Hearing Research 46 (1990) 161-170

Eisevier

161

HEARES 01376

Boise-induced threshold elevation as a function of peak sound pressure level

Jan Grenner ‘, Per O.L. Nilsson 2, Helen Sheppard 3 and Bharti Katbamna ’

Departments of t Audiology and Experimental Research, Malmij General Hospital, Sweden, 2 Occupational Audiology,

Sahlgren’s Hospital, Gothenburg, Sweden and 3 Institute of Physics, Lund University, Lund, Sweden

(Received 4 August 1989; accepted 17 January 1990)

Thirty-three groups of guinea pigs, consisting of five animals in each group, were exposed to a simulated impact noise with peak levels ranging between 119.5 and 134.5 dB SPL. By varying the repetition rate, different equivalent levels could be set at each peak level. The equivalent levels ranged from 96 to 117 dB SPL, and the exposure duration was 1.5 to 24 hours. The compound action potential thresholds were measured in l/3-octave steps between 1 and 20 kHz, one month after the exposure. Higher peak levels

resulted in a peak-shaped threshold elevation with a maximum around 8 kHz. For constant peak levels, the equal energy theory was supported. For exposures of equal energy but different peak levels, significantly higher threshold elevations resulted after exposure to higher peak levels.

Acoustic trauma; Equal energy hypothesis; Nr threshold; Guinea pig

Introduction

The equal energy hypothesis was developed as a model to predict the relative risk of noise-in- duced hearing loss at different sound levels. It assumes a trading relationship between sound pressure level and time in such a way, that when the exposure time is halved, an increase in the sound level of 3 dB is permitted (Eldred et al., 1955; Eldredge and Covell, 1958). The hypothesis has gained widespread support for continuous noise (Bums and Robinson, 1970; Grenner et al., 1989). Investigations have also supported the equal energy hypothesis for impact noise (Atherley and Martin, 1971; Martin, 1976; Smoorenburg, 1982; Stevin, 1986) and the same ‘3 dB-rule’ has been used in damage risk criteria (IS0 R 1999, 1971). In contrast to this, a 5-dB trading relationship has been used in another national standard (OSHA,

Correspondence to: J. Grenner, Dept. of Audiology, Malmij General Hospital, S-214 01 Malmo, Sweden.

1974). This relationship was adopted since inter- mittent noise was assumed to reduce the risk of hearing loss.

An earlier animal study by us (Grenner et al., 1988) on impact noise, supported the equal energy hypothesis for a constant peak level. In that study, the equivalent level was changed by adjusting the repetition rate of the impact noise. Other studies on impulse and impact noise have demonstrated departures from the equal energy hypothesis (Dancer and Franke, 1977; Henderson and Hamemik, 1986). The equal energy hypothesis may thus not be valid for all combinations of exposure time and sound level. Therefore, there is a need to investigate the effects of different aspects of noise exposure; exposure times, repetition rates, equivalent and peak levels. We have thus extended our study from 1988 with noise exposures at other peak levels.

The aim of the present study was to investigate the influence of different peak levels upon the resulting permanent threshold shift for impact noise. This type of study requires an animal model.

037%5955/90/%03.50 8 1990 Elsevier Science Publishers B.V. (Biomedical Division)

162

Material and Methods

Groups of pigmented guinea pigs, with five animals in each group, were exposed to noise from a loudspeaker (JBL 2441 + 2309) suspended above the animals’ cage in an anechoic chamber. Impact noise was simulated by a square wave click which was filtered in a l/3-octave-band filter to resem- ble a hammer blow on wood. The B-duration was 12 ms. The noise levels were monitored with Brtiel and Kjaer 2209 and 2218 meters. The peak levels

were set between 119.5 and 134.5 dB SPL (re 20 PPa). By adjusting the repetition rate between 0.15 and 33 impacts/second, equivalent levels between 96 and 117 dB SPL were obtained. The noise was the same as that used in an earlier study (Grenner et al., 1988). The spectral content is shown in Fig. 1. The durations of the noise exposures were be- tween 1.5 and 24 h.

28-35 days after exposure, the guinea pigs were anaesthetized with pentobarbital, droperidol and phenoperidine, as described by Evans (1979) and

TABLE I

EXPOSURE DATA OF THE GROUPS: PEAK AND EQUIVALENT LEVEL RE 20 pPa; EXPOSURE TIME IN HOURS; REPETITION RATE, EXPRESSED IN IMPACTS PER SECOND; AND TOTAL ENERGY OF THE EXPOSURE

L pea!% L, (dB) (dB) (f h)

Rep. rate

(s-l)

E

(dB re

1 J/m’)

T 1-20

(dB)

SD

(dB) * N=4

Fig.

119.5 102 6 15 25.3 4.1 1.6

125.5 96 12 0.88 22.4 4.1 1.3 *

125.5 102 12 4.0 28.4 8.2 4.1

125.5 102 24 4.0 31.4 17.8 5.1

125.5 108 6 IS 31.3 3.0 1.6

128.5 99 6 0.98 22.3 6.4 3.9

128.5 102 6 2.0 25.3 17.5 4.5

128.5 105 6 4.5 28.3 15.7 4.7

131.5 96 6 0.16 19.3 5.3 1.7

131.5 96 7 0.15 20.0 7.6 10.9

131.5 96 12 0.23 22.4 14.3 7.2 *

131.5 99 6 0.50 22.3 7.9 4.9

131.5 102 1.5 1.0 19.3 8.3 4.0 *

131.5 102 3 0.82 22.3 5.8 6.4

131.5 102 6 0.87 25.3 7.7 7.9

131.5 102 12 0.82 28.4 11.0 4.5

131.5 102 24 1.1 31.4 17.2 5.8

131.5 105 1.5 1.8 22.3 7.9 5.4

131.5 108 6 1.8 31.3 11.6 5.8

131.5 114 1.5 20 31.3 16.9 3.8

131.5 114 3 17 34.3 20.9 6.7

131.5 114 6 13 37.3 13.5 8.7

131.5 114 12 17 40.4 30.5 7.5

131.5 117 3 33 37.3 19.6 3.9

131.5 117 6 32 40.3 26.8 3.1 *

131.5 117 12 33 43.4 30.3 5.8

134.5 99 6 0.22 22.3 12.1 13.5

134.5 102 1.5 0.45 19.3 4.2 3.6

134.5 105 1.5 1.1 22.3 14.0 5.7

134.5 117 3 20 37.3 25.6 3.4

134.5 117 6 15 40.3 21.2 1.6

134.5 117 12 18 43.4 35.1 7.8 *

134.5 117 24 20 46.4 36.8 6.0

2D

2A, 5B

3A

3A

2E

2B. 4, 5B

2D, 4

4

_ 2A, 5A

2B. 5A

5A, 5B

2D

2C, 5A

2E

3B

3B

3B

3B

28

2C. 5B

3c

3c

3c

3c

Results with T,_ *,, = average threshold elevation l-20 kHz and standard deviation. Number of measured animals in each group (5 or

4). The groups with L,,, = 131.5 were also reported in Grenner et al., (1988).

163

a tracheostomy tube was inserted. The bulla was opened using the ventral approach, and an elec- trode was placed on the round window. The sig- nals were band-pass ‘ltered and led to an aver-

P ager (Nicolet 529). /3-octave clicks were pre- sented in the ear canal (external ear removed), using an electrostatic transducer (Briiel and Kjaer 4145). VIIIth nerve compound action potentials

were measured at fourteen frequencies in both ears, from 20 kHz to 1 kHz. A sixth animal from each group was kept as a control. The normalized thresholds were set with 19 selected control animals. Of the 33 groups which were exposed to noise, only four animals could be measured in five of these groups. Since we have shown that the thresholds in the right and left ears are positively correlated (Grenner et al., 1990) the right-left average was used for all group comparisons.

Results

The exposure data for the noise-exposed groups are given in Table I. In order to express the

exposure energy on a logarithmic scale, the energy values were calculated in dB relative to 1 J/m2, corresponding to 120 dB SPL for 1 s. The calcula- tion formula is thus:

dB(E) = L,, (in dB SPL) - 120

+ 10 log,,, t(in seconds)

The table also includes the average threshold elevation over fourteen frequencies from 1 to 20

power spectrum (ret. dB)

04 1

-25 _ lms

Frequency CkHz)

Fig. 1. Spectral distribution of the impact noise (Nicolet 444A) and waveform sample, as recorded from the microphone.

kHz for each group of animals. We have named this average T1_2,,.

Fig. 2 shows the mean threshold elevation as a function of frequency for a number of exposure conditions. In the panels A to C, the exposure energy was the same, 22.3 dB re 1 J/m’. In Fig. 2A, two different peak levels were used, but both

the equivalent level and the exposure time were the same. It can be seen that the higher peak value gave rise to a higher threshold elevation.

Figs. 2B-C show the same type of comparison. Again, the threshold elevation was greater after exposure to the higher peak level. The 8 kHz region was most sensitive to a high peak noise

level. In Fig. 2D, the exposure energy was 3 dB

higher than in A-C. The figure shows three ex- posures to 102 dB L, of impact noise for 6 h. The group with the highest peak level did not suffer the greatest losses in this case. In Fig. 2E, the

higher peak noise level again resulted in the higher

threshold elevation. In Fig. 3, the thresholds are shown as a func-

tion of frequency for the same peak and equiv- alent level, but for different exposure times. The

general configuration of the two curves in panel A is identical with a peak at 8 kHz, and a plateau at lower frequencies.

Figs. 3B and 3C also show the thresholds after different exposure times. After longer exposure times, the peak form of the curves disappeared and was replaced by progressively flatter curves. The figures also show the great variability that is considered typical for impact and impulse noise. In particular, the 6-hour group in Fig. 3B was less

damaged than could be expected. In Fig. 4, the exposure time was kept constant

as well as the peak level. The equivalent level was

changed by means of adjusting the repetition rate of the simulated impact noise. It can be seen again that 8 kHz was the frequency most affected by the

noise. Fig. 5A shows the threshold elevations in four

groups that were exposed to the same noise en- ergy, 22.3 dB re 1 J/m2. The peak level was also the same in all four exposures, 131.5 dB SPL. From the method of exposure, it can be concluded that the animals had been exposed to the same number of identical clicks, but spread out during

164

dB 30

i

7 I1 I I, mf 1 2 4 B 20 Frequency (ktiz)

Fig. 2. Threshold elevation as a function of frequency. Varying peak level, but constant exposure time and equivalent level within

each panel. Each point is the average of ten (in some cases, eight) ears.

different exposure times. The longest exposure time gave rise to the greatest threshold elevation. A theoretical explanation for this is not obvious. However, it should be noted that the group with the highest threshold elevation only contained four animals. Nevertheless, the difference at 8 kHz was significant (t = 2.51; d.f. = 7; P < 0.05).

Fig. 5B shows four other groups, also exposed

to the same noise energy. As opposed to fig. 5A, the peak value was not constant. Instead, the ‘crest factor’ was constant, i.e. the peak level was 29.5 dB above the equivalent level in all ex- posures. In this case, the highest threshold eleva-

tions were observed in the group exposed to a high peak level during a short period of time.

In order to analyse which sound components are the most harmful, a multiple linear regression was performed with T,_,, as the dependent varia- ble. Since the 8 kHz region in Figs. 2-5 seemed to be especially vulnerable to impact noise, the same analysis was applied to the threshold T, (right-left average at 8 kHz). We have shown earlier (Grenner et al., 1988) that the threshold elevation is best correlated with a linear function of the log of the energy. We therefore tried log energy, Leq, log time, and Lpeak as independent variables. The

165

I. = 134.5 dB 6h i

peak

\

\

Leq=117 dB m-m

Frequency (ktfz)

Fig. 3. Threshold elevation as a function of frequency. Varying exposure time within each panel, other variables constant.

results are presented in Table II. It can be seen that the log of the energy correlated well with the threshold elevation, and that nothing was gained by splitting the log of the E into L,, and log time. On the other hand, an improvement in the correla- tion was gained when the peak level was intro- duced into the regression equation. The Lpeak fac- tor was significant (P < 0.01). The magnitude of the difference may seem small. However, this only reflects the composition of the material, with more than half of the animals exposed to 131.5 dB. A much greater difference could be demonstrated if only the lowest and highest peak levels were con- sidered. For T,_zO, the following regression equa- tions were found in the 33 groups:

T ,_2,, = 0.99 log E - 14.6

T 1_20 = 0.93 log E + 0.70 Lpeak - 103.9

T 1_2,, = 0.87 L,, + 11.9 log t(hours)

+ 0.81 Lpeak - 193.0

The second equation thus shows that for equal exposure energy, but different peak level, different threshold elevations resulted. For application to damage risk criteria, the third equation can be converted to doubling rates (= permissible incre- ment in dB when the exposure time is halved). For constant peak level, the doubling rate is 4.1 dB, and for constant crest factor, 2.1 dB.

The regression analysis for T, showed the same pattern as did T1_2,,, but the correlations were generally lower. The correlation coefficients that are given in Table II were calculated from individ-

t=6 h

I 2 4 8 20

Frequency (kliz)

Fig. 4. Threshold elevations at different equivalent levels.

ual animals. When group data were entered, con- siderably higher values of r2 were obtained. For

T 1_20, the values were 0.75-0.82. When the material was divided according to the

peak noise level, separate regression lines were

TABLE II

CORRELATION BETWEEN THRESHOLD DATA AND

EXPOSURE VARIABLES FOR 160 ANIMALS

Coefficient of determinance (r’)

Function T* T l-20 Wag E) 0.31 0.56

f(log ET Lpeak ) 0.34 0.60

f(L,,, Iog t) 0.31 0.56

f(L peak, L,,, l"gt) 0.35 0.61

Thresholds at 8 kHz or l-20 kHz (14 frequencies), right-left average.

A

; 00 20 30 40

Exposure energy (dB re 1 J/m21

Fig. 6. Average threshold elevation as a function of exposure

energy. Regression lines for different peak levels. Endpoints

mark highest and lowest exposure energy for each regression

line.

obtained. These are shown in Fig. 6, demonstrat- ing higher threshold elevations after exposure to higher peak levels.

Discussion

The measured threshold elevations can be studied as a function of each variable of the impact noise, while the other variables are kept

unchanged. One may consider the elevations at each separate frequency, or the average threshold

dB

Frequency (kHz)

Fig. 5. Threshold elevations as a function of frequency. Equal exposure energy within each panel. Numbers at each curve refer to

equivalent level/exposure time. Left: constant peak level. Right: constant crest factor and repetition rate.

elevation. T,_ 20 is an unweighted measure of the

overall threshold elevation. However, this type of analysis does not take into account differences in the shape of the ‘audiogram’, although it can be seen that the threshold efevations were frequency

specific. There was a tendency for the short ex- posures to cause a localised, peaked loss around 8 kHz, whereas the long exposures, in generalized terms, resulted in a flat loss. T,_,, must therefore be interpreted with caution. The T8 analysis did not improve the significance or alter the results. We therefore saw no reason to use other limits than 1 and 20 kHz.

Peak ieuel When both the equivalent level and the ex-

posure time were kept constant, higher threshold

elevations were seen after exposure to higher peak levels, as shown in Fig. 2. The same tendency was found in a study on the chinchilla, where the animals were exposed to an impulse noise of a constant repetition rate (Patterson et al., 1986). The influence of the peak level did not show statistical significance. If, however, the results of that study are combined with those of another study (Hamemik et al., 1987) with additional animals exposed to the same noise, and the per-

manent threshold shift (PTS) values from the fig- ures are inserted in a multiple linear regression model, significance is obtained for the influence of the peak level (P = 0.013). The exposure energy then gives a multiple r2 of 0.91, rising to 0.97 after the peak level is also inserted (group data, not individual animals). In that experimental model, the influence of the peak level resembied our data, whereas the energy did not:

PTS (l-2-4 kHz) = 2.0 log E + 0.6 Lpeak - 63

In another study in the chinchilla, on impact noise, it was also concluded that short exposures to high peak levels gave rise to more hearing loss than long exposures to lower peak levels, although the total energy was the same (Henderson and Hamernik, 1986). The sig~fican~e level is not stated in their paper, but if the data from the figures are inserted in the same type of equation, the result at equal energy is:

PTS (0.5-2-8 kHz) = 1.3 Lpeak - 136

167

The peak level is significant (P i 0.001). The findings reported in these papers are thus in good agreement with the results obtained in the present study, since our equation predicts worse hearing loss after high peak levels, for the same exposure

energy. Our findings support the hypothesis that the

noise energy is a fundamental measure of the noise load. The energy corresponds to an integral of the square of the pressure over time. It has been described that the integral of pressure over time is a better descriptor of the noise load (Kraak, 1981). We have not measured that type of integral, and therefore we cannot test that hypothesis.

The linearity of the middle ear at high ampli- tudes has been measured by several authors. It has been shown in cadaver ears, that the mode of

stapes movement shifts at high amplitudes (von Btktsy, 1960). It has also been pointed out that the annular ligament of the stapes has a limited range of movement (Price, 1974a). This implies a peak clipping, and therefore less threshold eleva- tion than expected at extreme peak levels. This is contrary to our results, since we found more damage at higher peak levels. On the other hand, measurements of ossicular movements in vivo

(Guinan and Peake, 1967), have shown that the response of the middle ear of the cat is linear up to 140 dB SPL. If the same were true for the guinea pig, then the peak levels that we have used in our experiment would be below the clipping level. High peak levels may thus increase the hazard of noise-induced hearing loss below the level of clipping, but may be, relatively speaking, less harmful above the same limit.

With our method of exposure, a high peak level corresponded to a low repetition rate, for the same energy. For impact noise, the middle ear muscles can only protect the ear if the repetition rate is high, so that relaxation of the muscles does not occur between the impacts. Rather high repetition rates were used in some of the exposures, where the ‘impact noise’ sounded more or less continu- ous. It is thus possible that the action of the middle ear muscles may have affected the results. The exposures were not primarily designed to distinguish between a low repetition rate and a

168

high peak level as risk factors. However, the multi- ple stepwise correlation analysis gives the follow- ing results. After the influence of the (log) energy had been calculated, the partial correlation to threshold elevation was much greater for the peak level (0.32) than for the repetition rate (-0.07). Therefore, one must conclude that the peak level is a significant risk factor, whereas the low repe-

tition rate is not. In the paper on impact noise

(Henderson and Hamemik, 1986) it was reported that a low repetition rate was correlated with a high temporary threshold shift (ITS), but the ef-

fect on PTS was not so pronounced, and probably without statistical significance, according to our calculations. An earlier study on impulse noise in the guinea pig (Dancer and Franke, 1977) showed that the TTS was highest after a repetition rate of 3 per s, and less pronounced at both faster and slower rates. No such tendency was seen in our material.

equivalent levef and duration of exposure

For the Lpeak 131.5 dB exposures, we have

earlier shown that the exposure time and equiv- alent level follow the equal-energy hypothesis and thus are interchangeable, at least for the average threshold elevation (Grenner et al., 1988). In gen-

eral, longer exposure times gave rise to threshold elevations at all frequencies. The picture was more

uniform after intense exposures, creating a tend- ency towards ‘flat loss’. At lower peak and equiv- alent levels, as in Fig. 3A, the threshold elevations exhibited a peak at 8 kHz, more resembling the ‘noise dip’ in man. It has been shown that the resonance of the external ear and ear canal creates a ringing in the transient response, as recorded at the eardrum (Price, 1974b; Pfander, 1982). These resonances, coinciding with the maximum audi-

tory sensitivity, have been suspected to be of importance for the ‘noise dip’ at 3-6 kHz in man (Tonndorf, 1976). The threshold shift is thus asso- ciated with the most sensitive part of the audio- gram. In our study, the peak loss was greatest at 8 kHz. Since the m~mum sensitivity for the guinea pig is around 8-10 kHz (Cazals et al., 1980; Grenner et al., 1989), and the ear canal resonance lies at the same frequency (Saunders and Tilney, 1982), this upward shift in frequency, compared with man, is not surprising. The degree of peaking

of the curves is somewhat greater in our study than has been demonstrated in other animal stud- ies on impact noise. This may be due to our measurement technique, recording thresholds with l/3-octave intervals, rather than l/l or l/2 oc- tave intervals, which is more common.

The trading relationship between time and in- tensity is of interest for damage-risk criteria. In

the study on the chinchilla by Hamernik et al. (1987) the effects of an impulse noise (131-147 dB peak SPL) with a constant repetition rate were

investigated. For a tenfold increase in exposure time, a 10 dB reduction in the sound pressure was required to give the same hearing loss. Our results are of the same order of magnitude; if a tenfold increment in exposure time is inserted into equa- tion 3 above, a 7.1 dB reduction in intensity is required to produce the same threshold elevation. If, instead, the repetition rate is increased tenfold, and the exposure time kept constant, a 10.7 dB reduction is required.

In conclusion, the threshold elevation that fol- lows impact noise can be significantly better pre- dicted if the peak level of the stimulus is taken into account, and not merely the equivalent level. We also found that short exposures generally caused threshold elevations around 8 kHz, whereas

longer exposures caused a tendency towards a flat hearing loss. The results are a clear departure

from the equal energy hypothesis, stating that equal exposure energies give rise to equal hearing loss or threshold elevation.

Acknowledgement

This study was supported by the Swedish Work Environment Fund.

References

Atherley, G.R.C. and Martin, A.M. (1971) Equivalent-continu-

ous noise level as a measure of injury from impact and

impulse noise. Ann. Occup. Hyg. 14, 11-28.

Bums W. and Robinson, D.W. (1970) Hearing and noise in

industry. HMSO, London.

B&k&y G. von (1960) Experiments in hearing. MC Graw-Hill,

New York.

Cazals, Y., Aran, J.M. and Hawkins, J.E.J. (1980) Threshold

elevation at high frequencies of the auditory nerve action

potential in acute versus chronic recordings in guinea pigs.

Hear. Res. 2. 95-109.

169

Dancer, A. and Franke, R. (1977) Influence des divers

parambtres caracterisant un bruit impulsif sur l’audition.

Etude cha le cobaye.Revue d’Acoustique 41,127-136.

Eldred, K.M., Gannon, W.J. and Gierke, H. von (1955) Criteria

for short time exposure of personnel to high intensity jet

aircraft noise. USAF, WADC Technical Note 355.

Eldredge, D.H. and Covell, W.P. (1958) A laboratory method

for the study of acoustic trauma. Laryngoscope 68,465-477.

Evans, E.F. (1979) Neuroleptanaesthesia for the guinea pig.

Arch. Otolaryngol. 105, X85-186.

Grenner, J., Nilsson, P., Sheppard, H. and Katbamna, B.

(1988) Action potential threshold elevation in the guinea-pig

as a function of impact noise exposure energy. Audiology

27, 356366.

Grenner, J., Nilsson, P. and Katbamna, B. (1989) AP threshold

elevation in the guinea pig following exposure to a broad-

band noise. J. Acoust. Sot. Am. 86,2223-2228.

Grenner, J., Nilsson, P. and Katbamna, B. (1990) Right-left

correlation in guinea pig ears after noise exposure. Acta

Otolaryngol. (Stockh), 109, 41-48.

Guinan, J.J. and Peake, W.T. (1967) Middle-ear characteristics

of anesthetized cats. J. Acoust. Sot. Am. 41, 1237-1261.

Hamemik, R.P., Patterson, J.H. and Salvi, R.J. (1987) The

effect of impulse intensity and the number of impulses on

hearing and cochlear pathology in the chinchilla. J. Acoust.

Sot. Am. 81, 1118-1129.

Henderson, D. and Hamemik, R.P. (1986) A parametric

evaluation of the equal energy hypothesis. In: R.J. Salvi, D.

Henderson and R.P. Hamemik (Eds.), Basic and applied

aspects of noise-induced hearing loss, Plenum, New York,

pp. 369-377.

IS0 (1971) Acoustics: assessment of occupational noise ex-

posure for hearing conservation purposes. IS0 R 1999,

l-11.

Kraak, W. (1981) Investigations on criteria for the risk of

hearing loss due to noise. In: J. Tobias and E.D. Schubert

(Eds.), Hearing research and theory, vol. 1, Academic Press,

New York, pp. 189-299.

Martin, A. (1976) The equal energy concept applied to impulse

noise. In: D. Henderson, R.P. Harnemik, D.S. Dosanjh and

J.H. Mills (Eds.), Effects of noise on hearing, Raven, New

York, pp. 421-453.

OSHA, Department of Labor (1974) Occupational noise ex-

posure. Proposed requirements and procedures. Federal

Register 39, 207, X5-159.

Patterson Jr, J.H., Lomba-Gautier, I.M., Curd, D.L., Hamer-

nik, R.P., Salvi, R.J., Hargett Jr, C.E. and Turrentine, G.

(1986) The role of peak pressure in determining the audi-

tory hazard of impulse noise. In: R.J. Salvi, D. Henderson

and R.P. Hamemik (Eds.), Basic and applied aspects of

noise-induced hearing loss, Plenum, New York, pp. 405-

423.

Pfander, F. (1982) Damage risk criteria with and without ear

protection for impulse noise with high intensities regarding

ear, larynx and lungs. Stand. Audiol. Suppl. 16, 41-47.

Price, G.R. (1974a) Upper limit to stapes displacement: impli-

cations for hearing loss. J. Acoust. Sot. Am. 56, 195-197.

Price, G.R. (1974b) Transfo~ation function of the external

ear in response to impulsive stimulation. J. Acoust. Sot.

Am. 56, 190-194.

Saunders, J.C. and Tilney, L.G. (1982) Species differences in

susceptibility to noise exposure. In: R.P. Hamemik, D.

Henderson and R.J. Salvi (Eds.), New perspectives on

noise-induced hearing loss, Raven, New York, pp. 229-234.

Smoorenburg, G.F. (1982) Damage risk criteria for impulse

noise. In: R.P. Hamemik, D. Henderson and R.J. Salvi

(Eds.), New perspectives on noise-induced hearing loss,

Raven, New York, pp. 471-490.

Stevin, G.O. (1986) Mathematical simulation of the cochlear

mechanism applied to damage-risk criteria for impulse

noise. In: R.J. Salvi, D. Henderson, and R.P. Hamemik

(Eds.), Basic and applied aspects of noise-induced hearing

loss, Plenum, New York, pp. 603-611.

Tonndorf, J. (1976) Relationship between the transmission

characteristics of the conductive system and noise-induced

hearing loss. In: D. Henderson, R.P. Hamernik, D.S.

Dosanjh and J.H. Mills (Eds.), Effects of noise on hearing,

Raven, New York, pp. 159-177.