Two remnants of the cotton dust rulemaking are presented to us following extensive judicial and administrative consideration of the regulations. See AFL-CIO v. Marshall, 617 F.2d 636 (D.C.Cir.1979), aff'd in part sub nom. American Textile Manufacturers Institute, Inc. v. Donovan, 452 U.S. 490, 101 S. Ct. 2478, 69 L. Ed. 2d 185 (1981); see generally 50 Fed.Reg. 51,123-25 (1985). The first challenge, pressed by the National Cottonseed Products Association (NCPA), concerns Occupational Safety and Health Administration (OSHA) prescriptions for medical surveillance of workers exposed to cotton dust. The second challenge, framed by Minnesota Mining and Manufacturing Company (3M), concerns OSHA's effectiveness rating for the disposable respirators that 3M manufactures. We conclude that OSHA acted within its statutory authority and on a rational basis; we therefore deny the petition for review.

NCPA raises a question as to the scope of the Supreme Court's holding in Industrial Union Department, AFL-CIO v. American Petroleum Institute, 448 U.S. 607, 100 S. Ct. 2844, 65 L. Ed. 2d 1010 (1980) [hereinafter Benzene ], that OSHA, in promulgating standards for toxic substances under Sec. 6(b)(5) of the OSH Act, 29 U.S.C. Sec. 655(b) (1982), must make a threshold finding of significant risk. We hold that this requirement is substantially modified when the sole requirement imposed is one of monitoring employee health, and that the Secretary's findings here are sufficient. NCPA also claims that the monitoring requirements are not feasible for the cottonseed industry; we reject the contention.

Section 6(a) of the OSH Act, 29 U.S.C. Sec. 655(a) (1982), authorizes OSHA to adopt any "national consensus standard" as one of its own. In 1971 OSHA exercised this power as to cotton dust, adopting the 1000 ug/m3 permissible exposure limit ("PEL") that had been promulgated under the Walsh-Healey Act, 41 U.S.C. Sec. 35(e) (1982). Section 6(b) of the OSH Act authorizes independent promulgation of standards, and in 1978 OSHA exercised that grant. Determining that the dust generated by cottonseed mills posed a material risk to cottonseed workers' health, it set a PEL of 500 ug/m³ and required employers to adopt medical surveillance programs. 43 Fed.Reg. 27,350 (1978). On appeal, this court agreed that exposure to cotton dust presented a material risk of harm, but remanded for reconsideration or further explanation of the standard's economic feasibility. AFL-CIO v. Marshall, 617 F.2d 636, 666-73 (D.C.Cir.1979), aff'd on other grounds sub nom. American Textile Manufacturers Institute v. Donovan, 452 U.S. 490, 101 S. Ct. 2478, 69 L. Ed. 2d 185 (1981). Thus, the 500 ug/m³ PEL has never taken effect but the 1000 ug/m³ limit has remained in place continuously since 1971.

On remand, the agency reconsidered both the need for dust regulations in the cottonseed industry and their feasibility. 47 Fed.Reg. 5906 (1982). During this rule-making new studies of the domestic cottonseed industry came to light, indicating that, contrary to the Secretary's previous findings, "excess byssinosis and bronchitis are not present among U.S. cottonseed workers." 50 Fed.Reg. 51,120, 51,135 (1985). However, the record also indicated that a subset of hypersensitive workers suffers from respiratory ailments, as do a very high percentage of workers in foreign cottonseed mills (where dust levels are much higher), and that the precise causal link between cotton dust and respiratory harm remained unknown. Id.

From these findings the Secretary determined that the risk of material harm to cottonseed workers would not be "significant" even without a PEL, so long as medical surveillance was retained as a "backstop." The backstop mechanism would protect hypersensitive workers and safeguard against risks stemming from the current inability to pinpoint the exact link between cotton dust and serious respiratory ailments. Id. at 51,135-36. Finding the facilities and personnel necessary for medical surveillance to be available at a cost that appeared trivial in relation to the industry's gross revenues, the Secretary concluded that medical surveillance was technologically and economically feasible and required it. Id. at 51,171.

NCPA alleges that the Secretary's failure to find that the current level of dust in cottonseed mills presents a significant risk to workers' health precludes him, under Benzene, from imposing any Sec. 6(b) standard, including one limited to medical surveillance. We disagree; an unusally precise dictum in Benzene applies to this case and sanctions the Secretary's determination.

In Benzene, the Court reviewed an OSHA regulation reducing the PEL for benzene from 10 parts per million (ppm) to one ppm. In lowering the standard OSHA never adduced any evidence that exposures at 10 ppm presented a risk to workers. Rather, OSHA took the position that it was entitled (and possibly obligated) to lower the PEL to the maximum extent feasible simply because benzene was a carcinogen for which no level of exposure had been proven absolutely harmless. Benzene, 448 U.S. at 652, 100 S.Ct. at 2869. Under OSHA's interpretation of the law, this regulatory power would be constrained only if industry established, apparently "beyond a shadow of a doubt," that there was a safe level of exposure. Id.

The Court strongly rejected the notion that OSHA is entitled to regulate any risk, no matter how small or speculative, to the limits of feasibility. The OSH Act, the Court held, empowers OSHA to regulate only hazards presenting a "significant risk" of material harm to workers' health.1 Thus, before OSHA could reduce the existing PEL, it had "the burden ... to show, on the basis of substantial evidence, that it is at least more likely than not that long-term exposure to 10 ppm of benzene presents a significant risk of material health impairment." Id. at 653, 100 S.Ct. at 2869.

Here OSHA found that abandoning both the 1000 ug/m³ and the 500 ug/m³ PELs would not leave workers exposed to a significant risk. Accordingly, NCPA contends that under Benzene OSHA may issue no regulatory restrictions at all. We believe, however, that Benzene dictates a significant risk analysis for monitoring requirements considerably laxer than for other workplace standards.

The Benzene Court considered the possibility that OSHA might impose a standard but remain uncertain whether the residual risk was significant. It made clear that OSHA could in such a case impose monitoring requirements as a "backstop," in order to (1) check the validity of its assumptions in imposing the standard selected, (2) develop a sound evidentiary basis for decreasing the limit if it proved to have been set too high, and (3) ensure that unusually susceptible workers could be removed from exposure before they suffered permanent damage. 448 U.S. at 657-58, 100 S.Ct. at 2871-72.

NCPA appears to acknowledge that this discussion is to be taken seriously despite its technical status as dictum. It claims, however, that the statement is applicable only if (1) the substance is toxic and "at some reasonably attainable level" causes harm, and (2) there is a relationship between worker exposure and health effects (a "dose-response" relationship). The Court's discussion does not impose either of these conditions, but we take NCPA in essence to argue that the Court's approval of monitoring would be senseless otherwise: why monitor if there is no chance of harm at levels that may come about? We think the evidence of risk before OSHA justifies application of the Court's dictum.

In assessing whether harm exists at any "reasonably attainable" level, OSHA is hampered by the ambiguous relation between reality and the pre-existing regulation. Although it is agreed that the industry has not complied with the 1000 ug/m³ standard (nor a fortiori that of 500 ug/m³), see 50 Fed.Reg. 51,13 3 and 48 Fed.Reg. 26,968, there is no concession that the regulations of the past 16 years have been absolutely without effect. See 50 Fed.Reg. 51,133, 51,136; 48 Fed.Reg. 26,968; 43 Fed.Reg. 27,381; see also Joint Appendix ("J.A.") at 311 (study finding that only two of 18 mills had mean exposures greater than 1000 ug/m³). Accordingly, OSHA could fairly infer that removal of the existing PEL (and cancellation of the abortive one of 500 ug/m³) could lead to higher levels of exposure, to the detriment of workers' health.

The record indicates that even the levels of exposure prevailing with the standard in effect pose some risk to workers' health. For example, a recent National Institute for Occupational Safety and Health ("NIOSH") test that is heavily relied upon by all parties found that although cottonseed workers did not experience increased incidence of byssinosis, as a group they suffered from decreased lung functions; the smokers among them were afflicted with chronic coughs. Respiratory Disorders and Dust Exposure in Sectors of the Cotton Industry of the United States. Part 3: Cottonseed Oil Mills v, J.A. at 307, 311. The series of studies by the Tulane Group reached similar conclusions, Jones, Hammond, Butcher & Weil, Respiratory Health in Cottonseed Crushing Mills, J.A. at 263-66, and also indicated that 52% of the workers studied were current smokers, 20% were ex-smokers, and 15% had allergies, id., J.A. at 264-65. The health organizations and physicians testifying on record overwhelmingly felt that medical surveillance was necessary. E.g., J.A. at 311 (NIOSH), 968-70 (Dr. Jones of Tulane Group), 978 (Dr. Merchant), 987 (Dr. Engleberg). (Indeed, with one prominent exception, Dr. Jones, they recommended that a PEL be retained.)

While it is true that no dose-response relation can be affirmatively established under conditions of American cottonseed processing, 50 Fed.Reg. 51,135, broader evidence supports its existence. Foreign cottonseed workers, who are subject to markedly higher doses than American ones, evidently suffer significant health effects. See, e.g., Noweir, El-Sadeh & El-Dahhahny, Exposure to Dust in the Cottonseed Oil Extraction Industry, 19 Arch. Environ. Health 99 (1969), J.A. at 154 (35 of 110 workers examined in Egyptian cottonseed plants exhibited byssinotic symptoms); Barnes & Simpson, Ventilatory Capacity Changes on Exposure to Cotton Dust, Med. J. of Australia 897 (May 25, 1968), J.A. at 159 (study undertaken in response to worker complaints of wheezing and tightness of the chest found link to cottondust exposure). See also 53 Fed.Reg. 51,135.

Because of this evidence, OSHA rested its finding of no significant risk on "the assurance that retention of medical surveillance will provide a backstop if that judgment is incorrect and this surveillance will protect the health of the employees."2 50 Fed.Reg. 51,136. Of course this cannot turn the finding of no significant risk into its opposite; the agency simply did not find evidence of significant risk. We do take the statement, however, to invoke the conditions suggested by the Supreme Court in Benzene for a backstop monitoring requirement: a finding that the "less-than-perfect" evidence indicates that there is a real possibility of significant health risks under the other aspects of the standard adopted (here, no regulation at all). See Benzene, 448 U.S. at 657-58, 100 S. Ct. at 2871-72.

OSHA's standards must be technologically and economically feasible, see, e.g., American Textile Manufacturers Institute v. Donovan, 452 U.S. at 513 n. 31, 101 S. Ct. at 2492 n. 31; NCPA contends that the medical surveillance requirements are neither.

The standard requires that cottonseed workers be given an initial examination and follow-up examinations every two years thereafter if the employee manifests no signs of respiratory problems and every six months if he does. 29 C.F.R. Sec. 1910.1043(h)(2), (3) (1986). Each examination involves the compilation or updating of the subject's medical history, the completion of a standardized questionnaire, and a pulmonary function test. A licensed physician must supervise the program (but need not conduct all facets of the examination in person, id. Sec. 1910.1043(h)(1)). The physician must also issue an opinion summarizing the results of the examination, stating whether the employee has any medical conditions which would place him at increased risk of health impairment from exposure to cotton dust, and recommending limitations to be placed on the employee's exposure. Id. Sec. 1910.1043(h)(5).

NCPA finds technological infeasibility in OSHA's failure to establish that the consulting services, local clinics, and in-shop medical centers expected to conduct the examinations are currently in place and ready to serve. Of course, this deficiency is overwhelmingly likely; resources are unlikely to be allocated to such an activity until the requirement attracts them there. Accordingly, it is no surprise that the law requires OSHA to demonstrate only that it is reasonable to expect that such "technology" will develop in response to the standard's promulgation. United Steelworkers v. Marshall, 647 F.2d 1189, 1264-65 (D.C.Cir.1980), cert. denied, 453 U.S. 913, 101 S. Ct. 3148, 69 L. Ed. 2d 997 (1981). As NCPA suggests no reason to suppose that it will not develop, the sole question is whether compliance with the standard is economically feasible. See id.

We have indicated that a standard is economically feasible if the cost of compliance does not threaten the "competitive structure or posture" of the industry. Industrial Union Department v. Hodgson, 499 F.2d 467, 478 (D.C.Cir.1974); see also United Steelworkers, 647 F.2d at 1264-65. Thus if compliance were likely to disable the industry from competing with substitute products, or markedly to increase concentration within the industry, a finding of infeasibility would be appropriate. See Industrial Union, 499 F.2d at 478.

OSHA estimated the total cost of complying with the standard at $70,671. Comparing this estimate to the industry's annual gross revenues of $777.6 million, OSHA concluded that imposition of the standard was economically feasible.

Even for evaluating the industry's overall prospects for survival, the method is a crude one. If demand for industry products were highly price elastic, a very small price increase could force closure of a substantial segment of the industry. Nonetheless, on the hypothesis that very high price elasticities are rare, it is surely appropriate for OSHA to infer that a cost amounting to a tiny fraction of gross revenues (on OSHA's estimates, less than .01%) will not force a material segment of the industry out of business.

NCPA accepts the general approach, but offers some complaints about OSHA's figures. OSHA's cost estimate of $70,671 is based on an average cost per examination of $79. In deriving this figure, OSHA started with the median estimated cost per examination identified by an NCPA survey of 36 cottonseed mills, $60, a figure consistent with all other estimates on record. See 50 Fed.Reg. 51,170. Reliance on a median estimate might often be misleading in assessing a standard's impact on industry structure. It might, for example, conceal a significant group of firms laboring under a special disability and likely to fail as a result of enforcement. NCPA indeed contends that such a group exists--small firms located far from suitable testing spots. In fact, however, NCPA's own data, submitted to OSHA as part of the rulemaking proceeding, show no significant link between above-median estimated examination costs and firm size or remoteness. See J.A. at 640-41. Further, the record contains some evidence that the highest cost estimates might be reduced in practice. See J.A. at 165. On these facts, accordingly, we are not persuaded that OSHA's focus on median figures caused it to overlook a likely impact on competitive structure.

OSHA then increased its base estimate by adding $15 (three hours at $5/hour³) to account for lost production and $4 to cover transportation costs,4 yielding a total cost per examination of $79. 50 Fed.Reg. 51,170-71. It then multiplied by 817, the number of jobs in the industry, reflecting an assumption that there would be an annual "new hire" for every job. This produced $64,543, to which OSHA added the cost of retesting those workers that remained employed in the industry for two years and arrived at the $70,671 total figure. With the minor exception noted below, OSHA has accounted for every element of expense likely to flow from its standard.

NCPA notes correctly that OSHA's cost estimate omits the cost of providing follow-up examinations every six months to workers who manifest symptoms of respiratory problems. As hypersensitive workers are a major reason for requiring medical surveillance, it was clear error for OSHA to have ignored these costs. Nonetheless, we find the error harmless. See Greater Boston Television Corp. v. F.C.C., 444 F.2d 841, 851 (D.C.Cir.1970), cert. denied, 403 U.S. 923, 91 S. Ct. 2229, 29 L. Ed. 2d 701 (1971). Given the industry's high turnover rate, only a fraction of these workers will actually require a follow-up exam. Moreover, if employers implement physicians' recommendations to place hypersensitive workers in less dusty jobs, symptoms of respiratory ailments should decline, alleviating the need for future follow-up exams. We see no reason to believe that OSHA's error on this point could have had more than a trivial effect on the cost of compliance relative to the industry's total revenues.

OSHA's regulations place ceilings on cotton dust concentrations to which workers may be exposed. See, e.g., 29 C.F.R. Sec. 1910.1043(c)(1)(i) (1986) (limiting cotton dust exposure to 200 ug/m³ in yarn manufacturing and cotton washing operations). If worker exposure exceeds the OSHA decreed PEL, or if employees wish to reduce their exposure below the PEL, the employer is obliged to furnish respirators. Id. Sec. 1910.1043(f)(1). Two factors together determine respirator effectiveness: filter efficiency and the "fit factor," i.e., the extent to which leakage occurs between the respirator face and seal and the wearer's face.

OSHA's effectiveness ratings (or "protection factors") are tied to particular respirator styles. See id. Sec. 1910.1043(f) ("supplied air respirators" have a rating of ten; "high efficiency particulate filter respirators with a full facepiece" have a rating of fifty). The protection factor indicates OSHA's estimate of the amount of cotton dust filtered. The higher the number, the more successful the filter; a rating of ten means only one of ten dust particles is not filtered. Thus, a respirator with a protection factor of ten will permit work in environments laden with cotton dust concentrations up to ten times the PEL.

In 1978, OSHA adopted a protection factor of five for "single-use respirators."5 The absence of a reliable test for proper fit, not filter efficiency, accounted for the low rating. See 43 Fed.Reg. 27,386 (1978).6 This 1978 rating was not challenged in court.

Between 1978 and 1983, 3M and other respirator producers developed a "disposable respirator" similar in construction (material and dimensional characteristics) to the single-use respirator.7 On the basis of filter efficiency, NIOSH rated this respirator at ten. With no guideline in the 1978 OSHA regulations explicitly covering the newly developed disposable respirators, the cotton industry apparently treated them as having a protection factor of ten.

During its review of the cotton dust regulations between 1983 and 1985, OSHA concluded that testing for snug fit on a daily basis remained infeasible for any respirator constructed like the single-use respirator, i.e., one in which the filter constitutes the entire surface area of the respirator.8 Finding no justification for treating single-use and later developed disposable respirators differently, the agency rated both at five. 3M urges that OSHA acted arbitrarily in refusing to set the effectiveness rating for disposable respirators at ten.

OSHA asserts initially that 3M lacks standing to petition for review. FAIC Securities, Inc. v. United States, 768 F.2d 352 (D.C.Cir.1985), appears to us dispositive of this threshold issue. Under that decision's analysis, 3M is a proper petitioner for judicial review.

In FAIC Securities, a deposit broker9 and a national trade association whose members include deposit brokers successfully challenged as unlawful certain Federal Home Loan Bank Board and Federal Deposit Insurance Corporation regulations. The regulations in question altered federal insurance coverage of $100,000 per depositor, per financial institution by adding this qualification: in the case of funds deposited by or through a deposit broker, insurance coverage would be limited to $100,000 per broker, per financial institution. The deposit brokers alleged that the challenged regulations contravened the Federal Deposit Insurance Act ("FDIA") and the National Housing Act ("NHA"). They would be put out of business by the altered regulations, the brokers remonstrated, and their customers consequently would be deprived of the benefits of placing deposits through a broker. The Federal Home Loan Bank Board, as defendant-appellant in FAIC Securities, contested the brokers' standing; the Board argued that the brokers failed the prudential "zone of interest" test announced in Association of Data Processing Service Orgs., Inc. v. Camp, 397 U.S. 150, 156, 90 S. Ct. 827, 831, 25 L. Ed. 2d 184 (1970),10 because " '[t]he NHA and the FDIA are intended to protect the security of depositors, banks and thrifts, not the profits of deposit brokers.' " FAIC Securities, 768 F.2d at 356 (quoting Brief for Appellant Bank Board at 52).

Writing for the court in FAIC Securities, then Judge (now Justice) Scalia endeavored to analyze coherently "the confused field of jus tertii standing." Id. at 360. He concluded that, under current Supreme Court precedent, notably City of Revere v. Mass. Gen. Hosp., 463 U.S. 239, 103 S. Ct. 2979, 77 L. Ed. 2d 605 (1983), Carey v. Population Services Int'l, 431 U.S. 678, 97 S. Ct. 2010, 52 L. Ed. 2d 675 (1977), and Craig v. Boren, 429 U.S. 190, 97 S. Ct. 451, 50 L. Ed. 2d 397 (1976), vendors could meet the prudential requirement even if they did not independently fulfill the "zone" test; it would do for this purpose if their customers or potential customers passed the test. FAIC Securities, 768 F.2d at 358.

The depositors, all agreed, fit within the protective zone of the NHA and FDIA, and the broker-depositor relationship fit the vendor-vendee description. Supreme Court decisions, Judge Scalia observed, treat the interests of vendors and vendees as "two sides of the same coin." Id. at 359. High Court precedent, he determined, allows vendors to base their standing on their relationship to vendees, and to assert the interest of those vendees, even if no impediment exists to a suit by the vendees themselves. Id. at 360-61; see Block v. Meese, 793 F.2d 1303, 1309 (D.C.Cir.1986) (citing with approval the analysis in FAIC Securities ).

In a more recent decision, Haitian Refugee Center v. Gracey, 809 F.2d 794, 811 & n. 13 (D.C.Cir.1987), a divided panel questioned the reasoning, although not the result, in FAIC Securities. Judge Bork, writing for himself and Judge Buckley in Haitian Refugee Center, thought the analysis in FAIC Securities flawed because Judge Scalia's opinion did not advert to the Supreme Court's decisions in United States v. Payner, 447 U.S. 727, 100 S. Ct. 2439, 65 L. Ed. 2d 468 (1980), and California Bankers Ass'n v. Shultz, 416 U.S. 21, 94 S. Ct. 1494, 39 L. Ed. 2d 812 (1974). See Haitian Refugee Center, 809 F.2d at 811.

Payner rejected third party invocation of the exclusionary rule. The decision rests on fourth amendment, not standing law, analysis. See Rakas v. Illinois, 439 U.S. 128, 132-38, 99 S. Ct. 421, 424-28, 58 L. Ed. 2d 387 (1978); United States v. Salvucci, 448 U.S. 83, 87 n. 4, 100 S. Ct. 2547, 2551 n. 4, 65 L. Ed. 2d 619 (1980) ("In Rakas, this Court discarded reliance on concepts of 'standing' in determining whether a defendant is entitled to claim the protections of the exclusionary rule. The inquiry, after Rakas, is simply whether the defendant's rights were violated by the allegedly illegal search or seizure."). Rakas explained that the exclusionary rule is but one form of remedy afforded for fourth amendment violations; denial of this remedy to those invoking the fourth amendment rights of others, the Court reasoned, was appropriate in view of the "substantial social cost [of keeping] [r]elevant and reliable evidence ... from the trier of fact and [deflecting] the search for truth at trial." 439 U.S. 128, 134, 137, 99 S. Ct. 421, 425, 427, 58 L. Ed. 2d 387 (1978). Cf. Rohr, Fighting for the Rights of Others: The Troubled Law of Third-Party Standing and Mootness in the Federal Courts, 35 U. Miami L.Rev. 393, 459-61 (1981) (litigant is generally not positioned to seek damages for violation of a third person's rights).

California Bankers Ass'n also involved a fourth amendment challenge.11 Moreover, even if one read that less than crystalline 1974 decision to deny that a vendor-vendee relationship is enough to permit third-party standing, pre-1976 High Court precedent, as Judge Scalia pointed out, has been overtaken by the Court's later decisions. See FAIC Securities, 768 F.2d at 359.

The discussion of FAIC Securities in Haitian Refugee Center was both brief and unessential to the majority's decision.12 The suggestion that Payner and California Bankers Ass'n undermine the reasoning in FAIC Securities does not appear compelling in light of the special fourth amendment contexts in which those two Supreme Court dispositions are embedded. See Monaghan, Third Party Standing, 84 Colum.L.Rev. 277, 279 n. 9, 292 n. 88, 305 n. 149 (1984); see also Rohr, supra, 35 U. Miami L.Rev. at 461 n. 290 (case law shows uniquely firm denial of third-party standing to invoke the "exclusionary rule" of criminal procedure). We therefore conclude that FAIC Securities continues to state law of the circuit, binding upon us unless and until changed by the court sitting en banc, or shown to be incorrect by instruction from Higher Authority.13 If the FAIC Securities deposit brokers' and depositors' interests are "two sides of the same coin," 768 F.2d at 359, so too are 3M's interest in selling the disposable respirators it manufactures, and cotton processing plant operators' interest in purchasing the respirators.14 If the brokers had standing in FAIC Securities, then 3M has standing here; no tenable distinction can be drawn between the relationship of the litigant and third party in the two cases. Following FAIC Securities, we are constrained to recognize 3M's standing on the basis of "the vendor-vendee relationship alone." Id. at 361.15

3M features three bases for declaring arbitrary OSHA's assignment, in its cotton dust regulations, of a protection factor of only five to disposable respirators. First, 3M asserts that feasible tests of proper fit, if unavailable in 1978, are available now. Second, 3M points to the higher (ten) rating accorded disposable respirators by national standard-setting organizations. Third, and most weighty in 3M's presentation, OSHA itself, in standards the agency adopted for the lead industry in 1982, approved establishment of a protection factor of ten for disposable respirators. We examine each 3M position in turn.

Two procedures, the saccharin QLFT16 and the positive pressure fit check (PPFC),17 3M states, are available to test disposable respirators for face-fit; each, 3M contends, is independently adequate to do the job. If 3M were right about the adequacy of these tests, we would be obliged to rule in its favor; the sole reason OSHA gave for rating disposable respirators at five rather than ten is the inability of the cotton plant operator and worker to check reliably for proper fit.

OSHA acknowledges that the saccharin QLFT will detect improper fit. But the test is not proffered by 3M as one the plant operator will employ for each worker each day. As 3M conceded at oral argument, the saccharin test is intended for use every three or six months; the test, administered at these intervals, checks for alteration in a wearer's facial contours that might affect the fit of the respirator. OSHA observed that "it is not appropriate to require the employers to conduct the saccharin QLFT each time the respirator is worn since it is time consuming...." 50 Fed.Reg. 51,154 (1985). Unsurprisingly, 3M does not press for such a requirement, one likely to increase the cost, and reduce the attractiveness, of its product to employers.

The PPFC procedure is an effective daily check for the fit of a gas-mask style respirator. Respirators of that type confine intended air intake to valves that can be blocked off easily by the employee's hands. By contrast, the entire surface of a disposable respirator is intended to permit air intake. OSHA recognized that, in the case of disposable respirators, the worker's hands cannot effectively block intended air intake, and that intake only, while leaving unobstructed air taken in because of the respirator's improper fit. See supra note 6.18 We think it evident that OSHA did not rule without reason when it adhered to the view that no test appropriate for daily use adequately assured the proper fit for disposable respirators.

We consider next 3M's second point. Both NIOSH and the Committee on Safety Standards for Respiratory Protection of the American National Standards Institute ("ANSI") have rated disposable respirators at ten. But neither organization took account of respirator fit; both based their ratings on filter performance. Absent assurance of a respirator's proper fit, the NIOSH and ANSI ratings can reliably indicate only the efficiency of the filter, not the effectiveness of the entire respirator as it is used on the job.19 In 1978, OSHA rated single-use respirators at five because of the considerable risk of undetected leakage when worn at work. See 43 Fed.Reg. 27,386 (1978).20 With no evidence of significant improvement in a user's ability to thwart leakage, OSHA continued in force its prior judgment. See 50 Fed.Reg. 51,154 (1985) ("A protection factor of 5 for the class of disposable dust and mist respirators is the appropriate protection factor to provide an adequate margin of safety to overcome the fitting problem.").21 We see no necessity for change based on ratings of filters rather than disposable respirators in their entirety.

Finally, we evaluate the challenge to which 3M attaches greatest weight. While disposable respirators are rated at five for use in the cotton industry, the very same respirators can be rated at ten in the lead industry, 3M emphasizes. OSHA explains that this disparate treatment is attributable primarily to differences in the effectiveness of medical screening in the two industries. Medical screening for lead adequately substitutes for daily fit testing, OSHA maintains, but medical screening for lung diseases from cotton dust does not. Cf. 51 Fed.Reg. 22,735 (1986) (to be codified at 29 C.F.R. Sec. 1910.1001(g)(2)) (disposable respirators declared inadequate, hence unusable, as a means of affording any protection against exposure to asbestos).

Blood lead tests provide an immediate and direct indication of the presence of lead in a worker's body. Although the source of the lead detected (for example, whether it was inhaled as a result of an improperly fitted respirator) is not identifiable, a worker can be safeguarded promptly. He can be supplied with a gas-mask style respirator or removed from a lead-laden environment immediately upon screening, before adverse health effects develop. Workers in the lead industry are tested at six month intervals. OSHA determined that six months' exposure to lead between blood tests presents an acceptable health risk, and we have no reason, on the record before us, to question that determination. Because of the efficacy of monitoring blood-levels, OSHA found it suitable to allow disposable respirators, when used in the lead industry, a protection factor rating of ten.

By contrast, current medical screening of workers is incapable of detecting cotton dust inhalation immediately and unmistakably. Medical tests in the cotton industry rely, in part, on worker identification of symptoms. The administrative record here suggests that workers do not always acknowledge the presence of symptoms. J.A. at 49, 187. Further, the tests detect only actual lung impairment. Medical screening does not isolate cotton dust from other airborne contaminants that cause lung impairment; more significantly, medical tests expose the potential presence of a contaminant only by detecting actual adverse health effects. The administrative record does indicate that early manifestations of lung impairment are reversible. At least some health impairment, however, even if curable, develops before a cotton worker furnished with a disposable respirator can be informed of her plight through medical screening, and thus be alerted to the need to substitute a gas-mask style respirator or to transfer to a work station away from dusty areas.22

OSHA, we conclude, has adequately accounted for its position. We cannot set aside as irrational that expert agency's determination that the precision of medical testing for lead in comparison to the imprecision of testing for cotton dust inhalation warrants the differential rating of disposable respirators in the two industries.

For the reasons stated, the petitions for review are denied and the challenged regulations are affirmed.

It is so ordered.