dissenting.
The reason Congress attached “the most stringent criminal liability,” ante, at 283, to the violation of an emission standard for a “hazardous air pollutant” is that substances within that narrow category pose an especially grave threat to human health. That is also a reason why the Court should avoid a construction of the statute that would deny the Administrator the authority to regulate these poisonous substances effectively.
*294The reason the Administrator did not frame the emission standard for asbestos in numerical terms is that asbestos emissions cannot be measured numerically. For that reason, if Congress simultaneously commanded him (a) to regulate asbestos emissions by establishing and enforcing emission standards and (b) never to use any kind of standard except one framed in numerical terms, it commanded an impossible task.
Nothing in the language of the 1970 statute, or in its history, compels so crippling an interpretation of the Administrator’s authority. On the contrary, I am persuaded (1) that the Administrator’s regulation of asbestos emissions was entirely legitimate; (2) that if this conclusion were doubtful, we would nevertheless be required to respect his reasonable interpretation of the governing statute; (3) that the 1977 Amendments, fairly read, merely clarified his pre-existing-authority; and (4) that the Court’s reading of the statute in its current form leads to the anomalous conclusion that work-practice rules, even though properly promulgated, are entirely unenforceable. Accordingly, although I agree with the conclusions reached in Parts I, II, and III of the Court’s opinion, I cannot accept Part IV’s disposition of the most important issue in this case.1
I
The regulation which petitioner is accused of violating requires that asbestos insulation and fireproofing in large *295buildings be watered down before the building is demolished.2 The effect of the regulation is to curtail the quantity of asbestos which is emitted into the open air during demolition. Because neither the rule nor its limiting effect is expressed in numerical terms, the Court holds that the asbestos regulation cannot be a “standard” within the meaning of § 112 (b)(1) of the Clean Air Act.3 This conclusion is not compelled by the use of the word “standard” 4 or by Congress’ expectation *296that standards would normally be expressed in numerical terms; for the statute contains no express requirement that standards always be framed in such language. The question is simply whether § 112 (b), which directs the Administrator to adopt regulations establishing emission standards for hazardous air pollutants, granted him the authority to promulgate the asbestos standard challenged in this case.
Section 112 is concerned with a few extraordinarily toxic pollutants. Only three substances, including asbestos, have been classified as “hazardous air pollutants” within the meaning of § 112.5 These pollutants are subject to special federal regulation. In § 112, Congress ordered the Administrator to identify and to regulate them without waiting for the States to develop implementation plans of their own. Thus, the procedure under § 112 contrasts markedly with the more leisurely'and decentralized process of setting and enforcing the general ambient air standards.6 Congress was gravely concerned about the poisonous character of asbestos emissions when it drafted § 112.7 In fact, with regard to the hazardous air pollutants covered by this section, Congress expressed its willingness to accept the prospect of plant closings: “The standards must be set to provide an ample margin of safety to protect the public health. This could mean, effectively, that a plant would be required to close because of the absence of control techniques. It could include emission standards which allowed for no measurable emissions.” 8
*297In accord with Congress’ expectation, the Administrator promptly listed asbestos as a hazardous air pollutant,9 and published a proposed emission standard. As first proposed, the standard would have prohibited any visible emission of asbestos in connection with various activities, including the repair or demolition of commercial and apartment buildings.10
If that total prohibition had been adopted, it unquestionably would have conformed to the statutory mandate. It was not adopted, however, because industry convinced the Administrator that his proposal would prevent the demolition of any large building.11 At public hearings it was demonstrated that *298demolition inevitably causes some emission of particulate asbestos and, further, that these emissions cannot be measured. Accordingly, instead of the severe numerical standard of zero emissions — which might have put an entire industry out of business — the Administrator adopted a standard which would reduce the emission of asbestos without totally prohibiting it. Not a word in the Administrator’s long and detailed explanation of the standard indicates that anyone questioned his statutory authority to promulgate this type of emission standard.12
The promulgated standard is entirely consistent with congressional intent. Congress had indicated a preference for numerical emission standards.13 Congress had also expressed a willingness to accept the serious economic hardships that a total prohibition of asbestos emissions would have caused. But there is no evidence that Congress intended to require the Administrator to make a choice between the extremes of closing down an entire industry and imposing no regulation on the emission of a hazardous pollutant; Congress expressed no overriding interest in using a numerical standard when industry is able to demonstrate that a less drastic control tech-*299ñique is available,14 and that it provides an ample margin of safety to the public health.15
Admittedly, Congress did not foresee the Administrator’s dilemma with precision. But there is nothing unique about that circumstance. See, e. g., Mourning v. Family Publications Serv., Inc., 411 U. S. 356, 372-373. Indeed, there’ would be no need for interstitial administrative lawmaking if Congress could foresee every ramification of laws as complex as this.16 I am persuaded that the Administrator’s solution *300was faithful to his statutory authority and that he would have misused his power if he had either failed to regulate asbestos emissions at all or unnecessarily demolished an entire industry.
II
The precise question presented to this Court is not whether, as an initial matter, we would regard the asbestos regulation as an “emission standard” within the meaning of § 112. Rather, the issue is whether the Administrator's answer to the question of statutory construction is “sufficiently reasonable that it should have been accepted by the reviewing courts.” Train v. Natural Resources Defense Council, 421 U. S. 60, 75.
The Administrator, who has primary responsibility for carrying out the purposes of the Clean Air Act, interpreted the term “emission standard” to include the rule before us. Contrary to the Court’s implication, ante, at 287, the Administrator did not promulgate this rule “instead” of an emission standard. He unambiguously concluded that the rule was a proper emission standard.17
*301Because the statute is the Administrator’s special province, we should not lightly set aside his judgment. “When faced with a problem of statutory construction, this Court shows great deference to the interpretation given the statute by the officers or agency charged with its administration. 'To sustain the Commission’s application of this statutory term, we need not find that its construction is the only reasonable one, or even that it is the result we would have reached had the question arisen in the first instance in judicial proceedings.’ ” Udall v. Tallman, 380 U. S. 1, 16.18
*302The Administrator began the process of promulgating this rule within weeks of § 112’s enactment, 36 Fed. Reg. 23242 (1971). The wise teaching of Mr. Justice Cardozo, who spoke for the Court in Norwegian Nitrogen Co. v. United States, 288 U. S. 294, is therefore directly pertinent. He observed that an administrative “practice has peculiar weight when it involves a contemporaneous construction of a statute by the men charged with the responsibility of setting its machinery in motion, of making the parts work efficiently and smoothly while they are yet untried and new.” Id., at 315.
The Court holds that these well-established doctrines apply only in “ordinary circumstances.” Ante, at 288. I do not understand why thése rules of construction should be less applicable in the unusual than in the ordinary case. Indeed, it seems to me that the extraordinary importance of regulating a hazardous air pollutant in a way that is both fair and effective provides an additional reason for respecting the Administrator’s reliance on well-established doctrine, rather than a reason for reaching out to undermine his authority.19
In the Court’s view, however, the enactment of amendments to the Clean Air Act in 1977 was an extraordinary circum*303stance that justifies a departure from settled principles. The Court takes the novel position that the Administrator’s construction of the 1970 Amendments may be ignored because the legislative history of the 1977 Amendments did not produce an explicit endorsement of his construction. In my judgment this holding places an unwise limit on the deference which should be accorded to administrators’ interpretations of the statutes they enforce. It also misreads the history of the 1977 Amendments.
Ill
The Court’s conclusion ultimately rests on the 1977 Amendments. Even accepting the dubious premise that we can rely on the 95th Congress to tell us what the 93d had in mind, the 1977 Amendments do not support the Court’s interpretation of the statute.
The history of the Amendments is instructive. In late 1974, several wrecking companies successfully challenged indictments brought against them in the Northern District of Illinois for violating the wetting requirements.20 Six weeks after the first court ruling, the Administrator proposed an amendment that would expressly confirm his authority to establish design, equipment, or work-practice standards when numerical emission limitations were not feasible.21 A major bill to amend the Clean Air Act was proposed in the 94th Congress, but the House and Senate were unable to agree. In 1977, the Senate again proposed a major revision. It included the Administrator’s requested authorization. S. Rep. *304No. 95-127, p. 163. The Senate Report does not indicate whether the Senators considered the Illinois decisions correct or incorrect. Id., at 44. However, as introduced in the Senate, the bill clearly provided that a design, equipment, or operational standard was a species of “hazardous emission standard.” 22
When the bill emerged from conference, it no longer expressly stated that a work-practice rule was an emission standard. This change therefore lends support to the Court’s view. But it is most unlikely that the Conference Committee intended to express indirect disapproval of the Administrator’s reading of the 1970 Amendments. The Conference Report explained that the change in language was merely intended to “clarify” an aspect of the Senate version which was unrelated to the question whether a work-practice rule is, or had been a species of emission standard.23
*305There is only one relevant lesson that may be learned from this history: As soon as someone challenged the Administrator’s power to promulgate work-practice rules of this sort, Congress made it unambiguously clear that the Administrator had that power. As the Court notes, Congress preferred numerical standards; it accepted work-practice rules only as a last resort. But the same may be said of the Administrator, who instituted a wetting requirement only after becoming convinced that no other standard was practicable.
It is true, as the Court says, that the Senate Report “refrained from endorsing the Administrator’s view that the regulation had previously been authorized as an emission standard under § 112 (c).” Ante, at 289. It is equally true that the Senate Report refrained from criticizing the Administrator’s view. In short, what Congress said in 1977 sheds no light on its understanding of the original meaning of the 1970 Amendments. But what Congress did when it expressly authorized work-practice rules persuasively indicates that, if Congress in 1970 had focused on the latent ambiguity in the term “emission standard,” it would have expressly granted the authority that the Administrator regarded as implicit in the statute as written.24
*306IV
A reading of the entire statute, as amended in 1977, confirms my opinion that the asbestos regulation is, and since its promulgation has been, an emission standard. If this is not true, as the Court holds today, it is unenforceable, and will continue to be unenforceable even if promulgated anew pursuant to the authority expressly set forth in the 1977 Amendments.
The Clean Air Act treats the Administrator’s power to promulgate emission standards separately from his power to enforce them. While it is § 112 (b) that gives the Administrator authority to promulgate an “emission standard,” it is § 112 (c) that prohibits the violation of an “emission standard.” Presumably the Court’s holding that a work-practice rule is not an “emission standard” applies to both of these sections. Under that holding a work-practice rule may neither be enforced nor promulgated as an emission standard. This holding will not affect the Administrator’s power to promulgate work-practice rules, because the 1977 Amendments explicitly recognize that power. But Congress has not amended § 112 (c), which continues to permit enforcement only of “emission standards.” Accordingly, the Court’s holding today has effectively made the asbestos regulation, and any other work-practice rule as well, unenforceable.
Ironically, therefore, the 1977 Amendments, which were intended to lift the cloud over the Administrator’s authority, have actually made his exercise of that authority ineffectual. This is the kind of consequence a court risks when it substitutes its reading of a complex statute for that of the Administrator charged with the responsibility of enforcing it. More*307over, it is a consequence which would be entirely avoided by recognizing that the Administrator acted well within his statutory authority when he promulgated the asbestos regulation as an “emission standard” for hazardous air pollutants.
I would affirm the judgment of the Court of Appeals for the Sixth Circuit.
Nor can I join Mr. Justice Stewart’s opinion, because he does not explain what test he applies to determine that § 307 (b) precludes any challenge to the asbestos regulation in an enforcement proceeding. The preclusion provision applies only if the Administrator’s action could have been reviewed in the Court of Appeals for the District of Columbia Circuit; and review was not available there unless the Administrator’s “action” was the promulgation of an “emission standard” within the meaning of §307 (b). In short, Mr. Justice Stewart’s dissent rests either on the unarticulated premise that the asbestos regulation was an “emission standard” under §307 (b), or on the application of a test not to be found in the language of the statute.
The emission standard for asbestos provides, in pertinent part:
"(i) Friable asbestos materials, used to insulate or fireproof any boiler, pipe, or load-supporting structural member, shall be wetted and removed from any building, structure, facility, or installation subject to this paragraph before wrecking of load-supporting structural members is commenced. The friable asbestos debris shall be wetted adequately to insure that such debris remains wet during all stages of demolition and related handling operations.” 40 CFR §61.22 (d) (2) (i) (1975).
Section 112 (b)(1) provides:
“(A) The Administrator shall, within 90 days after the date of enactment of the Clean Air Amendments of 1970, publish (and shall from time to time thereafter revise) a list which includes each hazardous air pollutant for which he intends to establish an emission standard under this section.
“(B) Within 180 days after the inclusion of any air pollutant in such fist, the Administrator shall publish proposed regulations establishing emission standards for such pollutant together with a notice of a public hearing within thirty days. Not later than 180 days after such publication, the Administrator shall prescribe an emission standard for such pollutant, unless he finds, on the basis of information presented at such hearings, that such pollutant clearly is not a hazardous air pollutant. The Administrator shall establish any such standard at the level which in his judgment provides an ample margin of safety to protect the public health from such hazardous air pollutant.
“(C) Any emission standard established pursuant to this section shall become effective upon, promulgation.” 84 Stat. 1685, 42 U. S. C. § 1857c-7 (b)(1).
There is no semantic reason why the word “standard” may not be used to describe the watered-down asbestos standard involved in this case. Indeed, the Court itself has previously identified a “watered down standard” that is not expressed in numerical terms, see Benton v. Maryland, 395 U. S. 784, 796.
See 40 CFR § 61 (1975).
Compare § 112, 42 U. S. C. § 1857c-7, with §§ 109 and 110, 42 U. S. C. §§ 1857c-4 and 1857c-5 (1970 ed. and Supp. V).
See, e. g., National Air Quality Standards Act of 1970, S. Rep. No. 91-1196, p. 20 (1970).
This statement was made in a written summary of the conference agreement presented by Senator Muskie to the Senate, which then agreed to the Conference Report. Summary of the Provisions of Conference Agreement on the Clean Air Amendments of 1970, reprinted in Senate Committee on Public Works, A Legislative History of the Clean Air *297Amendments of 1970, 93d Cong., 2d Sess., 133 (Comm. Print 1974). See also id., at 150.
36 Fed. Reg. 5931 (1971). The three hazardous air pollutants — -asbestos, beryllium, and mercury — listed by the Administrator on March 29, 1971, were all identified in the legislative history.
The Administrator’s investigation fully supported Congress’ suspicion that asbestos was an intolerably dangerous pollutant. Among other risks, even low-level or intermittent exposure to- asbestos can cause cancer 20 or 30 years after the event. 38 Fed. Reg. 8820 (1973). For example, a form of cancer usually found almost exclusively in asbestos workers killed a woman whose only contact with the pollutant was washing the workelothes of her children, who worked for an asbestos company. See Horvitz, Asbestos and Its Environmental Impact, 3 Environmental Affairs 145, 146 (1974).
“(d) Visible emissions to the atmosphere of asbestos particulate matter resulting from the repair or demolition of any building or structure, other than a single-family dwelling are prohibited.” 36 Fed. Reg. 23242 (1971).
The Administrator explained:
“The proposed standard would have prohibited visible emissions of asbestos particulate material from the repair or demolition of any building or structure other than a single-family dwelling. Comments indicated that the no visible emission requirement would prohibit repair or demolition in many situations, since it would be impracticable, if not impossible, to do such work without creating visible emissions. Accordingly, the promulgated standard specifies certain work practices which must be followed when demolishing certain buildings or structures. The standard covers institutional, industrial, and commercial buildings or structures, including apartment houses having more than four dwelling units, which contain friable asbestos material.” 38 Fed. Reg. 8821 (1973).
There was no review of the emission standard for asbestos in the United States Court of Appeals for the District of Columbia Circuit. An untimely-petition for review was dismissed without any decision on the merits. Dore Wrecking Co. v. Fri, No. 73-1686 (CADC, Aug. 1, 1973). Contrary to the implication in n. 2 of the Court’s opinion, this case does not raise any question about fair notice to small businesses. The wrecking company prosecuted here was individually notified about the wetting requirement and individually responded to the notice by promising to comply fully with the regulation on all future jobs. Indeed, the company’s response specifically named the location, where, according to the indictment, it subsequently committed a knowing violation of the regulation.
Congress apparently believed that too frequent resort to work-practice rules or equipment specifications would discourage the private market’s pursuit of “the most economic, acceptable technique to apply.” S. Rep. No. 91-1196, at 17.
A summary of the conference agreement states that § 112 “could mean, effectively, that a plant would be required to close because of the absence of control techniques.” See text accompanying n. 8, supra. This statement implies that the Administrator should avoid setting emission standards that will require plant closings if alternative control techniques — including work-practice rules — can, provide an ample margin of safety. It is unlikely that Congress intended, by expressing a modest preference for numerical standards, see n. 11, supra, to mandate plant closings under a numerical standard when a work-practice rule would achieve the same level of protection with less economic disruption.
[T]he Administrator has determined that, in order to provide an ample margin of safety to protect the public health from asbestos, it is necessary to control emissions from major man-made sources of asbestos emissions into the atmosphere, but that it is not necessary to prohibit all emissions.” 38 Fed. Reg. 8820 (1973).
In Gemsco, Inc. v. Walling, 324 U. S. 244, this Court approved a much more dubious substitute for a regulation that Congress surely expected to be framed in numerical terms. In that case the Administrator of the Fair Labor Standards Act decided to ban industrial homework as a way of enforcing the minimum wage. If homework were allowed to continue, the Administrator concluded, industry could readily evade wage standards. Although the Administrator lacked any express authority to regulate industrial homework, this Court approved his action, saying: “The industry is covered by the Act. This is not disputed. The intent of Congress was to provide the authorized minimum wage for each employee so covered. Neither is this questioned. Yet it is said in substance that Congress at the same time intended to deprive the Administrator of the only means available to make its mandate effective. The construction sought would make the statute a dead letter in this industry.
“The statute itself thus gives the answer. It does so in two ways, by *300necessity to avoid self-nullification and by its explicit terms. The necessity should be enough. But the Act’s terms reinforce the necessity’s teaching. Section 8 (d) requires the Administrator to ‘carry into effect’ the committee’s approved recommendations. Section 8 (f) commands him to include in the order ‘such terms and conditions’ as he ‘finds necessary to carry out’ its purposes. . . . When command is so explicit and, moreover, is reinforced by necessity in order to make it operative, nothing short of express limitation or abuse of discretion in finding that the necessity exists should undermine the action taken to execute it.” Id., at 254-255.
In the present case, necessity also demanded the promulgation of a work-practice rule if Congress’ purposes were to be carried out at a cost acceptable to the Nation. Furthermore, the Administrator of the Environmental Protection Agency has similar powers “to prescribe such regulations as are necessary to carry out his functions under this chapter.” § 301, 42 U. S. C. § 1857g (a).
In promulgating the wetting requirement, the Administrator consistently referred to it as an emission standard:
“[T]he promulgated standard specifies certain work practices which must *301be followed when demolishing certain buildings or structures. The standard covers institutional, industrial, and commercial buildings or structures .... The standard requires that the Administrator be notified at least 20 days prior to the commencement of demolition.” 38 Fed. Reg. 8821 (1973).
In a recent case dealing with the proper construction of the Clean Air Act, the Court deferred to the view of the Administrator:
“Without going so far as to hold that the Agency’s construction of the Act was the only one it permissibly could have adopted, we conclude that it was at the very least sufficiently reasonable that it should have been accepted by the reviewing courts.” Train v. Natural Resources Defense Council, 421 U. S. 60, 75.
See also McLaren v. Fleischer, 256 U. S. 477, 480-481. The Court rejects the Administrator’s view because his “mere promulgation of a regulation” lacks power to persuade. Ante, at 288 n. 5. We have not previously required that judicial-style opinions accompany administrative actions or interpretations. In Train, supra, the Court deferred to the Administrator’s interpretation of the Clean Air Act even though his interpretation had been rejected by every Circuit to consider it, 421 U. S., at 72, and even though the interpretation was expressed and “supported” only by a single sentence in the Federal Register. 36 Fed. Reg. 22398, 22405 (1971). The Court’s “own 'analysis of the structure and legislative history,’ ” ante, at 288 n. 5, was limited to answering the question whether the Administrator’s construction was “sufficiently reasonable” to be permissible. 421 U. S., at 75. Similarly, in Norwegian Nitrogen Co. v. United States, 288 U. S. 294, the Court deferred to an administrative practice that apparently was formally justified only after the practice was challenged in court. Id., at 311, 314-315.
There is even more reason than usual to defer to the Administrator in the present case. Here we must decide whether the asbestos-wetting regulation is an emission standard within the meaning of a statute that allows prompt appellate review of such standards in a single court and precludes later challenges. § 307 (b), 42 U. S. C. § 1857h-5 (b) (1970 ed., Supp. V). Congress clearly wanted speedy, uniform, and final review of hazardous emission standards. Because this regulation is an attempt to control hazardous emissions on a nationwide basis, the need for speedy, uniform, and final review is just as great here as in the case of a numerical standard. If the reasons set forth in Part IV of the Court’s opinion are sufficient to sustain a collateral attack on this regulation, the preclusion statute has become almost meaningless. Of course, I do not suggest that the Administrator may take advantage of preclusion by simply "deeming” a regulation an emission standard. But when his characterization is challenged, we should try to understand the reason for the characterization before assuming that it was the product of a “Humpty Dumpty” thought process. See ante, at 283.
See United States v. National Wrecking Co., No. 74 CR 755 (Dec. 20, 1974); United States v. Nardi Wrecking Co., No. 74 CR 756 (Jan. 2, 1975); United States v. Harvey Wrecking Co., No. 74 CR 758 (Jan. 7, 1975); United States v. Brandenburg Demolition, Inc., No. 74 CR 757 (Jan. 31, 1975).
Letter from Environmentál Protection Agency Administrator to Senate Public Works Committee Chairman supporting proposed amendments to the Clean Air Act (Feb. 3, 1975), excerpted in Brief for United States, App. C.
The bill provided, in relevant part:
“(e) For purposes of this section the Administrator may promulgate a hazardous emission standard in terms of a design, equipment, or operational standard if he determines that such standard is necessary to control emissions of a hazardous pollutant or pollutants because, in the judgment of the Administrator, they cannot or should not be emitted through a conveyance designed and constructed to emit or capture such pollutants.” S. Rep. No. 95-127, p. 163 (1977).
The Conference Report characterized the original Senate version as follows:
“Amends section 112 of existing law to specify design, equipment, or operational standards for the control of a source of hazardous emissions, where an emission limitation is not possible or feasible to measure hazardous emissions or to capture them through appropriate devices for control.” H. R. Conf. Rep. No. 95-564, p. 131 (1977).
It described the conference substitute in these terms:
“The House concurs in the Senate provision with an amendment to clarify that the Administrator may specify a hazardous design standard if the emission of hazardous pollutants through a conveyance designed to emit or capture such pollutants would be inconsistent with any Federal, State or local law and minor clarifying modifications in the language.” Id., at 131-132.
This conclusion is buttressed by the recent amendment to the judicial review provision of the Clean Air Act. Ante, at 286 n. 4. At oral argument in the present case, Members of this Court pointed out that § 307 (b) applied by its terms only to “emission standards” and suggested that the words “emission standard” should be given a narrow reading. See, e. g., Tr. of Oral Arg. 20. That was on October 11. On November 1, a technical-amendments bill was introduced in both Houses to clarify “ambiguous language” and “technical problems” in the Clean Air Act. See 123 Cong. Rec. S18372 (Nov. 1, 1977) (statement of Sen. Muskie); see also id., at H11953 (reading of H. Res. 885). The bill, which passed both Houses and was signed into law on November 16, treated the Court’s present reading of “emission standard” as a simple error. To prevent future misreadings of the provision, Congress amended it to apply to “any emission standard or requirement” under § 112. See § 307 (b)(1), 42 U. S. C. §7607 (b)(1) (1976 ed., Supp. I), as amended and recodified *306by the Safe Drinking Water Amendments of 1977, § 14 (a) (79), 91 Stat. 1399 (emphasis added). The presence of a similar ambiguity in the enforcement provision was not pointed out at oral argument, and it was not corrected. This history indicates that Congress is patiently correcting judicial errors in construing “emission standard” narrowly.