Collier-Keyworth Company

Read the official release (oshrc.gov)

Plain-English summary

Collier-Keyworth manufactured office-chair mechanisms using power presses and lathes that produced high noise levels. The Commission initially held that the permissible-noise table did not cover impulse noise and planned to vacate related citation items unless the Secretary rebutted historical materials the Commission had officially noticed. The Secretary requested more time to gather evidence about the intent of the standard's drafters in 1969. The Commission granted that opportunity and remanded the case to the judge for an expedited supplemental decision. If the Secretary chose not to reopen the record, the judge was instructed to vacate the impulse-noise items.

Decision snapshot

• Cited standard(s): 29 C.F.R. § 1910.95(a); 29 C.F.R. § 1910.95(b)
• Outcome: Remanded for additional evidence and a supplemental decision on whether the standard covered impulse noise.
• Key point: The Secretary received an opportunity to rebut officially noticed historical materials concerning the original scope of the occupational-noise standard.

Full text (OSHRC public release)

Docket No. 80-2848

SECRETARY OF LABOR,

Complainant,

v.

COLLIER-KEYWORTH COMPANY,

Respondent.

OSHRC Docket No. 80-2848

ORDER OF REMAND

On April 6, 1987, the Commission issued a decision
holding that Table G-16 of the occupational noise standard at 29 C.F.R. � 1910.95(a)-(b)
did not cover impulse noise. The Commission based its holding on the language of the
standard, its legislative history and the Secretary's interpretation of the standard
shortly after its adoption. In weighing the legislative history of the standard and the
Secretary's interpretation of the standard shortly after its adoption, we referred to
documents that we officially noticed under section 7(d) of the Administrative Procedure
Act, 5 U.S.C. � 556(e). We also stated that we would afford the Secretary the opportunity
to rebut the officially-noticed documents with documents and other evidence of his intent
when he adopted the standard in 1969. The Commission therefore ordered that certain
citation items be vacated unless the Secretary requested an opportunity within 15 days to
rebut officially-noticed documents.

The Secretary has filed a motion for an extension of
time to gather, review and possibly submit documents relevant to the intent of the
standard's drafters with respect to the coverage of impulse noise. The Secretary asks for
extension of three weeks, through May 11, 1987. Ordinarily, we would simply grant this
motion, for it is supported by good cause. In this case, however, the Secretary's motion
would, if granted, cause the period of time to extend beyond the term of one of the
Commission's two current members. The Commission might for some time therefore be unable
to directly act on further motions in the case and might be unable to remand the matter to
the administrative law judge to receive further documents and supporting testimony.
Inasmuch as this issue is an important one and should be expeditiously resolved, we will
grant the motion and remand this case to the administrative law judge now with
instructions to afford the Secretary an opportunity to rebut officially-noticed documents
with evidence of the Secretary's intent when he adopted the standard in 1969. The judge
shall expeditiously prepare a supplemental decision affirming, modifying or vacating those
citation items relevant to the impulse noise issue in light of any additional evidence
that the parties submit. If the Secretary does not desire that the record be re-opened,
the judge shall issue a decision vacating the citation items in accordance with our
decision of April 6, 1987.

Accordingly, this case is remanded to the
administrative law judge for further proceedings.

FOR THE COMMISSION

Ray H. Darling, Jr.

Executive Secretary

DATED: April 22, 1987

SECRETARY OF LABOR,

Complainant,

v.

COLLIER-KEYWORTH COMPANY,

Respondent.

OSHRC Docket No. 80-2848

DECISION

Before: BUCKLEY, Chairman; WALL, Commissioner.

BY THE COMMISSION:

This case is before the Occupational Safety and
Health Review Commission under 29 U.S.C. � 661(j), section 12(j) of the Occupational
Safety and Health Act of 1970, 29 U.S.C. �� 651-678 ("the Act" or "the
OSH Act"). The Commission is in adjudicatory agency, independent of the Department of
Labor and the Occupational Safety and Health Administration ("OSHA"). It was
established to resolve disputes arising out of enforcement actions brought by the
Secretary of Labor under the Act and has no regulatory functions. See section 10(c)
of the Act, 29 U.S.C. � 659(c).

I. Introduction

Collier-Keyworth Company manufacturers swivel and
tilt mechanisms for office chairs in a plant in Gardner, Massachusetts. In April 1980,
OSHA industrial hygienists inspected Collier-Keyworth's plant to determine whether the
company was in compliance with the occupational noise standard at 29 C.F.R. � 1910.95
(a)-(b). From measurements made during the inspection, OSHA concluded that employees who
operated power presses and lathes at the plant were exposed to noise in excess of the
limits established by the standard. It therefore cited Collier-Keyworth for various
violations of section 1910.95. At the time of the alleged violations, the standard
provided:[[1]]

� 1910.95 Occupational noise exposure .

(a) Protection against the effects of noise exposure
shall be provided when the sound levels exceed those shown in Table G-16 when measured on
the A scale of a standard sound level meter at slow response. . . .

*

*

*

(b)(1) When employees are subjected to sound exceeding those listed in Table G-16,
feasible administrative or engineering controls shall be utilized. If such controls fail
to reduce sound levels within the levels of Table G-16, personal protective equipment
shall be provided and used to reduce sound levels within the levels of the table.

(2) If the variations in noise level involve maxima at intervals
of 1 second or less, it is to be considered continuous.

(3) In all cases where the sound levels exceed the values shown
herein, a continuing, effective hearing conservation program shall be administered.

Table G-16--Permissible Noise Exposures[[1]]

Duration per day, hours

Sound level dBA slow response

8
.......................................................................................90

6
.......................................................................................92

4
.......................................................................................95

3
.......................................................................................97

2
......................................................................................100

1� ...................................................................................102

1.......................................................................................105

�......................................................................................110

� or less ..........................................................................115

[[1]] When the daily noise exposure is composed of two or more periods of noise exposure
of different levels, their combined effect should be considered, rather than the
individual effect of each. If the sum of the following fractions: C 1 /T 1
+C 2 /T 2 [+ . . .] C n /T n exceeds unity, then,
the mixed exposure should be considered to exceed the limit value. C n indicates
the total time of exposure at a specified noise level, and T n indicates the
total time of exposure permitted at that level.

Exposure to impulsive or impact noise should not exceed 140 dB peak sound pressure
level.

The principal dispute in this case is whether Table
G-16 regulates a kind of noise called "impulse noise." The dispute arises in
this case because, in determining whether Collier- Keyworth's employees were exposed to
more noise than Table G-16 permits, OSHA employed measuring devices--sound level meters
and personal noise dosimeters--that registered impulse noises in Collier-Keyworth's plant.
Collier-Keyworth argues that Table G-16 was not intended to regulate impulse noise and
therefore that OSHA's measurements were unreliable.

A. Measuring Noise Exposure Under Table G-16

Before we set out and discuss these arguments, we
shall discuss generally the two methods of proving that an employer has violated the
limits for noise exposure in Table G-16 of section 1910.95(a)-(b). First, the Secretary
may show that an employee is subjected to a sound level equal to or greater than a level
listed in Table G-16 for a longer period of time than the table permits. For example, the
Secretary could meet his burden of proof by showing that an employee is exposed to 92 dBA
or higher for more than six hours, or to 105 dBA higher for more than one hour. If this
method is used, however, sound levels that last for less than their permissible periods
would have to be disregarded. Yet, sound levels in industrial plants often fluctuate and
continue for less than their permissible periods; some last for only a few minutes or
seconds. This method therefore can result in a substantial underestimate of an employee's
total noise exposure.[[2]]

The second way in which the Secretary can show
overexposure overcomes this difficulty but introduces others. The second method uses the
cumulation formula in the footnote to Table G-16. This formula explains how one may
calculate an employee's noise dosage when sound levels vary throughout the workday. Under
the cumulation formula, a fraction, C n /T n (C n being the
time the employee is exposed to that level and Tn being the exposure time permitted at
that level), is calculated for each different noise level to which an employee is exposed
during the workday. At the end of the day, all the fractions thus calculated are added
together. If the total is more than one ("unity"), the employee has been exposed
to noise exceeding permitted limits.

The cumulation formula yields a more accurate
determination of an employee's total noise exposure, for it counts as part of the
employee's noise exposure sound levels that continued for less than the permissible
period. This permits each variation in noise level to be given its proper weight and to be
cumulated. However, the cumulation formula is considerably more difficult to apply. An
OSHA compliance officer must hold the microphone of a sound level meter in an employee's
hearing zone; read each different sound level registered on the meter; with a watch,
determine the duration of that level; and record on a notepad each sound level and its
duration. Where the noise level is constantly varying, however, it is extremely difficult
to manually make the many measurements and calculations necessary to apply the formula.
And, because the task is so difficult to perform accurately, compliance officers cannot
survey the exposure of more than a few employees in a day.

To overcome these problems, the personal noise
dosimeter was developed. The dosimeter is an electronic instrument that contains the
circuitry of a standard sound level meter along with other circuitry. A dosimeter is small
enough to be worn by an employee throughout the workway as he goes about his duties. The
dosimeter microphone, which is part of the dosimeter's sound level meter, is placed in the
employee's hearing zone, where it continually detects the noise to which the employee is
subjected. The sound level meter circuitry within the dosimeter measures the varying sound
levels and sends that information to the other dosimeter circuitry, which automatically
calculates and sums the fractions as required by the cumulation formula.[[3]] At the end
of the workday, the dosimeter gives a reading in terms of the percentage of the permitted
daily dosage. A dosimeter reading of 100 percent means that the sum of the fractions is
one. A dosimeter reading exceeding 100 percent thus means that the permitted amount of
noise dosage has been exceeded. See New England Container Co. , 84 OSAHRC
55/A2, 12 BNA OSHC 1368, 1371  n. 7, 1984-85 CCH OSHD � 27,148, pp. 35,045-46 n. 7
(No. 78-1539, 1984).

The Secretary measured the noise exposure of
Collier-Keyworth's employees using dosimeters. Because the dosimeters indicated that
Collier-Keyworth's power press and lathe operators were exposed to more than 100 percent
of the permitted daily dosage, the Secretary alleges that the employees were exposed to
noise exceeding the limits permitted by the standard. Collier-Keyworth contends that the
dosimeter readings cannot be relied upon to find overexposure. The company raises a number
of potential inaccuracies with dosimeters, but its principal argument is that the
dosimeter readings obtained by the Secretary were unreliable because some of the noise
they measured was impulse noise, a type of noise the company argues is not intended to be
included in dosage calculations under Table G-16.

B. Impulse Noise

Impulse noise is noise of brief duration produced by
a short-lived physical phenomenon, such as a gunshot, a clap of thunder, or a power press
stroke. When such noise arises from the impact of two objects, it is called impact noise.
The distinction between impact noise and other impulse noise is not important to the
issues before us. A provision of the standard, however, requires that we distinguish
between impulse noises based on the time between their peaks or maxima. Subsection
1910.95(b)(2) states that "[i]f the variations in noise level involve maxima at
intervals of 1 second or less, it is to be considered continuous." This provision
reflects that much industrial noise consists of series of repetitive impulses and that the
impulses are often spaced so closely together as to sound continuous. One witness gave the
example of a gearbox, in which noise originates as the impact of gear teeth on gear teeth
but the impacts occur so frequently that the ear does not hear the individual impacts.
Subsection (b)(2) provides a criterion for classifying closely-spaced impulse noise as
continuous noise. Concomitantly, expert witnesses for both parties testified that noise is
considered to be "impulse noise" if it involves maxima that are less than one
second in duration and more than one second apart.

Collier-Keyworth's arguments focus both on the words
of the standard and its background. Collier-Keyworth points to a provision that is now
printed as the second paragraph of the footnote to Table G-16. That provision states that
"[e]xposure to impulsive or impact noise should not exceed 140 dB peak sound pressure
level." Collier-Keyworth argues that his provision demonstrates that the standard
treats impulse noise differently from other noise. It also points to subsection (b)(2) of
the standard. Because the one-second interval mentions in that subsection corresponds to
the interval that is universally recognized as the dividing line between impulse noise and
continuous noise, the company contends that this represents the standard's intent to draw
a line between impulse noise and continuous noise, with the latter regulates by Table G-16
calculation and the former excluded. It also relies on various documents and testimony to
support its argument that, when the noise standard was adopted, the Secretary did not
intend to regulate impulse under Table G-16.

Administrative Law Judge Foster Furcolo agreed with
Collier-Keyworth's argument that Table G-16 was not intended to cover impulse noise. On
that basis, the judge vacated the citations in their entirety. We agree with the judge's
conclusion that impulse noise must be excluded when determining whether Table G-16 limits
have been exceeded. We further find that the employees covered by items 1a(B) and (D), and
1b(B) and (D), of the citation were exposed to considerable amounts of impulse noise, and
as to those employees the Secretary did not prove overexposure to continuous noise.
Employees covered by citation items 1a(A) and (C), and 1b(A) and (C), however, were
exposed to noise defined as continuous by subsection (b)(2) of the standard, i.e., to
noise with impulses occurring more frequently than once per second. Although we find that
the Secretary proved that those employees were exposed to noise exceeding permitted
limits, we also find that the Secretary did not show that Collier-Keyworth failed to take
the steps required by the standard to protect the employees. We therefore affirm the
judge's disposition vacating all citation items.

II. Does Table G-16 Cover Impulse Noise?

The record establishes, as we shall discuss more
fully later, that at least some of the noise measured during OSHA's inspection of
Collier-Keyworth's plant was impulse noise, that is, noise with sharp energy peaks lasting
less than one second and spaced more than one second apart. If Table G-16 does not
regulate impulse noise--or, stated differently, if impulse noise must be excluded from the
calculation of dosage when one uses the cumulation formula in Table G-16--then dosimeter
readings that include impulse noises produce readings that are too high. Therefore, in
determining whether the readings obtained by OSHA show that Collier-Keyworth's employees
were exposed to excessive noise, we must first address Collier-Keyworth's argument that
section 1910.95(a)-(b) does not intend for impulse noise to be included in dosage
calculations employing Table G-16 limits.[[4]]

A. The Scope of Our Inquiry

The Secretary contends that Collier-Keyworth's
argument should be rejected because the record compiled in this litigation "contains
substantial expert testimony that impulsive...noise is no less damaging to hearing then
continuous noise." The Secretary cites the testimony of several witnesses, including
that of a physician specializing in otolaryngology, and argues that the degree of
permanent hearing loss is directly proportional to the amount of acoustic energy
transferred to the ear.[[5]] The Secretary reasons that because the energy associated with
prolonged exposure to impulse noise can cause irreversible hearing loss, it should be
regulated by Table G-16.

We do not dispute that there is now reason to believe
that impulse noise can cause hearing loss. But in the absence of an OSHA standard, we do
not have the authority to regulate exposure to impulse noise. New medical and scientific
evidence is not relevant if, as Collier- Keyworth claims, the Secretary did not intend
Table G-16 to cover impulse noise when he adopted it. As an adjudicative body, the
Commission must take the standard as it finds it and apply the standard in accordance with
the Secretary's intent at the time of promulgation. See Oscar Mayer & Co. v.
Evan , 441 U.S. 750, 758 (1979) (intent of Congress that enacted statute is
controlling). The Commission has no authority to "update" any standard through
interpretation according to the latest scientific findings. Such legislative authority
resides in the Secretary, who, in rulemaking, may consider whether new findings physical
agents make regulation necessary and what new duties should therefore be imposed on
employers. For the Commission to interpret a standard to produce what it believes will be
greater protection for employees than the Secretary originally intended both usurps the
Secretary's rulemaking authority and detracts from the statutory right of those affected
by a rule to participate in the rulemaking process.

B. The Language of Section 1910.95(a)-(b)

The Secretary argues that the plain language of the
standard indicates that he intended to regulate impulse noise under Table G-16. He notes
that section 1910.95(a) requires protection whenever "sound levels" exceed Table
G-16 values, and argues that the all-inclusive phrase "sound levels" draws no
distinction between impulse noise and non-impulse noise.

Standard must, however, be read as a coherent whole,
and Collier-Keyworth points to two provisions it contends indicate an intent to exclude
impulse noise from Table G-16.

Collier-Keyworth points to the statement, now printed
as the second paragraph of the footnote to Table G-16, stating that exposure to peak sound
pressure levels of impulse noise should be limited to 140 dB. It argues that this special
provision for impulse noise shows an intent to treat impulse noise differently from other
noise. The Secretary, on the other hand, contends that this provision demonstrates that
where the standard intended to distinguish impulse noise from other noise, it did so
explicitly.

We think that there is force in Collier-Keyworth's
argument. Although Table G-16 sets a ceiling of 115 dBA on sound pressure levels
regardless of the duration of exposure, Wheeling-Pittsburgh Steel Corp. , 83 OSAHRC
16/A2, 11 BNA OSHC 1292, 1294, 1983-84 CCH OSHD � 26,482, p. 33,673 (No. 15647, 1983),
the special provision for impulse noise warns against impulse noise that exceeds 140 dB.
This provision would be superfluous if, as the Secretary argues, the all-inclusive phrase
"sound levels" in section 1910.95(a) means that all sound, both impulse and
non-impulse, must stay within the limits in Table G-16. Thus, the impulse noise provision
suggests that the "sound levels" regulated by Table G-16 do not include all
sound levels and that impulse noise must be excluded from calculations under Table
G-16.[[6]]

It might seem somewhat odd that an answer to this
important question might be suggested by a brief second paragraph of a footnote to a
table. However, the drafters of the standard did not intend the impulse noise provision to
be relegated to such an obscure position. The provision was originally placed in the text
of the standard but later became located in the footnote as the result of a printer's
error.[[7]] We therefore ascribe no importance to the placement of the impulse noise
provision.

The second provision on which Collier-Keyworth relies
is subsection (b)(2), which defines noise as continuous if the peaks are closer than one
second apart. Collier-Keyworth contends that the Secretary's classification of
closely-spaced impulses as continuous noise must be viewed as evidence of an intent to
distinguish impulse noise from continuous noise and to exclude impulse noise from the
criteria established for continuous noise. The Secretary's brief does not address
subsection (b)(2) or suggest how it could be reconciled with his position that impulse
noise is subject to the limits of Table G-16.

We agree with Collier-Keyworth that subsection (b)(2)
appears to represent an attempt to distinguish impulse noise from continuous noise. Both
parties agree that continuous noise, i.e., noise of constant intensity, is included in
Table G-16 calculations. By defining noise with peaks closer than one second apart as
continuous, subsection(b)(2) obviously means that such noise is to be included in dosage
calculations. A provision that says something is included, however, implies that something
else is excluded, for there would otherwise be no reason for the provision. As the
one-second criterion in subsection (b)(2) represents the distinction between impulse noise
and non-impulse noise, the section suggests that the drafters of the standard intended to
exclude impulse noise from Table G-16.

On balance, both the 140 dB peak limit for impulse
noise and subsection (b)(2) suggest that impulse noise is excluded from dosage
calculations under Table G-16. Nevertheless, the standard is not entirely clear on its
face. It does not expressly exclude impulse noise from Table G-16. As the Secretary points
out, the standard uses the broad term "sound levels." In light of the ambiguity,
we look to evidence extrinsic to the standard to determine whether the drafters intended
to exclude impulse noise from such calculations.

C. Legislative History of the Standard

Section 1910.95 is derived from 41 C.F.R. �
50-204.10, a standard that was originally promulgated in 1969 by the Labor Department's
Bureau of Labor Standards to regulate work by government contractors under the
Walsh-Healey Government Contracts Act, 41 U.S.C. �� 35-45. Section 1910.99 (listing
sources of standards); see generally American Can Co. , 82 OSAHRC
5/A2, 10 BNA OSHC 1305, 1306-1308, 1982 CCH OSHD � 25,899, pp. 32,409-11 (No. 76-5162,
1982) (setting out course of promulgation of standard). Because 41 C.F.R. � 50.204.10
qualified as an "established federal standard" under section 3(10) of the
Occupational Safety and Health Act of 1970 ("the OSH Act"), the Secretary of
Labor was authorized to adopt it as an OSHA standard without further rulemaking
proceedings. Section 6(a) of the OSH Act, 29 U.S.C. � 655(a). The Secretary did so on May
29, 1971, adopting 41 C.F.R. � 50-204.10 verbatim as an OSHA standard and codifying it at
29 C.F.R. � 1910.95. 36 Fed. Reg. 10466, 10518. Because the Secretary did not, and--with
exceptions not pertinent here--could not substantively amend the standard when he
summarily adopted it as an OSHA standard, we must give the standard the same meaning it
had under the Walsh-Healey Act. See American Can , 10 BNA OSHC at 1310-11,
1982 CCH OSHD at pp. 32,413-14; Sherwin-Williams Co. , 84 OSAHRC 28/A2, 11 BNA OSHC
2105, 2109-10, 1984-85 CCH OSHD � 26,986, pp. 34,701-02 (No. 14131, 1984).

The promulgation of 41 C.F.R. � 50-204.10
represented the culmination of a long effort by the effort by the Labor Department and
other organizations to develop exposure limits for noise that would protect workers
against noise-induced hearing loss. Although the preamble that accompanied the standard
when it was adopted under Walsh-Healey Act sheds no light on the light on the impulse
noise question, the record in this case contains evidence of events leading up the
adoption of the Walsh-Healey standard, as well as evidence of the Secretary's
interpretation of the standard shortly after its promulgation. The parties presented
witnesses, James H. Botsford and Herbert H. Jones, who were personally familiar with many
of the events leading up to the standard's adoption. Their testimony is supplemented by
documentary exhibits introduced by the parties and other documents that we officially
notice under section 7(d) of the Administrative Procedure Act, 5 U.S.C. � 556(e).[[8]]
Our examination of the record and the material that we officially notice, together with
the words of the standard, lead us to conclude that the Secretary did not intend for Table
G-16 to regulate impulse noise when the standard was promulgated.

The documents we officially notice include only Labor
Department publications and scientific papers cited in documents that were introduced into
evidence. Our purpose in taking official notice is not to determine whether impulse noise
is harmful or whether the preponderance of scientific thinking in 1969 would have favored
the regulation of impulse noise under Table G-16. As we have said, weighing such evidence
is not our function but the Secretary's. Our purpose is instead to determine the intent of
the Secretary in 1969 by examining the body of knowledge on impulse noise that influenced
the adoption of the standard. Because our decision rests in part on these officially-noted
documents, we shall afford the Secretary an opportunity to rebut them with documents and
other evidence of his intent when he adopted the standard.

In 1960, the Walsh-Healey standard for noise stated
in its entirety: "Noise shall be reasonably reduced or eliminated as a means of
preventing fatigue or accident." 25 Fed. Reg. 13809, 13825 (1960). In early 1964, the
department proposed numerical guidelines that would establish weekly exposure limits for
"continuous steady noise" and also establish a peak limit of 135 dB for
"any exposure...however short in duration other than impact noises...." U.S.
Dept. of Labor, Bureau of Labor Standards, "Suggested Language for a Noise Control
Program," as printed in "Noise: Guidelines for control issued by Bureau
of Labor Standards," Safety Standards 18, 20-21, 24 (U.S. Dep't of Labor,
March-April 1964); see also U.S. Dept. of Labor, Bureau of Labor Standards, Suggested
Language for a Noise Control Program (1965).

The Technical Committee on Noise of the American
Industrial Hygiene Association ("AIHA") met in May 1964 with Jones (then of the
U.S. Department of Health, Education and Welfare) and Botsford (then employed by Bethlehem
Steel and its senior noise control engineer). The committee members believed that the
newly-published guidelines had technical errors in them and that better guidelines could
be developed. The AIHA therefore decided to offer to the Labor Department the formation of
a committee of noise experts to recommend different guidelines. The Labor Department
agreed to the formation of such a committee, which was called the Inter-Society Committee
on the Guidelines for Noise Exposure Control ("Inter-Society Committee"). The
committee consisted of two members from each of five technical societies: the AIHA, the
American Conference of Governmental Industrial Hygienists ("ACGIH"), the
American Academy of Ophthalmology and Otolaryngology, the American Academy of Occupational
Medicine, and the Industrial Medical Association. One of the ACGIH representatives was Dr.
Floyd Van Atta of the Labor Department, who, according to Jones, "had primary
responsibility in the noise area as far as the Department of Labor was concerned."

The Inter-Society Committee issued a report in 1967,
which it characterized as the "first attempt to extract and condense pertinent data
from various scientific literature into a meaningful and authoritative guide."
Inter-Society Committee, "Guidelines for Noise Exposure Control," 28 Am. Indus.
Hygiene J. 418, 419-22 (Sept.-Oct. 1967)(I-S). For this inquiry, it is noteworthy that the
committee's guidelines were directed at developing exposure criteria for only
"steady" noise.[[9]] The following are excerpts from the guidelines:

I. Forward

*

*

*

Noise-induced hearing loss increases with both the intensity of the noise and the
duration of exposure. Generally, many years of exposure to high noise levels are required
to produce significant permanent impairment in the exposed group; however, there will be
marked differences in the hearing of individuals and in their response to noise...These Guidelines
will be directed toward the prevention of that portion of the permanent hearing loss
resulting from exposure to steady noise.

*

*

*

III. Occupational Hearing Loss Control Program

*

*

*

A. Evaluation of the Noise Hazard

*

*

*

1. Noise Measurement .

Continuous or intermittent steady noise is readily measured by standard instruments;
impulsive noise requires special procedures not considered here. [Footnote omitted.]

*

*

*

2. Hazard Rating .

(a) Continuous Exposure . . . . The first two columns of Table 1 [omitted] indicate
the steady noise levels to which the various groups [of persons with hearing loss] were
exposed. . . .

(b) Intermittent and Part-time Exposure . The studies on which Table 1 and Figure 1
[omitted] are based, dealt with men exposed to noise during a normal workday, of eight
hours' duration. . . .

The report went on to explain that, although there
were no long-term studies of the effects of intermittent exposures, the results of some
studies implied the "simple rule"

that for each halving of daily exposure time, the noise levels may be increased by 5
db up to a maximum of 115 dB . . . without increasing the hazard of hearing impairment.

The application of this rule is illustrated in the
following table:

Permissible Increase in db for less than Eight Hours
Exposure

Daily Exposure Time

(Hours)
Permissible Increase

Above Eight-Hour

Criterion

(Decibels)

8
0

4
5

2
10

1
15

�
20

�
25

Notably, the Inter-Society Committee's report did not
propose any particular permissible exposure level; for this reason, the AIHA Board found
the report unacceptable and suggested that the Inter-Society Committee be disbanded.
Shortly thereafter, the ACGIH established its own committee to develop a permissible
exposure limit for noise as well as other physical agents. Jones was chairman of this
committee. The ACGIH Committee drafted a proposed standard that was published for public
comment and accompanied by an explanatory article written by Jones, "ACGIH's Proposed
Threshold Limit Value for Noise," 29 Am. Indus. Hygiene J. 537-40 (Nov.-Dec. 1968) ( ACGIH ).
See generally Appendix A.1 to this decision.

The most noteworthy aspect of the ACGIH Committee's
proposal is that it twice stated that its proposed limits would "not apply to impulse
or impact type of noise." ACGIH at 540 (proposed standard). Instead, the
proposed standard stated, "[i]t is recommended that exposure to this type of noise
should not exceed 140 db peak sound pressure level." Id . Jones' article
explained the ACGIH Committee's reasons for distinguishing between continuous noise and
impulse noise:

After considering [various data], the Committee
decided that at the present time it appears desirable to establish a limit of 92 dBA for 4
to 8 hours of exposure per day to broad band continuous noise. . . .

Laboratory data. . .and. . .field data indicate that, when exposure is for less than
a full 8-hour period or is intermittent in nature, the ear can tolerate more acoustical
energy per day than for a single exposure to continuous noise. Considering these two
factors, the limit is increased 5 decibels for each halving of exposure periods for the
work day regardless of whether this is a single exposure or an exposure which is
intermittent in nature.

*

*

*

Very little data is available upon which to base exposure to impact or impulsive
noise. It is known that exposure to a small number of 140 dB impulsive noises of short
duration will produce a temporary threshold shift. Until additional data is available, a
limit of 140 dB is being set impact or impulsive noise.

ACGIH at 538 (article) (references omitted).

Thus, the criteria proposed by both the Inter-Society
Committee and the ACGIH Committee excluded impulse noise, with the exception of the ACGIH
proposal for a peak limit of 140 dB for impulse noise. The reason they otherwise excluded
impulse noise was apparently because, as stated in Jones' article, "[v]ery little
data is available upon which to base exposure to impact or impulsive noise." Both the
Inter-Society Committee and the ACGIH Committee had cited a number of scientific studies
completed by the late 1960's attempted to measure the effect of noise on human hearing.
These studies, which are excerpted in the appendix to this decision, had measured hearing
loss caused by various degrees of noise exposure. Because most of them involved purely
steady-state noise, they formed a substantial body of evidence from which permissible
exposure limits and an exchange rate--the number of decibels by which sound levels could
rise if exposure time was halved--could be derived. However, there was considerably less
evidence on which to base permissible exposure limits for impulse noise. The scientists
who attempted to develop hearing protection criteria cautioned against applying them
impulse noise, and tended to regard impulse noise and continuous noise as two distinct
problems. The studies cited by the Inter-Society Committee and the ACGIH Committee dealing
with "steady" noise primarily attempted to determine a safe eight-hour exposure
level and an exchange rate that represented an appropriate trade-off of higher intensity
for shorter time. Studies of the harmful effects of impulse noise, by contrast looked for
hearing protection criteria in terms of factors such as the peak intensity, total number
of impulses, and duration of impulses. Thus, the criteria proposed for "steady"
noise cautioned against application to impulse noise while criteria for impulse noise
would not, by their very nature, apply to steady noise. See generally Appendix
A.2-A.7.[[10]]

On September 20, 1968, the Labor Department published
a proposed standard to regulate workplace noise exposure. 33 Fed. Reg. 12458, 14259-60.
This proposed standard bore very little resemblance to the present standard. It basically
established a weekly average exposure limit of 85 dB for "steady (or equivalent)
noise" and included impulse noise in its exposure calculations. It spoke of noise
with "intervals...more than one second and...maxima less than 1 second each,"
which fits the definition of impulse noise. The proposed standard provided that such noise
maxima are to be included in the exposure calculations as if they had durations of 1
second each. See generally Appendix A.8.

The proposal was unfavorably received in the
industrial hygiene community. Despite this, on January 17, 1969, a few days before
President Johnson's term expired, the Labor Department issued a package of Walsh-Healey
standards, including a noise standard very similar to that which had been proposed. 34
Fed. Reg. 788, 790-91. The promulgated standard retained the exposure limit of the
proposed standard and the provision that dealt with impulse noise.

Before the promulgated standard became effective, it
was stayed by the new Secretary of Labor. 34 Fed. Reg. 2207 (1969). On May 20, 1969, the
Labor Department issued a new noise standard that differed radically from the earlier one.
34 Fed. Reg. 7946, 7948-49. Except for some subsequent corrections (35 Fed. Reg. 1015
(1970)), the new standard is identical to the one at issue in this case. Major differences
between the earlier standard and the later standard include provisions relevant to the
impulse noise issue. The earlier standard explicitly stated that noise with maxima less
than one second in duration and greater than one second apart, i.e., impulse noise, was to
be included in weekly exposure calculations. However, the final standard contained no
similar provision and included a provision with no counterpart in the earlier standard--a
recommended limit of 140 dB for impulse noise.

The final standard strongly resembled that adopted by
the ACGIH on May 12, 1969, eight days before the final version of the Walsh-Healey
standard was adopted.[[11]] Jones was still the chairman of the ACGIH Committee; he stated
that Dr. Van Atta of the Labor Department had been aware of the workings of the Committee
and of its proposed standard, and that he had received advance copies of the final ACGIH
standard. Relevant excerpts from the ACGIH standard follow:

Threshold Limit Values

Noise

*

*

*

Continuous or intermittent

The sound level shall be determined by a sound level meter . . . operating on the
A-weighting network with slow meter response. Exposure shall not exceed that shown in
Table 1.

Table 1

Permissible Exposures

Duration per Day

Hours
Sound level

dB(A)*

8
90

6
92

4
95

3
97

2
100

1�
102

1
105

�
107

�
110

�
115-C**

*Sound level in decibels as measured on a standard
level meter operating on the A-weighting network with slow meter response.

**Ceiling Value

---

These values apply to total time of exposure per
working day regardless of whether this is one continuous exposure or a number of
short-term exposures but do not apply to impact or impulse type of noises.

When the daily noise exposure is composed of two or
more periods of noise exposure of different levels, their combined effect should be
considered, rather than the individual effect of each. If the sum of the following
fractions:

C 1 /T 1 +C 2 /T 1

• . . . . . . . C n /T n

exceeds unity, then, the mixed exposure should be considered to exceed the threshold
limit value, C1 indicates the total time of exposure at a specified noise level, and T1
indicates the total time of exposure permitted at that level . . .

Impulsive or impact noise

It is recommended that exposure to impulsive or
impact noise should not exceed 140 decibels peak sound pressure level-C.

This ACGIH standard is very similar to the
Walsh-Healey standard ultimately promulgated by the Secretary of Labor. Indeed, the final
Walsh-Healey standard was much closer to the ACGIH standard than it was to the earlier
Walsh-Healey standard that it supplanted. In particular, the most important part of the
final Walsh-Healey standard, the table listing the permissible exposure limits, is almost
identical to the corresponding table in the ACGIH standard. The passages on impulse noise
and the cumulation formula are also nearly identical to those in the ACGIH standard. In
light of the similarities between the two final standards, and of Jones' testimony that
Dr. Van Atta of the Labor Department was familiar with the proceedings of the ACGIH
Committee and had advance copies of its work products, we infer--as do both the Secretary
and Collier-Keyworth--that the final ACGIH standard heavily influenced the final
Walsh-Healey standard.[[12]]

The Walsh-Healey standard does not state explicitly
that impulse noise is excluded, as does the ACGIH standard, and the absence of such a
statement in the Walsh-Healey standard might be taken to mean that the Secretary rejected
this aspect of the ACGIH standard, deciding to include impulse noise in Table G-16
calculations. However, the final Walsh-Healey standard also does not state explicitly that
impulse noise is included in such calculations. In this critical respect, it departed from
the earlier Walsh-Healey standard, which expressly included impulse noise. Moreover, the
final Walsh-Healey standard contained the impulse noise provision and subsection (b)(2),
both of which suggest that impulse noise was to be excluded from Table G-16.

Although we cannot determine solely from the
standard's legislative history to this point whether the Secretary intended impulse noise
to be included in Table G-16 calculations, it does reveal several highly significant
facts: Many of the scientists and organizations whose attempts to develop criteria for
noise were cited by the ACGIH Committee and the Inter-Society Committee did not believe it
either necessary or appropriate for the criteria they developed for steady or continuous
noise to apply to impulse noise. Indeed, it appears that the preponderant thinking in the
scientific community that was cited by the ACGIH Committee and the Inter-Society Committee
was to exclude impulse noise from criteria for steady noise. The Labor Department was
evidently aware of this, for its technical expert on noise served on the Inter-Society
Committee and was aware of the views of and the limits developed by the ACGIH Committee.
The standard it adopted drew heavily on the work of the ACGIH Committee and, indirectly,
that of the Inter-Society Committee. There is also strong evidence that the Department of
Labor agreed with the body of scientific opinion cited by the ACGIH Committee and the
Inter-Society Committee.

On December 4, 1970, after the noise standard was
adopted under the Walsh-Healey Act, the Labor Department's Bureau of Labor Standards
published Bulletin 334, Guidelines to the Department of Labor's Occupational Noise
Standards for Federal Supply Contracts (Dec. 4, 1970), a detailed manual explaining to
government contractors what their duties were. The 1970 version[[13]] of Bulletin 334
states in part:

Table 1 [equivalent to Table G-16] indicates
[excessive noise] . . . . Employees must not be exposed to steady sound levels above 115
dBA, regardless of the duration.

*

*

*

. . . The "slow" response [on the sound level meter] is another setting . . .
which causes it to average out high level noise of brief duration (such as hammering),
rather than responding to the individual impact noises.

*

*

*

Impulse or Impact Noise

The last sentence in paragraph (d) of section
50-204.10 states:

"Exposure to impulsive or impact noise should not exceed 140 dB peak sound
pressure level."

This sets the upper limit of sound level to which a person should be exposed,
regardless of the brevity of the exposure.

In contrast with the 115 dBA upper limit for steady noise, the higher intensity for
impact noise is permissible because the noise impulse resulting from impacts, like hammer
blows or explosive processes, is past before the ear has time to react fully. Impact noise
levels are to be measured only with an impact meter or an oscilloscope.

*

*

*

Variable Noises

Paragraph (c) of section 50-204.10 states the final
consideration in determining whether or not a permissible sound level is being exceeded:

"(c) If the variations in noise level involve maxima at intervals of 1 second or
less, it is to be considered continuous."

This means that where the sound level meter on the A scale at slow response moves up from
a generally steady reading, say from 88 to 92 dB, at intervals of one second or less, the
high reading shall be taken as that to be used in Table 1 [equivalent to Table G-16].

As a corollary to this, intermittent sounds of brief duration at intervals greater
than one second should, as far as practical, be measured as to intensity and duration and
the total duration over a day be ascertained. This total should be entered in the equation
given in footnote 1, Table 1, to determine the permissible limit. These intermittent
sounds, which can be measured with a sound level meter, should not be confused with
impulse sounds of very short duration resulting from impacts or explosions .

Id . at 2, 6-7 (emphasis added).

Bulletin 334 went on to state that noise levels were
to be measured with a sound level meter, set for A scale and slow response, meeting the
specifications for sound level meters established by the American National Standards
Institute (ANSI) in ANSI S1.4-1961, "Specification for General-Purpose Sound Level
Meters." Bulletin 334 at 14. This ANSI standard contains the following provisions:

Introduction

*

*

*

It should also be recognized that the ballistics and other characteristics of the
indicating instrument are adapted mainly for measuring ordinary machinery noises and other
sounds of a reasonably constant character. For intermittent sounds, and more particularly
for repetitious sounds involving high peak-to-average ratios in sound pressure, and for
impact noises , other specialized equipment will be required in place of the ordinary
indicating instrument.

ANSI S1.4-1961 at p. 6 (emphasis added).

Several significant points emerge from Bulletin 334.
The explanation of why a higher peak level is permitted for impulse noise than for steady
noise--that the ear does not react fully to impulse noise--supplies a reason why impulse
noise would be excluded from Table G-16. The statement indicates that the Secretary was
aware of the scientific reasons why the investigators who had developed criteria by 1969
for steady noise did not apply their criteria to impulse noise. The instruction in
Bulletin 334 about including "intermittent sounds of brief duration" in dosage
calculations, coupled with the caveat against confusing such sounds with impulse noise, is
tantamount to an explicit statement that impulse noise is to be excluded from such
calculations. Finally, the provision in ANSI S1.4-1961 that standard sound level meters
cannot be used to measure impulse noise, together with the statement in Bulletin 334 that
sound level meters used for noise surveys must conform to ANSI S1.4-1961, is a further
indication that impulse noise is excluded from the standard except for the 140 dB peak
level provision.[[14]]

C. OSHA's Enforcement and Later Interpretations

The record also shows that it was OSHA's practice to
exclude impulse noise from Table G-16 calculations when enforcing the standard. Anne Hart,
the OSHA industrial hygienist who inspected ColIier-Keyworth's plant, answered
"no" when asked, "[y]ou are not supposed to include impulse noise in your
dosimeter or sound level meter measurements, are you?" Thomas Rockwell, an expert in
noise measurement and control and a consultant to numerous corporations, testified that he
wrote a letter to the Cleveland Area Office of OSHA asking if he was correct in
understanding that the standard excluded impulse noise, and received in response a phone
call from Fred Boelter, an industrial hygienist in the Chicago office, confirming that
understanding. We also note that Dr. John Barry, a noise expert employed by OSHA since
1976, wrote in a paper presented at Purdue University in 1979: "This standard [29
C.F.R. � 1910.95(a)-(b)] sets the permissible exposure level for non-impulse noise at 90
dBA for an 8-hour-per-day duration and less than or equal to 140 dB peak sound pressure
level for impulse noise irrespective of its duration."[[15]]

Dr. Barry also was the principal author of a document
instructing OSHA industrial hygienists in the proper techniques for determining compliance
with the standard. This document was first published on April 2, 1979, as OSHA Instruction
CPL 2-2.20 and was later incorporated in OSHA's Industrial Hygiene Field Operations Manual
("IHFOM") as Chapter IV.[[16]] Section C.1.e. provided:

Where both continuous and impulse noise are present,
measure the background on the dBA fast mode to determine if the continuous noise is above
the levels found in Table G-16, 29 C.F.R. � 1910.95.

Thus, OSHA instructed its inspectors to determine whether the continuous component alone
in a mixture of continuous and impulse noise exceeded the limits of Table G-16. Hart's
acknowledgment that she was not supposed to include impulse noise in dosage measurements
therefore reflected an official, agency-wide practice.[[17]]

Our understanding of OSHA's early enforcement
practices is further aided by OSHA's comprehensive proposal in 1974 to completely replace
section 1910.95 with a more precise standard. The standard that OSHA then proposed and the
explanation accompanying it indicate that OSHA did not view the existing standard as
including impulse noise in dosage calculations. The proposed standard would have retained
the 90 dBA limit in the existing standard but would have coupled that limit with
requirements for a stringent hearing conservation program when the eight-hour time
weighted average noise level exceeded 85 dBA. 39 Fed. Reg. 37773, 37774 (Oct. 24, 1974).
Although OSHA proposed to retain the 90 dBA limit at which engineering controls,
administrative controls, and personal protective equipment were required, it proposed to
draw a sharp distinction between steady-state noise and impulse noise, explicitly
excluding the latter from dosage calculations. The proposed standard provided in part:

� 1910.95 Occupational noise exposure .

*

*

*

(c) Permissible exposure limits .--

(1) Steady state noise -- single level . (i) The
permissible exposure to continuous noise shall not exceed an eight-hour time-weighted
average of 90 dBA with a doubling rate of 5 dBA. For discrete permissible . . . limits,
refer to Table G-16a [omitted, similar to Table G-16].

(ii) Where Table G-16a does not reflect
actual exposure times and levels, the permissible exposure to continuous noise at a single
level shall not exceed [an amount determined by a given formula] where "L" is
the workplace sound level measured in dBA on the slow scale of a standard sound level
meter . . . .

(2) Steady state noise -- two or more levels .
Exposures to continuous noise at two or more levels may not exceed [a dose computed by the
cumulation formula] where C is the actual duration of . . . a given steady state noise
level . . . .

(3) Maximum steady state noise level . Exposure to
continuous noise shall not exceed 115 dBA. . . .

(4) Impulse or impact noise . (i) Exposures to impulse or
impact noise shall not exceed a peak sound pressure level of 140 dB.

(ii) Exposure to impulses of 140 dB shall
not exceed 100 such impulses per day.  . . . For each decrease of 10 dB in the peak
sound pressure level of the impulse, the number of impulses to which employees are exposed
may be increased by a factor of 10.

39 Fed. Reg. at 37775. The preamble to the proposed
standard explained why OSHA proposed to treat impulse noise in this way:

The present OSHA standard recommends that impact or
impulse sounds not exceed a peak sound pressure level of 140 dB. The Advisory Committee
suggested that this limit be made mandatory. OSHA has made an addition to the Advisory
Committee's recommendation with respect to impulse noise exposure, because the actual
exposure is a summation of the peak sound levels of the impulses and the number of
impulses. OSHA proposes to limit exposure to impulses at 140 dB to 100 per day and to
permit a tenfold increase in the number of impulses for each 10 dB decrease in the peak
pressure of the impulse. For example, the number of impulses allowed at 130 dB would be
1,000 per day and the number of impulses allowed at 120 dB would be 10,000 per day.

39 Fed. Reg. at 37774. Although the proposed standard
thus specified separate exposure limits for steady-state noise and impulse noise, its
preamble gave no indication that OSHA considered the proposed standard a departure from
the existing standard. Indeed, its explanation of the proposed standard indicates that
OSHA viewed the 140 dB peak limit as being the only provision in the existing standard
addressed to impulse noise. While this falls somewhat short of being a formal
interpretation of the existing standard, it supplements the other evidence tending to show
that OSHA had interpreted the existing standard to exclude impulse noise from Table G-16.

The Secretary asserts that it was not his customary
practice to exclude impulse noise from Table G-16. He cites Turner Co. , 76 OSAHRC
108/A2, 4 BNA OSHC 1554, 1976-77 CCH OSHD � 21,023 (No. 3635, 1976), rev'd on other
grounds , 561 F.2d 82 (7th Cir. 1977), as an example of a case where a citation based
on impulse noise from power presses was issued and affirmed. Turner does not
support the Secretary's argument. The impulse noise issue was not raised in Turner
and the parties stipulated that the noise levels exceeded those permitted by Table G-16. 4
BNA OSHC at 1556, 1976-77 CCH OSHD at p. 25,274. Moreover, power press noise is not
inevitably impulse noise. If the noise maxima occur more frequently than once per second,
a situation that may well have been the case in Turner ,[[18]] the noise is defined
as continuous by subsection (b)(2) of the standard. In any event, even if Turner
did involve impulse noise, a single citation does not show an enforcement policy and does
not rebut the otherwise overwhelming evidence that OSHA interpreted the standard to
exclude impulse noise from Table G-16.

Bulletin 334 is nearly contemporaneous with the
adoption of the standard, consistent with OSHA's early enforcement policy, and therefore
likely to reflect the understanding of those in OSHA who drafted the standard. We cannot,
however, say the same of the arguments in the Secretary's brief. The Commission has
generally not considered interpretations of standards in the Secretary's brief to be
authoritative. E.g. , C. F. & I. Steel Corp. , 86 OSAHRC ____, 12 BNA OSHC
2067, 2074, 1986 CCH OSHD � 27,691, pp. 36,139-40 (No. 79-4786, 1986). It has Iong been
the experience of the Commission that interpretations advanced before the Commission in
the Secretary's briefs often are interpretations made to support litigation positions.
They are not necessarily interpretations traceable to the intent of the standard's
drafters. Id. ; see Investment Company Institute v. Camp , 401 U.S.
617, 626-28 (1971)(counsel's efforts in litigation are "hardly tantamount to an
administrative interpretation" of a statute). This is a vital distinction, for it is
the drafters of a standard who incorporate into it the policies and compromises that the
rulemaking record demonstrates are appropriate. To give weight to an administrative
interpretation not traceable to the standard's drafters would effectively permit the
Secretary to amend the standard without the safeguards mandated by Congress to afford
persons affected by the standard an opportunity to participate in the rulemaking process.

The brief the Secretary filed in this case gives us
no reason to believe that the interpretation it advances is traceable to the intent of the
drafters. The brief does not address subsection (b)(2) of the standard and its use of the
word "continuous." It does not mention Bulletin 334, or the 1974 proposed
standard and its preamble. It does not purport to have been written or reviewed by lawyers
who consulted with the drafters of the standard or who participated in the drafting of the
standard. The brief does address the 140 dB impulse noise provision but misapprehends the
events that lay behind its adoption. See note12 above. It also relies on the testimony
adduced in this litigation on the harmfulness of impulse noise and alludes to the findings
OSHA made in 1981 to support the hearing conservation amendment, though neither is
evidence of what the Secretary intended when he adopted section 1910.95(a)-(b). In sum,
the Secretary's brief reflects OSHA's current appreciation of impulse noise but provides
no insight into what the Secretary of Labor intended in 1969. It is therefore entitled to
no weight here.

Bulletin 334 and the Secretary's early enforcement
policy are entitled to considerable weight, however. Bulletin 334 is "a nearly
contemporaneous interpretation of the standard by its drafter." Wheeling-Pittsburgh
Steel , 11 BNA OSHC at 1294, 1983-84 CCH OSHD at p. 33,673. Its view that impulse noise
is to be excluded from Table G-16 calculations is also reflected in OSHA's original
enforcement policy and the understanding of OSHA officials in the 1970's, such as Dr.
Barry, who were responsible for the standard's enforcement. And it is consistent with the
rationale behind the 1974 proposed standard, a document that explicitly established
separate limits for impulse noise and steady-state noise and gave no hint that OSHA viewed
this as a departure from the current standard.

The Secretary argues that exclusion of impulse noise
"is fundamentally inconsistent with [the standard's] goal of safeguarding human
hearing." That a standard addresses a problem does not mean that it was intended to
safeguard employees from it entirely. Drafters may lack enough information to justify a
comprehensive treatment or to know how to write a comprehensive standard. They may
therefore decide to regulate a step at a time, going as far as their knowledge carries
them and leaving the rest for another day. That happened here. From its text and
legislative history, from Bulletin 334 and OSHA's early enforcement policy, we find that
the drafters of section 1910.95(a)-(b) intended to exclude from Table G-16 impulse
noise--that is, noise with sharp energy peaks lasting less than one second and spaced more
than one second apart. We therefore vacate citation items 1a(B) and (D), and 1b(B) and
(D), which are based on exposure readings of employees exposed to impulse noise.

III. Noise Exposure of Employees

The Secretary alleges that Collier-Keyworth employees
were exposed to noise in excess of the standard's limits and that the company did not take
the precautions required by the standard to protect the employees. The record shows that
the noise to which Collier-Keyworth's employees were exposed consisted of a mixture of
noise from the machines on which they were working and background noise that was present
even when their machines were not operating. The background noise was purely continuous
noise, i.e., noise of unvarying or slowly varying intensity. The machine noise basically
falls into two categories: (1) purely impulse noise, i.e., noise with sharp energy peaks
spaced at intervals greater than one second; and (2) noise resulting from impulses spaced
less than one second apart, which is defined as continuous by the standard and which, for
the sake of convenience, we shall call "quasi-continuous" noise. To prove
over-exposure, the Secretary introduced into evidence dosimeter readings that measure an
accumulated dosage from all noise present in the work environment, including impulse
noise. We have concluded, however, that pure impulse noise must be excluded from
measurements to determine compliance with Table G-16. Thus, in evaluating the dosimeter
readings introduced by the Secretary, we must determine whether they were contaminated by
sufficient impulse noise to render them unreliable measures of compliance with Table G-16
levels. We find, for the reasons discussed in below, that some samples were so
contaminated but that others were not, for they registrated quasi-continuous noise.
Collier-Keyworth also argues, however, that dosimeters do not reliably measure
quasi-continuous noise.

A. The Use of Dosimeters to Measure
Quasi-Continuous Noise

Collier-Keyworth argues, and Judge Furcolo agreed,
that in impulsive noise environments, the dosimeters used by OSHA may not be relied on to
prove violations of Table G-16 exposure limits because they indicate noise levels
significantly higher than the "true noise levels." In making this argument, the
company does not distinguish between pure impulse noise and quasi-continuous noise.

Such a distinction must be made, however. Although
pure impulse noise must be excluded in determining compliance with Table G-16,
quasi-continuous noise is included. Dosimeter readings in environments containing
significant amounts of pure impulse noise are inherently unreliable regardless of the
accuracy with which they measure impulse noise, for they should not be measuring such
noise at all. They should, however, be measuring quasi-continuous noise, and the accuracy
with which they measure this type of noise is critical. Thus, we must consider
Collier-Keyworth's argument to the extent it questions the accuracy with which OSHA's
dosimeters measure quasi-continuous noise.

We will assume in this discussion that
Collier-Keyworth is correct in its assertion that dosimeters in impulsive noise
environments register dosages as if noise levels were significantly higher than the
"true noise levels."[[19]] However, Collier-Keyworth's argument cannot logically
be seen as merely an objection to dosimeters but to the standard's prescription of the
slow response mode of measurement. As we have said, a noise dosimeter is in effect a sound
level meter with additional circuitry that automatically records dosages in the manner set
out in the cumulation formula of Table G-16. And, as we shall discuss below, the circuitry
of a dosimeter that arguably causes it to overcount "true noise" levels is not
its additional dosage- calculating circuitry but the slow response circuitry it shares
with conventional sound level meters. Thus, despite the emphasis in Collier-Keyworth's
brief on the inability of dosimeters to reliably detect "true noise levels," its
argument could apply with equal force to conventional sound level meters. Indeed,
Collier-Keyworth's acoustical engineering expert, Thomas Rockwell, admitted that he would
obtain the same results using a slow response sound level meter as a slow response
dosimeter.

Dosimeters and sound level meters employing slow
response seem to overcount the "true noise level" because slow response
instruments have an integration time of one second, that is, they register an impulse as
if it were spread out over a second. A burst of sound lasting one-half second would
therefore be displayed as if it were one second long. Because the slow response
integration time is longer than the duration of such a noise impulse, the slow response
mode distorts the "true" noise pattern, making impulses appear longer in
duration but lower in peak intensity.[[20]]

Collier-Keyworth argues that the net effect of the
slow response mode is to cause dosimeters to read higher than the "true noise
level." The problem with Collier-Keyworth's argument is that section 1910.95(a)-(b)
does not regulate "true noise levels." The standard regulates sound levels (1)
as measured by a standard sound level meter, (2) as detected by slow response, (3) as
weighed by frequency according to the A-weighting network (which weights each sound
frequency according to the car's response to it), (4) excluding impulses more than a
second apart, and (5) treating impulses less than a second apart as continuous. Indeed, it
is rather pointless to speak of "true noise levels" because they can never be
measured. To measure "true noise levels" would require an ideal instrument
having, among other things, an instantaneous response time, that is, an integration time
of zero. One could build a sound level detector with a short response time, perhaps an
eighth of a second, as fast response instruments have. But such an instrument would still
not be measuring "true noise levels."

Inasmuch as section 1910.95(a)-(b) does not regulate
"true noise levels" but levels as detected by slow response instruments,
Collier-Keyworth's argument is really complaint about the standard's prescription of slow
response. Collier-Keyworth seems to desire that sound level meters and dosimeters be
employed at fast response so as to more closely approximate the "true level" of
impulse noise.[[21]] However, the standard specifies the use of slow response and whatever
deviations from true noise levels that mode introduces were within the contemplation of
its drafters. That slow response instruments give greater weight to impulse noise than
fast response instruments must be regarded as one of the many compromises that the
Secretary struck when he adopted the standard. Just as the Secretary could provide that
impulses recurring at greater than one-second intervals not be counted at all, he could
provide for a measurement technique that gives additional weight to impulses closer than
one second together. Collier- Keyworth does not point to any illegality in the Secretary's
choice of slow response.[[22]] We therefore reject Collier-Keyworth's arguments that slow
response dosimeters may not be used to detect exposures greater than those permitted by
Table G-16.[[23]]

B. Evidence of Exposure

We now consider whether OSHA proved that any
Collier-Keyworth employee was over-exposed to quasi-continuous noise. On April 14, 1980,
OSHA industrial hygienist Anne Hart began an inspection of Collier-Keyworth's plant aimed
at determining whether the company was in compliance with the noise standard. Hart
attached General Radio Model 1954 dosimeters to five Collier-Keyworth employees, who wore
the dosimeters throughout their work day. One of the dosimeters malfunctioned, and Hart
returned on April 18 to obtain a dosimeter reading for that employee. Hart also measured
the noise levels in the employees' hearing zones at various times throughout the
inspection by using a sound level meter. The employees who were sampled operated power
presses and lathes in four different areas of the plant.

Presses 2208, 2209, and 2210 (citation items 1a(A))
and 1b(A)) are 60-ton Bliss open-back inclinable power presses used by Collier-Keyworth to
fabricate metal parts. The presses operate automatically. Sheet steel is fed into the
press bed by a feeder mechanism. A die attached to a ram above the bed descends and stamps
the part out of the stock. The part is removed from the bed by high-speed air blown out of
nozzles. The three presses were located within ten feet of each other.

The presses were operated by Paul LeBlanc and Norman
Melanson. The operators' duties including changing dies, making necessary adjustments,
installing coils of sheet steel stack, threading the stock into the die area, and
monitoring the operation of the presses. An operator had to be in the general area of the
machines while they were operating, but the operation of each machine did not have to be
monitored closely.

Using a sound level meter, Hart measured noise levels
in the vicinity of the presses ranging from 85 to 104 dBA, with readings below 90 dBA
occurring only when none of the presses were operating. The dosimeter worn by LeBlanc read
305% at the end of the day, and that worn by Melanson read 286%. These readings indicate
that both employees were exposed to approximately three times the permissible dose, or,
viewed a different way, that they were exposed to time-weighted average sound levels of
about 97.6 dBA and 98.0 dBA. See 29 C.F.R. � 1910.95, Appendix A, Table A-1 (hearing
conservation standard). Hart testified that she timed the presses using a watch with a
second hand and found that the impacts for each machine were less than one second apart.
She also stated that the strokes of the three machines were not synchronized, so that when
all three presses were running, the noise peaks were less than one-third second apart.

Collier-Keyworth's chief engineer, Robert Cochran,
testified that the presses had a maximum stroke rate of 80 per minute. The actual stroke
rate was controlled automatically by the feeder mechanism and was different for each part
being produced. Each machine was used to produce about 30 different parts. Cochran did not
know the stroke rates that corresponded to the various parts.

Following the inspection, noise measurements were
made in Collier-Keyworth's plant by experts for both parties: Thomas Rockwell for
Collier-Keyworth and John Barry for the Secretary. Both experts fed the output from sound
level meters into strip-chart recorders, which draw tracings showing how the noise levels
fluctuate with time. Tracings made by Rockwell on June 17, 1980, show press 2208 operating
at 57.5 stroke per minute, press 2209 at 48 strokes per minute, and press 2210 at 56
strokes per minute. Rockwell testified that Cochran told him that the presses were
producing some of the same parts they had been making on the day of Hart's inspection.

Under subsection (b)(2) of the standard, noise that
consists of a series of impulses less than one second apart is treated as continuous. This
quasi-continuous noise could be produced by a press operating at a stroke rate of more
than one per second, or sixty per minute. A press stroking slower than once per second
will, however, generate impulse noise. Collier-Keyworth's presses can stroke as fast as
eighty per minute, but the actual rate depends on the part being produced and for some
parts is slower than sixty per minute. Collier-Keyworth's presses therefore generate
quasi-continuous noise at certain times and impulse noise at other times.

The evidence shows that LeBlanc and Melanson were
exposed only to quasi-continuous noise on the day of the inspection. Hart testified that
she timed the presses and that the time between strokes was less than one second for each
press. There is no evidence directly contradicting Hart's testimony.[[24]]
Collier-Keyworth production records showing the parts produced on the day of the
inspection were introduced into evidence, but the company presented no evidence on the
stroke rates that corresponded to those parts.

Collier-Keyworth did show that on a later date, when
Rockwell measured the noise from its presses, the presses were stroking slower than once
per second, indicating that the noise at that time was impulse noise. Moreover, Rockwell
testified that Cochran had told him that the presses were producing some of the same parts
when he was in the plant as when Hart made her inspection. However, Rockwell spent parts
of three days in the plant, and the production records introduced into evidence show that
the presses typically produced different parts on different days. Therefore, during the
three days Rockwell was in the plant, the presses probably produced several different
parts each, including some that were made during Hart's inspection and others that were
not. In the absence of evidence that the parts produced when Rockwell made the strip chart
recordings showing the presses operating slower than once per second were the same as
those manufactured during Hart's inspection, Rockwell's testimony does not contradict
Hart's evidence that the presses were stroking faster than once per second.

We also note that it is not the noise generated by a
machine, but the noise heard by an employee, that the standard regulates. During most of
LeBlanc's and Melanson's workday, two or three presses were operating and generating
noise. As the presses were not synchronized, the noise maxima that they heard would have
occurred about twice per second with two presses running and about three times per second
with three presses operating. In any event, because each press individually produced
quasi-continuous noise, two or three presses operating at the same time would also produce
quasi-continuous noise.

Because LeBlanc and Melanson were exposed only to
quasi-continuous noise, their dosimeter readings did not include any measurements of
impulse noise. Those dosimeter readings--305% for LeBlanc and 286% for Melanson--facially
indicate exposure that exceeds permissible limits. Collier-Keyworth asserts that a number
of factors can affect dosimeter readings, including electromagnetic fields or radio waves,
wind, orientation of the microphone, chemicals, temperature effects, moisture, and noise
made by the employee wearing the dosimeter. There is no evidence, however, that any of
these factors significantly distorted the dosimeter readings made during the OSHA
inspection of Collier-Keyworth's plant. The dosimeter readings obtained by industrial
hygienist Hart are consistent with sound level meter readings she made on the day of the
inspection as well as with measurements made on later days by Rockwell and Barry. Hart
measured the noise level in LeBlanc's hearing zone to be 99-102 dBA when LeBlanc was
standing by press 2208 with presses 2208 and 2209, but not 2210, operating. When Melanson
was standing beside press 2209 with 2208 and 2209 running, the noise level was 100-104
dBA.[[25]] At certain times, Hart measured noise levels below 90 dBA for both LeBlanc and
Melanson with none of the presses running. Hart noted, however, that all three presses
were usually running during her inspection. Also, chief engineer Cochran testified that
Collier-Keyworth had no excess press capacity, indicating that the presses operated
full-time except when dies were changed or adjustments made.

As noted above, the dosimeter readings Hart obtained
for both LeBlanc and Melanson were approximately 300% of the permitted daily dose. An
employee exposed to a constant noise level of 100 dBA for 6 hours would receive a dosage
of 300%. The sound level meter readings made by Hart show that LeBlanc and Melanson were
sometimes exposed to noise levels less than 90 dBA, but were exposed to noise levels
around 100 dBA for most of their 8-hour shifts.

Thus, the sound level meter readings tend to show
that the dosimeters were accurately measuring and recording the noise levels to which
LeBlanc and Melanson were exposed.

In summary, we conclude that on April 14, 1980,
LeBlanc and Melanson, the operators of presses 2208, 2209, & 2210, were exposed to
noise in excess of the limits permitted by Table G-16.

Press 2216 (items 1a(C) and 1b(C)) is a 60-ton BIiss
press similar to the three presses discussed in the previous section but located some
distance from them. In April 18, 1980, industrial hygienist Hart attached a dosimeter to
press operator Alan Sund. After 202 minutes, or about 3.3 hours, the dosimeter read 167%.
Sound level meter readings made by Hart in Sund's hearing zone at four different times
over the period the dosimeter was operating were 103-107 dBA while the press was running.
Sund was then observing the machine from a distance of one to two feet. As with the other
presses, Hart testified that she timed the stroke rate and that the time between strokes
was less than one second. On June 17, 1980, Rockwell measured the stroke rate of the press
to be 45 strokes per minute, but there is no evidence that the press was producing the
same part, and thereby running at the same speed, as during Hart's inspection.

For much the same reasons as previously discussed, we
find that the evidence shows Sund was exposed to noise exceeding the standard's permitted
limits. Hart's testimony that the time between strokes was less than one second on the day
of the inspection establishes that the noise was continuous within the meaning of the
standard. The dosimeter reading facially indicates that Sund was exposed to excessive
noise. The sound level meter readings indicate that Sund was exposed to 103-107 dBA for
about 3.3 hours. At 103 dBA, the lowest sound level in this range, only about 1.3 hours of
exposure is permitted. Thus, the sound level meter readings corroborate the measurement of
excessive exposure made by the dosimeter.

Press 352 (items 1a(D) and 1b(D)) is a manually-fed
machine used for small bending jobs. On April 14, 1980, Hart obtained a reading of 256%
from a dosimeter attached to press operator Ernest Couture. According to Hart, the press
stroked once every two or three seconds. With the machine not operating, Hart measured the
noise in Couture's hearing zone to be from 84 to 88 dBA. With the machine operating, she
measured noise levels as high as 95 dBA.

We conclude that the Secretary did not prove Couture
was exposed to excessive noise. Because the press stroked only once every two to three
seconds, the noise that resulted from the press's operation was impulse noise, which is
not regulated by Table G-16. The only continuous noise to which Couture was exposed was
background noise measured to be 84 to 88 dBA, levels within the standard's limits for any
exposures time. We therefore find that Couture was not exposed to noise exceeding the
standard's limits and we will vacate items 1a(D) and 1(b)(D)).

Lathes 6040 and 3018 (items 1a(B)) and 1b(B)) are
used to cut metal tubing and chamfer the ends. Both lathes were operated by a single
employee, Ken Thompson. Lathe 6040 was operated manually, while lathe 3018 was automatic.
Therefore, when both lathes were operating, Thompson would be standing beside lathe 6040.
On April 14, 1980, a dosimeter attached to Thompson during his shift yielded a reading of
158.5%. According to sound level meter readings made by industrial hygienist Hart, the
noise level in Thompson's hearing zone ranged from 85 to 107 dBA.

Lathe 6040 was relatively quiet and, when operated
alone, produced noise levels below 90 dBA. Thus, the noise that caused the standard's
limits to be exceeded was generated by lathe 3018. That lathe sometimes emitted a very
loud high-pitched squeal, but at other times the squeal was missing. Hart recorded noise
levels of 107 dBA when the squeal was evident, but only 92 dBA with both lathes running
but no high-pitched squeal present. When Dr. Barry later visited Collier-Keyworth's plant,
he measured 118 dBA in the hearing zone of the operator of lathe 3018. Upon analyzing the
frequencies contained in the noise, Barry found that the excessive noise primarily
occurred at 8000 hertz and higher, frequencies which the human ear would perceive as very
high.

Both Barry and Rockwell made strip chart recordings
of the lathe noise. These recordings show very pronounced short bursts of noise, much like
the strip chart recordings for the power presses. The charts are not labeled with time
scales that would enable the time between bursts to be measured, but Rockwell testified
that the charts show a peak-to-peak separation greater than one second, indicating that at
least some of the noise is impulse noise.

We conclude that the Secretary failed to prove
Thompson was exposed to excessive noise. Some of the noise to which Thompson was exposed,
and which was detected by the dosimeter, was impulse noise that must be excluded under the
standard. There is no basis to conclude that Thompson's dosimeter reading would have shown
overexposure if the impulse noise had been excluded. Indeed, it is doubtful that the
Secretary proved overexposure even assuming that impulse noise were to be included.
Industrial hygienist Hart testified that Type 2 sound level meters and dosimeters, the
type she used in her inspection, generally have an inherent error factor of two dBA; when
that error is cumulated over time, it creates a potential error of 32% in the dosimeter
reading.[[26]] Thus, a dosimeter reading of 158.5%, as was obtained for Thompson, would
generally be outside the instrument's range of error and would show overexposure. However,
the record shows that the noise to which Thompson was exposed was of very high frequency,
predominantly 8000 hertz. At 8000 hertz, type 2 sound level meters and dosimeters have an
inherent error factor of 6.5 dBA.[[27]] This means that a dosimeter containing a type 2
sound level meter would have to read over 240% to show overexposure once the inherent
error of the instrument at 8,000 hertz is taken into account. Cf. 29 C.F.R. � 1910.95,
Appendix A, Table A-1 (hearing conservation standard)(entry for 96.3 dBA corresponds to
dose of 240 percent). Thus, where the noise is predominantly at 8000 hertz, a dosimeter
reading of 158.5%, the reading obtained for Thompson, does not show overexposure within
the accuracy limit of the instrument.

We therefore vacate citation items 1a(B) and (D), and
1b(B) and (D). Having found that OSHA proved over-exposure to quasi-continuous noise by
press operators LeBlanc, Melanson and Sund--the employees involved in items 1a(A) and (C),
and 1b(A) and (B)--we must now determine whether the Secretary showed that
Collier-Keyworth violated the standard by failing to take required precautions against
excessive noise exposure.

IV. Precautions Against Noise Exposure

The Secretary alleges in citation item 1a that
Collier-Keyworth violated section 1910.95(a) by failing to enforce the use of hearing
protection equipment. In item 1b, he alleges that the company violated section
1910.95(b)(1) by failing to implement feasible administrative or engineering controls to
reduce noise exposures.[[28]]

A. Personal Protective Equipment--Subsection
1910.95(a)

Citation items 1a(A) and (C) concern the alleged
failures by the three employees to wear personal hearing protectors. Industrial hygienist
Hart observed LeBlanc and Melanson wearing earmuffs and Sund wearing Swedish wool earplugs
during her inspection. Hart testified, however, that she also observed the employees
"at one time or another" not wearing hearing protection. The Secretary relies on
Hart's testimony that she observed the employees without hearing protection to support his
allegation under section 1910.95(a).

The evidence does not prove that Collier-Keyworth
violated section 1910. 95(a). Under the standard, employees may be exposed to noise levels
less than 115 dBA for some period of time without using protective equipment. Also, at
various times the three employees were subjected to noise levels less than 90 dBA, where
no protection is required regardless of the length of exposure. Thus, Hart's testimony
that she observed the employees at times without protection does not prove a violation of
the standard. If anything, the evidence that the employees were sometimes observed wearing
hearing protection tends to show that Collier- Keyworth complied with section 1910.95(a).
We therefore vacate citation items 1a(A) and (C).

B. Engineering Controls--Subsection 1910.95(b)(1)

The Secretary alleges in citation item 1b(A) and (C)
that Collier-Keyworth violated section 1910.95(b)(1) by failing to implement feasible
engineering controls on presses 2208, 2209, 2210, and 2216.[[29]] His expert witness on
engineering controls, Dr. Barry, proposed two basic noise reduction techniques he believed
were feasible: enclosing the die areas on the presses; and replacing the existing air
ejection nozzles with quieter ones.

Each press forms metal parts by the force exerted by
a rapidly descending die on sheet steel in the bed of the press. Barry testified that
there were two primary sources of noise associated with the operation of a press: (1)
noise from the impact of the die on the sheet metal stock; and (2) noise from the air
ejection nozzles used to expel the parts from the press bed. Barry recommended isolating
the press operators from the noise sources by installing acoustical enclosures around the
die areas of the presses.[[30]] He testified that the enclosures would have to be designed
"to minimize any operator encumbrances and allow for ease of die change." The
enclosures, in Barry's opinion, should be made of sheet metal, with transparent plastic
panels to permit visual observation of the die area. Doors and openings should be tightly
sealed, and the inside of the enclosure should be lined with acoustically absorptive
material to prevent sound buildup. Barry believed that each press enclosure would cost
$3,500 to $4,000 to design, fabricate, and install, and that a reduction in noise level of
10 dBA could be achieved. Barry had seen similar enclosures in use at another company's
plant and believed those enclosures achieved a 10 dBA reduction without inhibiting
production.

The enclosures Barry recommended would enclose the
air ejection nozzles as well as the die impact area, and would therefore reduce noise from
both sources. Barry also testified that the air noise could be reduced even if the die
area were not enclosed. Either quieter nozzles that were commercially available could
replace the existing nozzles, or the velocity of air from the existing nozzles could be
reduced to lessen the noise. Barry stated that the commercial nozzles would cost $5 to $20
each and would, in his opinion, produce a significant reduction in the noise level.

Curtis Holmer, a noise control engineer who testified
for Collier-Keyworth, did not believe that enclosing the die areas of the presses would
significantly reduce the impact noise emanating from the presses. Holmer testified that
enclosing the die area can significantly reduce the noise from a press only when the force
the press exerts is considerably below its capacity. However, when a press is used at or
near its capacity, as the presses were at Collier-Keyworth, then most of the energy of the
impact is transferred from the die area to the frame of the press, and most of the noise
the press produces radiates from its frame instead of from the point of impact. Holmer
therefore believed that enclosing the die area alone, as recommended by Barry, would not
reduce the noise resulting from the impact of the die on the stock. The enclosures would
reduce only the air ejection noise, and Holmer thought this noise reduction would be on
the order of 3 dBA or less.

Chief Engineer Cochran testified that die enclosures
would severely restrict the productivity of the presses by increasing the time necessary
for the operators to perform any duty requiring access to the die area. Cochran reviewed
the production reports prepared by the operators for a four-week period to determine how
often they needed access to the die area. Estimating that each instance of access would
require 30 extra minutes if the dies were enclosed, he calculated that the company would
have lost 72 hours of production over the four-week period, or about 950 hours for a
year.[[31]] Since the company had no excess press capacity, it would have to purchase a
new press to maintain its current level of productivity. According to Cochran, a new press
would cost $55,000. A plant addition to house the machine would cost an additional
$18,000, and a new operator would have to be hired to run the machine at a cost of $24,000
per year. Cochran also noted that the data on which he based these estimates might
understate the number of times the operators needed access to the dies because the
operators might not record situations that were now very easy to resolve, such as clearing
out a piece of scrap from the die area.

Regarding Barry's suggestion that the existing air
ejection nozzles be replaced with quieter ones, plant engineer English testified that the
company had tried using quieter nozzles in 1977 but that those nozzles reduced the air
velocity below what was needed to eject the parts from the press bed. Curtis explained
that quiet nozzles obtain their noise reduction by reducing the air velocity and by
spreading out the air stream over a larger area. However, for small parts such as
Collier-Keyworth made, much of the air from such nozzles would blow past the part and be
ineffective in moving it.

In Sherwin-Williams , 11 BNA OSHD at 2110,
1984-85 CCH OSHD at p. 34,702, the Commission stated what the Secretary must prove to show
a violation of section 1910.95(b)(1):

To prove a violation, therefore, the Secretary must
prove that proposed engineering and administrative controls are both technologically and
economically feasible. As the Ninth Circuit recognized in Castle & Cooke ,
"realism and common sense should dictate how the Secretary may meet his burden of
providing substantial evidence of feasibility." 692 F.2d at 650. After the Secretary
proves that controls are technologically feasible, the burden of producing evidence shifts
to the employer, who may raise the issue of economic feasibility and go forward with
evidence of the cost of controls and personal protective equipment. The burden of
producing evidence then returns to the Secretary, "who must establish that the
benefit of the proposed engineering controls justifies their relative cost in comparison
to other abatement methods." Id . The ultimate burden of persuasion on the
feasibility issue nevertheless remains with the Secretary.

The Secretary clearly did not prove the feasibility
of reducing the air ejection noise by installing quieter nozzles. Although quieter nozzles
are commercially available, Collier-Keyworth had tried such nozzles several years before
the alleged violation and found that the reduced velocity of air they produced provided
insufficient force to eject the parts from the die area. Thus, such nozzles would not
accomplish their intended purpose and would be technologically infeasible.

We also find that the Secretary failed to prove the
feasibility of die enclosures. The enclosures would produce a minimal benefit insufficient
to justify substantial cost.

The enclosures Barry recommended would surround the
die areas of the presses and isolate the operators from the noise produced by the impact
of the die on the stock and the noise from the air ejection nozzles. Barry believed that
such enclosures could reduce noise levels about 10 dBA. Collier-Keyworth's expert, Holmer,
testified that the enclosures would produce only about a 3 dBA reduction. In Holmer's
view, the die enclosures would only reduce the air ejection noise reaching the operators,
and not the noise from the die striking the stock.

We give considerable weight to Holmer's opinion.
Holmer had lengthy experience in the measurement and control of industrial noise and was a
member of the Institute of Noise Control Engineering, a national organization of competent
noise control professionals. He provided a reasoned explanation for his conclusion that
die enclosures on Collier-Keyworth's presses would not substantially reduce impact noise
and we find his testimony persuasive. Barry's opinion that enclosures could achieve a 10
dBA reduction is entitled to much less weight than HoImer's. Barry's training was
primarily in measuring the effects of noise on the human ear rather than the analysis and
reduction of industrial noise by engineering means. Barry had not designed or implemented
industrial noise controls and was not a member of the Institute of Noise Control
Engineering. Moreover, Barry's opinion was based only on his belief that die enclosures at
another company had achieved a 10 dBA reduction. Holmer's testimony indicates, however,
that the noise reduction produced by a die enclosure depends on the degree of capacity at
which the press is used. Thus, achievement of a 10 dBA reduction at another company, in
the absence of evidence of similar conditions, is a tenuous basis for believing that
Collier-Keyworth could obtain a comparable reduction. We find that the die enclosures
recommended by Barry would achieve a reduction of about 3 dBA in the noise levels to which
Collier-Keyworth's press operators were subjected.

A reduction of 3 dBA is significant ( see Continental
Can Co. , 76 OSAHRC 109/A2, 4 BNA OSHC 1541, 1543 n. 8, 1976-77 CCH OSHD � 21,009, p.
25,253 n. 8 (No. 3973, 1976)), but it must be balanced against the costs of engineering
controls and considered in light of other protective methods. See Sherwin-Williams ,
11 BNA OSHC at 2110, 1984-85 CCH OSHD at p. 34,702. The other method that we considered
here is the personal protective equipment that Collier-Keyworth's employees now wear. Dr.
Victor Hildyard, a medical doctor specializing in diagnosis and treatment of diseases of
the ear, testified that earplugs and earmuffs of the type used by Collier-Keyworth's
employees could reduce the noise reaching the employee's inner ear by 30 dBA if worn
properly. Even if loosely fitted, the equipment would reduce noise levels by 10-15 dBA.
Thus according to Dr. Hildyard, a noise reduction of 10 dBA is readily achievable even if
the performance of the equipment is far less than ideal. The highest time-weighted average
sound level to which Collier-Keyworth's press operators was exposed was about 98
dBA.[[32]] Thus, if engineering controls were not required at all, the personal protective
equipment would very likely meet Collier-Keyworth's obligation under section 1910.95(b) to
"reduce sound levels within the levels of the table." Yet, even if engineering
controls were installed, the employees would still have to wear the personal hearing
protectors because a 3 dBA reduction would not reduce noise to within Table G-16 limits.
They would simply not have to wear it for as long as they now must. Thus, installing die
enclosures would not eliminate the need for personal protective equipment but would only
decrease the amount of time employees must wear it.

The cost of the controls is, however, substantial.
According to Barry, each enclosure would cost $3,500 to $4,000 to design, fabricate, and
install. Moreover, to maintain its current level of production, Collier-Keyworth would
have to purchase a new press for $55,000, and would incur additional costs associated with
the new press, including a die enclosure for that press as well. The new press would add a
noise source to Collier-Keyworth's plant and could result in additional noise exposure for
some employees. On balance, we are not convinced that the benefits to be gained from the
enclosures justify these costs. We therefore find that the die enclosures recommended by
Barry were not shown to be feasible and will vacate citation items 1b(A) and (C).

Accordingly, we vacate citations item 1a(B) and (D),
and 1b(B) and (D), which involve employee exposure to impulse noise, unless the Secretary
requests an opportunity within 15 days of this decision to rebut officially-noted
documents with evidence of the Secretary's intent in 1969. We vacate items 1a(A) and (C)
because the Secretary failed to show that employees were not wearing personal hearing
protectors for the periods required. And we vacate items 1b(A) and (C) because the
Secretary failed to show that proposed engineering controls were feasible. The judge's
decision is therefore affirmed.

FOR THE COMMISSION

Ray H. Darling, Jr.

Executive Secretary

DATED: April 6, 1987

Appendix

1. The noise standard proposed by the ACGIH committee
   in 1968 stated in part:

Average
Sound Pressure

Levels of Octave Bands

with Center Frequencies

of 500, 1000, and 2000 Hz
A-Weighting

Network of

Sound Level

Meter

Duration of

Exposure per

Day

dB
dBA
Hours

85
92
4 - 8

90
97
2 - 4

95
102
1 - 2

100
107
less than

1 hour

These values apply to total time of exposure per working day regardless of whether this is
one continuous exposure or a number of short-term exposures but does [sic] not apply to
impact or impulsive type of noises.

When the daily noise exposure is composed of two or
more periods of noise exposure of different levels, their combined effect should be
considered, rather than the individual effect of each. If the sum of the following
fractions:

C 1 /T 1 + C 2 /T 2

• . . . C n /T n

exceeds unity, then, the mixed exposure should be
considered to exceed the threshold limit value[.] C 1 indicates the total time
of exposure at a specified noise level, and T 1 indicates the total time of
exposure permitted at that level.

*

*

*

The above limits do not apply to impose or impact type of noise. It is recommended
that exposure to this type of noise should not exceed 140 dB peak sound pressure level.

Appendix 1 of Herbert H. Jones, "ACGIH's
Proposed Threshold Limit Value for Noise," 29 Am. Indus. Hygiene J. 537-40 (Nov.-Dec.
1968)( ACGIH ).

2. A paper cited by both the Inter-Society Committee
   and the ACGIH Committee and that proposed criteria for steady noise concluded with the
   following observation:

We have purposely omitted any discussion of impulse noise up to now, for very sound
reasons. We know very little about the effects of impulse noise on the ear. We have just
started an intensive study of the relations between impulse noise and TTS [temporary
threshold shifts, a form of hearing loss].

The problem does not lend itself to easy solution,
for several reasons. Measuring impulse noise is difficult to begin with, but not nearly as
difficult to resolve as determining the effect the middle ear muscles have on the
transmission of impulsive noise across the middle ear. Limited laboratory research
indicates that the laws governing TTS from exposure to steady noise (non-impulse) do not
operate for impulsive noise exposure. Actually, this should not be surprising, since the
response of the basilar membrane to steady noise and to impulse noise should be completely
different.

For the present, we can only say that no one should
be habitually exposed to impulse noise of any considerable magnitude without the use of
hearing conservation measures.

*

*

*

Our very limited knowledge of the effects of impulse noise prevents any meaningful
discussion of impulsive-type noise exposure.

A. Glorig, W.D. Ward, & J. Nixon, "Damage
Risk Criteria and Noise-Induced Hearing Loss," 74 Archives of Otolaryngology
413, 423 (Oct. 1961), cited in I-S at 424 ref. 16; and ACGIH at 539
ref. 6 (article) and at 540 ref. 6 (proposed standard).

3. A later report by a subcommittee chaired by Dr.
   Glorig proposed criteria for hearing conservation with the following caveat:

Our knowledge of the relations of hearing loss to
noise-exposure permits us to propose guidelines for establishing standards for preventing
significant noise- induced hearing loss in the majority of exposed persons. These
recommended standards have been proposed by the International Organization for
Standardization. They are presented below in brief modified form. They do not apply to
exposure to impulsive noise but only to steady noise.

Subcommittee on Noise, Committee on Conservation of hearing, Am. Acad. of Ophtalmology and
Otolaryngology, Guide for Conservation of Hearing in Noise 13 (1964)(emphasis in
original), cited in I-S at 424 ref. 9.

4. A 1966 publication of the American Industrial
   Hygiene Association reviewed the criteria for recommended noise limits that had been
   proposed by various investigators up to that time. While noting that there was little
   disagreement as to the levels of continuous noise that were harmful, the publication went
   on to say:

Much more difficult will be the selection of
mandatory protection levels for intermittent, short duration, and impact noises . . . .
Because there is still much to be learned about the relationships between temporary and
permanent threshold shifts the criteria for brief exposures cannot be expected to have the
same reliability as those for continuous exposures. Until more is known about short-time
exposures, the criteria should be treated as temporary benchmarks and be used with
considerable judgment.

Am. Indus. Hygiene Ass'n., Industrial Noise Manual ,
ch. 7, at 59-60 (2d ed. 1966), cited in I-S at 424 refs. 3 & 13; and ACGIH
at 540 ref. 3 (proposed standard).

5. A 1967 paper by Coles et al. summarized studies
   that had tried to assess damage risk from impulse noise exposure. R. Coles, G. Garinther,
   D. Hodge, & C. Rice, Criteria for Assessing Hearing Damage Risk From Impulse-Noise
   Exposure , U. S. Army Tech. Memo. 13- 67 (Aug. 1961) , cited in ACGIH at
   539 ref. 12 (article).[[33]] These studies tended to seek permissible exposure criteria in
   terms of factors unique to impulse noise. Several of the studies demonstrated that
   exposure to 25-100 impulses in the 140-170 dB range at 6-30 impulses per minute could
   produce temporary threshold shifts ("TTS")[[34]] in a significant number of
   exposed persons. The variations in the TTS's of people exposed to such impulse noise was
   greater than among subjects exposed to steady-state noise, indicating that some people may
   be particularly susceptible to impulse noise. Responses were observed to vary depending on
   whether the noise reached the ear at normal or grazing incidence; in same instances, for
   example, noise from a gun fired by a person near the subject, which would reach the
   subject's ear at normal incidence, appeared more harmful than noise from a gun fired by a
   subject himself, which would reach his ear at grazing incidence. TTS's were greater for
   longer duration impulses of similar peak intensity. a result probably attributable to the
   greater energy associated with the longer impulses. The paper drew from the studies
   criteria for peak pressure level and duration limits that would not produce an excessive
   risk of hearing loss. It concluded that impulses in the 155-160 dB peak intensity range
   were safe for a duration of about 10 milliseconds, with higher intensity impulses
   permissible for shorter duration and lower intensity impulses permissible at longer
   durations.

6. In a slightly later paper, the same authors noted
   that it was customary in proposing damage risk criteria for steady-state noise to exclude
   impulse by implication rather than by direct statement. R. Coles, G. Garinther, D. Hodge,
   and C. Rice, "Hazardous Exposure to Impulse Noise," 43 J. Acous. Soc. 336
   (1968), cited in ACGIH at 539 ref. 11 (article). After discussing risk
   factors associated with common amounts of exposure to gunfire, the paper said, "the
   more industrial types of impulse-noise exposure as in riveting, pile driving. drop
   forging, etc., may require a separate criterion or may be treatable by steady-state noise
   criteria." Id . at 339. However, the authors concluded:

[w]hether a more general continuity between impulse
and steady-state noise criteria can ever be achieved seems very uncertain in view of the
evident differences in their ranges of TTS and of opinion such as that of Kryter and
Garinther that PTS from impulse noise may follow a different pattern from PTS due to
steady-state noise.

Id. at 343 (reference omitted).

7. Investigators attempted to determine the combined
   effect of impulse noise and steady-state noise in only one cited study. A. Cohen, B.
   Kylin, and P. LaBenz, Temporary Threshold Shifts in Hearing From Exposure to Combined
   lmpact/Steady-State Noise Conditions , 40 J. Acous. Soc. 1371 (1966), cited in ACGIH
   at 539 ref. 10 (article). An experiment was conducted on 15 subjects. who were exposed to
   a variety of combinations of impulse and steady-state noise. The study found that, under
   some circumstances, impulse noise in the presence of noise was less harmful to hearing
   than the impulse noise alone. The authors concluded that this was due to the
   "acoustic reflex," which is a contraction of the middle ear muscles that
   attenuate, the amount of sound transmitted to the inner ear. If the acoustic reflex is
   activated before exposure to a loud impulse, the inner ear receives protection from the
   impulse. Therefore, when the subjects were exposed to steady noise that was loud enough to
   activate the acoustic reflex, the impulse noise was not as damaging as it otherwise would
   have been.

8. The standard proposed by the Department of Labor
   under the Walsh- Healey Act on September 20, 1968, at 33 Fed. Reg. 14259-60, stated:

� 50-204.10 Occupational noise exposure .

(a) The maximum permissible steady (or equivalent)
noise level in the working environment shall not exceed 85 decibels. Every employer shall
utilize every feasible engineering method to control noise levels. Such methods of control
include reducing the amount of noise produced at the source, reducing the amount
transmitted through the air, and substituting quieter procedures or machinery.

(b) Where the noise is not steady the equivalent steady noise level is determined by
the following procedure: The duration over 1 week of each clearly distinguishable sound
level is located in column 1 of Table I following this paragraph (b) and the partial noise
exposure is read at the intersection of this row with the appropriate sound level column.
The partial noise exposures thus obtained are added arithmetically. The sum is the
composite noise exposure. The continuous noise exposure equivalent to the composite noise
exposure is then read from Table II next following Table I. If the variations in noise
level involve maxima at intervals of 1 second or less it is to be considered steady. If
the intervals are more than one second and the duration of the maxima are less than 1
second each, maximum is to be considered as 1 second.

THE PICTORIAL EXHIBIT REFERRED TO ABOVE IS NOT AVAILABLE IN THIS FORMAT.  PLEASE
TELEPHONE THE REVIEW COMMISSION PUBLIC INFORMATION OFFICE, 202-606-5398, TO REQUEST A
PAPER COPY; TTY: 202-606-5386; FAX: 202-606-5050; E-MAIL: [email protected]

SECRETARY OF LABOR,

Complainant,

v.

COLLIER-KEYWORTH CO.,

Respondent.

OSHRC DOCKET NO. 80-2848

APPEARANCES:

Robert Yetman, Esq., for Complainant

Douglas B. M. Ehlke, Esq., for Respondent

DECISION AND ORDER

This is a proceeding pursuant to the Occupational
Safety & Health Act of 1970, as amended (29 USC, sec. 651 et seq .)
hereinafter called the Act. The Complainant alleges that the Respondent has violated sec.
5(a)(2) of the Act (sec. 654) by not complying with the Occupational Safety & Health
standards.

The Respondent is a corporation engaged in the
business of manufacturing chair mechanisms and its business affects the commerce of the
United States.

The Respondent's worksite at Gardner, Massachusetts,
was inspected by the Occupational Safety & Health Administration (hereinafter called
OSHA) on or about April 9 - 18, 1980.

On or about April 28, 1980, the following Citation,
together with Notice of Proposed Penalty, was issued against the Respondent: Citation #1,
Items #1a - c, the serious violation of the standards at 29 CFR 1910.95(a), .95(b)(1), and
.95 (b)(3); respectively Items 1a, b, and c of Citation #1.

On May 16, 1980, the Respondent filed Notice of
Contest to Items #1a - c of Citation #1 and the penalty proposed therefor.

In his post-hearing brief, the Complainant withdrew
item 1(c).

The pertinent words of the standards and the Act are
appended to this decision under appropriate titles. The basic issue is whether or not the
Respondent's employees were exposed to noise exceeding the limits of Table G-16 of the
standard at 29 CFR 1910.95(a).

SINCE THE HEARINGS WERE HELD ON MARCH 23 - 27, JUNE
16 - 19, DECEMBER 8 - 11 [ALL IN 1981] AND MARCH 23 and 24, 1982, THE 1981 TRANSCRIPT WILL
BE REFERRED TO BY MONTH AND PAGE NUMBER [E.G., MARCH 225] AND THE 1982 TRANSCRIPT BY YEAR,
MONTH, AND PAGE NUMBER BECAUSE SOME OF THE TRANSCRIPT PAGES ARE NUMBERED THE SAME EVEN
THOUGH THEY COVER DIFFERENT DATES.

It was a very long and hard-fought case, involving
many scientific treatises and expert testimony. Both attorneys are to be commended for
their excellent preparation and the zeal with which they tried the case.

PRELIMINARY

Unless otherwise indicated, the word
"dosimeter" [as used herein] refers to the Gen Rad 1954 type 2 instrument used
in the April inspection of the Respondent's jobsite. The dosimeter has an integrator that
"counts" the sound waves that constitute noise; and it has a 5dB rate of
exchange as distinguished from a 3dB one. A "decibel" is roughly understood as a
unit for expressing the relative intensity of sound. The integrator may be
"slow" [a time constant of 1 second] or "fast" [a time constant of
1/8th of a second]. The slow integrator is considered to be less accurate than the fast in
following the quick noise...Tr. June 1147-1150.

The jobsite was inspected in April of 1980. The
machines cited were punch presses [#'s 2208, 2209, 2210 on the second floor annex, #0352
on the 4th floor, and #2216 on the 2nd floor of the steel warehouse] and two lathes [#6040
and #3E 3018, both in the torsion department]. The sound levels were obtained through the
use of a dosimeter operated by or under the supervision of OSHA Industrial Hygienist Anne
Hart.

In measuring sound for the purposes of Table G-16 of
the standard at 29 CFR 1910.95(a) - (b)(3), a dosimeter at slow response with a 5dB rate
of exchange is used because the standard mandates a "slow" response. The
evidence indicated that a different race of exchange [e.g., 3dB] would not qualify as a
"slow" response; but, on the other hand, a different rate of exchange might be
more accurate for certain types of sound. One of the issues in the case was whether the
5dB rate of exchange would disclose the true level of sound if fluctuating or impulse
[impact] noise were also present. "Exchange rate" is the relationship between
noise level and duration. To illustrate: a 5dB exchange rate means that the sound level
can be increased by 5dB with every halving of exposure duration.

There are two sets of dates on which sound
measurements were taken by the Complainant: those in April, to which OSHC Compliance
Officers Goyda and Hart testified; and those in September, to which sound expert Barry
testified. The witness Rockwell took sound measurements in February of 1981; and the
witnesses Jones, Kamperman, Holmer, Kundert, and Botsford apparently did not participate
in actual measurements of the cited machines but testified as sound experts.

IMPULSE NOISE

The case required considerable evidence of a highly
technical nature involving sound level meters, dosimeters, 3bB and 5dB rates of exchange,
various kinds of noise, etc. A basic issue concerned impulse noise.

Several questions about impulse [or impact] noise
must be resolved: What is it? Did it exist in the cited machines? Was it accurately
measured by the dosimeters used in the inspection? Should it have been included or
excluded in the dosimeter readout to determine compliance with Table G-16?

TESTIMONY

Goyda

The Witness Goyda, who is an experienced and
competent Compliance Officer, was a member of the OSHA inspection party. He monitored the
dosimeter testing in April of the employee, Couture . . . Exh. C-1. I was impressed with
his competence and honesty. I find that he accurately reported the readout registered by
the dosimeter, as recorded in Exh. C-1.

Hart

The witness Hart, who is an experienced and competent
Industrial Hygienist, was in charge of the April inspection of the jobsite. She calibrated
the dosimeters used, directed the placing of dosimeters on employees, and recorded or
supervised the recording of their readouts. . .Tr. June 672 - 685. I was impressed with
her competence and honesty. I find that she followed proper procedures in testing and that
she accurately reported the readouts registered by the dosimeters, as recorded in Exh. C1

• 5.

She testified that the testing instrument was a Gen
Rad 1954 Type 2 dosimeter, slow response, A weighted. . .Tr. June 730, 733. She said that
impulse noise is not supposed to be included in the dosimeter measurement because the
reading will be distorted and will inaccurately register higher. . .Tr. June 768-770. She
also testified that there was impulse noise from press #0352. . .Tr. June 715, 716.

Exh. C-1 - 5 reflect that the sound levels recorded
during the April inspection ranged from 94 to 101.3dBA as related to Table G-16.

Barry

The expert Barry, who was called by the Complainant,
was well-qualified. . .Tr. March 319 - 324, Exh. C-20. He testified that the effects of
impulse noise are not supposed to be excluded; and that Table G-16 has no relation to
impulse noise and does not exclude it. . .Tr. March 512, 513, June 913, 914, December 902.
He emphatically stated that the findings of Compliance Officer Hart were reliable with
correction factors that would not change her conclusions; and that the type dosimeter used
in her inspection was a suitable instrument for measuring noise levels to determine OSHA
compliance. . .Tr. June 952 - 958. He concluded that, even applying error factors,
employees were still exposed to excessive noise. . .Tr. December 984.

However, Barry also testified that the dosimeter has
been a controversial instrument in recent years. . .Tr. March 517, 518. He said that
dosimeters tend to register higher readings than actual noise levels; and that, for the
sounds typical here, the reading could be higher by 4 or 8 decibels. . .Tr. June 953,
December 939, Exh. R-60. He also said that the more impulse and the higher, the greater is
the dosimeter's error. . .Tr. June 1008. He testified that a dosimeter's readout might be
inaccurate due to the slow response setting required by the standard because, where there
are both steady and impulse noises, the dosimeter distorts on the high side; and that the
use of both fast and slow modes of response would be better testing procedure. . .Tr. June
925 - 928, 933, December 909, 910, 917, 948. He suggested that, "to measure it really
properly, the standard would have to be rewritten.". . .Tr. June 927. It had been his
opinion [as he wrote in the OSHA Field Officer's Manual] that, where both continuous and
impulse noise are present, the proper way to determine compliance with Table G-16 is to
use the fast mode setting and record the reading between the impulses. . .Tr. June 922,
923, 1015 - 1019. He had also authored an article in which he stated that the standard at
29 CFR 1910.95 "sets the permissible exposure level for non-impulse noise.". .
.Tr. June 920. He acknowledged that he had said that Table G-16 was intended primarily for
non-impulse noise. . .Tr. December 904, 906, Exh. R-59, R-62.

As concerns the 3dB and 5dB rates of exchange, he
agreed that the readout would differ depending on which was used, resulting in a possible
dosimeter overstatement inaccuracy of 100 - 200 percent. . .Tr. June 934 - 937. He had
also recommended a change in the dosimeter from a 5dB exchange rate to the use of a 3dB
one "and making no differentiation between the type of noises. In other words,
integrate all noise as noise. And, use that for determining compliance or
non-compliance.". . .Tr. June 943. While that suggestion could be explained as
referring to possible ambiguities in the standard, it could also be interpreted as a
belief that the standard does exclude some noise; otherwise, why the need to specify that
all. noise is included?

Barry's testimony also indicated that there was a
discrepancy of several decibels between the slow and fast modes when he measured the
machines in question in his September visit at the Respondent's jobsite. . .Tr. June 975 -
999, Exh. C-22. [On this point, the testimony is germane to the difference between the
slow and fast modes, if not on the merits of the citation itself.] He also testified that,
in that same visit, he found impulsive components of noise in all the cited machines
except one lathe. . .Tr. June 913, 955.

Rockwell

The expert Rockwell, who was called by the
Respondent, was well- qualified. . .Tr. June 1027 - 1031, 1131 - 1146, Exh. R-22. He
testified that impulse noise [which is less than 1 per second and fewer than 60 per
minute] should be handled separately; and that it has been OSHA policy since 1972 to
exclude or edit out such noise in testing. He also quoted an OSHA regional office and
Chapter 4 of the OSHA Field Manual to the same effect. . .Tr. June 1150 - 1154, 1291. He
said that the type of dosimeter used in the April inspection erroneously read out several
decibels higher because of impulse noise that should have been excluded. . .Tr. June 1159,
1160. He pointed out that recognized authorities took the same position. . .Tr. June 1161

• 1169, Exh. R-23. He also testified about tests he had conducted in the Respondent's
  plant in February, 1981, using both slow and fast response dosimeters. The slow response
  one [the type used in the instant case] registered 5 - 14 decibels higher than the fast. .
  .Tr. June 1174 - 1207, Exh. R-25 - 33. He concluded that the dosimeter overestimates
  exposure in "this" type of environment [the Respondent's jobsite] by 10 - 12
  decibels over the actual exposure level. . .Tr. June 1195 - 1202. He said that the use of
  a 5dB rate of exchange dosimeter resulted in inflated exposure readings, grossly
  overestimated the true noise, and was an inaccurate instrument for impulsive or rapidly
  fluctuating noise. . .Tr. June 1200 - 1202, 1209, 1234, 1236. He specified errors of 10 -
  15 decibels at each location that he tested. . .Tr. June 1287, Exh. R-33. It was his
  opinion that the instrument used could not substantiate the allegations, and that impact
  noise levels should be measured only with an impact meter or an oscilloscope. . .Tr. June
  1252, 1296, Exh. R-35.

Holmer

The expert Holmer, who was called by the Respondent,
was well-qualified. . .Tr. June 1298 - 1310, 1479 - 1481, Exh. R-36. He testified that
impulse noise should be eliminated in determining compliance with the G-16 table. . .Tr.
June 1456, 1457, 1515. He said that, because of the difference between the 3dB and the 5dB
exchange rates, the Gen Rad 1954 dosimeter used in the inspection would read higher than
the actual noise level. . .Tr. June 1447, 1455, 1466, 1467. He said that such dosimeters
invariably overestimate signals fluctuating at the rate of 1 second or less; and that no
dosimeter accurately records impulse noise in the Table G-16 sense because the standard
requires a 5dB exchange rate. . .Tr. June 1441 - 1443, 1456. He also said that, although
dosimeters read lower than theory would project, and a Gen Rad 1954 dosimeter does not
overestimate as badly as an "ideal" dosimeter would, it still overestimates by a
factor of 10 whereas an "ideal" one would overestimate by 12. . .Tr. June 1467,
1471.

Kundert

The expert, Kundert, who was called by the
Complainant, was well-qualified. . .Tr. December 778 - 780, Exh. C-27, C-28. He testified
that the Gen Rad 1954 dosimeter was designed specifically for testing against the OSHA
standard, is excellent for that purpose, and can process short signals even though most
dosimeters cannot. Although he said it was designed to measure both continuous and impulse
noise, he acknowledged that it cannot accurately measure very short impulses but tends to
produce low readings. . .Tr. December 795 - 797, 871 - 874. Concerning the 3dB and 5dB
rates of exchange, he testified that the dosimeter responds to impulsive sound as a 5dB
instrument...Tr. December 863, 864. Concerning the question of whether Table G-16 is
interpreted to include or exclude impulse sound, he testified that the OSHA field guide
says to exclude it its did ANSI S-1.25. He concluded that, if impulse noise is to be
excluded, the way to do it is by reading between the pulses, as the witness Rockwell
did...Tr. December 888 - 890. He also said that the authorities cited in Exh. R-23 did not
support Rockwell's position but, on the contrary, disagreed with it. . .Tr. December 798 -

803. Recalled by the Complainant in rebuttal, he said that, for the type of impulsive
     noise here, the sound level meter and dosimeter give almost identical results; and that
     his in-court demonstration with the dosimeter anti sound level meter (measuring a
     continuous sequence of impulses) produced the same results. . .Tr. 1982 March 1728 - 1730,
     1871 - 1876.

Botsford

The well-qualified expert, Botsford, who was called
by the Respondent, testified that the Department of Labor had recommended the exclusion of
impact noise from Table. G-16, and impact noise was not included in Table G-16. . .Tr.
December 1038 - 1040, Exh. R-66.

His testimony tended to establish that impact noise
could not be accurately measured by a sound level meter but should be measured by an
impact meter against a limit of 140 decibels; while continuous noise should be measured
against Table G-16, using a time-weighted instrument with a 5 decibel exchange rate and
then reading between the impact peaks. . .Tr. December 1076 - 1080. He said that the OSHA
field manual recommends this procedure. . .Tr. December 1077, Exh. R-76.

Hildyard

The Respondent, seemed to place great reliance on the
testimony of Dr. Hildyard to the effect that, although the advisory committee on OSHA
standards made a final recommendation of priority for administrative and engineering
controls over personal protective devices, the committee members had actually voted not to
accord priority to administrative and engineering controls over personal protective
devices [such as ear plugs]...Tr. March 565, 583. However, even if there had been no
objection to such testimony [and the Complainant did object to it], it would be given no
weight because it violates the parole evidence rule [See Evidence �1022, 1027 of 30 Am.
Jur. 2nd; sec. 77.5 of Federal Trial Handbook; and Del Prete v. Board of
Selectmen , 351 Mass. 344]. The official record of the advisory committee on OSHA
standards which said administrative and engineering controls must be tried before personal
protective equipment -- governs and not evidence tending to contradict it. Accordingly, I
am limiting Dr. Hildyard's testimony about the committee vote to whatever bearing it may
have on the Respondent's good faith.

Jones

The expert Jones, who was called in rebuttal by the
Complainant, was a well-qualified industrial noise consultant. He testified that the sound
level meter measures noise in the slow response better than the fast and that there is no
basis for editing out impulse noise. . .Tr. 1982 March 1629 - 1639. However, he also
stated that he agreed with the statement that only steady noise is to be measured against
Table G-16. . .Tr. 1982 March 1695. Although his name also appeared on a document
apparently stating that Table G-16 values do not apply to impulse noise, he testified that
such use was unauthorized and did not represent his opinion. . .Tr. 1982 March 1651 -
1658, 1668 - 1670. Because of his sworn testimony to that effect, I do not place any
reliability on that document as far as it concerns any opinion by him.

Jones also testified that Bosford had written an
article in which he said impulse noise should be included in measurements of industrial
noise. . .Tr. 1982 March 1637, 1638.

Kamperman

The expert Kamperman, who was called in rebuttal by
the Complainant, was a well-qualified industrial noise expert. He testified that the
dosimeter is a very accurate Instrument for measuring noise in an industrial setting; that
it always underestimates noise; and he had found no dosimeter that read high. . .Tr. 1982
March 1902, 1947, 1984. He also said that impulse noise should not be excluded from
measurements and that industry custom is to include it in computing noise for purposes of
Table G-16. . .Tr. 1982 March 1913 - 1917. He pointed out that if impulse noises are
edited out of a punching operation, there would be no noise left because they are all
impulsive signals. . .Tr. 1982 March 1898, 1911. He said that his field testing of the
1954 Gen Rad dosimeter and a sound level meter produced approximately the same results,
whether with or without impulse noise. . .Tr. 1982 March 1900, 1901. He also said that the
fast response is not as good as the slow in determining noise per Table G-16. . .Tr. 1982
March 1902. He concluded that he had never said that impulse noise should only be measured
in accordance with the 140dB peak, and had never said that impulse noise should be
excluded from measurements. . .Tr. 1982 March 1984, 1985.

However, when he was asked if he had said that
"it is not possible for a simple dosimeter to correctly compute the noise dose using
a 5dB exchange rate for noises that vary in levels of less than 10 seconds," his
answer included statements that "problems come because of the OSHA 5dB exchange rate.
. .because the 5dB exchange rate violates basic laws of physics". . . Tr. 1982 March
1960-1966. He also testified that the dosimeter "has a hard time correctly measuring
anything other than steady random noise" and that he is "on the committee to
revise standard now to make dosimeters capable of correctly measuring impulsive
sound.". . .Tr. 1982 March 1945. He also conceded that there is "confusion"
over the measurement of impulsive noise in the industrial environment and that he had
written an article stating that the standard only addresses the measurement of continuous
noise. . .Tr. 1982 March 1948, 1949, Ex. R-110. He also testified that he had made a
written statement that, for impulsive noise, "the dosimeter will always indicate a
much higher dose that the sound level meter" and that the disparity would be 100% or
a difference of 5dB. . .Tr. 1982 March 1953, 1954, 1968, 1970. While he also testified
that this was an incorrect interpretation of his view, he nevertheless conceded that he
had told a seminar that "dosimeters used in the noise environment typically found in
metal fabricating facilities would read between 5 and 13dB too high, two to six times the
expected noise dose relative to the OSHA 5 decibel exchange rate,". . .Tr. 1982 March
1953 - 1957, 1963-1970.

DISCUSSION

With very minor differences, the experts all seemed
to be in general agreement that impulse noise is a sound of brief duration that occurs
less frequently than one a second and less than 60 a minute; and includes such examples as
hand clapping, dropping a book, firing a rifle, the ram of a punch press, etc. It is also
clear that all the cited machines [with the possible exception of the lathes had impulsive
components in the noise. . .Tr. March 311, June 913, 921, 955, 956, 1150, 1251.

As indicated above in this decision, the experts
differed in their opinions on the accuracy of the measurements by the dosimeter. Rockwell,
Holmer, and Botsford were firm in their conclusions that the dosimeter had to be
inaccurate because of the standard's requirements that only the slow response may be used.

Although Barry, Kundert, Jones and Kamperman
concluded that the dosimeter readouts were accurate, they did make many concessions
[specified above] that tended to establish there were distortions by the dosimeters used
in the April inspection.

In my opinion, the weight of the evidence clearly
establishes that the dosimeters used were certain to overstate the employees' exposure to
noise. I so find.

The most troublesome question is whether impulse
noise should be included or excluded from the dosimeter readout. The standard itself
merely uses the words "noise" or "sound" except for the word
"continuous" in paragraph (b)(2). Table G-16, with figures going only as high as
115, has a footnote: "Exposure to impulsive or impact noise should not exceed 140dB
peak sound pressure level". The standard at .95(b) 1 - 3 clearly mandates that sound
levels be "measured on the A scale. . .at slow response". Table G-16 also
specifies: "Sound Level dBA slow response". The Compliance Officer has no
choice: the standard must be obeyed. The dilemma is that the standard mandates the use of
the slow response mode of testing, but the evidence establishes beyond doubt that, when
impulsive noise is present, the slow response mode results in an inaccurate readout in
excess of the true noise level,

Even if the use of the slow response dosimeter
results in distorted readings, is the question of compliance to be based on such figures?
It is hard to believe that any branch of government would intend such a result. Certainly
a judicial forum could not condone any decision based on inaccurate facts.

The Complainant has the burden of proving that the
Respondent's employees were exposed to noise that exceeded the amounts specified in Table
G-16. In my opinion, that calls for the Complainant to prove that there were actual and
true noise levels in the amounts stated in Table G-16, and not merely readout figures
provided by the mandated slow response dosimeter. I find that that is the Complainant's
burden of proof even if it means that the Compliance Officer has to use two sets of
instruments: one to comply with tile mandate of the OSHA standard and a second to satisfy
the demands of proof and due process of law.

A third alternative, of course, is for some
modification of the standard (or Table G-16) in conformity with the most acceptable
technology available [a need recognized by the Complainant's very experienced expert,
Barry. . .Tr. June 927]. I strongly recommend such modification because it is a scandalous
situation when the various entities concerned with safe working conditions OSHA, the
employer and employees, the Complainant and the Respondent, the attorneys, and the
judicial forum itself -- are all at the mores of a system that practically guarantees
controversy and uncertainty because even the experts themselves are in almost total
disagreement.

If the present state of the law requires the use of
instruments in addition to any that are mandated by the standard, or calls for measures to
correct the readout figures, or makes the Complainant's task more difficult in any way,
that is unfortunate; but the Complainant has the burden of proving its allegation.

On the question of the accuracy of the slow response
dosimeter when impel - impart noise is present at the jobsite, there may be a conflict of
testimony between the experts for the Complainant and the experts for the Respondent.
Assuring that all are equally honest and equally well-qualified [and 1 that them so to be]
has the Complainant carried his burden of proof?

I find that the dosimeter's accuracy has not been
established; in fact, quite to the contrary, its inaccuracy was proven. The testimony of
the Respondent's experts to that effect was not shaken; and the Complainant's experts
[Barry, Jones, Kundert, and Kamperman] also made several statements tending to the same
conclusion. The inspection officer herself (Hart) also agreed that the dosimeter readout
would be inaccurate when impulse noise was included. In that state of the evidence, I find
that the dosimeter's accuracy was not proven.

On the question of whether to include or exclude
impulse noise in determining compliance with Table G-16: I find it should be excluded. In
the first place, its inclusion was the basic cause of the dosimeter's inaccuracy in the
instant case. Secondly, I find that it was the customary practice for OSHA to exclude
impulse noise in determining the possibility of noise exceeding the figures in Table G-16.
The uncontradicted testimony of the expert Rockwell was that it has been OSHA policy since
1972 to exclude such noise; and that Chapter 4 of the OSHA Field Manual is to the same
effect. . .Tr. June 1150 - 1154, 1291. Hart's testimony that impulse noise was not
supposed to be included corroborates that conclusion.

The Respondent has a constitutional right to be
treated the same as any other employer cited under the same standard. If the customary
practice of OSHA was to exclude such impulse noise, it should have been excluded in the
instant case.

Lastly, I find that any reasonable interpretation of
Table G-16 leads to the conclusion that impulse noise must be excluded. An analysis of the
standard can lead to no other conclusion. We begin with the assumption that it intended to
have a workable standard with accurate figures in Table- G-16 -- and the standard should
be interpreted to achieve those ends. The standard clearly and explicitly states that the
slow response mode must be used -- but that mode cannot accurately measure impulse noise.
The standard is so ambiguous about impulse noise that (in my opinion) reasonable persons
would agree that two interpretations are possible: (1) that impulse noise should be
included in the dosimeter readout; or (2) it should be excluded. The first interpretation
[including impulse noise] results in an inaccurate readout. Consequently, the more
reasonable conclusion is to exclude impulse noise since that is the only interpretation
that gives accurate figures in the readout while still complying with the slow response
mode, as mandated by the standard. That is the only interpretation that gives a workable
standard with accurate figures in Table G-16.

In the final analysis, the Complainant has not proven
its case whether impulse noise is supposed to be included or excluded. If it should have
been excluded, that was not done here; and if it should have been included, it results in
an inaccurate readout. In either event, the dosimeter used was not a reliable instrument
on which to base the citation.

The difference between the true sound level and the
dosimeter's inaccurate readout was not minor but amounted to at least several decibels.
When the dosimeter's readout ranged from 94 to 101.3 decibels, an error of several
decibels could easily bring the true figures to those in compliance with Table G-16. While
an error of several decibels might not be of any significance if the dosimeter readout
were greatly in excess of the figures in Table G-16, it becomes of paramount importance
when [as here] the readout ranges from 94 to 101.3 decibels.

Where the noise is so excessive that an error of
several decibels in the dosimeter is comparatively trivial, the inaccuracy becomes
immaterial. For example, in the Frank Nutty case [5 OSHC 1727], where the recorded
noise was 8 times more than that permitted in Table G-16, the Review Commission affirmed
without review the decision of the Administrative Law Judge that the Respondent had not
complied with the noise standard concerned. [Although the dosimeter was held to be
adequate in that case, I find it easily distinguishable from the instant case]. Here,
where the alleged excessive noise is comparatively slight and the dosimeter error
comparatively great, it must be found that the Complainant has not proven excessive noise.

Offers of Proof by the Complainant

Even assuming the admissibility of the evidence in
the Complainant's various offers of proof, the quantity and quality of evidence casting
doubt on the reliability of the dosimeter is not diminished. If the proffered evidence had
been admitted, it would not change my findings.

Barry's testimony of September inspection

In seeking to establish the admissibility of Barry's
testimony concerning his September, 1980, visit to the jobsite, the Complainant offered
the testimony of Compliance Officer Goyda that the conditions prevailing at that time were
similar to those of the official inspection in April, 1980. The substance of Goyda's
direct examination was that the machines and the parts produced appeared to be the same on
both dates and he observed no difference in them. . .Tr. March 29 - 31, 299 - 302. He took
no measurements but described the parts in fairly general terms. He testified that he
heard the noise emitted by the machines and it was "the same" on both dates. .
.Tr. March 303.

Cochran, the Respondent's chief engineer, testified
that the punch presses used some 20 - 30 different dies of varying weights, shapes, and
thicknesses in producing 30 different components or parts. The weights vary from 200 to
one thousand pounds in sizes ranging from 8 x 10 inches to 22 x 16 inches. . .Tr. March
103 - 108, 122. He said the manner in which the machines are operated changes every day;
and the number of parts and the length of the run vary with the particular job. . .Tr.
March 112 - 114, 171. That is also true of the machine's number of strokes per minute. .
.Tr. March 130, 149 - 151. He testified that the parts made by machines 2208 - 2210
"are all different" and that the two parts made on September 15th [the date of
Barry's visit] but were "4 absolutely different parts". . .Tr. March 287. He
explained that, although the machines themselves were identical, the operation was not
"because it had a different tool in it and every tool runs differently. . .not the
same steel strip.. a different width and fed a different length.". . .Tr. March 289.
He also pointed out that there would be "a wide variation" in the noise because
the gauge of the material is "one factor that is involved in producing noise when
that die hits the stock". . .Tr. March 291, 292. He also testified that there were
different sounds on the lathe on some days. . .Tr. March 128.

Bergeron, the Respondent's assistant director of
manufacturing, testified that dies may or may not be changed frequently because it is a
job shop. . .Tr. March 226 - 229.

There was no contradiction of the testimony of
Cochran and Bergeron on the differences in dies, parts, strokes, and length of run.
Moreover, I was favorably impressed by the candor and honesty of both witnesses; and both
certainly were in positions to know about the operations they described.

It was ruled that, on the basis of Goyda's testimony
that conditions during the April and September visits were similar, Barry's testimony
about his September visit was technically admissible; but both parties were alerted to the
fact that it would be given very little weight. . .Tr. March 303 - 307. My opinion was
[and is] that in a matter as sensitive and delicate as the determination of decibels of
noise, more is required than Goyda's mere eye and ear observation of the appearance and
sound of machines and parts. The testimony of Barry about his September visit was admitted
because, technically, Goyda's testimony furnished a foundation of similarity. I have given
it very little weight because, substantively, Cochran's uncontradicted testimony
established that the difference in dies, parts, strokes, and lengths of run may well have
resulted in different noise levels.

REVIEW COMMISSION

In view of my findings concerning the inaccuracy of
the dosimeter, I have not reached items 1(b) and 1(c) of Citation #1. I purposely have not
made alternative findings because it seems to me to be of over-riding importance for the
Review Commission to clear up the uncertainties arising out of the possible
interpretations to which the standard at 29 CFR 1910.95 is susceptible. I respectfully
request that my decision be reviewed to that end.

[As cited above. I am aware of the Frank Nutty
case interpreting a virtually identical noise standard. However, I think that decision is
not only distinguishable in several aspects from the instant case but, of great
importance, the point raised here was not discussed there.]

FINDINGS OF FACT

Having heard the testimony, observed the witnesses,
and examined the exhibits, the following Findings of Fact are made:

1. At all times concerned, the Respondent regularly
   received, handled or worked with goods which had moved across state lines.

2. As concerns Items #1(a) of Citation #1, the
   dosimeter used in the inspection did not accurately record the true sound levels.

3. The true sound levels did not exceed the figures
   of Table G-16.

4. The sound levels exceeded those shown in Table
   G-16 when measured on the A scale of a standard sound level meter at slow response.

CONCLUSIONS OF LAW

1. At all times concerned, the Respondent was an
   employer engaged in a business affecting commerce within the meaning of the Act; and the
   Occupational Safety & Health Review Commission has jurisdiction over the subject
   matter and the parties.

2. The Complainant has not sustained the burden of
   proving the Respondent violated the standard at 29 CFR 1910.95(a).

ORDER

The whole record having been considered, it is
ordered that Citation #1, and the penalty proposed therefor, be vacated.

SO ORDERED.

FOSTER FURCOLO

JUDGE, OSHRC

Dated: June 28, 1982

Boston, Massachusetts

APPENDIX

THE ACT

Section 654 [section 5(a)(2)] Employer ".
. .shall comply with occupational safety and health standards. . ."

Section 666 [section 17(b)] ". .
.employer who has received a citation for a serious violation. . .of this Act. . .shall be
assessed a civil penalty of up to $1,000 for each such violation."

Section 666 [section 17(k)] ". . .a
serious violation shall be deemed to exist. . .if there is a substantial probability that
death or serious physical harm could result. . . unless the employer did not, and could
not. . .know of the presence of the violation."

THE STANDARDS

29 CFR 1910.95(a) - (b)(3):

� 1910.95 Occupational noise exposure.

(a) Protection against the effects of noise exposure shall be provided
when the sound levels exceed those shown in Table G-16 when measured on the A scale of a
standard sound level meter at slow response.

(b)(1) When employees are subjected to sound exceeding those
listed in Table G-16, feasible administrative or engineering controls shall be utilized.
If such controls fail to reduce sound levels within the levels of Table G- 16, personal
protective equipment shall be provided and used to reduce sound levels within the levels
of the table.

(2) If the variations in noise level
involve maxima at intervals of 1 second or less, it is to be considered continuous.

(3) In all cases where the sound levels
exceed the values shown herein, a continuing, effective hearing conservation program shall
be administered.

TABLE G-16 - PERMISSIBLE NOISE EXPOSURES[[1]]

Sound level

dBA slow

Duration per day, hours

response

8..................................................................90

6..................................................................92

4..................................................................95

3..................................................................97

2.................................................................100

1�..............................................................102

1.................................................................105

�................................................................110

� or less ....................................................115

FOOTNOTES:

[[1]] The standard is quoted in its entirely/except for provisions dealing with
determining noise levels by octave band analysis, a measurement technique not involved in
this case.

[[2]] This problem was noted early in the
Commission's experience with the noise standard. See Weyerhaeuser Co. , 74
OSAHRC 57/F4, 2 BNA OSHC 1152, 1153-54, 1974-75 CCH OSHD � 18,468, pp. 22,485-86 (No.
2116, 1974)(lead opinion); Sun Shipbuilding & Drydock Co. , 74 OSAHRC 61/A2, 2
BNA OSHC 1181, 1182-83, 1974-75 CCH OSHD � 18,537 (No. 268, 1974) (lead and concurring
opinions); Weyerhaeser Co. , 77 OSAHRC 9/A2, 4 BNA OSHC 1972, 1974-75, 1976-77 CCH
OSHD � 21,465, pp. 25,748-49 (No. 1231, 1977), aff'd in pertinent part , sub
nom . Noblecraft Industries, Inc. v.Secretary , 614 F.2d 199 (9th Cir. 1980).

[[3]] Collier-Keyworth argues that the use of a
dosimeter is inherently inconsistent with the standard, which specifies that noise levels
are to be measured with a standard sound level meter. As we have explained, the dosimeter
itself contains the circuitry of a standard sound level meter. It also contains other
circuitry, but the other circuitry only performs automatically calculations that would
otherwise have to be done manually. Thus, the dosimeter does precisely what the standard
requires: measures sound levels with a standard sound level meter and performs the
calculations required by the cumulation formula. Whether it performs these functions with
sufficient accuracy is another question, which we will address later. For now, we note
only that the use of a dosimeter is consistent with the standard, and reject
Collier-Keyworth's contrary argument Love Box Co. , 76 OSAHRC 45/D5, 4 BNA OSHC
1138, 1140 n. 2 1975 76 CCH OSHD � 20,588, p. 24,628 & n. 2 (No. 6286, 1976).

[[4]] In 1981, section 1910.95(b)(3) was replaced
with a new standard, known as the "hearing conservation standard" or
"hearing conservation amendment," which lists in detail the requirements a
hearing conservation program must have. 46 Fed. Reg. 4078 (Jan. 16, 1981), now codified at
29 C.F.R. � 1910.95(c)-(p). The hearing conservation standard requires an employer to
implement a hearing conservation program whenever employee noise exposures equal or exceed
an eight-hour time-weighted average level of 85 dBA. 29 C.F.R. � 1910.95(c)(1). It also
provides that, in determining whether this limit is exceeded, "[a]ll continuous,
intermittent and impulsive sound levels from 80 decibels to 130 decibels shall be
integrated into the noise measurements." 29 C.F.R. � 1910.95(d)(2)(i). Thus, impulse
noise is explicitly included in the hearing conservation standard. However, in
promulgating the hearing conservation standard, the Secretary did not purport to amend any
provision of the old standard except subsection (b)(3). The hearing conservation standard
therefore does not affect whether impulse noise is included under the standard involved in
this case, section 1910.95 (a)-(b)(2), which require administrative and engineering noise
controls and personal protective equipment.

[[5]] The Secretary's expert witness, Dr. Burton
Jaffe, testified that noise is a "mechanical transfer of energy that really shakes
the hair cells and damages them." The manner in which this testimony is summarized in
the Secretary's brief echoes what has become known as the "equal energy rule," a
theory that equal amounts of noise and energy are equally harmful regardless of their form
or duration. This theory was discussed at length in the preamble to the Secretary's new
hearing conservation standard and expressly applied to impulse noise. See 46 Fed. Reg. at
4096. Its application to impulse noise seems to have been upheld on the basis of the
rulemaking record compiled by OSHA to support the hearing conservation standard. See
Forging Industry Ass'n v. Secretary of Labor , 773 F.2d 1436, 1451 (4th Cir. 1985) ( en
banc ).

[[6]] The Secretary attempts on the basis of a
witness's testimony to reconcile Table G-16 with the 140 dB provision. We shall consider
this testimony after we have completed our review of the text of the standard. See note 12
below.

[[7]] When the original version of the standard was
published by the Labor Department's Bureau of Labor Standards under the Walsh-Healey
Government Contracts Act, 41 U.S.C. � 35-45, this provision was set in the same size type
as other text and placed after the footnote to the table. See 34 Fed. Reg. 7949 (1969) and
35 Fed. Reg. 1015 (1970) (corrections to standard). When the standard was adopted and
reprinted under the OSH Act in 1971, the Secretary did not purport to--and was in any
event not empowered to--make any substantive change in it. Yet, the 140 dB provision was
then printed in the same small type as the footnote to the table, apparently because a
typesetter mistook it to be a continuation of the footnote to the table immediately above.
36 Fed. Reg. 10466, 10518 (1971). We are confirmed in this impression by four Labor
Department documents. The Labor Department's own edition of OSHA standards, I General
Industry Standards and interpretations 144 (OSHA No. 2077, 1984), prints the 140 dB
impulse noise provision in, and in the same size type as, the text. The 140 db provision
is still part of the text of the original Walsh-Healey standard, 41 C.F.R. � 50-204.10. A
Labor Department bulletin interpreting the Walsh-Healey standard characterized the
provision as "[t]he last sentence in paragraph (d) of section 50-204.10. . . ."
Bureau of Labor Standards, U.S. Dept. of Labor, Bulletin 334, Guidelines to The
Department of Labor's Occupational Noise Standards for Federal Supply Contracts , 4
(Dec. 4, 1970). Finally, the corresponding provision in OSHA's noise standard for the
construction industry--which was adopted at about the same time as the original general
industry standard (see 36 Fed. Reg. 7340, 7348 (1971))--was printed in a separate
subsection, now codified as 29 C.F.R. � 1926.52(e).

[[8]] That provision states in part that "[w]hen
an agency decision rests on official notice of a material fact not appearing in the
evidence of the record, a party is entitled, on timely request, to an opportunity to show
the contrary."

[[9]] The record contains various terms used to
describe non-impulsive noise, such as steady noise, continuous noise, steady-state noise,
and intermittent noise. Except for intermittent noise, these terms suggest noise that is
constant in intensity over a relatively long period of time. Intermittent noise varies in
intensity, but not as rapidly as impulse noise. For example, noise from a machine that is
alternately on and off for several seconds at a time would be classified as intermittent
noise.

[[10]] Even in 1981, when the Secretary promulgated
the hearing conservation amendment to the noise standard, there was a relative dearth of
information on impulse noise. The Secretary noted: "In contrast to the studies of
continuous noise mentioned above, dose-response relationships for impulse noise are not so
easily defined." 46 Fed. Reg. at 4096.

[[11]] ACGIH, "Threshold Limit Values of
Physical Agents" (1969), reprinted in National Safety Council, Fundamentals
of Industrial Hygiene , Appendix A, pp. 739-41 (1971). Collier-Keyworth introduced as
Exhibit R-101 the 1971 version of the ACGIH standard, which appears to be identical.

[[12]] See also a chapter by Jones, "Standards
and Threshold Limit Values for Noises," in National Safety Council, Industrial
Noise and Hearing Conservation ch. 11, p. 309 (Olishifski & Harford eds.
1975)(ACGIH limits "incorporated in" Walsh-Healey standard). That the ACGIH
standard substantially influenced the drafter of the Walsh-Healey standard leads the
Secretary to rely heavily on testimony by Jones, who was chairman of the ACGIH committee,
that the table in the ACGIH standard corresponding to Table G-16 "would include
anything that the sound level meter would respond to, be it a continuous[,] impact [or]
impulsive noise." Jones further testified that the 140 dB impulse noise provision was
intended to be an additional restriction on impulse noise, included because of the
possibility that high intensity impulses could be harmful even if the limits in the table
were not exceeded. The Secretary states that Jones "was, in effect, the author of the
[140 dB impulse noise] provision" because the limits in the table adopted by Jones's
ACGIH Committee were "issued in virtually identical form by the Department of Labor
as a standard under the Wash-Healey Act. . . ." We are unable to credit this part of
Jones's testimony because the ACGIH standard expressly stated that the exposure limits in
its table do not apply to impulse noise. We therefore disagree with the argument in
the Secretary's brief that this part of Jones's testimony is a reliable guide to the
intent of the drafters of the Walsh-Healey standard.

[[13]] The bulletin was revised on June 8, 1971,
after the Walsh-Healey standard was adopted under the OSH Act. The 1971 version was
introduced into evidence.

[[14]] A passage written by Dr. Van Atta in Industrial
Noise and Hearing Conservation , n. 12 above, at 322, also suggests this. The passage,
which was read into the record, states:

This regulation provides for a basic level of 90 dBA
for continuous noise exposure with a tradeoff of 5 dBA for each halving of the
noise exposure time. This includes a rough allowance for interruptions of noise exposure.
Impact noises must be limited to less than 140 dB peak sound pressure level.

It is important to note that this is a double
requirement. For example, in a power-press department, it is necessary that both the
instantaneous peaks of the impact noises arising out of the operation of large presses be
below 140 dB and that the continuous ambient background noise be below 90 dBA.
[Emphasis added.]

[[15]] J. Barry, "Problems in Enforcement of the
Occupational Noise Standard," in Proceedings of Noise-Con 79, Machinery Noise
Control , 11 (Sullivan & Crocker eds. 1979). At the hearing in this case, Dr. Barry
first testified that Table G-16 "has no relation to impulse noise." He later
stated, however, that the standard did intend for impulse noise to be included in dosage
calculations.

[[16]] The chapter is printed in Occup. Safety and
Health Admin., U.S. Dep't of Labor, VI OSHA No. 3058, Industrial Hygiene Field
Operations Manual (1980), and in CCH Employ. S. & H. Guide, Edition No. 419 (May
24, 1979).

[[17]] At the hearing in this case, Dr. Barry stated
that he recommended the use of fast response in this instruction to enable the inspector
to analyze the components of the noise to assist in evaluating the feasibility of
engineering controls. (Fast response, like slow response, is a setting on a sound level
meter. For the moment it is enough to observe that where impulse noises are closely
spaced, fast response is more accurate than slow response in displaying the true
background noise level.) When the Secretary promulgated the hearing conservation amendment
in 1981, he said that this IHFOM provision "is clearly at variance with the present
standard and is being deleted." 46 Fed. Reg. 4078, 4137 & n. 11(Jan. 16, 1981).
If, as Dr. Barry stated, the purpose of the provision was to assist inspectors to evaluate
the feasibility of engineering controls, there would have been no need for the Secretary
to disavow it in this manner. In sum, we read the instruction as OSHA did--to signify that
impulse noise was not to be included in determining whether employee exposure exceeds
Table G-16 levels.

The Secretary also claims that the prescription of
slow response in the standard is inconsistent with any intent to exclude impulse noise
from Table G-16. We are unconvinced of any logical inconsistency here. The ACGIH
standard--from which the Secretary acknowledges the drafters of the Walsh-Healey
standard--from which the Secretary acknowledges the drafters of the Walsh-Healey standard
drew heavily--both prescribed slow reposes and excluded impulse noise. The same is true of
the noise standard proposed by the Secretary in 1974. See 39 Fed. Reg. 37773, 37775 (Oct.
24, 1974), proposing new section 1910.95(c)(1)(ii).

[[18]] Turner involved three presses in the
same room. One stroked at a rate of 27 per minute, and a second at 33 to 38 strokes per
minute. 4 BNA OSHC at 1556, 1976-77 CCH OSHD at pp. 25,273-4. Thus, these two presses
alone would produce 60 or more noise impulses per minute, i.e., continuous noise.

[[19]] Collier-Keyworth occasionally alludes to this
point in a confusing fashion, speaking of dosimeters having an "effective"
exchange rate of 3 decibels. A dosimeter set to employ a 3 decibel exchange rate would
register the same sound levels as would one set to employ a 5 decibel exchange rate. It
would record a higher dosage, however, because the lower exchange rate shortens
permissible durations. Yet, Collier-Keyworth does not argue that OSHA set its dosimeters
here to employ a 3 decibel exchange rate. Instead, it maintains, based on the testimony of
expert witnesses, that in impulsive noise environments dosimeters set to use a 5 decibel
exchange rate and slow response will so overcount "true noise levels" as to
approximate the results of employing a 3 decibel exchange rate with fast response.
Collier-Keyworth's argument rests on the assumption that under Table G-16 sound levels
must be measured with a fast response instrument. We will discuss the validity of this
assumption in the text below.

[[20]] This effect is illustrated by Ex. R-26, which
contains several strip chart recordings made by Rockwell, and by C-22L, a strip chart
recording made by Barry. One recording in Ex. R-26 shows noise patterns for a power press
operating at 100 strokes per minute. Because the peaks are less than one second apart, the
noise is quasi-continuous. With the sound level meter set for fast response, which
corresponds to an integration time of an eighth of a second, the individual impulses are
readily visible and have peak intensities of 108-109 dBA superimposed on a background of
about 99 dBA. On slow response, however, the noise appears relatively continuous at a
level of 104-105 dBA. Thus, compared to fast response, the sound level meter at slow
response reads higher troughs but lower peaks, that is, the individual impulses appear to
be of longer duration but lower peak intensity.

The reason for the difference is best explained by
discussing the behavior of the needle of a sound level meter: At slow response, the needle
does not have time to return to the continuous, background noise level before the next
impulse comes along and causes it to swing back up. For example, in Ex. C-22, the sound
peaks are about two to three seconds apart, and hence much farther apart than in Ex. R-

26. The noise levels indicated by slow response are still different from that of fast
    response but not as different as in Ex. R-26, where noise peaks only six-tenths of a
    second apart gave the needle little time to swing back before the next impulse arrives.

[[21] ]Its argument is echoed in the observation of
one dosimeter manufacturer that "[i]t is the requirement of a slow response time
constant in a dosimeter that causes a dosimeter to appear to read high." Ex. R-24,
statement by Quest Electronics.

[[22]] See American Mining Congress v.
Marshall , 671 F.2d 1251, 1255-7 (10th Cir. 1982) (discussing whether measuring method
prescribed by mine health regulation is "arbitrary, capricious, or an abuse of
discretion" within the meaning of the Administrative Procedure Act).

[[23]] The brief at the Chocolate Manufacturers
Association, which the Commission permitted to appear on review as an amicus curiae,
argues that dosimeters generally are unreliable. Attached to its brief are studies by Dr.
Paul Hess of Hershey Foods that argue in support of this thesis and that criticize OSHA's
evaluation of Dr. Hess's studies in OSHA's preamble to revisions in 1983 of the hearing
conservation standard. The Chocolate Manufacturers Association also asks the Commission to
take official notice of statements in OSHA's Industrial Hygiene Technical Manual ,
p. B-47 (March 30, 1984) ("the IHTM), reprinted in CCH Employ. S. & H.
Guide, Extra Edition No. 680 (May 24, 1984), and in 2 BNA Occup. S. & H.
Rep. 77:8001, 77:8925 (June 14, 1984). The Secretary of Labor objects to the Chocolate
Manufacturers Association's brief, arguing that it improperly introduces evidence. We
deferred ruling on this motion and on the taking of official notice of the IHTM material.

We take official notice of the lHTM but we are not
persuaded that accepting its statements as true mandates a result here. The portion of the
IHTM relied on by the Chocolate Manufacturers Association warns OSHA field personnel that
"[s]hort-duration pulses with a low repetition rate may be completely eliminated from
the dosimeter readings simply because the averaging circuits in the instruments cannot
respond fast enough to the change in level." However, we understand this passage to
mean that dosimeters sometimes undercount the noise level when certain impulses are
present; moreover, if "low repetition rate" refers to impulses more than one
second apart, then the supposed inaccuracy is irrelevant because we hold that such
impulses may not be counted at all. Another passage relied on by the amicus states that
"[c]onversely, pulses with a higher repetition rate may be overestimated by the
dosimeter because the averaging circuits will build up to the high level but will not have
time to drop back down between pulses." We are puzzled by this passage, for it could
be read to complain of the slow response mode mandated by OSHA's own standard. In any
event, if it refers to impulses more than one second apart, it is irrelevant for the
reasons we have already stated. And if it refers to impulses less than one second apart
then it is also irrelevant because under subsection (b)(2), the dosimeter does not have to
drop back down between impulses at all. Thus, we are not persuaded that the IHTM requires
rejection of the dosimeter readings here.

On the other hand, we grant the Secretary's motion to
strike those portions of the Chocolate Manufacturers Association brief that introduces the
studies of Dr. Hess. We decline to take official notice of this material. The material
attached by the amicus to its review brief is largely evidentiary, consisting of papers
and studies by Dr. Hess challenging the accuracy of dosimeters generally. As the Secretary
argues, however, the author of this material has not been subjected to cross- examination
and the Secretary's counsel has had no opportunity to evaluate, impeach or rebut his
studies. Thus, if this new material were to be considered, a remand to Judge Furcolo would
be in order. Yet, this reopening of the record would be on the implied invitation of an
amicus, not a party. In these circumstances, we adhere to our rule against the admission
of evidentiary material in review briefs. E.g. , Anoplate Corp. , 86 OSAHRC
____, 12 BNA OSHC 1628, 1683 n. 6, 1986 CCH OSHD � 27,519, pp. 35,681-2 n. 6 (No.
80-4109, 1986).

[[24]] Collier-Keyworth points out that on
cross-examination Hart testified she did not know the "strokes per minute" for
the presses and suggests that this casts doubt on her testimony that she measured the
speed of the presses. However, Hart's testimony that she did not know the "strokes
per minute" does not contradict her unequivocal testimony that she timed the presses
and found that the time between strokes was less than one second. Under the standard, the
noise is deemed continuous if the time between strokes is less than one second and, once
this is known, it is not necessary to determine the exact time between strokes or the
exact stroke rate.

[[25]] The strip chart recordings of noise levels
made by Barry and Rockwell show noise levels in the vicinity of the presses ranging from
95 to 105 dBA. Although these measurements were made on different days than Hart's, they
lend credence to Hart's measurements indicating that the sound levels near the presses
were typically in this range.

[[26]] See Table 2 of Field Service Memorandum No.
11 , in Appendix B, p. B-36 of the IHTM, note 23 above, Cf . 29 C.F.R. �
1910.95, Appendix A, Table A-1 (hearing conservation amendment) (entry for 91.9 dBA
corresponds to dose of 130 percent).

[[27]] The error factors for Type 2 sound level
meters are found in ANSI S1.4-1971, "Specification for Sound Level Meters," at
11, Table 3. Some of the information in that table, including the error factor at 8000
hertz, was included in the paper Dr. Barry presented at Purdue University, which appears
in the record as Exhibit R-62. OSHA applied that information to Type 2 dosimeters in its
IHTM. See Table 1 of Field Service Memorandum No. 12 , in Appendix B, pp. B-44--B-45
of the IHTM, note 23 above.

[[28]] Because he vacated all citation items on
finding that the standard excluded impulse noise, Judge Furcolo did not make findings on
the merit. Normally, we would remand to him for such findings. However, in view of the age
of the case, we will make our own findings.

[[29]] The Secretary did not attempt to show that
administrative controls for these machines were feasible.

[[30]] Barry believed that erecting acoustical
enclosures entirely around the presses would be impractical due to space limitations and
production demands.

[[31]] The dies for each press were changed an
average of about once per day. Some of Collier-Keyworth's dies weighed as much as 1,000
pounds and had to be changed using a chain hoist. When the chain hoist was used, the die
could swing into the enclosure and damage it. Barry testified that this could be avoided
if the operator was careful. Holmer testified, however, that another company experienced
severe damage to similar enclosure after 14 to 20 die changes. Holmer also estimated that
enclosures would add 20 to 40 minutes to the time required for a die change.

[[32]] The highest dosimeter reading the Secretary
obtained for any of the press operators was 305%. This is equivalent to an eight-hour
time-weighted average exposure of 98 dBA. See section 1910.95, Appendix A, Table A-1.

[[33]] Most of the studies relied on in the paper
were performed with gunshot noise, but the authors considered the information sufficiently
general to be extended to other types of impulse noise.

[[34]] Threshold shift is a measure of hearing loss.
A person's hearing threshold is the lowest sound level in decibels that person can detect.
If the person suffers hearing loss, his threshold increases, i.e., he can no longer hear
decibel levels as low as he previously could. The number of decibels by which his
threshold increases is his threshold shift. A permanent hearing loss is referred to as a
permanent threshold shift or PTS. Short exposures to high noise levels can cause a
temporary hearing loss, or temporary threshold shift. When a person suffers a temporary
threshold shift, his hearing recovers after the exposure to high noise levels ceases.
However, repeated exposures sufficient to cause temporary threshold shifts can lead to a
permanent threshold shift. Therefore, to test whether noise exposures give rise to the
possibility of permanent hearing loss, scientists measure temporary threshold shifts in
exposed individuals. If an exposure produces measurable temporary threshold shifts,
repetitive exposures of the same magnitude will eventually cause permanent threshold
shifts. See 46 Fed. Reg. 4078, 4080 (1981)(preamble to hearing conservation amendment).