CHEVRON U.S.A., INC., PETITIONERS V. NATURAL RESOURCES DEFENSE COUNCIL, INC., ET AL. Nos. 82-1005, 82-1247, and 82-1591 In the Supreme Court of the United States October Term, 1983 On Writ of Certiorari to the United States Court of Appeals for the District of Columbia Circuit Reply Brief For the Administrator of the Environmental Protection Agency The issue posed by this case is whether the 1977 Amendments to the Clean Air Act -- which created the New Source Review (NSR) program for nonattainment areas -- impose a statutory prohibition precluding the Administrator of EPA from adopting a regulation that allows -- but does not require -- a state to adopt a plant-wide ("bubble") definition of the term "source" for purposes of new source review. In order to prevail, respondents must establish that Congress, in the 1977 Amendments, itself decided to require new source review whenever new equipment is added or existing equipment is modified at a plant even though resulting increased emissions are offset by corresponding (or even greater) reductions elsewhere in the plant, and even though EPA has found that the State Implementation Plan (SIP) is consistent with the Act and contains adequate assurance of progress toward attainment of national air quality standards. /1/ Respondents must further establish that Congress foreclosed the "bubble" approach in a statute one of whose "main purposes," according to Congress itself, was "to allow reasonable economic growth to continue" during the attainment process (see H.R. Rep. 94-294, 95th Cong., 1st Sess. 211 (1977) (emphasis added), reprinted in 4 Staff of Senate Comm. on Environment and Public Works, 95th Cong., 2d Sess., A Legislative History of the Clean Air Act Amendments of 1977, at 2678 (Comm. Print 1978) (hereinafter Leg. Hist.)) -- notwithstanding the wide consensus that the "bubble" approach represents an innovative and creative method precisely for accommodating the goals of efficient growth and environmental protection. /2/ And, finally, they must show that the Act denies the Administrator discretion to give the states any flexibility whatever in how to define "source" for purposes of the NSR program, even though Congress stated that the second of the two "main purposes" of the nonattainment provisions of the 1977 Amendments was "to allow States greater flexibility" in achieving reasonable economic growth while making reasonable further progress to assure attainment (ibid. (emphasis added)). The history of this case consists of a series of attempts to invent new justifications for the foregoing, implausible, propositions. Respondents' brief in this Court is no exception. Respondents have dramatically abandoned any defense of the artificial "bright-line" test adopted by the court of appeals -- a test which led that court to the extraordinary conclusion that the term "source" is to be given a fixed judicial definition notwithstanding the court's own conclusion that the text and legislative history of the Act provide no such definition. Instead, respondents renew the contention that the plantwide definition is "prohibited" by the Act itself because Congress "specifically intended" new source review to apply to every individual piece of equipment (Br. 6). /3/ The principal vehicle advanced to buttress this contention is that the definition of "source" provided by Congress in Section 111 of the Clean Air Act -- a definition applicable by its terms only to the NSPS (new source performance standards) program (42 U.S.C. 1957c-6(a)(3) and 7411(a)(3)) /4/ -- in fact applies to the NSR (new source review) program as well. However, the argument that the Section 111 NSPS definition "applies" to new source review receives an entirely new twist in respondents' merits brief. Respondents purport at the eleventh hour to have discovered that the new source review program created by Congress in 1977 is not a new program at all. It is now asserted to be simply the continuation of an old "new source review" program, instituted by Congress in the 1970 Act, and contained in Sections 110(a)(2)(D) and (4) of the Act, 42 U.S.C. 1857c-5(a)(2)(D) and (4) (App., infra, 1a-2a). Respondents argue that those sections have always required (what respondents characterize as) "new source review." And, since these sections contain an explicit cross-reference to the Section 111 NSPS definition, the suggested conclusion is that, ever since 1970, the NSR program has applied to sources as defined in Section 111. It follows, respondents say, that in perpetuating the new source review program in 1977, Congress obviously assumed that the program would continue to apply to all individual pieces of process equipment. In fact, this argument is a concoction presenting an entirely distorted picture of the architecture of the Clean Air Act. The program created by Congress in 1970 in Sections 110(a)(2)(D) and (4) -- usually called locational review -- has nothing to do with the new source review program first created in 1977 in order to deal with the problem of nonattainment. The cross-reference to the Section 111 NSPS definition contained in those sections /5/ thus does not in any way suggest that the Section 111 definition was to apply to the nonattainment program created in 1977. The 1970 Clean Air Act did not contemplate nonattainment at all. It envisioned that all areas of the country would attain National Ambient Air Quality Standards (NAAQS) by the statutory deadline of 1975. To achieve this, the 1970 Act relied primarily on two programs. First, emissions from existing sources were to be reduced, in accordance with the state implementation plan, to meet NAAQS (see Gov't Br. 4). Second, new "sources" were subject to new source performance standards to be promulgated by EPA pursuant to Section 111, 42 U.S.C. 1857c-6. In connection with the process of attaining and maintaining NAAQS, new "sources" were required to undergo a locational review process (to be specified in the SIP) pursuant to Sections 110(a)(2)(D) and (4). This was to be a "procedure" for "review (prior to construction or modification)" of the "location of new sources" so as to "prevent the construction or modification of any new source" that the state determines "will prevent the attainment or maintenance" of NAAQS. See 42 U.S.C. 1857c-5(a)(2)(D) and (4). The "procedure" was to cover all new sources to which a "standard" under the NSPS Section 111 program "will apply." The purpose of locational review was to create a review process -- applicable in "clean" and "dirty" air areas alike -- to make sure that the installation of new plant and equipment would not prejudice either the attainment or the continuing maintenance of national air quality standards. /6/ It looked to the net effect on air quality of a new or modified "source." For this reason, its reference to the NSPS definition of "source" has no independent functional significance. The only substantive question in connection with locational review is whether construction or modification of a new "source" -- however defined -- should be prevented because it will adversely affect net air quality. "Netting" is implicit in the only substantive judgment called for in locational review: whether the addition or modification should be vetoed because it will lead to air quality that does not satisfy NAAQS. If the addition or modification produces no net increase in emissions, locational review has, by hypothesis, no bite. In contrast, the whole point of respondents' submission about NSR is to impose elaborate requirements (e.g., LAER technology, see Gov't Br. 7) on individual pieces of machinery even if there is no net effect on air quality. Importing the NSPS definition into the 1977 NSR program would thus have a radically different substantive result than is created by the cross-reference to that definition in the 1970 locational review provisions. Beyond this, the fact is that the 1977 Amendments marked a major change in the over-all architecture of the Clean Air Act. In contrast to the 1970 Act, which had assumed that air quality standards would be timely attained, the 1977 Act was established in response to the failure of a number of regions to attain air quality standards by 1975. A major concern was that existing law might completely prohibit the construction or modification of facilities in such areas. /7/ Following EPA's promulgation of an administrative solution to this problem in its "Emissions Offset Ruling" (41 Fed. Reg. 55524 (1976)), Congress passed the NSR program specifically to deal with the issue of accommodating industrial growth to the problem of nonattainment. The NSR statute was intended to be a "comprehensive scheme." /8/ It was designed "to reconcile the() conflicting concerns" between air quality attainment and continued economic growth. /9/ Congress made clear that it was adopting new legislation to address new problems not foreseen in 1970. There exists no support for the novel assertion that the elaborate NSR provisions considered at length in 1977 are nothing more than an extension of the locational review requirements. Indeed, NSR and locational review both continue to exist in the current version of the Act (see 42 U.S.C. 7410(a)(2)(D) and (4)); such coexistence would hardly be expected if NSR were simply the current version of location review. At the same time Congress passed the nonattainment provisions in Part D of the 1977 Amendments, it also adopted Part C. This dealt with a converse problem: preventing the significant deterioration of air quality in areas where the air was cleaner than required by federal standards. This program -- called the "Prevention of Significant Deterioration" (PSD) program -- was considered by Congress together with the NSR program in 1977. See, e.g., H.R. Conf. Rep. 95-564, supra, at 128, 3 Leg. Hist. 508. The test for what is to be considered "major" in connection with the NSR term "major stationary source" and the PSD term "major emitting facility" is provided in the same section of the Act (42 U.S.C. 7602(j)). And, as noted in our opening brief (at 11 n.19), it is not unusual for a given plant to be subject to NSR review for one pollutant and PSD review for another. All these considerations suggest that "source" should have the same meaning in both the NSR and PSD programs. And it has of course been held that the plant-wide definition of "source" is the proper one for purposes of the PSD program. See Alabama Power Co. v. Costle, 636 F.2d 323 (D.C. Cir. 1979). In sum, the assertion that the Section 111 definition of "source" applies to the 1977 NSR program because that program simply continues the requirements of the 1970 locational review sections -- which do refer to the Section 111 definition -- is a last-minute invention that ignores the Act's history and twists it into an unrecognizable shape. II. Respondents' remaining arguments concerning the applicability of the Section 111 NSPS definition to the 1977 NSR program were generally anticipated in our principal brief and can be dealt with summarily. As pointed out in that brief (at 42-43), Section 111 explicitly provides that its definition is "(f)or purposes of this section" (emphasis added). Respondents object that the specification does not provide that it is "'for the (sic) purposes of this section only'" (Br. 30 n.57). But why should Congress, in enacting the NSPS definition in 1970, add the (redundant) word "only" to exclude the possibility of its applying to the NSR program not created until 1977? /10/ In fact, Congress took affirmative steps to avoid tying the new source review program to the NSPS definition of "source": as discussed in our opening brief (at 43-44), it deleted from the 1977 Act a provision expressly adopting the NSPS definition for the NSR program, while at the same time incorporating by reference other NSPS terms. Finally, the NSR and NSPS programs are entirely different in function and structure (see Gov't Br. 13 n.24, 42-43 n.55). There is therefore no substantive reason to import the NSPS definition into the NSR program. III. Respondents' further argument is that the legislative history of the 1977 Act shows that Congress intended the NSR program to apply to every individual piece of process equipment. Respondents misconceive that history. In fact, the tangled legislative history, compiled during the course of two different Congresses, does not directly address -- let alone resolve -- the question presented here. Respondents' interpretation of the legislative history hinges on the Senate version of a 1976 bill that proceeded through the Conference Committee but was not enacted into law because of a filibuster. The 1976 Senate bill explicitly incorporated the NSPS definition of "source" for purposes of the NSR program; /11/ in addition, as respondents emphasize (Br. 10, 33), its NSR section referred to "a facility proposed for construction or modification at an existing site or plant," /12/ thereby suggesting a difference between "facilities," on the one hand, and "sites" and "plants," on the other. As previously discussed (Gov't Br. 45-46 & nn.58, 60), however, both of these provisions were deleted from the 1977 NSR statute enacted by Congress. /13/ Respondents' contention is that, nevertheless, the Senate assumed that the NSPS definition would apply to NSR and that this assumption persisted throughout the consideration and ultimate enactment of the NSR provisions. Not a shred of evidence supports this argument. Indeed, it ignores the critical fact that the NSR program eventually enacted was based principally on the House rather than the Senate bill. /14/ The Conference Report plainly states: All implementation plan revisions must contain a permit program for new or modified major facilities similar to the program required in the House bill. H.R. Conf. Rep. 95-564, 95th Cong., 1st Sess. 157 (1977), 3 Leg. Hist. 537. /15/ Most significantly, neither the House bill nor the NSR statute incorporated the NSPS definition of "source" or referred to a "facility * * * at an existing site or plant" as the entity subject to new source review; both applied new source review to a "new or modified major stationary source" and left the term "source" undefined. Respondents argue -- in a confusing paragraph -- that the 1977 House bill did not need to adopt the Section 111 definition of "source" because the "plain reading" of "source" is the Section 111 definition ("'building, structure, facility, or installation'") (Br. 31). /16/ But this is nothing more than ipse dixit. The question is whether the House intended, in the absence of an explicit cross-reference, to import the NSPS definition of "source" into the NSR program. It begs that question simply to assert that such an explicit cross-reference was unnecessary, in the absence of some independent evidence that the language or structure or functional purposes of the statute suggest such a cross-reference. /17/ The sole "evidence" in fact adduced by respondents is their quotation (Br. 31-32) (footnote omitted) from the Conference Report of the sentence, "The bill '(d)efines major stationary source and major emitting facility as defined in the Senate bill.'" But the quoted sentence has nothing to do with the question whether "source" means a plant or an individual piece of machinery. As a glance at the context will make clear, it refers exclusively to the fact that the House and Senate bills differed in their definitions of the term "major." The House bill was based on "design capacity to emit 100 tons annually," while the Senate bill referred to "an annual potential to emit 100 tons." The Conference adopted the Senate's definition of "major." H.R. Conf. Rep. 95-564, supra, at 172, 3 Leg. Hist. 552. The quoted sentence, referring to this action on the quantitative standard for measuring "major" emissions, in no way suggests that the Conference also intended -- at the same time that it deleted the Senate's explicit definition of "source" -- to endorse the Senate's understanding of the industrial unit that, if "major," would be covered by the NSR program. IV. Respondents next argue that EPA has, since 1971, applied the NSPS definition of "source" to new source review (Br. 4, 6, 7, 15-19, 42) and that Congress ratified that approach in enacting the NSR program in 1977 (Br. 11, 33, 42, 43). This argument is without merit. It is EPA's Emissions Offset Ruling of 1976 that constitutes the proper point of departure for analyzing EPA's approach to the issue in this case (see Gov't Br. 5). /18/ There is no quarrel over respondents' contention (Br. 16-17, 34) that the Offset Ruling, in referring to a "major source," essentially tracked the language of the NSPS definition (41 Fed. Reg. 55528 (1976)). Nor can it be disputed that Congress was well aware of the Offset Ruling when it passed the 1977 Act. /19/ But we find no evidence of congressional intent to ratify -- indeed, to freeze -- this definition of "source" for purposes of the NSR program. The critical fact is that in 1977 Congress determined that the Offset Ruling should not be enshrined into permanent national policy. It was adopted as a temporary measure to apply to nonattainment areas until July 1, 1979. See 42 U.S.C. 7502 note, 3 Leg. Hist. 251. /20/ Even in this interim period, the one specific feature of the Ruling that Congress addressed -- the "baseline" to be used for calculating the level of available offsets (ibid.) -- was amended to make it easier to build or alter a "source" in nonattainment areas. And with respect to all other aspects of the Ruling during the pre-1979 period, the statute expressly provided that they "may be modified by rule of the Administrator" (ibid.). Surely this limited interim continuation of the Offset Ruling is a giant step from the proposition that Congress intended in perpetuity to freeze its approach to new source review. In any event, Congress made it clear that it intended no such thing. After the interim period, Congress specified its own approach to the problem of nonattainment. And, in doing so, it stated explicitly that one of the "two main purposes" of the NSR program was "to allow States greater flexibility for the former purpose (i.e., allowing 'reasonable economic growth to continue in an area while making reasonable further progress to assure attainment') than EPA's present interpretative regulations afford" (H.R. Rep. 95-294, supra, at 211, 4 Leg. Hist. 2678 (Emphasis added)). Against this background, it is hardly plausible that Congress should have intended, sub silentio, to freeze into place an approach to new source review that would prevent the Administrator from giving states any discretion to adopt a definition of the (statutorily undefined) term "source" precisely calculated to achieve the statute's "main purposes." We submit that the most significant fact about the "administrative" history of the question at issue is not the pre-1977 history at all. It is that, ever since the 1977 Amendments were adopted, EPA has taken the position that the statute does not itself mandate one authoritative definition of "source" for purposes of new source review. As our principal brief recounts in detail (at 9-14), from the first time that it addressed the 1977 Amendments the agency has asserted the power to allow states to adopt a plantwide approach. (Respondents' argument (Br. 18) that EPA's 1979 proposals were somewhat narrower than the rules adopted in 1981 is, of course, irrelevant on the question at issue: whether EPA has power to permit states to use the "bubble" concept at all.) In fact, when EPA in August 1980 adopted the dual definition, it did not assert that it has no discretion in the premises; its action constituted an exercise of that discretion. This case is, then, one for the application of the conventional rule of administrative law that a court should defer to the agency's view of its own authority under a statute committed to it for enforcement. See Gov't Br. 22. V. Respondents deploy various policy arguments to show that the "bubble" approach will subvert the purposes of the Clean Air Act. They contend, first, that the EPA regulation is invalid because it does not require a greater-than-even offset and therefore conflicts with the purpose of "(m)aking the construction of major new sources a force for reducing pollution" (Br. 38). As shown in our opening brief (at 26-27, 28, 35), however, the purpose of new source review is to ensure that emissions from new or modified sources do not prejudice the transition to attainment. New source review is only one part of an integrated and complex scheme; it was not designed to carry by itself the burden of improving air quality in nonattainment regions. By looking to the net impact on pollution levels of a new or modified source, the plantwide definition serves the preventive objective of the NSR program. Respondents also object (Br. 4, 18, 20, 40 n.69) that the EPA regulation does nto require new source review even when there is not a total offset in emissions at a plant. This objection, however, does not bear on the issue of the NSR definition of "source." Rather, it is an attack on the significance levels that EPA has set pursuant to the decision in Alabama Power Co. v. Costle, supra, /21/ which were not challenged in a proceeding for judicial review and whose validity is not before the Court. It should also be noted that respondents' argument applies to the dual as well as the plantwide definition: the significance levels are applicable under both definitions (see Gov't Br. 6 n.10, 9 n.14, 28 n.36). In fact the increases permitted by the significance levels are truly de minimis (see Gov't Br. 28 n.36). The NSR statute itself excludes -- as "minor" -- sources that emit less than 100 tons per year of a pollutant (see Gov't Br. 6 n.10, 9 n.14); surely it should not be read to require a dual definition of "source" in order to trigger review of every marginal increase in emissions. The existing EPA regulation fairly effectuates the statutory purpose of new source review: to ensure that emissions resulting from economic growth and industrial expansion do not interfere with the goal of attainment. /22/ Respondents further urge (Br. 2, 4, 32, 40-41) that it is illogical and contrary to the purposes of the Act to require new source review when a large piece of equipment is installed at a new site, but not to require it when that same piece of equipment is installed (and offsetting pollution reductions are made) at an existing plant. But, once again, respondents have overlooked the "hold harmless" purpose of new source review in the overall context of the nonattainment program. In the case of the existing plant, the plantwide definition ensures that the industrial modernization or expansion does not interfere with air quality attainment. On the other hand, the project at the new site, by hypothesis, results in a new increase in emissions that significantly worsen the level of pollution, thus interfering with the goal of attainment. To make new source review a function of a project's net impact on air quality and its consequences for attainment, as the plantwide definition does, is in full accord with the design and purpose of the NSR program. /23/ VI. Respondents lastly contend (Br. 4-5, n.11, 19-20, 40 n.69, 42, 44) that the EPA regulation, even if valid with respect to construction or modifications in areas in which there is an approved state implementation plan, is improper in areas that do not have an approved plan and are subject to the construction moratorium (see Gov't Br. 7). Respondents suggest no reason why EPA should be required to give two different meanings to the term "source" in the NSR program, /24/ and to do so would exacerbate the regulatory complexities and burdens that the 1981 regulations were intended to reduce (see Gov't Br. 12-14, 35 n.45). Furthermore, the regulation is consistent with the purposes of the moratorium: by allowing industrial growth that does not adversely affect air quality while prohibiting both the construction of all new facilities and the renovation of existing facilities for which there is a non-de minimis emissions increase, it prevents a deterioration in air quality during the period an approved implementation plan is not in place and provides a strong incentive for the prompt submission of a plan that complies with federal requirements (see Gov't Br. 27-28 n.35, 29 n.37, 34-35 n.44). Accordingly, application of the EPA regulation in areas subject to the construction moratorium is not invalid. VII. Even if the NSPS definition of "source" -- "any building, structure, facility, or installation which emits or may emit any air pollutant" (42 U.S.C. 1857c-6(a)(3) and 7411(a)(3)) -- were formally applicable to new source review, it would not resolve the question presented in this case: the industrial unit or entity that is subject to such review. Consistent with general principles of administrative law, the D.C. Circuit has recognized, in this case /25/ and in other cases involving the Clean Air Act, /26/ that EPA has discretion to construe the constitutent terms of the definition in light of the language, history, and purposes of the particular program under consideration; as these decisions "illustrate, different delineations may be appropriate for different statutory programs" (82-1005 Pet. App. A9 n.22). Once the "bright line" test adopted by the court below is set aside -- as even respondents do not contest it must be -- there is no reason why the terms "building," "structure," "facility," and "installation" cannot be understood to embrace an entire plant rather than a specific, individual item of equipment. /27/ Respondents offer two arguments in response to this analysis. First, they assert (Br. 29-30 n.56) that a dual definition applies throughout the Clean Air Act and that Alabama Power supports that conclusion. But Alabama Power does no such thing. As the court below understood (82-1005 Pet. App. A2-A3, A13, A16), Alabama Power clearly holds that the plantwide definition is required in the PSD program and that the dual definition is impermissible (636 F.2d at 396-398, 401-402). /28/ Second, respondents argue (Br. 8, 27, 41-42) that in the NSPS program itself every individual piece of pollution-emitting equipment must be deemed to be a "source." Again, this is incorrect. The NSPS program calls upon the Administrator to "publish * * * a list of categories of stationary sources * * * (that) in his judgment * * * cause(), or contribute() significantly to, air pollution" (42 U.S.C. 1857c-6(b)(1)(A)) and 7411(b)(1)(A)), for to establish "Federal standards of performance for new sources within (each) such category" (42 U.S.C. 1857c6(b)(1)(B) and 7411 (b)(1)(B)); in establishing such standards the Administrator is authorized to "distinguish among classes, types and sizes within categories of new sources" (42 U.S.C. 1857c-6(b)(2) and 7411(b)(2)). Thus, in the NSPS program itself, the Administrator exercises discretion and judgment in listing categories of sources and in determining the appropriate "source" to be subject to a given standard. In contrast, if the NSPS meaning of "source" were simply carried over to new source review, such flexibility would be lost. Instead -- and this is the very result respondents seek -- the NSR program would rigidly encompass any and all industrial units that fall within the expanse of the NSPS definition. Such an outcome wrenches the NSPS definition from its substantive and statutory context; it is neither good law nor good sense. This brings us back to where we began. Our submission is that the definitional issue in this case is, like many other questions brought to the courts, a matter of sound policy, not legal validity. Respondents bear a heavy burden in challenging that submission. To prevail, the must demonstrate that the Act precludes the regulation and that the Administrator has no discretion -- either in promulgating a definition of "source" for NSR or in interpreting the constituent terms in the NSPS definition. Nothing in the text, history, or policies of the Act, or in its administrative implementation, warrants such an extreme conclusion. For the foregoing reasons and those stated in our opening brief, it is respectfully submitted that the judgment of the court of appeals should be reversed. REX E. LEE Solicitor General FEBRUARY 1984 /1/ There is no basis whatever for responents' assertion (Br. 4, 37, 41, 42) that EPA is seeking to dispense with or waive the statutory requirements prescribed by Congress. The issue in this case is the definition of the "source" that is subject to those requirements; it does not avoid or waive the statutory standards not to apply new source review to an industrial unit that is not a "source." As is common fare for administrative action, EPA has done no more here than define a term for which no legislative definition is provided (see Gov't Br. 40-41). /2/ In addition to the authorities cited in our opening brief, see also Rhinelander, The Proper Place for the Bubble Concept Under the Clean Air Act, 13 Envtl. L. Rep. 10406 (1983); Comment, The Bubble Concept -- A Feasible Emissions Reduction Alternative?, 9 Dayton L. Rev. 65 (1983). /3/ "Br." refers to respondents' brief in this Court. /4/ As in our opening brief (see Gov't Br. 4 n.4), statutory citations include both the original and recodified versions of the Act. References to the current version of the Act are to "Supp. V" of the 1976 edition of the United States Code; the designation "Supp. V" will be omitted from such citations. /5/ In fact, the locational review sections do not simply incorporate the NSPS definition of "source" or apply to all "sources" within the NSPS definition. They require locational review for those sources for which an NSPS standard has in fact been promulgated by the Administrator. With respect to the NSPS process, see 42 U.S.C. 1857c-6(b) and 7411(b); page 20, infra. /6/ In practice locational review did not turn out to be a program of great significance; "(f)or the first 4 years after the 1970 amendments, States and EPA conducted rather spotty review of new sources in dirty air areas." 123 Cong. Rec. 18017 (1977) (remarks of Senator Muskie), 3 Leg. Hist. 713. /7/ See Gov't Br. 4-5; H.R. Conf. Rep. 95-564, 95th Cong., 1st Sess. 155-158 (1977), 3 Leg. Hist. 535-538; S. Rep. 95-127, 95th Cong., 1st Sess. 55 (1977), 3 Leg. Hist. 1429; H.R. Rep. 95-294, 95th Cong., 1st Sess. 207-211 (1977), 4 Leg. Hist. 2674-2678. /8/ S. Rep. 95-127, supra, at 55, 3 Leg. Hist. 1429. /9/ H.R. 95-294, supra, at 210, 4 Leg. Hist. 2677. /10/ Both in the original enactment and in subsequent amendments, when Congress intended to apply the NSPS definition to other programs, it said so and made such modifications as were appropriate. See 42 U.S.C. 1857c-7(a)(3) and 7412(a)(3); Energy Supply and Environmental Coordination Act of 1974, 42 U.S.C. 1857c-10, repealed, Pub. L. No. 95-95, Sec. 112(b)(1), 91 Stat. 709, 3 Leg. Hist. 215. /11/ S. 3219, 94th Cong., 2d Sess. Sec. 302(m), at 79 (1976), 6 Leg. Hist. 4688, 4918. /12/ S. 3219, supra, Sec. 113(g)(1), at 43, 6 Leg. Hist. 4652, 4869. /13/ "Where Congress includes limiting language in an earlier version of a bill but deletes it prior to enactment, it may be presumed that the limitation was not intended" (Russello v. United States, No. 82-472 (Nov. 1, 1983), slip op. 7), and "(i)t is surely not for (the Court) to add to the statute what Congress * * * decided to omit" (Secretary of Interior v. California, No. 82-1326 (Jan. 11, 1984), slip op. 21). /14/ It also ignores the fact, noted in our reply brief at the petition stage (at 5 n.6), that the 1977 Senate bill, although introduced in the same form as the 1976 bill, was itself significantly amended in the Senate committee and followed the NSR provisions of the 1976 bill only for an interim period until July 1, 1979. See S. 252, 95th Cong., 1st Sess. Sections 110(a)(3)(A) and 113(g)(1), at 58-60, 61-62 (1977), 3 Leg. Hist. 632-634 and 635-636, 1520 and 1540. /15/ In fact an examination of the House bill will show that its permit program was largely the same as the one ultimately adopted. See Gov't Br. 5-7, and H.R. 6161, 95th Cong., 1st Sess. Section 127, at 347-353 (1977), 4 Leg. Hist. 2332-2338, 2854-2857. /16/ The confusion is caused, in part, by the use of the word "another" in the sentence: "In the House bill, there was simply no need for another provision to define "major stationary source'" (Br. 31 (emphasis added)). This suggests what is not the case -- that the just quoted Section 111 definition is made applicable to the NSR program by some other provision. In this connection, respondents refer (Br. 31 n.59) to the fact that the 1976 House bill -- unlike the 1977 bill -- included a definition of "major stationary source" that refers to the NSPS definition of "source." See H.R. 10498, 94th Cong., 2d Sess. Sections 121(a)(3) and 127(c)(3), at 159, 236 (1976), 7 Leg. Hist. 5895 and 5972, 6880 and 6896. However, this provision is not found in either the 1977 House bill or the final statute, and no significance can be attached to it. In addition, the 1976 provision also encompasses "a source which is a major stationary source within the meaning of * * * (the section for the) prevention of significant deterioration of air quality * * *," which, under the decision in Alabama Power Co. v. Costle, 636 F.2d 323 (D.C. Cir. 1979), requires a plantwide definition of "source." Finally, the 1976 House bill envisioned a new source review program involving a system of variances that would control emissions from new and existing sources -- a fundamentally different approach than that taken in the 1977 House bill or the NSR statute enacted by Congress; the definitional provision cannot be separated and viewed in isolation from the substantive program to which it applied. /17/ Respondents rely (Br. 36) on part of a statement made by Senator Muskie during the debates over the Conference Report. Even assuming that great weight can be accorded to a single floor remark (see, e.g., Pennhurst State School v. Halderman, 451 U.S. 1, 20 (1981); Chrysler Corp. v. Brown, 441 U.S. 281, 311 (1979)), this statement does not carry respondents' argument. Fairly read in its entirety, it appears to do no more than explain the difference between "new" and "modified" sources -- both subject to new source review -- and makes clear that a "new" source does not escape review simply because it replaces an older facility that had higher emissions. See 123 Cong. Rec. 26847 (1977), 3 Leg. Hist. 355-356. In addition, respondents (Br. 38) and amicus United Steelworkers of America cite to the history of the Steel Industry Compliance Extension Act of 1981 (42 U.S.C. 7413) to show that Congress contemplated a process-unit definition of "source." The purpose of the Steel Extension Act was to afford assistance to the steel industry by extending certain compliance dates. Neither the Act nor its legislative history focused on the quite unrelated issue of the definition of "source" under existing law. When the Steel Extension Act was under consideration in May and June of 1981, the EPA regulations then in force provided for a dual definition (see Gov't Br. 11-14); thus, it is not surprising that the working assumption of those involved might reflect a process-unit orientation. Cf. United States v. Clark, 445 U.S. 23, 33 n.9 (1980). And, given that in March 1981 EPA had published proposed rules allowing states to adopt the plantwide definition and had invited comment on the matter, the only reasonable assumption is that, if Congress had intended to resolve the issue, it would have spoken with far greater focus than is indicated in the cited legislative record. In these circumstances, deductions drawn from the views expressed in 1981 "'form a hazardous basis for inferring the intent'" of Congress in enacting the NSR statute in 1977. See CPSC v. GTE Sylvania, 447 U.S. 102, 117 (1980) (citation omitted); see also, e.g., Regional Rail Reorganization Act Cases, 419 U.S. 102, 132 (1974). /18/ Respondents' reference to EPA's practice back to 1971 is based on the 1970 locational review provisions, which (as we have shown) have no bearing on the question presented here. Furthermore, since those provisions expressly referred to the NSPS section, the fact that EPA in 1971 and 1973 construed or applied the locational review program in light of NSPS is without significance. In fact, EPA's pre-1977 regulations recognized state flexibility to determine the "sources" subject to such review. See 36 Fed. Reg. 22398, 22400, 22406 (1971); 38 Fed. Reg. 15834, 15835, 15836 (1973). /19/ However, EPA did not explain or elaborate on its use of the NSPS language in defining "major source." It is not surprising, therefore, that respondents cannot point to any evidence that Congress knew about or focused on the specific definitional issue. /20/ See also H.R. Conf. Rep. 95-564, 95th Cong., 1st Sess. 157 (1977), 3 Leg. Hist. 537; H.R. Rep. 95-294, supra, at 14, 211, 4 Leg. Hist. 2481, 2678. /21/ See Gov't Br. 6 n.10; 40 C.F.R. 51.18(j)(1)(x); 40 C.F.R. 52.24(f)(10); 45 Fed. Reg. 52676, 52705 (1980). /22/ As discussed in our opening brief (at 13, 14, 29-34), the plantwide definition can facilitate reductions in emissions by encouraging the replacement of older, heavily polluting equipment with newer, cleaner equipment. (Respondents' claim that this contention was abandoned in the court of appeals (Br. 22, 44, 45, 46) is unfounded; see EPA C.A. Br. 12, 37, n.34.) Respondents echo (Br. 42, 44-47) the court of appeals' criticism (82-1005 Pet. App. A18 n.41 (emphasis in original)) that, "in abandoning (in 1981) its earlier position (in the 1980 regulation), EPA did not cite, nor have we found in the record, any study, survey, or support for the opposite position, now tendered by EPA, that the dual definition would indeed retard improvement of air quality in the aggregate * * * by discouraging replacement of older processes or equipment with newer, cleaner varieties." This fails to recognize that the Administrator's conclusion involved a prediction of how business and investment decisions would differ in response to the plantwide and dual definitions. "(T)o the extent that factual determinations were involved * * *, they were primarily of a judgmental or predictive nature * * *. In such circumstances complete factual support in the record for the (agency's) judgment or prediction is not possible or required * * * ." FCC v. National Citizens Committee for Broadcasting, 436 U.S. 775, 813-814 (1978). In fact, EPA had as much record support in promulgating the plantwide definition as it had in adopting the dual definition -- a definition that had been in place for only seven months when the proposed change was announced. The critical point is that the agency gave a full and reasoned explanation for the change in the rule. Compare Motor Vehicle Manufacturers Association v. State Farm Mutual Co., No. 82-354 (June 24, 1983), slip op. 10-11, 26. /23/ For the same reason, there is no merit to respondents' argument (Br. 40-41) that the EPA regulation improperly discriminates between existing and new business competitors in a nonattainment area. In much the same vein, the amici states argue that the EPA regulation, by not requiring a uniform national definition of "source," will allow a state to be pressured into adopting the definition that minimizes its environmental controls in order to attract new businesses; in effect, amici ask the Court to relieve them of the hard choices about accommodating growth and environmental protection (see State Am. Br. 11) that Congress desired in 1977 to leave to the states. Their argument -- which would be satisfied by any nationwide definition and therefore lends no support to the dual definition as such -- overlooks the fact that the Act requires the states to have an EPA-approved implementation plan that conforms to the requirements of the Act and provides for reasonable progress toward and timely attainment of national air quality standards. That states have flexibility to design the detailed methods and enforcement procedures for achieving attainment, consistent with federal law, and to make economic choices between existing and new sources or between present and future growth, is exactly what Congress intended in 1977 (see Gov't Br. 23, 35-41); indeed, the statutory provision that allows a state to budget, at its option, an emissions allowance for growth (42 U.S.C. 7502(b)(5)) clearly reflects this intent. Amici have cited nothing in the legislative history to indicate that Congress was so concerned about economic competition among states that it decided to eliminate the very state flexibility it said it was trying to increase. In fact, where Congress spoke to that issue, it did so in other programs, such as NSPS (see H.R. Conf. Rep. 95-564, supra, at 129, 3 Leg. Hist. 509; H.R. Rep. 95-294, supra, at 11, 184, 4 Leg. Hist. 2478 2651; 123 Cong. Rec. 26846 (1977) (remarks of Sen. Muskie), 3 Leg. Hist. 353). Finally, given the numerous other differences among states in environmental and economic conditions, we do not believe that state flexibility regarding the definition of "source" will have a significant impact on the complex considerations that underlie the decisions of states and businesses; in fact, a recent study has concluded that environmental regulation has virtually no effect on a company's decision where to locate its operations. See C. Duerksen, Environmental Regulation of Industrial Plant Siting 56-71 (Conservation Foundation 1983.) /24/ As the court below noted (82-1005 Pet. App. A17 n.40), the moratorium "has been cast, not as a discrete program, but as an integral part of a larger production, the nonattainment scheme." /25/ 82-1005 Pet. App. A9 n.22. /26/ Alabama Power Co. v. Costle, 636 F.2d at 396-398, 410; ASARCO, Inc. v. EPA, 578 F.2d 319, 324 n.17 (1978). /27/ We note that in any event the NSPS definition in no way requires a dual definition of "source." /28/ In Alabama Power, the court stated (636 F.2d at 397) that, in order "(t)o allow an entire plant or other appropriate grouping of industrial activity to be subject as a single unit to PSD, as Congress clearly intended, EPA should devise regulatory definitions of the terms 'structure,' 'building,' 'facility,' and 'installation' to provide for the aggregation, where appropriate, of industrial activities according to considerations such as proximity and ownership." APPENDIX APPENDIX MATERIAL IS NOT AVAILABLE ON JURIS.