak amplitude (arbitrary units) for the data in Fig. 7.

the mouth temperature. This may explain the small amount of hysteresis in the relationship between frequency and temperature. The emission frequency is probably a better estimate of the inner ear temperature than the probe temperature. Consequently all later analyses will be related to lYequency and not to temperature. The height of the emission is correlated ( r = 0.73 for the combined data) with the bandwidth (see Fig. 8): the higher the emission peak the narrower the bandwidth, with the total energy (proportional to peak height times bandwidth) staying essentially constant with changing bandwidth, (the correlation is very low r = 0.10). Since the estimate of bandwidth is least contaminated by variations in the characteristics of the acoustic coupling of the microphone to the ear and the calibration of the microphone, bandwidth was used for further analysis of the data. If the

200

I oo

,~ 40 m ~ 20

10

4

500

600

700

800 900 11~00

2000

Frequency (Hz)

Fig. 9. Bandwidth in Hz of spontaneous otoacoustic emissions as a function of frequency for the data plotted in Fig. 7.

The temperature dependence of the frequency of each emission in this paper is compatible with other research on the effect of temperature on OAEs in frogs (Whitehead et al., 1986; Van Dijk et al., 1989). Van Dijk et al. (1994) were unable to identify a consistent pattern of emission level on temperature. In addition to the frequency data, this paper describes a complex but consistent pattern of emission level on temperature. From the data presented in this work, it appears that there are four potential emissions in each frog. Two lowfrequency emissions are present around 600 Hz. In addition a high-frequency emission is present above 1000 Hz. This high-frequency contribution could be interpreted as being the resultant of two separate emission oscillations. This interpretation is based on (1) the fact that the spectrum of the emission is in many cases best fit by a dual-peaked Lorenzian curve (see Fig. 5), and (2) the observation that an external tone splits the high-frequency contribution in two separate peaks (Fig. 6). The complicated interaction between a spontaneous emission and an external tone was also shown by Wit (1989). In Wit's experiment, a 1 kHz emission shifted down to half the frequency of a 1790 Hz external tone. At the same time, he observed a quadratic distortion product of the two frequencies. Van Dijk et al. (1989) found at most two emissions in each frog ear. They reported one exception, where they found three emissions. Van Dijk et al. (1994) could have missed the distinction between the two low-frequency emissions because they used a more rapid rate of temperature change. Whereas Van Dijk et al. (1989) typically used 0.5°C/min, the rate used in this experiment was as low as 0. l ° C / m i n (see Fig. 2 between 5 and 20 min when the two low-frequency emissions are very obvious). In principle, the asymmetry observed for the highfrequency emission contribution could be the result of correlated amplitude and frequency fluctuations of a single sinusoidal signal (Middleton, 1987). However, the fact that the peak splits in two when an external tone is presented, is not consistent with a single-oscillator interpretation. The shape of all emissions with frequencies above 500 Hz appears to be determined by a separate temperature-independent mechanism: the width of each emission is minimal, while the peak height is maximal near 750 Hz (see

G.R. Long et al. / Hearing Research 98 (1996) 22-28

Fig. 9). The tallest narrowest emission is always the emission that is closest to 750 Hz and any emission becomes taller and narrower as its frequency approaches this frequency. This seems to suggest that the emissions are filtered by a temperature-independent filter. The existence of a second resonator to enhance the emissions over a range of frequencies is similar to the claim that distortion product otoacoustic emissions in humans are filtered by the tectorial membrane (Allen and Fahey, 1993) or the reticular lamina (Neely, 1993). However, note that the peak height, rather than the peak sound pressure level (e.g., peak area) is maximal near 700 Hz. W e found no consistent correlation between emission peak area and frequency. So, the second filter apparently does not filter the emission itself. Rather, the filter appears to attenuate the internal noise in the ear. Noise, which interacts with the emission oscillator, causes the emission peak to spread over a certain bandwidth (e.g. the emission has a finite width) (Bialek and Wit, 1984; Van Dijk et al., 1994). Our results indicate that the internal noise is minimal near 700 Hz, possibly because it is filtered by a separate temperature-independent filter in the ear. In the frog amphibian papilla, the tectorium has a complicated morphology (Lewis, 1991). The rostral portion of the papilla is triangular in shape and is covered by a massive tectorial structure. The caudal extension of the papilla is S-shaped, and is covered by a tectorial membrane. A curtain-like structure is strung across the entire amphibian recess and is connected to the tectorial membrane, which sits on the caudal extension. This extension is only present in advanced frog species. Both the tectorial curtain and tectorial membrane are well designed to transmit any fluid vibration to the hair cells. Possibly, the mechanics of the tectorium contributes to the behaviour of emission peak level and emissions peak width as function of temperature, which we report here. The greater complexity of the amphibian papilla means that it shares many characteristics of the basilar papilla of birds and the organ of Corti of mammals and this could make it a better candidate to be the source of the emissions. The basilar papillar has no efferents, no documented temperature sensitivity, no sensitivity to anoxia or ototoxic drugs and no two tone suppression (reviewed in Zakon and Wilczynski, 1988). The amphibian papilla has all of these characteristics and also has electrically tuned hair cells (Pitchford and Ashmore, 1986). As is shown in Fig. 7, the frequency of emissions below about 600 Hz are less temperature sensitive than the frequency of emissions above 600 Hz. Nerve fiber recordings revealed also differences below and above 600 Hz, Lewis et al. (1982) showed that the amphibian papilla is tonotopically organized, i.e. fibers with characteristic frequency below 600 Hz innervate the portion of the papilla rostral, to the tectorial curtain while those above 600 Hz innervate the portion caudal to it. In addition, the lowfrequency fibers exhibit two-tone suppression and f2-fJ

27

response, while the high-frequency fibers do not (Capranica and Moffat, 1980; Narins and Benedix, 1994). Finally, electrical tuning has been shown in hair cells from the low-frequency, rostral, portion of the papilla (Pitchford and Ashmore, 1986), and it appears to be limited to lower frequencies (Eatock et al., 1993; data from alligator lizard). If we assume that a low-frequency emission is generated at the low-frequency portion of the papilla, then the specialized hair cell and nerve fiber properties of this area may be related to the lower temperature sensitivity of this emission.

5. Conclusions Manipulation of the frequency of spontaneous otoacoustic emissions by temperature provides a useful tool to examine the pattern of interaction between spontaneous otoacoustic emissions in this frog. What had previously been seen as two independent emissions were revealed to be a complex pattern of at least four interacting emissions that were apparently being filtered by a temperature-independent process.

Acknowledgements The research of Glenis Long was supported by Purdue University, the Groningen Graduate School for Behavioral and Cognitive Neurosciences (BCN), and the N I D C D DC00307. The research of Pim van Dijk was supported by the Royal Netherlands A c a d e m y of Sciences and Arts. The ENT department in Groningen is part of and supported by the Groningen Graduate School for Behavioral and Cognitive Neurosciences (BCN).

References Allen, J.B. and Fahey (1993) A second cochlear-frequency map that correlates distortion product and neural tuning measurements. J. Acoust. Soc. Am. 94, 809-816. Biatek, W. and Wit, H.P. (1984) Quantum limits to oscillator stability: theory and experiments on acoustic emissions from the human ear. Phys. Rev. 104A, 173-178. Brownell, W.E. (1990). Outer hair cell electromotility and otoacoustic emissions. Ear Hear., 82-92. Capranica, R.R. and Moffat, A.J.M. (1980) Nonlinear properties of the peripheral auditory system of anurans. In: A.N. Popper and R.R. Fay (Eds.), Comparative Studies of Hearing in Vertebrates, Springer, New York, pp. 139-165. Eatock, R.A., Saeki, M., Hutzler, M.J. (1993) Electrical resonance of isolated hair cells does not account for acoustic tuning in the freestanding region of the alligator lizard's cochlea. J. Neurosci. 13, 1767-1783. KiSppl,C, (1995) Otoacoustic emissions as an indicator for active cochlear mechanics: a primitive property of the vertebrate ear. In: G.A. Manley, G.M. Klump, C. K~Sppl,H. Fastl and H. Oeckinghaus (Eds.), Advances in Hearing Research. World Scientific Press, Singapore, pp. 207-216.

28

G.R. Long et al. / Hearing Research 98 (1996) 22-28

Lewis, E.R. (1981) Suggested evolution of tonotopic organization in the frog amphibian papilla. Neurosci. Lett. 2, 131-136, Lewis, E.R., Leverenz, E.L. and Koyama, H. (1982) The tonotopic organization of the bullfrog amphibian papilla, an auditory organ lacking a basilar membrane. J. Comp. Physiol. 145, 437-445. Lewis, E.R. and Lombard, R.E. (1988) The amphibian inner ear. In: B. Fritzsch, M.J. Ryan, W. Wilczynski, T.E. Hetherington and W. Walkowiak (Eds.), The Evolution of the Amphibian Auditory System. Wiley, USA, pp. 93-124. Lift, H., Goldstein, M.H., FrishkopL L.S. and Geisler, C.D. (1968) Best inhibitory frequencies of complex units in the bullfrog. J. Acoust. Soc. Am. 44, 635-636. McFadden, D. and Mishra, R. (1993) On the relation between hearing sensitivity and otoacoustic emissions. Hear. Res. 71, 208-213. Middleton, (1987) An Introduction to Statistical Communication Theory, Peninsula Publ., Los Altos, CA. Narins, P.M. and Benedix Jr., J.H. (1994) Temperature dependence of two-tone suppression in the auditory nerve of the frog. J. Acoust. Soc. Am. 96, 2738-2745. Neely, S.T. (1993) A model of cochlear mechanics with outer hair cell motility, J. Acoust. Soc. Am. 94, 137-146. Pitchford, S. and Ashmore, J.F. (1986) An electrical resonance in hair cells of the amphibian papilla of the frog, Rana temporaria. Hear. Res. 27, 75-83. Probst, R., Lonsbury-Martin, B.L. and Martin, G.K. (1991) A review of otoacoustic emissions. J. Acoust. Soc. Am. 89, 2027-2067. Ronken, D.A. (1990) Basic properties of auditory-nerve responses from a 'simple' ear: the basilar papilla of the frog. Hear. Res. 47, 63-82. Talmadge, C.L., Long, G.R., Murphy, W.J. and Tubis, A. (1993) New off-line method for detecting spontaneous otoacoustic emissions in human subjects. Hear. Res. 71, 170-182.

Van Dijk, P., Wit, H.P. and Segenhout, J.M. (1989) Spontaneous otoacoustic emissions in the European edible frog (Rana esculenta): spectral details and temperature dependence. Hear. Res. 42, 273-282. Van Dijk, P , Wit, H.P., Tubis, A., Talmadge, C.L. and Long, G.R. (1994) Correlation between amplitude and frequency fluctuations of spontaneous otoacoustic emissions. J. Acoust. Soc. Am. 96, 163-169. Van Dijk, P. and Wit, H.P. (1995) Speculations on the relation between emission generation and hearing mechanisms in frogs. In: G.A. Manley, G.M. Klump, C. K~Sppl, H. Fastl and H. Oeckinghaus (Eds.), Advances in Hearing Research. World Scientific Press, Singapore, pp. 105-115. Wever (1985) The Amphibian Ear, Princeton University Press, Princeton. NJ. Whitehead, M.L., Wilson, J.P. and Baker, R.J. (1986) The effect of temperature on otoacoustic emission tuning properties. In: B.C.J. Moore and R. D Patterson (Eds.), Auditory Frequency Selectivity, Plenum Press, London, pp. 39-46. Wilson, J.P. (1986) The influence of temperature on frequency-tuning mechanisms. In: J.B. Allen, J.L. Hall, A. Hubbard, S.T. Neely and A. Tubis (Eds.), Peripheral Auditory Mechanisms. Springer-Verlag, Berlin, pp. 229-236. Wit, H.P. (1989) Comment on 'Otoacoustic evidence for nonlinear behaviour in frogs" hearing: suppression but no distortion' by R.J, Baker, J.P. Wilson and M.L. Whitehead. In: J.P. Wilson and D.T. Kemp (Eds.), Cochlear Mechanisms, Structure, Function, and Models. Plenum, London, p. 357. Zakon, H.H. and Wilczynski, W. (1988) The physiology of the anuran eighth nerve. In: B. Fritzsch, M.J. Ryan, W. Wilczynski, T.E. Hetherington and W. Watkowiak (Eds.), The Evolution of the Amphibian Auditory System. Wiley, USA, pp. 125-183.