The Forging Industry Association (FIA) petitions for review of the Secretary of Labor’s promulgation of a hearing conservation amendment (amendment) to its occupational noise exposure standard, 29 C.F.R. § 1910.95 (1984). The FIA contends that the Department of Labor’s Occupational Safety and Health Administration (OSHA) exceeded its authority in adopting the amendment, that there was not substantial evidence to support OSHA’s determination of a significant risk of harm in the workplace justifying the amendment’s requirements, that the amendment was not reasonably necessary or appropriate to provide *1440safe employment, and that the amendment’s requirements are not feasible. In-tervenor, National. Arborist Association (NAA), contends that the Agency arbitrarily refused to exempt the tree care industry from the amendment’s requirements.
I.
An occupational noise exposure standard has existed since OSHA’s inception in 1971. The current standard, which is found at 29 C.F.R. § 1910.95, was originally promulgated under the Walsh-Healey Public Contracts Act, 41 U.S.C. §§ 35-45 (1982) for the purpose of protecting employees from workplace exposure to damaging levels of noise. The Walsh-Healey standard was adopted by OSHA pursuant to Section 6(a) of the Occupational Safety and Health Act, which allowed the Secretary to promulgate any established Federal standard within two years of the effective date of the Act without following normal rulemaking procedure.
The standard established a permissible workplace limit of 90 decibels (db)1 calculated using an 8-hour time-weighted average (TWA).2 29 C.F.R. § 1910.95(a). If the 90 db exposure limit is exceeded, the employer must reduce noise to or below this level by using feasible engineering or administrative controls.3 Id. at § 1910.-95(b)(1). If such controls are infeasible, employers may use hearing protectors, such as ear muffs or plugs, to reduce employee noise exposure to permissible limits. Id. Prior to amendment, the standard also contained a generally phrased requirement that employers administer “a continuing effective hearing conservation program” in workplaces where sound levels exceeded the permissible exposure level (PEL).4 Id. at § 1910.95(b)(3) (1980).
When studies revealed that many employees suffered significant hearing impairment at noise levels below the 90 db threshold, OSHA began the process of collecting and evaluating the information necessary to issue a comprehensive new regulation with a reduced permissible exposure level of 85 db. As an interim measure, OSHA adopted in 1983 a hearing conservation amendment to replace the general conservation program requirement.
The amendment requires employers to determine, through implementation of a monitoring program, which employees are exposed to an “action level” of 85 db or above measured as an 8-hour TWA. 29 C.F.R. 1910.95(d). Such employees must be notified of the amount of sound they are exposed to and be provided with an audiom-etric test to determine their hearing level. Id. at 1910.95(e), (g)(1). At least annually thereafter, the employer must provide the exposed employee with an additional au-diometric test to determine whether the employee has suffered an average loss of hearing of 10 db or more in either ear *1441known as a standard threshold shift (STS).5 Id. at 1910.95(g)(6). If there has been an STS, the employer must take follow-up measures to prevent the employee from reaching the material impairment stage.6 These measures include fitting the employee with hearing protectors, providing training, and requiring the employee to use the protectors. Id. at 1910.95(g)(8). The protectors must reduce the employee’s exposure to an 8-hour TWA of 85 db or less. Id. at 1910.950(3).
In addition, the employer must institute a training program on audiometric testing, hearing protectors, and effects of noise on hearing for all employees who are exposed to noise at or above an 8-hour TWA of 85 db. Id. at 1910.95(k). The employer must also retain records of employee exposure measurements and audiometric tests. Id. at 1910.95(m).
The provisions of the amendment apply to all employees covered by the Act, except those in construction, agriculture, and oil and gas well drilling and servicing. 46 Fed.Reg. 42622; 29 C.F.R. § 1910.95(o). OSHA estimates the annual cost of compliance for the amendment at $210.3 million. In terms of industry costs, it estimated the costs at $41 annually for each employee. Regulatory Impact and Regulatory Flexibility Analysis of the Hearing Conservation Amendment, U.S. Department of Labor, Occupational Safety and Health Administration, Office of Regulatory Analysis (February 1983), part IV.
II.
Promulgation of the amendment was preceded by almost ten years of administrative proceedings. In 1974 OSHA published a notice of proposed rulemaking and invited interested persons to submit written data, views, and arguments regarding the proposal. Hearings were held in 1975, in which over ninety parties participated. An economic impact analysis prepared for OSHA by Bolt, Beranek & Newman, Inc., was made publicly available in June 1976, and interested persons were afforded the opportunity to comment on this study. In October 1976 an informal public hearing was held on the economic analysis. OSHA reopened the record in 1980 for the purpose of introducing additional comments, letters, and reports that had been received by the Agency, and interested persons were given an opportunity to submit comments on this added information.
OSHA first promulgated the hearing conservation amendment on January 16, 1981, requiring noise exposure monitoring, au-diometric testing, use of hearing protection devices, education of employees, posting of warning signs, and retention of records.7
In response to this promulgation, OSHA received numerous requests for clarification as well as petitions for reconsideration and administrative stay. Additionally, petitions for judicial review under section 6(f) of the Occupational Safety and Health Act of 1970, 29 U.S.C. § 655(f) (1982)8 were *1442filed in this court by the Chocolate Manufacturers Association, Chamber of Commerce, and American Iron and Steel Institute. In January 1982 the Agency announced that additional public hearings would be held on all aspects of the amendment.9
After completion of these additional hearings and substantial revision of the amendment, the Agency, on March 8, 1983, published the amendment as a final rule. 29 C.F.R. §§ 1910.95(c)-(p).
From the evidence presented during these proceedings, the Agency concluded that noise is one of the Nation’s most pervasive occupational health problems. Further, the Agency determined that 10 to 15 percent of employees exposed to an eight-hour TWA of 85-90 db will suffer a material hearing impairment. The Agency ultimately concluded that enforcement of the amendment’s requirements would result in a significant prevention of hearing loss by the 2.2. million workers in American production industries with eight-hour TWAs between 85-90 db.10
The Agency also found that compliance with the amendment was technologically feasible, reasoning that the requirements did not create any engineering problems and that the necessary equipment and specialized personnel were available to implement the amendment’s requirements. 46 Fed.Reg. at 4117. Finally, OSHA concluded that the standard was economically feasible in that the cost of compliance, $41 per employee, was small relative to both sales (averaging less than 0.0148 percent) and profits (averaging less than 0.1932 percent).
III.
The Agency’s authority to promulgate standards is established in sections 3(8) and 6(b) of the Act, 29 U.S.C. §§ 652(8) and 655(b) (1982). Section 3(8) defines “occupational safety and health standard” as “a standard which requires conditions, or the adoption or use of one or more practices, means, methods, operations, or processes, reasonably necessary or appropriate to provide safe or healthful employment and places of employment.” 29 U.S.C. 652(8). In promulgating standards dealing with toxic materials or harmful physical agents, section 6(b)(5) requires the Agency to
set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life.
29 U.S.C. § 655(b)(5).
FIA asserts correctly that, under either section, OSHA’s authority is limited to ameliorating conditions that exist in the workplace. Industrial Union Department v. American Petroleum Institute, 448 U.S. 607, 642, 100 S.Ct. 2844, 2864, 65 L.Ed.2d 1010 (1980) (Secretary required to “make a threshold finding that a place of employment is unsafe.”); Atlas Roofing Co. v. OSHARC, 430 U.S. 442, 445, 97 S.Ct. 1261, 1264, 51 L.Ed.2d 464 (1977) (“The Act created a new statutory duty to avoid maintaining unsafe or unhealthy working conditions.”) From this language, FIA constructs its first argument that because hearing loss may be sustained as a result of activities which take place outside the workplace — such as listening to loud music, age, or engaging in certain recreational activities — OSHA acted beyond its statutory authority by regulating non-occupational conditions or causes. There is simply no merit to FIA’s argument that the *1443amendment seeks to regulate non-occupational hearing loss.
OSHA received and relied on a panoply of scientific reports and studies indicating that ten to fifteen percent of the workers exposed to an eight-hour TWA of 85 db at the workplace would suffer material hearing impairment. OSHA derived these conclusions from evidence including studies done by the Environmental Protection Agency, the National Institute for Occupational Safety and Health, the International Organization for Standardization, and Dr. W. Baughn of the General Motors Corporation. Additionally, studies performed by Bolt, Beranek, and Newman, Inc., the Center for Policy Alternatives and the Agency indicated that substantial benefits in terms of diminution of hearing impairments would result from instituting the amendment’s requirements. There is considerably more than substantial evidence to support the conclusion that the amendment is directed toward abating workplace noise that has been shown to adversely affect employees. Further, the additional steps necessitated by a finding that a worker has suffered from an STS, i.e., providing hearing protectors and training, áre also properly directed towards the employees’ workplace environment. OSHA determined, on the basis of the evidence before it, that workers who have suffered an STS face even greater risk from continued exposure to high levels of workplace noise than workers whose hearing is unimpaired, and that “noise induced- hearing loss' ... progresses with increased exposure.” Thus, the amendment does nothing more than ensure that a hearing-endangered worker is provided with protection in the workplace in order to decrease the risk of a hearing impairment. Having identified employee susceptibility to noise, “[t]he Act does not wait for an employee ... [to] become injured. It authorizes the promulgation of health and safety standards ... in the hope that these will act to prevent ... injuries from ever occurring.” Whirlpool Corp. v. Marshall, 445 U.S. 1, 12, 100 S.Ct. 883, 890, 63 L.Ed.2d 154 (1980).
Regulation of the industrial cause and effect of hearing loss obviously is not as simple as regulation of most mechanical hazards. Dangers inherent in the operation of moving machinery such as a table saw are easily perceived, and their causal relation to mutilated human limbs or eyes are readily understood and frustrated. Prevention of more subtle hazards requires more sophisticated solutions. ' OSHA relied on the extensive and thorough research of several scientific institutions in defining the problems related to industrially-caused hearing loss and designing its proposal.
Section 6(f) of the Act, 29 U.S.C. § 655(f) provides that “[t]he determination of the Secretary shall be conclusive if supported by substantial evidence in the record considered as a whole.” The Supreme Court has defined “substantial evidence” in the context of this section as “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” American Textile Manufacturers Institute v. Donovan, 452 U.S. 490, 522, 101 S.Ct. 2478, 2497, 69 L.Ed.2d 185 (1981) (quoting Universal Camera Corp. v. NLRB, 340 U.S. 474, 477, 71 S.Ct. 456, 459, 95 L.Ed.2d 456 (1951)). Further, where, as here, the Agency’s findings are based upon complex scientific and factual data or involve speculative projections, our application of this test is particularly deferential. Baltimore Gas & Electric Co. v. National Resources Defense Council, Inc., 462 U.S. 87, 103, 103 S.Ct. 2246, 2255-2256, 76 L.Ed.2d 437 (1983); United Steelworkers of America v. Marshall, 647 F.2d 1189, 1206-07 (D.C.Cir.1980), cert. denied sub nom. National Ass’n of Recycling Industries, Inc. v. Secretary of Labor, 453 U.S. 913, 101 S.Ct. 3149, 69 L.Ed.2d 997 (1981). In these circumstances, the Agency’s findings must only be within a “zone of reasonableness.” United Steelworkers, 647 F.2d at 1207. The scientific data supporting the Agency’s conclusions regarding the occupational risk of noise exposure as well as the importance and efficiency of the amendment are clearly within the zone of reasonableness.
*1444To be sure, some hearing loss occurs as a part of the aging process and can vary according to non-occupational noise to which employees are exposed. The amendment, however, is concerned with occupational noise — a hazard of the workplace. The hazard is identified as sustained noise of great intensity — 85 db and above. Nonoccupational noise of that intensity sustained over a period of eight hours each day is hard to imagine.
The amendment provides that non-occupationally caused hearing loss be excluded from its regulation. See 29 C.F.R. §§ 1910.95(g)(8)(ii), 1910.95(g)(10)(ii) (1984). Assuming, however, that some loss caused by aging or smaller amounts of noise sustained for shorter periods also aggravates the hearing loss incurred by an individual employed in a high noise-producing industry, that is scant reason to characterize the primary risk factor as non-occupational. Breathing automobile exhaust and general air pollution, for example, is damaging to lungs, whether healthy or not. The presence of unhealthy lungs in the workplace, however, hardly justifies failure to regulate noxious workplace fumes. Nor would there be logic to characterizing regulation of the fumes as non-occupational because the condition inflicted is aggravated by outside irritants.
FIA further argues that the amendment cannot be justified by reference to section 655(b)(5) because it does not establish or change the permissible limits of workplace noise and because noise is not a harmful physical agent' but rather an effect. The legislative history of the Act, however, reveals that excessive noise was one of the harmful physical agents that Congress anticipated would be subject to health standards:
In 1967, the Surgeon General of the United States studied six metropolitan areas, examining 1700 industrial plants which employed 142,000 workers. The study found that 65% of the people were potentially exposed to harmful physical agents, such as severe noise or vibration, or to toxic materials. The Surgeon General further examined the controls that were in effect to protect workers from such hazards, and found that only 25% of the workers were adequately covered
S.Rep. No. 1282, 91st Cong., 2d Sess., reprinted in 1970 U.S. Code Cong. & Ad. News 5177, 5179. (Emphasis added). Additionally, section 655(b)(5) clearly states that “[development of standards ... shall be based upon research, demonstrations, experiments, and such other information as may be appropriate” and directs the Agency to consider “the latest available scientific data in the field.” In view of the legislative history, as well as the overwhelming scientific evidence presented to the Agency, it is manifest that noise is a harmful physical agent.
Moreover, the amendment, while leaving temporarily unchanged the 90 db PEL, establishes requirements whose aim is to reduce the level of noise that reaches the ears of particularly susceptible employees. As such, it satisfies the mandate of section 655(b)(5) by assuring, to the extent feasible, “that no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life.” (emphasis supplied).
IV.
Having decided that the Agency acted within its statutory authority in promulgating the amendment, we next consider whether substantial evidence exists to support the Agency’s finding that the preamendment noise standard exposed workers to a significant risk of material health impairment. Industrial Union Department v. American Petroleum Institute, 448 U.S. 607, 653, 100 S.Ct. 2844, 2869-2870, 65 L.Ed.2d 1010 (1980). We conclude that the Agency’s determination easily passes this test.
Prior to the amendment, the noise standard set a PEL of 90 db with requirements to protect employees’ hearing through noise reduction or hearing protectors limited to those instances when that standard *1445was exceeded. 29 C.F.R. § 1910.95(b)(3) (1980). The amendment retains the maximum permissible exposure level of 90 db but requires hearing conservation measures when the noise level reaches 85 db. With regard to the issue of the effect of noise on hearing, the Agency especially relied on studies performed by the Environmental Protection Agency (EPA), the National Institute for Occupational Safety and Health (NIOSH), and the International Organization for Standardization (ISO).
The EPA study, using the data of Dr. Baughn of the General Motors Corporation,11 Dr. Passchier-Vermeer,12 Drs. Burns and Robinson,13 and others concluded that 22 percent of those workers exposed to an average noise exposure of 90 db at 500, I, 000, and 2,000 Hz over a 40-year working lifetime would suffer material hearing impairment as would 12 percent exposed to 85 db and 5 percent at 80 db.
The ISO Recommendation,14 published in 1971 and approved by twenty member bodies, concluded that 21 percent of those workers subjected to an average noise exposure of 90 db at 500, 1,000, and 2,000 hz over a 40 year working lifetime would suffer material hearing impairments, as would 10 percent of those exposed to 85 db.
The NIOSH study15 reported that 29 percent of employees exposed to an average noise exposure of 90 db at 500, 1,000 and 2,000 hz over a 40 year working lifetime would suffer material hearing impairments as would 15 percent of those exposed to 85 db and 3 percent subjected to 80 db.
Other independent studies were consistent with these reports. These included studies by Drs. Martin, Gibson, and J.N. Lockington,16 and Drs. Royster, Thomas and Elliott Berger.17
The results of these studies demonstrated that between 21 and 29 percent of those workers exposed to an average noise level of 90 db would suffer material hearing loss. Further, 12 percent to 15 percent of workers exposed to an average noise level of 85 db would suffer similar impairment as would up to 5 percent of those workers exposed to 80 db. The results easily consti*1446tute substantial evidence to support the Agency’s conclusion that “even assuming compliance with the current occupational noise exposure standard, many workers will still be at increased risk of suffering material impairment of functional capacity from noise in the workplace.” 46 Fed.Reg. 4082 (January 16, 1981).
FIA counters that the retention of the 90 db PEL constitutes an Agency admission that exposure to noise below that level does not present a significant risk of material health impairment. This contention merits little discussion. Not only is the 90 db standard not written in stone, but the Agency demonstrated that the 85 db action level for implementing the hearing conservation measures logically should be lower than the maximum permissible noise level. Moreover, from the inception of the administrative proceedings, OSHA evidenced interest in lowering the PEL from 90 db to 85 db but hesitated pending “further empirical data and information about the health risk, feasibility, and economic impact indicatpng] the practicality and necessity of the 85 d[b][] requirement.” 39 FR 37773 (1974). In 1981, the Agency explained:
[F]or the present, OSHA will leave the permissible exposure level and compliance mechanisms of the • current noise standard unchanged and continue its enforcement. The Agency will defer the final decision on methods of compliance and the permissible exposure level until it has obtained and evaluated the necessary information.
46 Fed.Reg. at 4078 (1981). From this, it is apparent that the Agency is still studying the 90 db PEL for possible future amendment. Id. at 4105. See also 46 Fed.Reg. 42623 (August 21, 1981). Indeed, the studies previously cited constituted substantial evidence that 90 db is not a safe exposure level. Significantly, the legislative history of 29 U.S.C. § 655(a) states that “it is essential that such standards be constantly improved and replaced as new knowledge and techniques are developed.” S.Rep. No. 1282, 91st Cong., 2d Sess. 6, reprinted in 1970 U.S.Code Cong. & Ad.News 5177, 5182-83. In light of the foregoing, the temporary retention of the 90 db standard hardly diminishes the impact of the substantial evidence supporting the Agency’s determination that exposure to 85 db poses a significant risk of material hearing impairment.
V.
Likewise, substantial evidence exists to support the Agency’s conclusion that the requirements of the amendment will eliminate or reduce employees’ risk of suffering a material health impairment. Industrial Union Department v. American Petroleum Institute, 448 U.S. 607, 643, 100 S.Ct. 2844, 2864-2865, 65 L.Ed.2d 1010 (1980); United Steelworkers of America v. Marshall, 647 F.2d 1189, 1246 n. 85 (D.C.Cir.1980). The Agency’s conclusion that the amendment will reduce hearing loss resulted primarily from the analysis conducted by Bolt, Beranek, and Newman, Inc. (BBN), and the Center for Policy Alternatives (CPA).
The 1976 BBN study, commissioned by OSHA, analyzed, among other things, the expected benefits resulting from imposing a noise limit of 85 db rather than 90 db and estimated the number of industrial workers that would be protected against hearing impairments by that change. BBN figures indicated that as of 1975, approximately 19 percent of the production work force were exposed to sound levels above 90 db and approximately 34 percent were exposed to levels above 85 db. Using the results of the Baughn and Burns-Robinson studies, BBN concluded that the probability of a 40-year-old worker exposed to workplace noise of 90 db incurring a material hearing impairment was between 2.7 percent and 16 percent. The probability of such loss posed by a sound level of 85-90 db was between 1.5 percent and 12.0 percent. The BBN report further indicated that lowering the PEL from 90 db to 85 db, absent other requirements imposed by the amendment, would result in the protection of up to 243,900 additional workers. Finally, the BBN study concluded that a standard consisting of lowering the PEL from 90 db to *144785 db, requiring monitoring and audiome-tric tests for those employees exposed to an average of 85 db or more, and providing training and hearing protectors for exposed employees who experience an STS of 10 db would prevent material impairment in up to 324.000 employees over a 20-year exposure period.
The BBN results were confirmed in a related study performed by CPA, under contract to the Environmental Protection Agency, which concluded that a reduction in the PEL from 90 db to 85 db would save 396.000 workers from material hearing impairment caused by occupational noise over a 40-year working lifetime. OSHA revised and updated these estimates 18 and concluded that implementation of the amendment would reduce the number of hearing impairments by 212,000 in the tenth year, 477.000 in the twentieth year, 696,000 in the thirtieth year, 799,000 in the fortieth year, and 898,000 in the seventieth year.
VI.
Having found substantial evidence to support the Agency’s determination that there existed a significant risk of hearing loss to industrial employees and that the amendment, as a whole, will reduce that risk, we are next confronted with FIA’s argument that the individual requirements of the amendment are not “reasonably related to the purposes of the enabling legislation.” Mourning v. Family Publications Service, Inc., 411 U.S. 356, 369, 93 S.Ct. 1652, 1661, 36 L.Ed.2d 318 (1973); United Steelworkers of America v. Mar-shall, 647 F.2d 1189, 1237 (D.C.Cir.1980); see also 29 U.S.C. § 652(8).
The basic purpose of the legislation is to assure a safe environment in our industries. As the Supreme Court has said: “[W]hen Congress passed the Occupational Safety and Health Act in 1970, it chose to place pre-eminent value on assuring employees a safe and healthful working environment, limited only by the feasibility of achieving such an environment.” American Textile Manufacturers Institute v. Donovan, 452 U.S. 490, 540, 101 S.Ct. 2478, 2506, 69 L.Ed.2d 185 (1981). A review of the record convinces us that each of the requirements of the amendment are reasonably related to this congressional purpose.
A. Audiometric Testing
Subsections 1910.95(g) and (h) of the amendment, 29 C.F.R. §§ 1910.95(g), (h), require employers to establish and maintain an audiometric testing program for all employees exposed to an eight-hour TWA of 85 db or more. FIA argues that (1) an audiometric test is merely an indicator of hearing acuity and not a method of alleviating occupational noise levels, (2) testing performed at frequencies of 500, 1000 and 6,000 Hz is not relevant to the determination of hearing loss, (3) the requirement of section 1910.95(g)(1) making employers responsible for testing conducted off-premises is contrary to the rationale that employer responsibility is based on employer control over the workplace, (4) the requirement of section 1910.95(g)(5)(iv), that an employer notify employees of the need to avoid non-oceupational noise for fourteen hours prior to testing, is impermissible, and (5) the requirement that audiometric tests be performed annually, 29 C.F.R. § 1910.-95(g)(6), is not necessary because hearing remains constant for a period of three to five years. The record simply does not support these assertions.
The usefulness of audiometric testing in evaluating an employee’s hearing is obvious. Testing establishes an individual baseline standard to which an employee’s *1448hearing later may be compared in order to ascertain whether a hearing impairment is imminent. If hearing loss is not identified, ameliorative steps might never be taken. Additionally, audiometric testing is important to determine whether hearing protectors are operating properly. Although an audiometric test does not, in itself, alleviate hearing loss occasioned by noise exposure at an employee’s workplace, the requirement relates to the purpose of protecting an employee’s hearing.
FIA argues, nonetheless, that because the Agency defined an STS as a 10 db loss of hearing activity at frequencies of 2000, 3000, and 4000 Hz relative to the baseline audiogram, testing at 500, 1000 and 6000 Hz is unnecessary and not reasonably related to the purposes of the Act.
Substantial evidence, however, also supports the Agency’s decision to regulate in this manner. With respect to testing at 500 Hz, Dr. Gene Del Polito, Director of the Audiology Program of the American Speech and Hearing Association, explained:
500 Hertz can add information that is useful in detecting most medical problems affecting the function of the outer and middle ears. Detecting these problems by way of hearing measurement could provide an additional measure for assuring the safety and health of the worker. Also, industrial accidents resulting in trauma to the auditory mechanism, such as punctured eardrums or disarticulated ossicles, may affect hearing more in the low frequency range than in the higher frequencies.
This evidence supports the Agency’s determination that “[t]esting at all frequencies is important to have a complete record of the employee’s hearing ability and may be invaluable to the audiologist or physician upon referral.” 48 Fed.Reg. 9763 (March 8, 1983). Further, the Agency explained that “[h]earing loss at 500 Hz may be indicative of job related conductive hearing loss, which could possibly result from an adverse reaction to hearing protection, or from upper respiratory infections due to inhalation of irritating dust.” 46 Fed.Reg. 4147 (January 16, 1981).
Testing at 1000 Hz is also warranted in view of the fact that the Agency defined material impairment as an average loss that exceeds 25 db for the frequencies 1000, 2000, and 3000 Hz. Additionally, numerous groups and individuals, including NIOSH, the Academy of Ophthamology and Otolaryngology, Dr. Gage Miller of the Karolinska Institute in Stockholm, researcher William C. Sperry, the American Medical Association, the Environmental Protection Agency and the Center for Policy Alternatives, identified 1000 Hz as a suggested criterion for determining hearing impairments.
Similarly, record evidence demonstrates that testing at 6000 Hz furthers the purpose of the legislation. A 1973 report by the Environmental Protection Agency concluded that “[t]he typical pattern of NIPTS [Noise Induced Permanent Threshold Shift] seen in the audiogram is maximum loss in the range 4000 to 6000 Hz, with a somewhat smaller loss (initially) at higher and lower test frequencies.” Consequently, the Agency decided to incorporate the 6000 Hz frequency in the audiometric testing requirement as an early warning device sig-nalling the onset of hearing damage. Additionally, the Agency relied on evidence that individuals with significant hearing loss in the 4000-6000 Hz frequencies have difficulty hearing consonant sounds.
FIA next objects to the requirement of section 1910.95(g)(1) that employers provide audiometric testing, arguing that this places employers in the position of guaranteeing the accuracy and reliability of those who are not employees if the testing is performed at clinics or mobile vans. As we have indicated, audiometric testing is an integral aspect of the amendment and is reasonably related to the purposes of the Act. Simple logic requires that employers see to the effective administration of this requirement. No part of the amendment dictates that employee testing be performed away from the job site and the *1449decision whether to offer such tests at the workplace is solely that of the employer.
FIA also argues that the requirement that the employer notify employées of the need to avoid high levels of non-occupational noise in the fourteen hours proceeding the baseline audiometric examination, 29 C.F.R. § 1910.95(g)(5)(iv), is improper in that it makes the employer responsible for conditions' existing outside the workplace. We disagree. The baseline audiogram is important as the reference against which future audiograms are to be compared and we view this requirement as a reasonable demand designed to protect the validity of the baseline audiogram.
Finally, in this context, FIA asserts that the requirement of an annual audiogram is unreasonable, arguing that because hearing loss occurs gradually, there is no need to perform an audiogram that frequently. Although NIOSH and witnesses from Tenneco Co. recommended less frequent testing, numerous other organizations and witnesses, including Noise Pollution Consultants, Inc., and Dr. Larry Royster of the Center for Acoustical Studies of the North Carolina State University, commented that annual testing was appropriate to aid in identifying temporary threshold shifts before they become permanent and to prevent significant hearing loss in particularly susceptible individuals. On the basis of this evidence, OSHA concluded that annual audiograms were warranted. We find this policy decision to be supported by persuasive considerations, Industrial Union Department v. Hodgson, 499 F.2d 467, 476 (D.C.Cir.1974), and is reasonably related to the ameliorative purposes of the Act.
B. Monitoring Requirements
29 C.F.R. § 1910.95(d) requires an employer to monitor workplace noise whenever information indicates that an employee’s exposure may equal or exceed an eight-hour TWA of 85 db. Section 1910.-95(d)(2)(l) directs that noise at 80 db be integrated into the noise measurements. Section 1910.95(e) provides that the employer notify each employee exposed at or above the 8-hour TWA of 85 db of the results of the monitoring while § 1910.95(f) directs that employees or their representatives be provided with an opportunity to observe noise measurements.
FIA argues that each of these requirements is not reasonably related to the purposes of the Act. Again, we disagree. As the Agency explained, the inclusion of the 80 db level in the monitoring was compelled by the comments of representatives from the Institute of Noise Control Engineering and the American National Standards Institute’s Sound Level Meter Standards Working Group which indicated that the expected range of error for sound measuring instruments is 5 db and thus that setting the device at 5 db below the 85 db level would ensure the proper identification of affected workers.
The basis for the notification and observation requirements is found in 29 U.S.C. § 657(c)(3). It directs that, standards addressed to harmful physical agents “shall provide employees or their representatives with an opportunity to observe such monitoring or measuring, and to have access to records thereof” and that “[e]ach employer shall promptly notify any employee who has been or is exposed to ... harmful physical agents in concentra-tions or at levels which exceed those prescribed by an applic-able occupational safety and health standard promulgated under section 655.” Having concluded that noise is a harmful physical agent, we find that section 657(c)(3) mandates that employers notify employees of their exposure to excessive noise levels and provide an opportunity to observe monitoring. By including these requirements in the amendment, the Agency merely implemented the congressional determination that these requirements advance the purposes of the Act.
C. Standard Threshold Shift (STS)
A standard threshold shift is defined as “a change in hearing threshold relative to the baseline audiogram of an average of 10 db or more at 2000, 3000, and *14504000 Hz in either ear.” 29 C.F.R. § 1910.-95(g)(10). FIA argues that the 10 db shift is not reasonable in light of the fact that material impairment does not occur until the shift reaches 25 db. Here, too, this argument ignores the protective purpose of the amendment. The amendment is concerned with protecting workers before they sustain an irreversible 25 db shift. Consequently, it was incumbent upon the Agency to select a trigger level that would protect workers by providing an early warning yet not be so low as to be insignificant or within the range of audiometric error. We find that the Agency struck a reasonable balance between these concerns and adequately explained its reasoning.
D. Employee Information Requirements
The amendment also requires that employers inform all employees of the effects of noise on hearing, the purpose and characteristics of various types of hearing protectors, and the purpose and methods of audiometric testing. 29 C.F.R. § 1910.-95(k)(3) FIA contends that this requirement is unreasonable because it is not specifically directed towards those workers exposed to excessive noise.
This misreads the amendment. Section 1910.95(k) is entitled “Training program” and section 1910.95(k)(l) specifically states that the training program is limited to those “employees who are exposed to noise at or above an 8-hour time-weighted average of 85 db.” Thus, section 1910.95(k)(3) merely provides one aspect of that training program and, when read in conjunction with section 1910.95(k), indicates that it is limited to workers exposed to significant occupational noise.
E. Ensuring Compliance
Various provisions of the amendment state that employers shall ensure compliance with its requirements. See 29 C.F.R. § 1910.95(i)(2) (“ensure that hearing protectors are worn”); 29 C.F.R. § 1910.-95(i)(5) (“ensure proper initial fitting” of hearing protectors); 29 C.F.R. § 1910.-95(k)(3) (ensure that employees are informed of effects of noise, purpose and advantages of hearing protectors); § 1910.-95(k)(l) (ensure employee participation in training program.) FIA contends that these requirements are unreasonable and beyond the scope of OSHA’s authority.
The legislative history of the Act, however, recognized that “[ejmployers have primary control of the work environment and should insure that it is safe and healthful,” S.Rep. No. 91-1282, 91st Cong., 2d Sess., reprinted in 1970 U.S.Code Cong. & Admin.News 5177, 5196 (emphasis added). These provisions of the amendment are in harmony with this legislative concern. While it has been held that an employer may not be considered the guarantor of an employee’s compliance, e.g., General Electric Company v. OSHARC, 540 F.2d 67, 69 (2nd Cir.1976), we do not interpret the requirement that an employer ensure compliance to mean that the employer is a guarantor but rather that the employer must take all reasonable steps to accomplish the standard’s requirements, including imposing work rules, communicating the rules to employees, and providing training, supervision and disciplinary action designated to enforce the rules. See Barton’s, Inc., 10 O.S.H.C. (BNA) 1462, 1464 (1981). Despite these steps, if an employee disobeys the requirements of the standard, the employer has available the defense of “unforeseeable employee misconduct.” See, e.g., Ocean Electric Corp. v. Secretary of Labor, 594 F.2d 396 (4th Cir.1979). So viewed, we believe these requirements are reasonably related to the purpose of providing safe and healthful employment.
F. Eight-hour Time-Weighted Average
Under the amendment, measurement of noise exposure is expressed in terms of an 8-hour TWA of a stated level of decibels. FIA urges that the 8-hour criterion is unreasonable because hearing loss occurs only after prolonged exposure and will not result after only eight hours of exposure. The purpose of the 8-hour TWA, however, is to express noise exposure in a standardized form and as such is employed internationally. It is thus rea*1451sonably related to the function of the Agency, enabling OSHA to conform to accepted practice and communicate with the scientific community and the public.
G. Impulse Noise19
29 C.F.R. § 1910.95(d)(2)(i) provides that, in measuring noise for purposes of monitoring, “[a]ll continuous, intermittent, and impulsive sound levels from 80 decibels to 130 decibels shall be integrated into the noise measurements.” FIA objects to the combination of impulse noise and continuous noise on the ground that the Agency possessed insufficient information that impulse noise posed a significant risk of harm. However, scientific investigation has found that impulse noise can be extremely damaging. A study performed by Drs. Ceypek, Kuzniarz, and Lipowczan on 213 drop forge workers indicated that impulse noise combined with continuous noise produced a more rapid development of permanent hearing loss than continuous noise alone.20 Another study undertaken on a group of chinchillas by Drs. Hamernik, Henderson, Croisley, and Salvi indicated that while continuous noise exposure and impulse noise exposure, when measured separately, did not produce a permanent threshold shift, combination exposure produced sizeable amounts of shift. Further, a study performed by Drs. Perkins, Hamernik, and Henderson exploring the effects of the time interval between impulses on the magnitude of hearing loss found that the group with the shorter interval between exposures had a greater incidence of both permanent threshold shift and temporary threshold shift than did the group tested at longer intervals.
Relying on the synergistic effect of impulse noise superimposed over continuous noise as revealed by the above studies, the Agency determined that impulse noise and continuous noise should be combined in sound measurements rather than treated separately. We find this to be a reasonable exercise of Agency discretion based on findings supported by substantial evidence. In view of the evidence indicating the hazard of exposure to impulse noise, over and above exposure to continuous noise, the inclusion of impulse noise is reasonably related to the purpose of ensuring safe and healthful employment.
H. Employer Payment for Hearing Protectors
Section 1910.95(i)(l) of the amendment provides that employers must provide hearing protectors at no charge to all employees exposed to an 8-hour TWA of 85 db or more. 29 C.F.R. § 1910.95(i)(l). FIA asserts that this requirement is beyond OSHA’s authority, noting that who shall provide safety equipment is a traditional collective bargaining matter in which the Agency should not meddle. We disagree.
The legislative history clearly indicates “that Congress understood that the Act would create substantial costs for employers, yet intended to impose such costs when necessary to create a safe and healthful working environment.” American Textile Manufacturers Institute v. Donovan, 452 U.S. 490 at 519-20, 101 S.Ct. 2478, 2495-96, 69 L.Ed.2d 185 (1981). In view of this clear statement, it is only logical that OSHA may require employers to absorb such costs. FIA’s argument that this is a traditional area of collective bargaining is unpersuasive. As the court observed in United Steelworkers of America v. Mar-shall, 647 F.2d 1189 (D.C.Cir.1980):
In passing a massive worker health and safety statute, Congress certainly knew it was laying a basis for agency regulations that would replace or obviate worker safety provisions of many collective bargaining agreements. Congress may well have viewed collective bargaining *1452agreements along with state worker’s compensation laws as part of the status quo that had failed to provide workers sufficient protection.
Id. at 1236.
VII.
Section 6(b)(5) of the Act states that a standard imposed to deal with toxic materials or harmful physical agents shall be that “which most adequately assures to the extent feasible,” that no material health impairment will ensue. 29 U.S.C. § 655(b)(5). Although 29 U.S.C. § 652(8) does not by its terms require a feasibility analysis, the Supreme Court has stated that “any standard that [is] not economically or technologically feasible would a fortiori not be ‘reasonably necessary or appropriate’ under [29 U.S.C. § 652(8)].” American Textile Manufacturers Institute v. Donovan, 452 U.S. 490, 513 n. 31, 101 S.Ct. 2478, 2492 n. 31, 69 L.Ed.2d 185 (1981). Although feasibility in the § 655(b)(5) context has been held not to impose a cost-benefit analysis,21 American Textile Manufacturers Institute v. Donovan, id. at 512-13, 101 S.Ct. at 2492-93, the feasibility analysis involves an inquiry to determine whether a standard is both technologically and economically feasible. American Textile Manufacturers, 452 U.S. at 513 n. 31, 101 S.Ct. at 2492 n. 31. American Federation of Labor v. Mar-shall, 617 F.2d 636, 655 (D.C.Cir.1979).
A. Technological Feasibility,
Technological feasibility has been construed to mean that there exists “a reasonable possibility that the typical firm will be able to develop and install engineering and work practice controls that can meet [the standard] in most ... operations,” United Steelworkers of America v. Marshall, 647 F.2d at 1272; see also American Iron and Steel Institute v. OSHA, 577 F.2d 825, 834 (3d Cir.1978), not that “the standard [is] certainly feasible for all firms at all times in all jobs____” United Steelworkers v. Marshall, 647 F.2d at 1270 (emphasis in original). We must decide simply on the basis of the substantial evidence test whether the Agency has proven “a reasonable possibility that the typical firm will be able to develop and install work practice controls that can meet the [requirements of the amendment] in most of its operations.” United Steelworkers v. Marshall, 647 F.2d at 1272. Applying that test, we have little difficulty upholding the Agency’s determination that the amendment is technologically feasible. The amendment does not require any modification of plants or equipment, so no engineering problems are created. Hence, our review is limited to ascertaining whether there existed substantial evidence to support the conclusion that necessary equipment and personnel are available.
The sole argument raised by FIA regarding technological feasibility is that dosimeters capable of measuring noise at an 80 db level, as required by the amendment, are not currently available. This contention is not borne out by the record. A report by Dr. W. Burns published in 1975 stated that at that time there existed a dosimeter “able to handle steady, varying, intermittent and impulse noise and possessing] a dynamic range of 80 db____ The instrument is now available commercially, and has been in general use for some time.” A catalogue from Glen-Rad, a manufacturer of acoustic instruments, lists a “1983 Sound-Level Meter” capable of measuring levels from 70 db to 120 db. Other comments and information submitted to the Agency indicated that an adequate number of dosimeters, sound level meters, audiometers, and hearing protectors are available. As to the availability of necessary personnel, reports from the American Speech and Hearing Association (ASHA), the American Speech-Language-Hearing Association (ASLHA), the American Council of Otolaryngology *1453(ACO), and the Council for Accreditation in Occupational Hearing Conservation (CAOHC) indicated that sufficient qualified personnel are available to conduct audiome-tric testing.22 Clearly the Agency has established that the amendment’s requirements are technologically feasible.
B. Economic Feasibility.
A standard is economically feasible if the industry’s “long-term profitability” is not threatened. American Textile Manufacturer, 452 U.S. at 531 n. 55, 101 S.Ct. at 2501 n. 55. In reviewing the Agency’s determination that the standard will not impose insuperable financial burdens on the affected industries, however, we “do not review [the agency’s] cost figures de novo, but accord [the agency] discretion to arrive at a cost figure within a broad zone of reasonable estimate.” Weyerhaeuser v. Costle, 590 F.2d 1011, 1049 (D.C.Cir.1978). Our ultimate inquiry is whether OSHA has constructed “a reasonable estimate of compliance costs and demonstrate[d] a reasonable likelihood that these costs will not threaten the existence or competitive structure of an industry, even if it does portend disaster for some marginal firms.” United Steelworkers v. Marshall, 647 F.2d at 1272.
In estimating the compliance costs imposed by the amendment, the Agency quantified costs for five separate requirements, including monitoring ($18.7 million), au-diometric testing ($100.6 million), hearing protectors ($51.3 million), training ($30.2 million), and recordkeeping ($9.4 million), to reach a total annual cost of approximately $210.3 million dollars or $41.00 per worker. With respect to each component, the Agency exhaustively analyzed and evaluated data submitted by the industry and scientific community and fully explained the basis of its computations. From this, we conclude that these figures constitute a reasonable estimate of compliance costs that is amply supported by record evidence.
The Agency also demonstrated that the amendment poses no insurmountable burden to the continued viability of any of the affected industries. In January 1981 the Agency published its economic analysis of the proposed amendment, which concluded that if the costs of the proposed amendment were passed on to consumers, prices would increase by 0.0148 percent while if the industries absorbed compliance costs, they would expend only 0.1932 percent of their profits.23 Because the final amendment is significantly less expensive than the proposed amendment, $210.3 million as compared to $295.3 million, the Agency was justified in concluding that the final amendment is also economically feasible.
FIA does not take issue with the methodology employed or conclusions reached by the Agency. It argues, however, that the agency failed to quantify the cost of compliance with certain specific provisions of the amendment including the requirements for annual audiometric testing, that the employer ensure that employees wear hearing protectors and participate in a training program, and that employers provide affected employees or their representatives the opportunity to observe noise monitoring.
The Regulatory Impact Analysis indicated that the total cost of audiometric testing, including equipment and personnel, was $101.6 million annually. The Agency calculated the cost of audiometric testing *1454by multiplying the number of workers exposed to noise over 85 db by the hours away from work required to take the test. This figure was multiplied by the hourly industry labor cost. The estimated fee charged for the audiometric test also was added to the computation.24 OSHA further estimated that 5 percent of those workers tested would need to be referred to a physician for medical reasons at $35.00 per referral. The Agency also quantified the equipment costs necessary to institute in-house annual audiograms. These figures included $2,000 for a test booth, $1,630 for an audiometer and accessories, $250 plus three days of an employee’s time to obtain an operator’s certificate, and $100 to calibrate the equipment every other year. Adjusting all of these figures to account for the 20 percent of employees in the largest industries who already receive audiometric testing, the Agency arrived at an estimated annual cost of $100.6 million. We find the estimate to be well within a zone of reasonable estimate, a conclusion which FIA apparently concedes. The included cost of performing annual audiograms is also economically feasible.
Having concluded that ensuring compliance entails no more than taking reasonable steps to see that the amendment’s requirements are met, we believe that “ensuring” costs, if any, are subsumed under the costs for training, monitoring, testing, and providing hearing protectors, all of which were quantified by the agency and shown to be economically feasible.
FIA asserts that requiring an opportunity for employee observation poses the possible scenario where a few employees are being measured for noise while all other employees are exercising their right to observe, thereby adversely affecting production schedules. This argument is undercut by the Agency’s statement in the preamble to the amendment that the employee’s right to observe monitoring “must be tempered with a ‘standard of reasonableness,’ and observation which seriously disrupts production or the sampling itself is not permitted.” 46 Fed.Reg. 4156 (Jan. 18, 1981).25 In view of this clear interpretation, coupled with the mandatory observation and notification requirements of 29 U.S.C. § 657(c)(3), we find no reversible error in OSHA’s failure separately to assess the economic impact of employee observation of monitoring. Moreover, as the total cost of monitoring is only $18.7 million, the cost of employee observation must be only a fraction of that and therefore may be considered de minimis.
VIII.
Finally, intervenor National Arborist Association, Inc. (NAA) contends that the Agency arbitrarily and capriciously refused to exempt the tree care industry from compliance with the amendment.26 The basis of its argument is that OSHA decided to exempt the oil and gas drilling industry in view of factors such as variations in working conditions, highly mobile operations, high employee turnover rates, and limited accessibility to worksites, that also exist in the tree care industry. Consequently, NAA asserts, the tree care industry was arbitrarily denied equal treatment.
Even assuming that these factors do exist in both industries, we nonetheless hold that the Agency’s action was proper. The *1455Agency explained that this exemption was granted “because of the unique characteristics of the industry, which characteristics have also convinced the agency to initiate rulemaking that addresses the unique hazards for the gas and oil well drillers and servicers.” 48 Fed.Reg. 9775 (March 8, 1983). This exemption makes clear that, rather than escaping the requirements of the amendment, the oil and gas drilling industry is to be the subject of a comprehensive vertical standard designed to regulate numerous aspects of the industry including hearing conservation. See United Steelworkers v. Marshall, 647 F.2d at 1309-10. Indeed, on December 28, 1983, OSHA published a proposed rule embodying wide-ranging safety requirements for that industry. The Agency specifically presented a three-paragraph alternative27 to the hearing conservation amendment applicable to the industry and invited comments on the alternative as well as requesting other data regarding noise levels, monitoring, employee turnover rate, audiograms and present hearing conservation measures. See 48 Fed.Reg. 57,205-57,207 (December 28, 1983).
It is settled that “[njothing in the Act prevents the Agency from exercising discretion in delaying specific standards according to the unique problems of specific industries.” United Steelworkers v. Marshall, 647 F.2d at 1310; see also National Congress of Hispanic American Citizens v. Usery, 554 F.2d 1196, 1199 (D.C.Cir.1977); National Roofing Contractors Association v. Brennan, 495 F.2d 1294, 1299 (7th Cir.1974), cert. denied, 419 U.S. 1105, 95 S.Ct. 775, 42 L.Ed.2d 801 (1975); cf. Florida Citrus Packers v. California, 549 F.Supp. 213, 217 (N.D.Cal.1982). Here the Agency has determined that a particular industry should be made the subject of a vertical standard specifically addressed to that industry. That decision was not arbitrary or capricious under any definition of those terms.28 Nor does the use of a comprehensive vertical standard amount to a prohibited special treatment.
Conclusion
For the foregoing reasons, we uphold the Secretary of Labor’s promulgation of the hearing conservation amendment to the occupational noise exposure standard. Accordingly, the petitions are denied.
DENIED.
. Decibels are a measure of sound loudness. The entire spectrum of audible sound pressure can be compressed into a logarithmic scale of 0 to 140 db. Hertz ("Hz"), in contrast, measure the frequency of sound. Frequency is determined by the number of times that a complete cycle of compressions and expansions occurs in a second. The audible range of frequencies for people with good hearing is 20 Hz to 20,000 Hz.
. The time-weighted average combines noise level and duration of exposure to measure the accumulation of noise levels experienced by an employee over a workshift. OSHA computes the relationship between noise level and exposure time by using a 5 db "exchange rate." This means that for each 5 db increase in noise level, the exposure time must be cut in half. For example, an employee who works for 4 hours in continuous noise of 95 db would be exposed to an 8-hour TWA of 90 db, as would an employee exposed to 100 db for 2 hours. 46 Fed.Reg. 4080; 29 C.F.R. § 1910.95, Table G-16.
. Engineering controls involve modification of plant, equipment, processes, or materials to reduce noise; for example, adding a muffler to a vehicle. Administrative controls involve modification of work assignments to reduce employee exposure to noise; for example, rotating employees so that they work in noisy areas for a short time. 48 Fed.Reg. 7473.
. This broadly phrased requirement was declared unenforceably vague in Kropp Forge Company v. Secretary of Labor, 657 F.2d 119 (7th Cir.1981).
. Hearing loss is measured by an audiometer. Audiometers produce pure tones at specific frequencies (e.g„ 250, 500, 1000, 2000, 3000, 4000, 6000, and 8000 Hz) and at specific sound levels. The record of a given individual’s hearing sensitivity is called an audiogram. An audiogram shows hearing threshold level measured in decibels as a function of frequency in hertz. It indicates how intense or loud a sound at a given frequency must be before it can be perceived. Thus under the amendment, follow up measures are required whenever the quietest sound an employer can hear at 2000, 3000 and 4000 Hz is an average 10 db louder than it was when the baseline audiometric test was performed.
. A material hearing impairment occurs when an employee suffers an average hearing loss of 25 db for the frequencies 1000, 2000, or 3000 Hz. 46 Fed.Reg. 4083 (January 18, 1981).
. At that time, the Agency decided to defer consideration of lowering the permissible exposure limit standard from 90 db to 85 db pending further study and additional public comment. 46 Fed.Reg. 4071 (January 16, 1981).
. Section 655(f) provides that:
Any person who may be adversely affected by a standard issued under this section may at any time prior to the sixtieth day after such standard is promulgated file a petition challenging the validity of such standard with the United States court of appeals for the circuit wherein such person resides or has his principal place of business, for a judicial review of such standard____
*144229 U.S.C. § 655(f) (1982).
. The consolidated § 6(f) petitions were held in abeyance pending completion of this reopened rulemaking proceeding. These petitions were subsequently withdrawn.
. The Agency determined that the number of hearing impairments prevented would be at least 212,000 in the 10th year after implementation; 477,000 in the 20th year; 696,000 in the 30th year; 799,000 in the 40th year; and 898,-000 in the 70th year. 46 Fed.Reg. at 4112-4113.
. Dr. Baughn’s study, performed between 1960 and 1965, examined the effects of average noise exposures of 78 db, 86 db, and 92 db on 6,835 workers employed in Midwestern plants producing automobile parts. Interpolating his findings so as to provide estimates for exposures to 80 db, 85 db and 90 db, Dr. Baughn found that among workers 56 years of age, 25% of those exposed to an 8-hour TWA of 90 db suffered material hearing loss, while 13% of those exposed to an 8-hour TWA of 85 db suffered such loss and 4% of those exposed to an 8-hour TWA of 80 db.
. Dr. Passchier-Vermeer, in 1968, conducted a review of hearing loss as a function of exposure to average noise levels of 80 db through 102 db. Dr. Passchier-Vermeer's study consisted of an analysis and correlation of data provided by laboratory and field studies conducted by British, Dutch, Swedish, and American investigators. According to her analysis, employees exposed to 90 db over a working lifetime of 40 years suffered a median hearing loss of 16 db at 3000 Hz while those exposed to 85 db had a median loss of 8 db, and those exposed to 80 db suffered a 3.5 db loss. The results of Dr. Pass-chier-Vermeer’s study were presented in a report by Col. Daniel Johnson of the U.S. Air Force.
. This study of 759 British factory workers was conducted between 1963 and 1968. The study resulted in a mathematical formula designed to predict hearing losses in the 500-6000 Hz frequency range due to specific levels and durations of noise.
. The ISO Recommendation was derived from Dr. Baughn’s study.
. The NIOSH report utilized studies performed by Dr. Passchier-Vermeer, Drs. Taylor, Pearson, Moir, and Burns, and Drs. Burns, Hinchcliffe and Littler.
. This study, conducted on 228 Canadian industrial workers ranging from 18 to 65 years of age, related the degree of hearing loss to average noise levels of 85 and 90 db. The study group excluded non-occupational hearing loss and concluded that the risk of hearing loss at 500, 1,000, and 2,000 Hz increased significantly between 85 and 90 db, leaving up to 22 percent of the population at risk by a 90 db noise exposure standard.
. This study examined 42 men and 58 women working in North Carolina industry. The study concluded that the observed hearing losses were compatible with the data of Dr. Baughn, Dr. Passchier-Vermeer, Drs. Burns and Robinson, and NIOSH.
. OSHA’s revisions included:
1. Enlarging the size of the workforce to reflect current numbers.
2. Including loss of hearing through pres-bycusis (loss due to aging) in the calculations.
3. Determining a material loss of hearing as a 25 db loss at frequencies of 1000, 2000, and 3000 Hz.
4. Correcting the amount of hearing attenuation provided by hearing protectors.
5. Including retired workers or older workers who have been exposed to noise levels over a period greater than 20 years.
. Impulse noise is characterized by a rapid rise time, high peak value, short duration, and rapid decay.
. The report concluded that material hearing loss at 6000 Hz occurred after less than two years exposure. Next in frequency was loss at 4000 Hz which, after five years, equalled losses at 6000 Hz. Loss at 2000 Hz equalled approximately 10 db per year during the first two years and an additional 5 db in 10 years.
. Cost benefit analysis contemplates a "systematic enumeration of all benefits and all costs, tangible and intangible, whether readily quantifiable or difficult to measure, that will accrue to all members of society if a particular project is adopted.” American Textile Manufacturers, 452 U.S. at 507 n. 26, 101 S.Ct. at 2489 n. 26 (quoting E. Stokey & R. Zeckhauser, A Primer for Policy Analysis 134 (1978)).
. A 1975 report of the ASHA estimated that 3,500 audiologists existed at that time with the number expected to double in five years. A 1980 ASHA report indicated 6,052 current audiologist members, 2,000 additional licensees, and 1,600 students about to receive Masters’ or Doctoral degrees. ACO estimated that there were almost 5,000 practicing otolaryngologists as of 1973 while the CAOHC report indicated that approximately 2,000 trained technicians would be available.
. The Agency also compared the amount of money needed for compliance to cash on hand, outstanding short term loans, and net working capital in order to determine the ability of the affected industries to raise sufficient funds through normal commercial channels. This analysis revealed that the cost of the proposed amendment would be equivalent to 0.59% of cash on hand, 0.59% of short term debt, and 0.19% of working capital. 46 Fed.Reg. 4126 (Jan. 16, 1981).
. The fee charged per test was estimated by the Agency at $15.00 per worker for testing performed in a clinic, $10.00 per worker for testing performed at a mobile unit, and $2.50 per worker for in-house testing. The amount of production time lost was estimated at two hours for tests performed at a clinic, and one-half hour for mobile unit testing and in-house testing.
. FIA also observes that the benefit of employee monitoring is comparable to that accruing by allowing a layman to observe a medical operation to assure that it was done correctly. FIA thus invites us to undertake a cost-benefit analysis which, as noted above, we find unnecessary. See American Textile Manufacturers, 452 U.S. at 512-13, 101 S.Ct. at 2492-93.
. NAA relied on section 706(2)(A) of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A) (1982), which provides, in pertinent part, that "[t]he reviewing court shall ... (2) hold unlawful and set aside agency action, findings, and conclusions found to be — ... (A) arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”
. The alternative requires that employers conduct annual audiograms on their workers exposed to an 8-hour TWA of 85 db or more which shall be reviewed by a qualified otolaryngologist, audiologist, or physician and that the employers provide training and protective devices to employees.
. In this context we note that numerous of the provisions of the amendment were modified in response to those concerns noted by the tree care industry, including (1) permitting area sampling, and (2) extending the period within which to administer baseline audiograms to six months after hiring for clinic and in-house programs and to one year for mobile van testing.