CHEMICAL MANUFACTURERS ASSOCIATION, ET AL., PETITIONERS V. NATURAL RESOURCES DEFENSE COUNCIL, INC., ET AL. UNITED STATES ENVIRONMENTAL PROTECTION AGENCY, PETITIONER V. NATURAL RESOURCES DEFENSE COUNCIL, INC., ET AL. No. 83-1013, No. 83-1373 In the Supreme Court of the United States October Term, 1984 On Writs of Certiorari to the United States Court of Appeals for the Third Circuit Reply Brief for the United States Environmental Protection Agency I. In our opening brief (at 18-23), we argued that EPA is authorized under the Clean Water Act to grant "fundamentally different factors" (FDF) variances from the pretreatment standards applicable to indirect dischargers, as well as from the currently practicable BPT and the potentially more stringent BAT limitations that govern direct dischargers. Respondent answers, first, that this broad question is not before the Court and that the only point in dispute is the meaning of Section 301(l) of the Act, 33 U.S.C. 1311(l). In the remainder of its brief, however, respondent repeatedly relies on the assertion that FDF variances are proper only in the case of BPT limitations. Resp. Br. 11-14, 24, 26-27, 34. As we explained in our opening brief (at 18-23), respondent's position is wrong and is contrary to E.I. duPont de Nemours v. Train, 430 U.S. 112 (1977), and EPA v. National Crushed Stone Ass'n, 449 U.S. 64, 72 (1980). Contending that duPont and National Crushed Stone sanctioned FDF variances only in the case of BPT limitations, respondent maintains (Br. 12) that FDF variances are proper in that context because Congress did not create an applicable statutory variance. By contrast, respondent contends (Br. 12), FDF variances may not be granted from BAT limitations because "congress created its own BAT variance in Section 301(c) of the Act * * * and deemed this the sole deviation from otherwise uniform national standards." If anything, this argument supports our position, rather than respondent's. Respondent appears to have forgotten that Section 301(c) modifications are not available for indirect dischargers. /1/ Section 301(c) authorizes modifications of "the requirements of subsection (b)(2)(A) of this section with respect to any point source for which a permit application is filed after July 1, 1977 * * * " (emphasis added). Indirect dischargers are not subject to the permit program established under the Act; only direct dischargers apply for and obtain such permits. /2/ Section 301(c) is therefore inapplicable on its face to indirect dischargers. Consequently, even if respondent were correct that FDF variances may not be granted from BAT direct discharger requirements due to the availability of Section 301(c) modifications, /3/ that would not support respondent's argument that FDF variances may not be granted to indirect dischargers. On the contrary, FDF variances with respect to the pretreatment standards for indirect dischargers would be just as appropriate as FDF variances with respect to BPT limitations -- a practice that respondent concedes was approved in duPont. /4/ Respondent relies (Br. 13-14) on duPont's holding that FDF variances may not be granted from new source direct discharger requirements. Respondent contends that the reasons for this holding apply equally to pretreatment standards for existing sources. With two exceptions discussed below, all of respondent's arguments (Br. in Opp. 13) in this regard were addressed in our opening brief (at 20, 21 n.15, 22), and we will not repeat those points here. Respondent argues that both pretreatment standards and new source standards for direct dischargers, unlike BPT limitations, are intended to ensure national uniformity and maximum feasible control" and to "force technology to the level of the best possible performer." Resp. Br. 13 & n.18. However, respondent's suggestion that BPT limitations are not intended to promote uniformity conflicts with duPont's holding (and with respondent's own decade-long insistence) that BPT no less than BAT requirements are to be established through national categorical regulations. The FDF variance does not conflict with the goal of uniformity in the BAT or pretreatment context any more than it does in the BPT context, where it has been explicitly upheld by this Court. Respondent's reliance on the technology-forcing purposes of pretreatment, BAT, and new source requirements likewise is invalid. It is true that these requirements reflect the technological capabilities of the best performers within each subcategory. But, as we have explained (Gov't Br. 10-11, 30-31), the granting of an FDF variance signifies that the plant or plants in question constitute a separate subcategory and hence that the requirements for these plants are not properly determined by reference to the "best performer" in some other subcategory of the industry in question. /5/ Consequently, granting FDF variances is not at odds with Congress's intent to improve the technology for industrial categories and subcategories. Respondent draws an analogy (Br. 13-14) between new source direct dischargers (which may not obtain FDF variances) and all indirect dischargers. The more appropriate comparison is between new source direct and indirect dischargers (neither of which may obtain FDF variances) or between existing direct and indirect dischargers (both of which, in our view, are eligible for FDF variances). A new plant, whether a direct or indirect discharger, can be designed to conform to the national rule and thus has no need for the FDF variance. See Gov't Br. 22-23. By contrast, an existing source, whether direct or indirect, does not have this flexibility, and the availability of FDF variances is therefore useful and important. Ibid. II. A. Respondent next turns to the meaning of Section 301(l) of the Act, but respondent ignores the cardinal principle of statutory construction articulated by this Court in a similar context just last Term (Chevron U.S.A., Inc. v. NRDC, No. 82-1005, (June 25, 1984), slip op 4-5 (footnotes omitted): When a court reviews an agency's construction of the statute which it administers, it is confronted with two questions. First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress. If, however, the court determines Congress has not directly addressed the precise question at issue, the court does not simply impose its own construction on the statute, as would be necessary in the absence of an administrative interpretation. Rather, if the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency's answer is based on a permissible construction of the statute. Here, respondent certainly cannot show that its interpretation of Section 301(l) is supported by Congress's clear intent as evidenced by either the statutory language or the legislative history. Nor can respondent demonstrate that EPA's interpretation is impermissible. Accordingly, EPA's interpretation of the Act it administers must prevail. B. Respondent begins (Br. 14-20) by contending that the plain language of Section 301(l) prohibits FDF variances for toxic pollutants and that it is therefore unnecessary to look any further. Respondent's argument clearly lacks merit. Section 301(l) provides that EPA may not "modify" any pretreatment standard or other requirement of Section 301 with respect to any toxic pollutant. In our opening brief (at 24-27), we pointed out that the statutory term "modify" may not be given its broad, literal meaning, i.e., to change or alter, because Congress could not have intended to prohibit EPA from amending its regulations as warranted, and because such an interpretation would bring Section 301(l) into conflict with Section 307(b)(2), 33 U.S.C. 1317(b)(2), which commands EPA to "revise" its standards, including those for toxic pollutants (see 33 U.S.C. 1317(b)(1)), on a periodic basis. /6/ Respondent answers our argument (Br. 14-16) by maintaining that the term "modify" as used in Section 301(l) -- unlike in other provisions of the Act (Br. 15) -- signifies only a change in one direction: to "make more temperate and less extreme: lessen the severity of: moderate." Respondent seems to suggest that application of this definition gives Section 301(l) a reasonable meaning and eliminates any conflict with Section 307(b)(2). Respondent is wrong. Under respondent's interpretation, Section 301(l) would still place severe and unlikely restrictions on EPA's ability to amend its own rules. Once EPA promulgated a limitation or standard for a toxic pollutant, EPA could not amend its rules to make that requirement "more temperate or less extreme." As we noted in our opening brief (at 25), neither the discovery of an error in the Agency's previous analysis, the compilation of better or more accurate data, nor subsequent scientific developments would allow the Agency to change the standard. Unless Congress acted, the requirement could never be relaxed. Since Congress entrusted the EPA with the responsibility for issuing the requirements for toxic pollutants in the first place, it seems most unlikely that Congress intended to restrict EPA's customary rulemaking authority in this way. Thus, respondent's definition of the term "modify" fails to eliminate the conflict between Sections 301(l) and 307(b)(2), since the periodic revisions required by the latter provision may well result in the relaxation of requirements shown by subsequent research or experience to have been unreasonably stringent. Needless to say, when two statutory provisions are facially conflicting, it can hardly be argued that the statutory language is plain and that extrinsic aids to construction may not be consulted. Furthermore, respondent points to nothing in the language of Section 301(l) that shows which of the two concededly accepted meanings of the term "modify" Congress intended to adopt. Losing sight of the fact that it is attempting to advance a "plain meaning" argument, respondent finds it necessary to rely on other provisions of the Act (Br. 17, 19) and even the legislative history (Br. 16) to show what Section 301(l) means. Obviously this is no "plain meaning" argument at all. /7/ Respondent appears to shrink from the unavoidable implications of its own construction of Section 301(l). It is noteworthy that respondent refrains from arguing that Section 301(l) prohibits the amendment of categorical limitations or standards for toxic pollutants. Instead, respondent states (Br. 17; emphasis added) that Section 301(l) means that EPA "cannot * * * make these national standards less stringent for individual plants, at least where toxics are concerned." Where in the plain language of Section 301(l) does respondent discern this distinction between modifications that affect a broad category of plants and those that affect only a smaller class? /8/ III A. Because the statutory language does not answer the question presented by this case, we turn to the legislative history. The relevant legislative history is sparse, but it clearly provides far more support for EPA's interpretation of Section 301(l) than for respondent's. There are three relevant aspects of the legislative history: (1) the evolution of Section 301(l), (2) what Congress said about this provision, and (3) the inference that may be drawn from what Congress did not say in light of the then existing state of the law and EPA's established practices. 1. The origin and evolution of Section 301(l) are set out in Chemical Manufacturers Association's (CMA's) opening brief (at 29-30). In short, what is now Section 301(l) was first proposed in 1977 as a provision of Section 301(c) that prohibited the granting of Section 301(c) modifications for toxic pollutants. When Congress decided to add Section 301(g), which authorizes the granting of water quality modifications, a virtually identical provision banning such modifications for toxic pollutants was included in that section. The Conference Committee deleted the prohibitions in Section 301(c) and (g) and consolidated them into a single provision, Section 301(l). The Conference Committee did not explain the reason for this amendment, but Representative Roberts, the House floor manager stated: Due to the nature of toxic pollutants, those identified for regulation will not be subject to waivers from or modification of the requirements prescribed under this section, specifically, neither section 301(c) waivers based on the economic capability of the discharger nor 301(g) waivers based on water quality considerations shall be available. Senate Comm. on Environment and Public Works, 95th Cong., 2d Sess., Legislative History of the Clean Water Act of 1977, 328-329 (Comm. Print 1978) (emphasis added) (hereinafter Leg. Hist.). Putting aside for the moment what Representative Roberts said, we think that the evolution of Section 301(l) supports EPA's interpretation of that provision rather than respondent's. In our view, the prohibitions originally included in Section 301(c) and (g) were consolidated into a separate Section 301(l) for stylistic purposes. /9/ For this reason, the Conference Committee had no cause to explain the change. Respondent argues (Br. 19), however, that the conferees created Section 301(l) because they wanted to make a substantive change not contained in the bill passed by either House. According to respondent, Section 301(l) was created to prohibit other modifications in addition to those available under Section 301(c) and (g). But if the conferees intended to make such a substantive change, one would expect them to have provided notice to their colleagues and some explanation of their action. None, however, was provided. There is another glaring defect in respondent's interpretation of the drafting process. If the Conference Committee intended for Section 301(l) to apply to modifications in addition to those available under Section 301(c) and (g), the Committee must have had some other type or types of modifications specifically in mind. But according to respondent (Br. 21), "Congress did not believe other variances (i.e., other than Section 301 (c) and (g) modifications) were authorized." /10/ What, then, does respondent think the conferees sought to accomplish? We of course submit that Congress was well aware of EPA's practice of granting FDF variances, especially since that practice had been prominently discussed and approved in duPont. And as previously noted, if the conferees had intended to restrict the issuance of such variances, some reference to FDF variances would most likely have been made. None was. 2. EPA's interpretation is also supported by Congress's discussion of Section 301(l). Like the remarks of Representative Roberts previously noted, the legislative history cited by respondent only confirms the accuracy of EPA's interpretation. For example, the quoted statement of Senator Muskie -- that "there are no potential wiavers or modifications" (Resp. Br. 21) -- was made specifically in the context of discussing Section 301(c) and (g). Leg. Hist. 458. Similarly, the Senate Report excerpt quoted by respondent (Br. 21), which states that "the (Senate) bill prohibits variances for (toxic) pollutants," comes specifically in a discussion of the Senate version of what became Section 301(g). See Leg. Hist. 673-677 (discussing water-quality based modifications). The only other portions of the legislative history claimed by respondent to support its position are references to Section 301(c) and (g) as "variance" provisions (see Br. 16 & n.21). Contrary to respondent's suggestion (Br. 16-17), we do not dispute the authoritativeness of these pronouncements. What we contest is their bearing on the issue in this case. Although not spelled out, respondent's argument is based on the following chain of reasoning: 1. Section 301(l) prohibits "modifi(cations)" for toxic pollutants. 2. Section 301(c) and (g), which authorize "modifi(cations)," were described by Congress as "variance" provisions. 3. An FDF variance is obviously a "variance." 4. Therefore, Section 301(l) prohibits FDF variances. It is not necessary to be a logician to see what is wrong with this reasoning. If A includes B, and B resembles C, it does not follow that A also includes C. Here, just because a Section 301(c) or (g) modification may be described as a variance, it does not follow that an altogether different sort of variance (i.e., an FDF variance) is a "modification" in the legal sense intended by Section 301(l). The fallacious reasoning used by respondent can be employed to "prove" any number of untruths -- for example, that EPA is required by Section 307(b)(2) to issue Section 301(c) and (g) modifications and FDF variances on a periodic basis: 1. Section 307(b)(2) requires EPA to "revise" its requirements as developments warrant. 2. A revision may be described as a "correction" or alteration." Webster's Third New International Dictionary 1944 (1976 ed.) Random House Dictionary of the English Language 1227 (unabr. ed. 1967). 3. Section 301(c) and (g) modifications and FDF "variances" may be described as "corrections" or "alterations." 4. Therefore, Section 307(b)(2) requires EPA periodically to issue Section 301(c) and (g) modifications and FDF variances. /11/ 3. The remaining question with respect to the legislative history is what may be inferred from the undisputed fact that Congress never mentioned FDF variances during its consideration of Section 301(l). We think that Congress must have been aware of and understood EPA's practice of granting FDF variances and that Congress's failure to mention those variances in connection with Section 301(l) suggests that Congress did not intend to affect that practice. Respondent contends (Br. 21) that Congress had no reason to mention FDF variances as among those modifications to be prohibited by Section 301(l) because "Congress did not believe other variances (besides section 301(c) and (g) modifications) were authorized, 1972 Leg. Hist. 172, 309, and EPA had not yet promulgated an FDF variance for either pretreatment or BAT standards." Respondent's reliance on 1972 legislative history is clearly misplaced. It is Congress's intent in 1977, not 1972, that is at issue here. By the time Congress took up the 1977 amendments to the Act, EPA had regularly promulgated FDF variances as part of national BPT effluent limitations. Moreover, these FDF variances were always made available for pollutants that Congress designated as toxic in 1977. And variances for such pollutants had been granted. /12/ That EPA had not yet incorporated the FDF variance into BAT effluent limitations or pretreatment standards is beside the point, /13/ because the issue is whether Section 301(l) bars FDF variances for all requirements of Section 301. Congress also must have been aware of this Court's opinion in duPont, which discussed and approved EPA's practice of granting FDF variances. Respondent suggests (Br. 26-27) that Congress was not familiar with the discussion in duPont, but in respondent's own words, duPont was a "pivotal" case in the history of the Clean Water Act. It had just been decided. And the discussion of FDF variances was closely related to the Court's holding. We doubt that this discussion could have escaped the attention of a Congress that was then in the process of amending the Act. Respondent argues that duPont approved FDF variances only with respect to BPT limits. For reasons already discussed (see Gov't Br. 20-23; pages 1-5, supra), we disagree. But in any event, since Section 301(l) applies to BPT limitations, as well as all other requirements of Section 301, respondent's argument is beside the point. Respondent contends, however, that Congress did not "focus()" on BPT limitations when it enacted Section 301(l) (Br. 27). But Section 301(l) indisputably applies to BPT limitations; and in 1977, those limitations were scheduled to remain in effect for the next seven years. Respondent is accusing Congress of a major and implausible oversight. In sum, we think that Congress was aware of EPA's practice of granting FDF variances and that if Congress had intended to restrict that practice it would have said so. Congressional intent to overturn established law should not be inferred in the absence of an express indication. /14/ Edmonds v. Compagnie Generale Transatlantique, 443 U.S. 256, 266-267 (1979). /15/ The remainder of respondent's brief (at 30-43) is devoted to arguing that EPA's practice of granting FDF variances is not sound policy and that EPA's objectives can be better achieved by alternative means. We disagree with respondent's analysis. But, in any event, the issue in this case is not whether, in the Court's view, EPA has adopted the best practice. Chevron, slip op. 26-28. Rather, the issue is whether EPA's interpretation of the Act is permissible, and on that question respondent's argument falls far short. Despite its rhetoric, respondent's disagreement with the Agency is really quite narrow. There is no dispute that in setting pretreatment standards and other requirements of Section 301 EPA is justified in making special allowance for those facilities that face factors fundamentally different from other facilities in the industry. EPA has chosen to deal with this problem by attempting to issue categorical requirements as quickly as possible, while permitting atypical facilities to seek FDF variances at a later date. Respondent, on the other hand, while contesting the issuance of FDF variances, does not dispute EPA's authority to create special subcategories for an atypical plant or plants. On the contrary, respondent observes (Br. 41; emphasis added) with apparent approval that "requests to establish separate subcategories for one or a few plants which are somehow 'different' from others in the same category are a common feature of (EPA) rulemakings." Nor does respondent dispute EPA's authority to amend its rules after issuance to create such subcategories. In other words, respondent disagrees, not with the substance of what EPA is doing, but with the procedure. But if deference to an administrative agency's expertise and construction of the statute it administers means anything, it must mean that the agency is permitted to make such procedural choices absent a clear expression of contrary congressional intent. Particularly in view of the small number of FDF variances that are sought and granted, there was no justification for the court of appeals' interference with the procedure chosen by the Agency to achieve a concededly valid end. Arguing that the FDF variance procedure conflicts with Congress's broad intent in enacting the 1977 amendments, respondent emphasizes that the 1977 amendments reflect a congressional intent to achieve control of toxic pollutants through national categorical regulations that are "as uniform as possible" (Br. 22, 23, 32-33 (emphasis added)). The FDF variance, however, is fully consistent with this goal, because Congress recognized that EPA must be allowed to take into account the diversity within each industry by establishing appropriate subcategories (Leg. Hist. 455). The categorical approach to setting standards necessarily means that some individual plants within a category must incur costs greather than other plants in that category (See. Resp. Br. 34). But the point of the FDF variance is to allow a recategorization of those plants that must incur costs wholly disproportionate to the range of costs considered by EPA during the rulemaking for that category. /16/ Moreover, the FDF variance mechanism does not give EPA "carte blanche" to establish case-by-case requirements "without limit" (Resp. Br. 32). FDF variances may be granted only under very narrowly circumscribed conditions -- conditions that would have warranted creation of a separate subcategory in the first place. And contrary to respondent's contention (Br. 32), an FDF variance application is not considered in isolation from the larger category or subcategory in which the plant was originally classified. Instead, the Agency compares the plant in question to the category as a whole to determine if the plant's factors are "fundamentally different" from those considered by the Agency. /17/ Respondent also contends that the procedures for granting FDF variances differ from those followed when subcategories are created during initial rulemaking or subsequent revisions. The latter procedures are clearly more to respondent's liking. Resp. Br. 36-37. But the issue in this case is the meaning of Section 301(l), and there is no basis for contending that the purpose of that provision was to regulate the procedures by which EPA can accomplish the same permissible end. Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. 519 (1978). Respondent, echoing the court of appeals, claims that it "makes no sense" to conclude that Congress intended to prohibit Section 301(c) modifications but not to prohibit FDF variances. Resp. Br. 31-32. The fact is, however, that Section 301(c) modifications and FDF variances do not serve the same function. A Section 301(c) modification is available to a facility able to show that the national categorical standard would cause it severe economic hardship and that the facility meets the other statutory prerequisites. A Section 301(c) modification may be granted even though EPA took into account all of the appropriate factors in establishing the national standard. The FDF variance, by contrast, rectifies EPA's oversight or omission in establishing the national rule and has nothing to do with a plant's ability to afford compliance. If Section 301(c) modifications and FDF variances were indistinguishable, it would have been pointless for this Court to approve FDF variances for BPT requirements in duPont, yet disallow modifications to BPT requirements based on Section 301(c) factors in National Crushed Stone. /18/ Therefore, simply because Section 301(l) prohibits Section 301(c) modifications for toxic pollutants, it does not follow that FDF variances are also prohibited. See Gov't Br. 30-33. Finally, the FDF variance mechanism does not thwart Congress's desire that national technology-based controls for categories of industries be established as expeditiously as possible. On balance, the FDF variance mechanism advances that goal by helping the Agency to establish sound national categorical regulations that are less subject to attack on the ground that an atypical plant or plants were not adequately considered. Respondent has the luxury of calling the FDF variance "nothing more than a fig leaf for inadequate agency rulemaking" (Br. 38), because it is EPA, not respondent, that is faced with the difficult task of implementing and construing the Clean Water Act. Respondent's bald assertion that the FDF variance makes national rules more vulnerable to challenge (Br. 38) is difficult to understand. The FDF variance has in fact helped to protect against precisely this risk (see e.g., Kennecott Copper Corp. v. EPA, 612 F.2d 1232, 1243 (10th Cir. 1979)), and respondent cannot cite one case arising under the Clean Water Act to support its contrary contention. The FDF variance does not put all of EPA's "eggs in (one) basket" (Br. 38); the FDF variance mehcanism constitutes an infrequently used but nevertheless important safeguard that courts have relied on in rejecting challenges to national rules. The Agency has promulgated categorical regulations covering tens of thousands of industrial dischargers, and the fact that a handful of individual plants make use of the FDF variance process does not constitute a threat to the goals of the Act. /19/ EPA's interpretation of the Act it administers is a "permissible" one (Chevron, slip op. 5) and should be sustained. For the foregoing reasons and those set forth in our opening brief, the opinion of the court of appeals should be reversed. Respectfully submitted. Rex E. Lee Solicitor General A. James Barnes General Counsel Environmental Protection Agency October 1984 /1/ Respondent incorrectly states (Br. 29 & n.41) that EPA has made Section 301(c) modifications available for pretreatment standards. The cited regulation dealt with BAT direct discharger requirements, not pretreatment standards. 40 Fed. Reg. 55770, 55776 (1975). /2/ Section 402 of the Act, 33 U.S.C. 1342, established the "National Pollutant Discharge Elimination System" (NPDES) permit program only for direct dischargers, including publicly owned treatment works (POTWs) to which indirect dischargers send their waste water. Effluent limitations guidelines are applied to direct dischargers through this permit program. EPA v. California ex rel. State Water Resources Control Board, 426 U.S. 200, 205 (1976). Pretreatment standards, by contrast, apply directly to the indirect discharger without the issuance of any permit. See Pet. App. A17. /3/ As previously noted, we disagree with this argument. /4/ Respondent incorrectly asserts (Br. 24) that duPont "upheld" BAT rules that did not contain an FDF variance clause. In fact, the validity of the BAT regulations was not as issue in duPont. Respondent also suggests (Br. 24) that in duPont this Court "appropriately assumed" that only Section 301(c) modifications, and not FDF variances, would be available for BAT regulations. Respondent does not, and cannot, point to anything in the opinion to support such a reading. /5/ Respondent also ignores the fact that, for many industries -- as a result of a consent decree authored in relevant part by respondent -- EPA is required to promulgate BPT-level pretreatment standards as an interim measure pending development of potentially more technology-forcing BAT-level standards. See NRDC v. Train, 8 Env't Rep. Cas. (BNA) 2120, 2128 (D.D.C. 1976). The electroplating pretreatment standards unsuccessfully challenged below were one such regulation. /6/ EPA has not changed its position, regarding the interpretation of Section 301(l), as respondent contends (Br. 15). The Agency's position has always been straightforward -- that an FDF variance is not a "modification" as Congress used that term in Section 301(l). /7/ It is also interesting to note that respondent's definition does not support the court of appeals' holding that EPA may not issue any FDF variances for toxic pollutants. An FDF variance can result in a requirement that is more stringent than the national standard, if the facility's fundamentally different factors so warrant. 40 C.F.R. 403.13(b) (Pet. App. A122). Respondent calls this argument "disingenuous" (Br. 14 n.19) because no applicant has requested more stringent requirements. The point is, however, that respondent and any other interested person could initiate the FDF variance process to seek a more stringent requirement. That respondent has never seen fit to invoke the regulation for this purpose proves nothing. /8/ Furthermore, respondent's argument that EPA may not make any plant-specific modifications, cannot be squared with respondent's own recognition that EPA may establish subcategories of one or a few plants. Resp. Br. 41 & n.63; Transcript of Oral Argument in court of appeals 252. EPA thus has the authority to promulgate national pretreatment standards in the first instance for a subcategory of one plant, and it has commensurate authority under Section 307(b)(2) to "revise" a national pretreatment standard for such a subcategory. /9/ When the Conference Committee created Section 301(l) by consolidating the prohibitions previously included in Section 301(c) and (g), the Committee added a clause to Section 301(g) prohibiting water quality modifications for conventional pollutants and thermal discharges, as well as for toxic pollutants. See CMA Br. 30 n.38. Because this clause is broader than Section 301(l), which applies only to toxic pollutants, it is not redundant, as respondent claims (Br. 20 n.31). /10/ Respondent elaborates that at the time of the 1977 amendments, Congress did not "focus" on BPT limitations (Br. 27) and had no way of anticipating that EPA would issue FDF variances for pretreatment and BAT standards (Br. 24-30). /11/ The present Congress did not reject the concept of FDF variances for toxic pollutants, as respondent suggests (Br. 26 n.37). The Senate bill amending the Clean Water Act was reported out of committee on the day after the court of appeals' decision (see Resp. Br. 26 n.37) and thus naturally did not contain a provision to overrule that decision. In the House, EPA did advert to the need to overrule that decision, but no bill to that effect was proposed. This history does not show that the present Congress endorsed the court of appeals' decision. And in any event, "'the views of a subsequent Congress form a hazardous basis for inferring the intent of an earlier one.'" Jefferson County Pharmaceutical Ass'n v. Abbott Laboratories, 460 U.S. 150, 165 n.27 (1983), quoting United States v. Price, 361 U.S. 304, 313 (1960). /12/ See EPA Br. 30. /13/ During the first five years after initial passage of the 1972 Act, EPA's main objective was to issue the first tier of national technology-based regulations, which established BPT requirements for categories of dischargers. These regulations covered discharges of both toxic and non-toxic pollutants, and FDF variances were available regardless of the type of pollutant. Prior to 1977, the FDF variance was not yet included in BAT regulations or pretreatment standards because the Agency had not yet turned its full efforts and refined its approach to those requirements. When it did so after enactment of the 1977 amendments, it recognized that FDF variances were equally appropriate for these regulations. The fact that the Agency's thinking evolved only confirms that the Administrator has broad leeway to construe the statute. Chevron, slip op. 25. /14/ The pre-existing law in Edmonds does not appear to have been as well-settled as respondent would like to believe. See Edmonds, 443 U.S. at 275 (Blackmun, J., dissenting). /15/ Zuber v. Allen, 396 U.S. 169 (1969), relied on by respondent (Br. 25), supports EPA. There, the Court held that the Secretary of Agriculture had exceeded his authority in issuing milk pricing orders that included an adjustment based on a milk producer's location. The Court concluded (396 U.S. at 180-181) that such location differentials were fundamentally inconsistent with the statute's "plain thrust" and conflicted with Congress's intent as expressed in the legislative history. In that context, the Court concluded that "(l)egislative silence is a poor beacon to follow in discerning the proper statutory route" (id. at 185) and that Congress's failure in ensuing years to reverse the Secretary's practice could not be construed as an endorsement (id. at 185-186 n.21). Our position in this case is fully consistent with Zuber. The FDF variance does not conflict with either the "plain thrust" of the Clean Water Act or with the legislative history. More fundamentally, however, in Zuber, as well as in TVA v. Hill, 437 U.S. 153 (1978), on which respondent also relies (Br. 9, 23, 26), this Court refused to infer congressional approval of a particular agency practice in the absence of an express indication of such an intent. But here, respondent asks the Court to infer the opposite -- a congressional intent to disapprove a longstanding agency practice, in the absence of any specific disapproval of that practice, and even though the statutory language is not clear on its face. Even if Congress did not specifically embrace the FDF variance, there is certainly no evidence that Congress sought to invalidate that variance as to toxic pollutants. Particularly where, as here, the pre-existing administrative practice is not at odds with the underlying structure and purpose of the Act, such an intent on the part of Congress should not be inferred. /16/ For example, if during the rulemaking EPA finds that the maximum cost to be incurred by a plant in a subcategory is $1 million, then any plant in that category clearly can be required to incur such costs even if other plants may need to spend much less to comply. But a plant that must spend $5 million to meet the requirement because it is fundamentally different from those considered by the Agency in establishing the rule may be entitled to argue that it should be treated as a separate subcategory. /17/ Nor has EPA improperly incorporated BPT criteria into variance decisions regarding BAT, as respondent claims (Br. 34-35). The Act clearly requires the Agency to consider costs in establishing BAT effluent limitations (33 U.S.C. 1314(b)(2)(B)) and thus an individual plant's costs -- as opposed to its ability to afford those costs -- is also an appropriate factor in determining whether to grant an FDF variance. In any event, the issue in this case is not the content of any specific variance, but EPA's authority to promulgate any FDF variance for pretreatment standards. See duPont, 430 U.S. at 128 n.19. /18/ Contrary to the interpretation of both respondent (Resp. Br. 31) and the court of appeals (Pet. App. A42), this Court did not conclude in National Crushed Stone Ass'n that Section 301(c) modifications and FDF variances "serve the same function." See Gov't Br. 33 n.25. /19/ Respondent cites (Br. 43) an excerpt from the Conference Report reflecting congressional concern that steam electric power plants had been abusing a provision found in an entirely separate section of the Act, Sec. 316, 33 U.S.C. 1326, with respect to the discharge of heat. This is utterly irrelevant to the proper construction of Section 301(l), which does not even apply to Section 316.