Opinion for the Court filed PER CURIAM.
TABLE OF CONTENTS
Page
I. Ripeness................................................................189
II. The NEPA-Related Regulations..........................................191
A. Permit Conditions Unrelated to Effluents — Ripeness..................192
B. Permit Conditions Unrelated to Effluents — Merits ....................192
C. Effluent-Related Permit Conditions..................................194
D. Admissibility of Evidence in Permit Proceedings......................195
III. State Program Requirements ............................................196
A. Penalties and Participation — Ripeness................................197
B. Penalties and Participation — Merits................................ 197
1. Regulatory uniformity and state autonomy......................198
2. Public participation.............................................199
3. Maximum penalties.............................................202
C. EPA Veto Authority — Ripeness......................................205
D. EPA Veto Authority — Merits........................................206
IV. Toxicity Limitations .....................................................213
A. Statutory Authority.................................................213
B. Technical Feasibility.................................................213
C. Intrusion on State Authority ........................................214
D. Procedural Claims...................................................214
V. Non-Adversary Panel Procedures........................................215
A. The Merits of the Procedures .......................................216
1. Oral testimony and cross examination...........................216
2. The panels’ composition ........................................217
B. Defects in the Mode of Adoption....................................218
VI.The Antibacksliding Controversy.........................................219
A. The NSPS Issue....................................................220
B. Attack on the BPJ Regulation — Ripeness.............................220
C. Attack on the BPJ Regulation — Merits...............................221
D. Application of the New WQA Rules.................................228
VII. Net/Gross Limits........................................................228
VIII. Upset Defense....................................................... 229
A. Ripeness............................................................229
B. Merits..............................................................230
IX. APA Continuance.......................................................235
A. Ripeness............................................................235
B. Merits..............................................................236
1. Consistency with Clean Water Act scheme......................237
2. Consistency with section 558(c) .................................238
PER CURIAM:These consolidated cases arise out of the Environmental Protection Agency’s issuance of regulations implementing the National Pollution Discharge Elimination System (NPDES) permit program, established under the Clean Water Act, 33 U.S. C. §§ 1251 et seq. (1982). The events leading up to these challenges are set forth fully in our prior opinion, which disposed of several issues. See NRDC v. EPA, 822 F.2d 104 (D.C.Cir.1987). This opinion addresses the remaining challenges, mounted by both environmental and industry petitioners.1
Although numerous issues are raised, they fall into eight general areas or categories. The first set of challenges involves regulations promulgated by EPA in furtherance of its obligations under the National Environmental Policy Act (“NEPA”), 42 U.S.C. §§ 4321 et seq. (1982).2 Also challenged are regulations that (2) establish the rules for transfer of authority over the NPDES permitting program to the states and supervision of the state programs; 3 (3) permit the establishment of permit limits in terms of “toxicity”; (4) permit the use of non-adversary panel procedures (“NAPP” or “NAP procedures”) for the issuance of initial permits and variances; (5) prohibit “backsliding” from permit limits when subsequent guidelines have become more lax;4 and (6) define permit limits in “gross” terms, giving credit for pollutants in intake waters only in certain limited circumstances. Industry challenges EPA’s refusal to provide an “upset defense” for noncompliance with water-quality-based permit limits; that is, a defense based on a showing that noncompliance was due to factors beyond the permittee’s control. Finally, NRDC challenges on several fronts EPA’s regulation providing for continuances of out-of-date permits pending renewal by the agency.
For the reasons that follow, we deny the petitions for review on all issues, save for those pertaining to the permit conditions unrelated to effluent limits and the upset defense.
I. Ripeness
In its initial brief and argument before this court, the agency questioned the • ripeness of several of the challenges to its NPDES regulations. Because of doubt as to the ripeness of other issues as well, and in the interests of consistency, we asked for briefing on ripeness with respect to all. We note that we may find a want of ripeness even where the agency raises no such defense. This is clearly so where the issue is so unripe that the petitioner is in substance asking for an advisory opinion, in violation of the case-or-controversy requirement of Article III. See Muskrat v. United States, 219 U.S. 346, 31 S.Ct. 250, 55 L.Ed. 246 (1911); Golden v. Zwickler, 394 *190U.S. 103, 89 S.Ct. 956, 22 L.Ed.2d 113 (1969); see also Regional Rail Reorganization Act Cases, 419 U.S. 102, 138, 95 S.Ct. 335, 356, 42 L.Ed.2d 320 (1974) (noting Article III component of ripeness); cf. Gene R. Nichol, Jr., Ripeness and the Constitution, 54 U.Chi.L.Rev. 153 (1987) (generally arguing against constitutionalization of ripeness law but noting the advisory opinion element). Even in the absence of any impingement on Article III, the fact that ripeness rests in part on concerns of judicial economy suggests a need for independent judicial scrutiny, as occurs under the doctrine of exhaustion. See ASARCO, Inc. v. EPA, 578 F.2d 319, 320-21 n. 1 (D.C.Cir.1978) (weighing “the interests of judicial efficiency” as a factor in exhaustion analysis).
Ripeness of course goes to when an issue may be raised. The Supreme Court has succinctly identified the functions of the doctrine in Abbott Laboratories v. Gardner, 387 U.S. 136, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967), stating that
its basic rationale is [1] to prevent the courts, through avoidance of premature adjudication, from entangling themselves in abstract disagreements over administrative policies, and also [2] to protect agencies from judicial interference until an administrative decision has been formalized and its effects felt in a concrete way by the challenging parties.
Id. at 148-49, 87 S.Ct. at 1515. From those two purposes the Court derived a two-part test (its parts nowadays usually referred to as “prongs,” making it one of the many pronghomed creatures in the modern legal bestiary), consisting of “[1] the fitness of the issues for judicial decision and [2] the hardship to the parties of withholding court consideration.” Id. at 149, 87 S.Ct. at 1515. As this court has put it, for a claim to be ripe, “the interests of the court and agency in postponing review until the question arises in some more concrete and final form [must be] outweighed by the interests of those who seek relief from the challenged action’s ‘immediate and practical impact' upon them.” Continental Air Lines, Inc. v. CAB, 522 F.2d 107, 125 (D.C.Cir.1974).
The hardship element has considerably evolved in this circuit. As formulated in Toilet Goods Association, Inc. v. Gardner, 387 U.S. 158, 87 S.Ct. 1520, 18 L.Ed.2d 697 (1967), decided the same day as Abbott Laboratories, the test was satisfied only if the challenged administrative action would have a “direct and immediate effect” on the “primary conduct” of the petitioner. Id. at 164, 87 S.Ct. at 1524-25; accord Office of Communication of United Church of Christ, 826 F.2d 101, 109 (D.C.Cir.1987). The paradigmatic hardship situation is where a petitioner is put to the choice between incurring substantial costs to comply with allegedly unlawful agency regulations and risking serious penalties for non-compliance. Abbott Laboratories, 387 U.S. at 152, 87 S.Ct. at 1517. A petitioner cannot show hardship by positing a speculative or hypothetical future harm. Tennessee Gas Pipeline Co. v. FERC, 736 F.2d 747, 750 (D.C.Cir.1984). Further, neither the possibility that the petitioner may have to make capital budgeting decisions under a cloud of uncertainty, Diamond Shamrock Corp. v. Costle, 580 F.2d 670, 673 n. 1 (D.C.Cir.1978), nor the fact that it may incur future expense in challenging the regulations in a later permit or enforcement proceeding, Arkansas Power & Light v. ICC, 725 F.2d 716, 726 (D.C.Cir.1984), will qualify as hardship.
In Eagle-Picher Industries, Inc. v. EPA, 759 F.2d 905, (D.C.Cir.1985), this court adjusted the hardship analysis to reflect the possibility that institutional concerns might militate in favor of early review despite the absence of conventional hardship. In that case the court was asked to rule upon the legality of the “Hazardous Ranking System” adopted by the EPA pursuant to the Comprehensive Environmental Response, Compensation and Liability Act of 1980 (“CERCLA”), 42 U.S.C. §§ 9601 et seq. Apart from the issues raised being pure ones of law and the agency’s position being final and fully crystalized, we noted special circumstances favoring early review. First, CERCLA’s statutory review provision, 42 U.S.C. § 9613(a), required parties *191to challenge newly promulgated regulations within 90 days and barred such challenges at the enforcement stage. We read the review provision to bespeak a congressional desire that the courts “provide prompt, uniform ‘preenforcement’ review of CERCLA regulations ... in order to avoid needless delays in the implementation of an important national program.” 759 F.2d at 916. Second, the agency urged the court not to postpone review. It sought an early and final resolution of legal attacks on decisions that formed the keystone of a complex regulatory program. Id. at 916-17. Third, there was no “compelling” judicial interest in deferring review. Id. at 917. Accordingly the court concluded that “mechanical application” of the hardship element of Abbott Laboratories “could work mischief in such a situation,” causing postponement of review even where all the institutional interests sought to be served by the doctrine, and an express congressional determination, militated in favor of early review.
Here as in Eagle-Picker the congressional provision for judicial review, 33 U.S.C. §v1369(b), evinces a strong will that it occur at the time of promulgation. It provides a 90-day deadline and bars a party from making at the time of enforcement claims that it could have brought at the time of promulgation.5 We cannot, however, apply Eagle-Picker across the board to the welter of claims raised here. They are not by any means all so purely legal as to be clearly fit for review; the agency in some instances asserts that its institutional interests favor deferred review; and in some cases there is real doubt whether the issue will ever have material significance. Accordingly, where there are institutional benefits for court or agency in deferring review, we must consider the hardship to the challenging parties from delay, and proceed to the merits only when the latter outweighs the former.
II. The NEPA-Related Regulations
As is well known, the National Environmental Policy Act (“NEPA”), 42 U.S.C. §§ 4321 et seq., requires environmental impact statements (“EISs”) for “major Federal actions” affecting the environment, and normally authorizes the relevant agency to make some adjustment in its conduct as a result of what it learns from the necessary inquiry. 42 U.S.C. § 4332; NRDC v. Berklund, 609 F.2d 553, 558 (D.C.Cir.1979). In the area of water quality regulation, § 511(c) of the CWA, 33 U.S.C. § 1371(c), structures and limits the role of NEPA.6 It makes clear that the provision of federal financial assistance for the construction of publicly owned treatment works and the issuance of discharge permits to new sources are the only actions taken by the Administrator under the CWA that will trigger a NEPA duty.
Section 511(c)(2) of the Act further limits NEPA’s effect. Subsection (B) states that NEPA does not authorize any federal per*192mitting authority “to impose, as a condition precedent to the issuance of any license or permit, any effluent limitation other than any such limitation established pursuant to [the Act].” 33 U.S.C. § 1371(c)(2)(B).
The interaction of the CWA and NEPA has led EPA to adopt regulations in three separate areas, all challenged by parties to this suit. First, it took the view that it could, on the basis of NEPA, subject permits to conditions not related to water quality. Industry asserts that it may never do so, and we find this pure legal issue ripe. To the extent, however, that industry attacks the agency’s general predictions of when it may do so, we find the challenge unripe.
Second, the agency’s regulations abjure any authority under NEPA to impose conditions relating to water quality. Because the impact of this position depends entirely on nuances of the agency’s interpretation of the CWA, we find it unripe.
Finally, the regulations set up a general framework for limiting evidence in permit proceedings on the basis of prior EPA proceedings relating to the same source. Environmentalists object to the restriction. Because the significance of these rules turns on their application, we find the claim unripe.
We address the matters in that order.
A. Permit Conditions Unrelated to Effluents — Ripeness
Industry challenges 40 C.F.R. §§ 122.29(c)(3), 122.44(d)(9) and 122.49(g), on the grounds that they unlawfully assert a power to impose such conditions. One may question whether they literally do so. Their assertions of power are oddly qualified. 40 C.F.R. § 122.29(c)(3) provides that:
[t]he Regional Administrator, to the extent allowed by law, shall issue, condition ... or deny the new source NPDES permit following a complete evaluation of any significant beneficial and adverse effects of the proposed action and a review of the recommendations contained in the EIS or finding of no significant impact.
(emphasis added). See also 40 C.F.R. § 122.44(d)(9) (providing that new source NPDES permits shall “[ijncorporate ... appropriate requirements, conditions, or limitations ... to the extent allowed by [NEPA] and section 511 of the CWA ”) (emphasis added); 40 C.F.R. § 122.49(g) (calling for consideration of EIS-related permit conditions as provided in § 122.29(c)).
It would be sheer casuistry, however, to read the qualification as completely undoing the assertion of power. First, it would make the regulation meaningless or self-contradictory. Second, the agency in its preamble alludes to its prior imposition of such conditions and flatly asserts its intention to do so again. 49 Fed.Reg. 38,034/2. Third, the agency's supplemental brief, while denying ripeness, explains that the regulations “reflect[] EPA’s definitive position that it possesses authority to impose [such] conditions____” Supplemental Submission of EPA on Ripeness at 12. Given the purely legal character of the claim and the apparent certainty that it will arise in the future, we think the institutional concerns underlying ripeness militate in favor of immediate resolution. To the extent, however, that industry merely attacks the agency's hints in the preamble as to when it may impose such conditions, see, e.g., Joint Industry Petitioners’ Brief on Ripeness at 12, the challenge is completely fact-dependent and therefore unripe.
B. Permit Conditions Unrelated to Effluents — Merits
We thus proceed to the merits of Industry’s challenge. As the issue is one of statutory interpretations, our analytical framework is provided by Chevron U.S.A., Inc. v. NRDC, 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), and its progeny. To restate briefly these now-familiar principles: On a pure question of statutory construction, we attempt to discern congressional intent, using “traditional tools of statutory construction.” INS v. Cardoza-Fonseca, 480 U.S. 421, 107 S.Ct. 1207, 1221, 94 L.Ed.2d 434 (1987). If Congress’ intent is clear, “then that interpretation must be given effect, and the regulations at issue *193must be fully consistent with it.” NLRB v. United Food & Commercial Workers Union, 484 U.S. 112, 108 S.Ct. 413, 98 L.Ed.2d 429 (1987). But where “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.” Chevron, 467 U.S. at 843, 104 S.Ct. at 2782. See generally Continental Airlines, Inc. v. Dep’t of Transportation, 843 F.2d 1444 (D.C.Cir.1988); Rettig v. Pension Benefit Guaranty Corp., 744 F.2d 133 (D.C.Cir.1984).
For the reasons that follow, we conclude that neither the Clean Water Act nor NEPA authorizes EPA’s imposition of non-water quality permit conditions. Accordingly, the challenged regulations, 40 C.F.R. §§ 122.29(c)(3), 122.44(d)(9), 122.49(g), cannot stand.
In arguing that it is empowered to impose permit conditions unrelated to effluents, EPA relies primarily on NEPA itself. As noted above, section 511(c)(1) of the CWA provides that EPA’s issuance of new source permits shall be considered a “major Federal action” for NEPA purposes and therefore subjects to NEPA review. 33 U.S.C. § 1371(c)(1). EPA contends that as part of this review, NEPA authorizes the agency to make decisions based on environmental factors not expressly identified in the agency’s underlying statute. With this general proposition, we fully agree.
Where we part company with EPA is in the nature of the decision NEPA authorizes the agency to make. As EPA sees it, NEPA provides the agency with “supplemental authority.” This authority, it is said, permits EPA not only to consider additional environmental factors, but to act on these factors by imposing any condition necessary to account for the environmental effects of the entire new facility. EPA Brief at 24, 26, 30.7 This, we believe, misapprehends the nature and import of NEPA. True enough, NEPA instructs the agency to consider all environmental effects of any “major Federal action” and to incorporate this information into its final decision. 42 U.S.C. § 4332. NEPA does not, however, expand the range of final decisions an agency is authorized to make. As we reiterated in our earlier decision (in holding that EPA lacked authority to ban construction of new sources pending permit issuance), NEPA does not expand an agency’s substantive powers. NRDC v. EPA, 822 F.2d 104, 129 (D.C.Cir.1987). Any action taken by a federal agency must fall within the agency’s appropriate province under its organic statute(s). See, e.g., id.; Park County Resource Council, Inc. v. Dep’t of Agric., 817 F.2d 609, 620 (10th Cir.1987); Winnebago Tribe v. Ray, 621 F.2d 269, 272-73 (8th Cir.), cert. denied, 449 U.S. 836, 101 S.Ct. 110, 66 L.Ed.2d 43 (1980); Gage v. United States Atomic Energy Comm’n, 479 F.2d 1214, 1220 n. 19 (D.C.Cir.1973).
Here, the “major Federal action” subject to NEPA review is identified in section 511(c)(1): the issuance of a permit to discharge pollutants. 33 U.S.C. § 1371(c)(1). In fulfillment of its NEPAmandated duty, EPA is to consider all “direct, indirect, and cumulative” environmental effects of the discharge of pollutants by the potential permittee. See 40 C.F.R. §§ 1508.7-.8, 1508.18 (1987). Having done so, EPA can properly take only those actions authorized by the CWA — allowing, prohibiting, or conditioning the pollutant *194discharge. 33 U.S.C. § 1342; see also NRDC v. EPA, 822 F.2d at 129 (“EPA’s jurisdiction [under the CWA] is limited to regulating the discharge of pollutants____”).8 And, contrary to EPA’s assumption, the CWA does not empower the agency to regulate point sources themselves; rather, EPA’s jurisdiction under the operative statute is limited to regulating the discharge of pollutants. Thus, just as EPA lacks authority to ban construction of new sources pending permit issuance, so the agency is powerless to impose permit conditions unrelated to the discharge itself. See NRDC v. EPA, 822 F.2d at 126-31.
This is not to say that EPA lacks authority to take action in the wake of its NEPA-mandated review. To the contrary, it can deny a new-source permit on NEPArelated grounds or impose NEPA-inspired conditions on discharges that the agency determines to allow. Indeed, Industry does not contest this authority. See Joint Industry Petitioners’ Brief on Common Issues at 47 n. 1; Joint Industry Petitioners’ Reply Brief at 32-33 & n. 1. EPA may not, however, under the guise of carrying out its responsibilities under NEPA transmogrify its obligation to regulate discharges into a mandate to regulate the plants or facilities themselves. To do so would unjustifiably expand the agency’s authority beyond its proper perimeters. See e.g., Board of Governors v. Dimension Financial Corp., 474 U.S. 361, 374, 106 S.Ct. 681, 689, 88 L.Ed.2d 69 (1986); Schwabacher v. United States, 334 U.S. 182, 209-10, 68 S.Ct. 958, 972, 92 L.Ed. 1305 (1948) (Frankfurter, J., dissenting); American Mining Congress v. EPA, 824 F.2d 1177, 1186-87 (D.C.Cir.1987).
Perhaps mindful of NEPA’s procedural nature, EPA also seeks succor in the CWA itself. In particular, the agency relies on section 306(b)(1)(B), which provides: “In establishing or reviewing Federal standards of performance for new sources under this section, the Administrator shall take into consideration the cost of achieving such effluent reduction, and any non-water quality environmental impact and energy requirements.” 33 U.S.C. § 1316(b)(1)(B). To comply with this provision, EPA claims, the agency “routinely considers the effect of the regulations imposing effluent discharge limitations on air pollution, solid waste generation, water scarcity and energy consumption.” EPA Brief at 33. EPA’s authority under NEPA to consider the environmental effects of the entire facility, it is said, is fully consistent with this express mandate under CWA to consider the non-water quality impacts of the discharge. Id. at 34.
The response to EPA’s argument is found in the language of. section 306(b)(1)(B) itself. To be sure, the statute instructs EPA to consider non-water quality related environmental factors in connection with new sources. This consideration, however, obtains only in connection with the establishment of national new source standards, not in the issuance of individual permits. Indeed, this provision says nothing about the issuance of individual permits, much less authorizes the imposition of conditions relating to the entire facility. And even if the provision had anything to do with the issuance of particular permits (which it does not), requiring the Administrator to consider certain factors is a far cry from authorizing the imposition of any condition the Administrator finds desirable in particular permits.
For the foregoing reasons, we conclude that the statutory provisions relied upon by EPA as authorizing the challenged regulations do not supply the requisite authority. That being so, the challenged regulations cannot stand.
C. Effluent-Related Permit Conditions.
In 40 C.F.R. §§ 122.29(c)(3) and 122.-44(d)(9), the agency appeared to exclude *195effluent limitations from the category of permit conditions that might be adjusted in individual NPDES permits as a consequence of an EIS. The environmental petitioners contend that this position illegally renounces EPA’s authority to impose effluent limitations more stringent than those expressed in the national standards when necessary to satisfy NEPA’s policy goal of mitigating environmental damage.
Both EPA and the environmentalists at points appear to frame the issue as the purely legal one of whether EPA is among the agencies covered by § 511(c)(2)(B)’s restrictions on NEPA-conditioning of effluent limitations. Those agencies are identified in § 511(c)(2)(A) as including “any Federal agency authorized to license or permit any activity which may result in the discharge of a pollutant” into navigable waters. 33 U.S.C. § 1371(c)(2)(A). Both sides recognize that Congress adopted § 511 at least in part in response to this court’s decision in Calvert Cliffs Coordinating Committee, Inc. v. AEC, 449 F.2d 1109 (D.C.Cir.1971), which held that NEPA required the Atomic Energy Commission to reconsider the water quality impacts of projects already subject to CWA review. NRDC Brief at 74. The environmentalists suggest that § 511 was intended only to preclude federal agencies other than the EPA from reconsidering de novo the water quality impacts of projects already subject to review under the Act; the agency discerns broader purposes, including that of preventing even EPA from disrupting the comprehensive scheme set forth in the CWA with essentially ad hoc NEPA intrusions. EPA Brief at 54; see also Industry Intervenor Brief at 25-26.
If the issue were as self-contained as this summary suggests it would be fit for review. But the agency acknowledges that under § 302(a) of the CWA, 33 U.S.C. § 1312(a), it must supplement its national technology-based effluent limits with ones derived directly from concerns for local stream quality. EPA Brief at 52-53. Under § 302(a), the Administrator must impose these more stringent limits when necessary to “assure public water supplies, agricultural and industrial uses, and the protection and propagation of a balanced population of shellfish, fish and wildlife, and allow recreational activities in and on the water.” 33 U.S.C. § 1312(a). Moreover, the agency has made clear that it will consider NEPA-generated information when making its § 302(a) determinations. 49 Fed.Reg. at 38,033/3. All this leaves us frankly unsure just what NEPA-related factors (if any) the agency is purporting to ignore. Cf. Industry Intervenor Brief at 29.
The upshot of all this is to suggest strong institutional reasons against immediate review. (1) The ultimate legal issue will turn on the agency’s as-yet uncrystallized position on the relation between its § 302(a) duties and § 511(c)(2)(B). (2) The harm foreseen by the environmentalists may never come to pass. (3) Our ability to resolve whatever legal issues may arise will be greatly enhanced by their appearing in the context of a concrete factual situation. We conclude that the issues are not fit for review. ACLU v. FCC, 823 F.2d 1554, 1577 (D.C.Cir.1987) (“[Fjitness for review is not satisfied when ‘judicial appraisal ... is likely to stand on a much surer footing in the context of a specific application of this [regulation] than could be the case in the framework of this generalized challenge made here,’ ” quoting Toilet Goods Association, 387 U.S. at 164, 87 S.Ct. at 1524-25).
D. Admissibility of Evidence in Permit Proceedings.
40 C.F.R. § 124.85(e) directs the presiding officer in a permit hearing to exclude new evidence (and disallow cross-examination) as to specific types of environmental effects, where these effects “were considered or could have been considered” in a previous permit hearing under specific statutes and the applicant received the type of permit in question.9 But the regulation *196authorizes the officer to receive without cross-examination relevant portions of the record from such prior proceedings, and it in no way restricts the type of evidence that a party may submit for consideration during the informal comment period that precedes the permit hearing. EPA Brief at 56. EPA’s purpose is to eliminate duplication of effort and “to avoid turning NPDES hearings into ‘a forum for reexamination of decisions under statutes to which NEPA does not apply.' ” EPA Brief at 58-59 (quoting from 49 Fed.Reg. at 38,033/1). NRDC concedes that the “EPA’s goal of streamlining hearings is legitimate,” NRDC Reply Brief at 32, but claims that the agency’s regulations prevent admission of relevant evidence and cross-examination, thus violating the APA, CWA, NEPA and due process.
The interpretive discretion that the rule affords the presiding officer, and the uncertain character of its impact on any real proceeding, persuade us that the issue is unripe. The phrase “which were considered or could have been considered” is far from self-defining. Construed expansively (i.e., in such a manner as to increase the preclusive effect), the evidentiary restrictions might have the effects feared. Construed cautiously, they would be altogether benign.10
Given this uncertainty, there is no issue fit for review. It is not our task to render opinions on a series of hypotheticals. See Toilet Goods Ass’n v. Gardner, 387 U.S. 158, 163-64, 87 S.Ct. 1520, 1524, 18 L.Ed.2d 697 (1967); Air New Zealand, Ltd. v. CAB, 726 F.2d 832, 837 (D.C.Cir.1984). Review is no more appropriate here than it would be for some general evidentiary rule barring repetition.
The unfitness of the issue for review of course means that ripeness doctrine’s concern for the proper use of the courts militates against review. So do the other institutional values, and there is no offsetting hardship. The agency has, and here asserts, EPA Brief on Ripeness at 15, an interest in working out the functional meaning of its regulations free from premature judicial interference. Abbott Laboratories, 387 U.S. at 148-49, 87 S.Ct. at 1515. The only hardship suggested by petitioners — the existence of the regulations since 1984 — is patently inadequate. They assert no effect whatever on primary conduct. In the event that a presiding officer unlawfully excludes evidence under the regulation, and such exclusion harms petitioner by influencing the outcome of the hearing, they have a remedy. 40 C.F.R. § 124.85(d)(3) provides that a party wishing to challenge an exclusion can make an offer of proof for the record. This will preserve the dispute for appeal and allow a court to judge the validity of the regulations as applied in a concrete factual situation. See Toilet Goods Ass’n v. Gardner, 387 U.S. at 164-65, 87 S.Ct. at 1524-25.
III. State Program Requirements
Because of Congress’s desire to “recognize, preserve, and protect the primary re*197sponsibilities and rights of States to prevent, reduce, and to eliminate pollution,” 33 U.S.C. § 1251(b), the CWA provides for state assumption of the NPDES permit program. Id. § 1342(b). It specifies some prerequisites to states’ assuming permitting responsibilities, id., authorizes the Administrator to supplement them, id. §§ 1361(a), 1314(i)(2), and requires him to approve a state’s application once satisfied that these standards have been met, id. § 1342(b). Through this mechanism the agency has thus far transferred its permitting authority to 39 of 50 states. EPA Brief on Ripeness at 12 n. 9.
Environmentalist and industry petitioners attack different aspects of EPA’s rules for transfer of authority and for supervision of the state programs. We find all these attacks ripe.
A. Penalties and Participation — Ripeness.
Environmental petitioners challenge two regulations clarifying the prerequisites for transfer of EPA’s authority. Citizens for a Better Environment (“CBE”) contests the agency’s standards for state allowance of public participation, set forth in 40 C.F.R. § 123.27(d). It contends that 33 U.S.C. § 1251(e) requires states to provide for citizen suits and intervention rights just as the CWA does at the federal level. See 33 U.S.C. § 1365. NRDC complains that the agency’s regulations for what a state must provide by way of penalties provision, 40 C.F.R. § 123.27, is illegal because it fails to require the states to set maximum penalties as high as those provided in the CWA itself. 33 U.S.C. § 1319, as amended.
We believe these challenges to be ripe. Both are pure issues of law: if the challengers’ legal position be correct, the regulations are invalid. The agency position is final and the regulations' vulnerability to these attacks is not materially dependent upon further interpretations. The agency urges prompt resolution of both disputes. Although the issues are not as key as those involved in Eagle-Picher, continued uncertainty as to each will cast a shadow over the agency’s approvals of state applications for permitting authority. Thus delay would inconvenience both the agency and state authorities. Although we can discern no hardship to the petitioners from postponing review, the clear tilt of institutional factors in favor of ripeness permits us to dispense with that requirement.
B. Penalties and Participation — Merits.
A unique federal-state division of responsibilities is at the structural core of the Act. This is made crystal clear by the congressional declaration of the policy and goals pervading the Act:
It is the policy of the Congress to recognize, preserve, and protect the primary responsibilities and rights of States to prevent, reduce, and eliminate pollution, to plan the development and use [ ] of land and water resources, and to consult with the Administrator in the exercise of his authority under this chapter.11
Accordingly, the NPDES provides for state assumption of the permit program,12 and Congress repeatedly articulated its expectation that the states would play a major role in administering it.13 The Act specifies prerequisites for state assumption of the program,14 empowers the Administrator to further refine these standards,15 and *198commands the Administrator to approve the state permit system once he determines that the statutory requirements and administrative guidelines are met.16
1. Regulatory Uniformity and State Autonomy. Petitioners challenge two regulations implementing the Act’s provisions on state assumption of the permit program. Citizens for a Better Environment (CBE) attacks the standards for minimum public participation at the state level. NRDC complains of the absence of state authority to impose a given maximum penalty. Both protests rest on the assumption that congressional emphasis on uniformity was directed to procedural as well as substantive standards, and that as a result federal requirements respecting public participation and penalties must be mirrored on the state level.
Uniformity is indeed a recurrent theme in the Act, a direct manifestation of concern that the permit program be standardized to avoid the “industrial equivalent of forum shopping”17 and the creation of “pollution havens” by migration of dischargers to areas having lower pollution standards.18 The desired uniformity, however, is spoken of almost exclusively19 in relation to effluent limitations.20 Moreover, Congress’ quest for homogeneity is in tension with its independent emphasis on state autonomy, which is repeated throughout the legislative history of the Act,21 is enshrined in the Act as the basic policy to “recognize, preserve, and protect the primary responsibilities and rights of States,”22 and is the very foundation of the permit program.23 Congress made even clearer its intent
that the act be administered in such a manner that the abilities of the States to control their own permit programs will be developed and strengthened. [Members of Congress] look for and expect State and local interest, initiative, and personnel to provide a much more effective program than that which would result from control in the regional offices of the Environmental Protection Agency.24
In fashioning its guidelines on both participation and penalties, EPA endeavored to reconcile the competing objectives of regulatory uniformity and state autonomy by establishing a floor for citizen participation and state enforcement authority, while ensuring that states have the maximum possible independence.25 We are fully mindful *199of the rule that an agency is entitled to special deference when it harmonizes competing policies; “ ‘[i]f this choice represents a reasonable accommodation of conflicting policies that were committed to the agency’s care by the statute, we should not disturb it unless it appears from the statute or its legislative history that the accommodation is not one that Congress would have sanctioned.’ ”26 The regulations at issue are classic examples of such administrative balancing, and we find that they fall well within the permissible bounds of the agency’s discretion.
2. Public Participation. CBE assails on two grounds EPA’s regulations specifying the minimum level of public participation that states must afford. CBE first claims that the regulations are inadequate because they do not include all of the protections built into the federal permit program. Alternatively, it contends that they fail to provide any meaningful right.
The requirement of public participation in efforts to control water pollution is established in the' congressional declaration of policy and goals of the Act:
Public participation in the development, revision, and enforcement of any regulation, standard, effluent limitation, plan, or program established by the Administrator or any State under this chapter shall be provided for, encouraged, and assisted by the Administrator and the States. The Administrator, in cooperation with the States, shall develop and publish regulations specifying minimum guidelines for public participation in such processes.27
The statutory text does not, however, elaborate on the extent of public participation contemplated by Congress.28 The legislative history of the Act repeatedly echoes the desire “that its provisions be administered and enforced in a fishbowl-like atmosphere.” 29 CBE asserts that in Section 1365, which governs citizen suits, Congress spelled out the elements of the public participation envisioned.30 This contention is based on the characterization in the legislative history of that section as a “provision ... that would provide citizen participation in the enforcement of control requirements and regulations.”31 We find no indication, however, that these statements, made in the context of committee reports on Section 1365, were intended as definitional cross-references to Section 1251(e).
As CBE insists, Congress considered the citizen suit provision to be of dual importance, serving both as a method of prodding the agency,32 and as a backup *200means of enforcing the Act.33 On the other hand, Congress also expressed reservations about potential abuses of citizen suits, and expressly noted its concern “with protecting local government, industry, and individuals from harassing law suits while still keeping the courts open to responsible parties concerned with environmental protection.” 34 Nowhere in either the Act or its legislative history is there any express statement that the provisions of Section 1365 extend to states, nor do we find persuasive any equivocal intimation in that direction. It would have been very easy for Congress to say so if that was what it had in mind; instead, even intervention as of right was restricted to litigation in federal courts.35 It is difficult for us to believe that if Congress wanted the states simply to copy the federal minimum public participation requirements, it would limit the elaborate regulatory system to definition of federal-level public participation, and leave the specifications of the minimum acceptable state-level public participation for an exercise of the Administrator’s rulemaking authority. CBE invokes In re Permanent Surface Mining Regulation,36 which we affirmed en banc, requiring citizen suit authority in state programs governing coal surface mining.37 Not only does that decision rest on both statutory language and legislative history plainly contemplating state-level citizen suits, but it also affirms the agency’s interpretation of the statute.38
CBE would also bootstrap a prescription for state-level citizen suits from the statutory mandate that “[t]he permit program of the Administrator ... shall be subject to the same terms, conditions, and requirements as apply to a State program ... under subsection [1342(b) ].”39 As the Seventh Circuit has noted,40 however, this provision on its face applies in only one direction: the federal program must meet specific requirements set out in subsection (b), such as a five-year fixed permit term and incorporation of effluent limitations.41 Furthermore, the section cross-references only “terms, conditions, and requirements ... under subsection (b).” Subsection (b) already applies to states, rendering CBE’s argument empty, and refers only generally to public participation without any requirement of citizen suits.42
Finally, we note that EPA maintains that “[njothing in the Act or its legislative history indicates that Congress intended that states be required to provide identical rights to those Congress specified for citizens in Federal court.”43 Because we have found that “Congress has not directly addressed the precise question at issue,” and determined that “the agency’s answer is based on a permissible construction of the *201statute,” we defer to EPA’s reading.44 We therefore hold that state-level citizen suits are not commanded by the Act, and find no impropriety in the Administrator’s failure to require state programs to afford them.45
CBE asks that, should we decide -that state public participation specifications need not match the federal requirements, we hold that the present regulations are incapable of producing meaningful public involvement.46 We decline this invitation. The pertinent regulations read:
Any state administering a program shall provide for public participation in the State enforcement process by providing either:
(1) Authority which allows intervention as of right in any civil or administrative action ... by any citizen having an interest which is or may be adversely affected; or
(2) Assurance that the State agency or enforcement authority will:
(i) Investigate and provide written responses to all citizen complaints ...;
(ii) Not oppose intervention by any citizen when permissive intervention may be authorized by statute, rule, or regulation; and
(iii)Publish notice of and provide at least 30 days for public comment on any proposed settlement of a State enforcement action.47
Implementation of either of these options suffices,48 but it must be adequate enough to meet the “minimum guidelines for public participation.”49 Congress contemplated that these regulations would do more than pay lip service to public participation; instead, “[t]he public must have a genuine opportunity to speak on the issue of protection of its waters”50 on federal, state, and local levels.51 “[CJitizen groups,” it was said, “are not to be treated as nuisances or troublemakers but rather as welcome participants in the vindication of environmental interests.”52
We might be somewhat more hospitable to CBE’s claim, especially with regard to the second option, were it not for two statements by EPA in interpreting the regulations.53 First, in promulgating the regulations and again in its brief before this court, the agency indicated that the first option — provision of intervention as of right — called for state intervention rights similar to those accorded by the federal rules.54 Although, given the Act’s confinement of mandatory intervention as of right *202to federal courts,55 we would not insist upon such a provision, its existence does undercut CBE’s argument.
More importantly, however, EPA asserted at oral argument that the second option, to the extent that it is based on a state’s agreement not to oppose permissive intervention, will not be available in states that do not provide some means of intervention. This interpretation is critical to our decision to uphold the agency. Were the second option open where permissive intervention is impossible, public participation would be limited to that flowing from the state’s agreement to respond to citizen complaints and to entertain citizen comments on proposed settlements of state enforcement actions56 — rights dismissed by the Seventh Circuit as “no more than a legalistic articulation of a common courtesy and hardly ... satisfaction of the EPA’s statutory duty to issue regulations promoting public participation in state enforcement.”57
With this caveat, however, we conclude that the regulations, as interpreted,58 provide meaningful and adequate opportunity for public participation consistent with the statutory mandate. The regulations reasonably accommodate conflicting statutory prescriptions by “establish[ing] requirements which ensure the benefits of public participation, while intruding less into the States’ management of their judicial and administrative systems.”59 We accordingly reject CBE’s challenge.
3. Maximum Penalties. Section 1319, as amended by the Water Quality Act of 1987, specifies the penalties assessable on the federal level. A negligent violation is punishable by a penalty of “not less than $2,500 nor more than $25,000 per day of violation” for a first offense, and “not more than $50,000 per day” for any subsequently occurring.60 A knowing violation invites a fine of “not less than $5,000 nor more than $50,000 per day of violation” for a first offense, and “not more than $100,-000” for additional offenses.61 A violator who knowingly places another person in imminent danger of death or serious bodily injury is subject to a fine of “not more than $250,000” if the violator is an individual and “not more than $1,000,000” if an organization.62 Falsification of reports, or monitoring devices or methods calls for a fine of “not more than $10,000” for a first offense, and “not more than $20,000 per day of violation” for recurring offenses.63 Civil penalties for permit violations are “not to exceed $25,000 per day for each violation.”64
States are required to have “adequate authority" “[t]o abate violations of the permit or the permit program, including civil and criminal penalties and other ways and means of enforcement.”65 The Administrator was charged with the responsibility of fashioning guidelines defining the minimum enforcement provisions deemed adequate.66 Pursuant to this mandate, the Administrator promulgated the regulations here in question, which require state authority
*203[t]o assess or sue to recover in court civil penalties and to seek criminal remedies, including fines, as follows:
(i) Civil penalties ... shall be assessable in at least the amount of $5,000 a day for each violation.
(ii) Criminal fines [for permit or filing violations] ... shall be assessable in at least the amount of $10,000 a day for each violation.
(iii) Criminal fines [for falsification of reports or monitoring] ... shall be recoverable in at least the amount of $5,000 for each instance of violation.67
Petitioner NRDC contends that these regulations are invalid because they do not compel the states to provide authority to levy the maximum penalties assessable in federal enforcement programs.68
Throughout its consideration of the Act, Congress reiterated the important role penalties play in enforcement of water pollution standards,69 and emphasized the need for substantial penalties, for example:
[Sanctions under existing law have not been sufficient to encourage compliance____ Therefore, the Committee proposes to increase significantly the penalties____ [I]f the timetables established throughout the Act are to be met, the threat of sanction must be real, and enforcement provisions must be swift and direct.70
In this articulation of congressional purpose, coupled with the congressional expectation that states would bear the primary enforcement burdens of the Act,71 NRDC would find a mandate for state ability and willingness to assess the federally required maximum penalties.72
NRDC also relies on Section 1370, which provides that states “may not adopt or enforce any effluent limitation, or other limitation, effluent standard, prohibition, pretreatment standard, or standard of performance which is less stringent than the effluent limitation, or other limitation, effluent standard, prohibition, pretreatment standard, or standard of performance under this chapter.”73 NRDC asserts that this means that states also may not adopt penalty provisions less stringent than those that the Act prescribes for federal enforcement.74 We note, however, that this section refers, not to enforcement powers, but only to effluent limitations and similar standards.75
The Water Quality Act of 1987 contained a provision amending Section 1319 that weighs heavily against NRDC’s .contentions. The section laying out increased maximum civil penalties states:
INCREASED PENALTIES NOT REQUIRED UNDER STATE PROGRAMS. —The Federal Water Pollution Control Act shall not be construed as requiring a State to have a civil penalty for violations described in section 309(d) of such Act which has the same monetary amount as the civil penalty established by such section, as amended by paragraph (1). Nothing in this paragraph shall affect the Administrator’s authority to establish or adjudge by regulation a minimum acceptable State civil penalty.76
NRDC concedes that this section relieves state programs from any obligation to *204match the maximum federal civil penalties set forth in the Water Quality Act. But, NRDC says, the quoted language should be read as a narrow exception applicable only to the federal penalties to which it expressly refers, with the result that state programs must adopt the federal civil penalties in force prior to the passage of the Water Quality Act and, more clearly than ever, must incorporate the maximum federal criminal penalties into their programs. EPA’s reply is that the new provision should be understood as simply a confirmation of the broad authority the Administrator already enjoyed in crafting state program requirements.77 We accept EPA’s position.
This challenge by NRDC exposes the same logical infirmity flawing the attack leveled by CBE.78 It presumes an unexpressed congressional intent that state requirements must mirror the federal ones, a presumption inconsistent with the elements of the statutory scheme limiting operation of the provisions to enforcement efforts at the national level and explicitly empowering the Administrator to set the prerequisites for state plans.79 Nothing in the Act or its legislative history supports a reduction of the Administrator’s discretion to activity purely ministerial, ’ as approval of NRDC’s thesis would do, and the Administrator’s conclusion to the contrary is eminently reasonable.
The Water Quality Act amendments to Section 1319 support this reasoning. The provision concerning federal civil penalties states specifically that “[njothing in this paragraph shall affect the administrator’s authority to establish or adjust by regulation a minimum acceptable State civil penalty.” 80 This language is indicative of the understanding that Congress was not enlarging the Administrator’s authority for a narrowly defined purpose, as NRDC argues, but was confirming his existing authority. Surely the provision quoted is conclusive with respect to the treatment of federal civil penalties, and holding that federal civil penalties antedating the Water Quality Act must be included in state programs would run afoul of this clear intent.
Similarly, NRDC’s contention that federal máximums for criminal violations must be incorporated into state programs must also be discarded here. NRDC maintains that congressional silence with respect to the applicability of federal criminal penalties to state programs, when contrasted with the treatment accorded civil penalties, should be read as a mandate for their adoption.81 We cannot agree. Absent any stronger indication to the contrary, it is patently unwise to assume that Congress intended sharply to limit the preexisting discretion of the Administrator in this area. Criminal penalties, even more than civil penalties, are traditionally under the control of the individual states. It is much more reasonable to assume that the Administrator would have far broader discretion to respect state autonomy in the criminal sector than in the civil area, and that Congress did not intend to divest the Administrator of this authority by its silence.
The rationale EPA offers for its disinclination to adopt the statutory maxima also buttresses this conclusion. The proposed regulations would have required the states to exert enforcement authority virtually identical with the federal, including the same levels of minimum and maximum fines.82 In final structure, however, the regulations, changed largely in response to state comments,83 reflect the balancing of *205uniformity and state autonomy contemplated by the Act:
The Agency has determined that it is necessary to set specific minimum levels of fines and penalties which States must have the authority to recover in order to ensure effective State enforcement programs. Without such minimum levels, EPA would often be forced to take its own enforcement action in approved States because the State action imposed inadequate penalties. Such EPA action, while available as a backup, is not intended to be relied upon as the prime enforcement mechanism in approved States. Accordingly, the Agency has set minimum levels of fines and penalties. However, it has reduced the levels below those available to EPA based on the large volumes of comments from states requesting such relief.84
We will not disturb this “reasonable accommodation of manifestly competing interests,” 85 and consequently we uphold the agency’s penalty regulations.
C. EPA Veto Authority — Ripeness.
Congress provided that even when the EPA vested a state with permitting authority, it should retain a measure of control. One such control — not at issue here — is the power to withdraw approval of the state’s program. 33 U.S.C. § 1342(c)(3). Because of the administrative burden on EPA and the increased state-federal friction, the remedy is so drastic that EPA cannot be expected to use it except in egregious cases. Accordingly, Congress complemented it with a finer-tuned instrument: the power to require a state to submit a proposed permit for preissuance review and to veto the permit if the Administrator finds it to be “outside the guidelines and requirements of [the Clean Water Act].” 33 U.S. C. § 1342(d)(1).
The agency promulgated 40 C.F.R. § 123.44(c) to identify the types of permit defects that should justify exercise of the veto power. Two of the six types are challenged by industry. A permit would be subject to veto if
(5) Any provisions of the proposed permit relating to the maintenance of records, reporting, monitoring, sampling, or the provision of any other information by the permittee are inadequate, in the judgment of the Regional Administrator, to assure compliance with permit conditions, including effluent standards and limitations required by CWA, by the guidelines and regulations issued under CWA, or by the proposed permit; or, (6) In the case of any proposed permit with respect to which applicable effluent limitations ... have not yet been promulgated by the agency, the proposed permit, in the judgment of the Regional Administrator, fails to carry out the provisions of CWA or of any regulations issued under CWA____
Id. §§ 123.44(c)(5) and (6).
Industry asserts that § 402(d)(1) of the Act, 33 U.S.C. § 1342(d)(1), allows the Administrator to object to a state-issued permit only if the state has failed to abide by a formally promulgated effluent limitation regulation or has, in some other manner, ignored an explicit requirement of the Clean Water Act itself. It characterizes §§ 123.44(d)(5) and (6) as authorizing essentially ad hoc agency interference whenever a Regional Administrator believes, for example, that a state agency has erred in identifying “best practicable control technology” (“BPT") where the CWA requires that standard. 33 U.S.C. § 1311(b)(1)(A). Unless the relevant BPT were established in a formal federal guideline, id. at § 1311(b)(1), industry believes that the state agency’s deviation from EPA’s reading of the statute could not be “outside ... the requirements” of the CWA. Industry also asserts a procedural flaw — namely, that EPA failed to adequately answer its critiques of the agency’s regulatory proposals. We find these challenges ripe.
The primary dispute between the parties is purely one of law: the scope of *206§ 402(d)(l)’s reference to “requirements” of the CWA. The agency’s position has crystalized. Each exercise of the veto power will entail discretion, which often militates against immediate review. See Air New Zealand, Ltd. v. CAB, 726 F.2d at 837. But there appears no prospect whatever that such exercises of discretion might eliminate the legal issue or even seriously alter its character. Contrast, for example, the rules aimed at preventing use of duplicative evidence, pp. 171-72, supra. The issues are thus thoroughly fit for review.
The institutional interests also point to immediate review. The agency favors that course, as resolution of the issue will guide states’ efforts to frame their permits so as to avoid EPA veto. EPA Brief on Ripeness at 19. As the agency has in fact used its veto authority in the manner that the petitioners find objectionable, see e.g., Ford Motor Co. v. EPA, 567 F.2d 661 (6th Cir.1977); State of Washington v. EPA, 573 F.2d 583 (9th Cir.1978), it is improbable that our treatment of the issue would waste judicial resources.
Finally, petitioners can point to some likelihood of hardship. At least if we assume that EPA will generally incline to greater severity than state agencies, the breadth of the terms for veto exercise is likely to be reflected in harsher permit conditions. Years ago the Supreme Court noted the obvious point that “[i]t is common experience that men conform their conduct to regulation by governmental authority,” Columbia Broadcasting System, Inc. v. United States, 316 U.S. 407, 418, 62 S.Ct. 1194, 1201, 86 L.Ed. 1563 (1942), and we have little doubt that the same is true of state decisionmakers with federal officials hovering over them and offering “suggestions.” Moreover, a permittee receiving a very demanding permit would have a hard time establishing that its severity derived from state fear of the federal veto.
D. EPA Veto Authority — Merits.
Whether the CWA authorizes EPA’s veto authority regulations is a question of statutory interpretation, to be resolved under Chevron principles. See supra p. 168 (Permit Conditions Unrelated to Effluents— Merits). Applying those principles to the issue at hand, we conclude that Congress’ intent in § 402(d)(1) of the Act, 33 U.S.C. § 1342(d)(1), is not clear; because the agency’s interpretation of the statute is reasonable, we uphold the challenged regulations.
1
(a)
Initially, we consider Industry's challenge to 40 C.F.R. § 123.44(c)(6) (subsection 6). That regulation provides that the Regional Administrator may veto a proposed permit if, in the absence of formally promulgated effluent limitations, the proposed permit fails in the Administrator’s judgment to comport with the Act or regulations promulgated thereunder. In Industry’s view, this state-veto regulation is inconsistent with both the terms and the spirit of the Act. Congress intended, Industry maintains, for the States to serve as the primary permitting authorities, and for federal oversight of state issuance of permits to be minimal. By allowing Regional Administrators to reject a state-issued permit on the basis of ad hoc judgments that the permit does not comport with desirable effluent limitations, subsection 6 is, as Industry sees it, fundamentally at odds with the Congressionally crafted structure governing the respective roles of EPA and state authorities.
In Industry’s view, the sole basis for rejecting a proposed permit by virtue of improper effluent limits is if those limits are set pursuant to formally promulgated effluent limitations guidelines under section 1314(b). That section requires EPA to promulgate, pursuant to formal notice-and-comment rulemaking, guidelines (initially by October 18,1972, and if necessary annually thereafter) to be used in setting effluent limits in particular permits. The standards to be achieved (by different successive dates) are based variously on “best practicable control technology” (BPT), “best available technology” (BAT), and *207“best conventional control technology” (BCT). See 33 U.S.C. § 1311(b).
Industry argues that these § 1314(b) guidelines are the “guidelines and requirements” referred to in § 1342(d). The agency may not, in Industry’s view, reject permits based on noncompliance with effluent guidelines unless those limitations are set pursuant to § 1314. Industay contends, first, that until EPA has formally promulgated these guidelines, the States are not required to adhere to the vague technology-based standards set forth in the statute. It argues that the language of § 1311(b), which mandates the achievement of increasingly stringent technology-based standards as defined by the Administrator under the procedure laid out in § 1314(b), see, e.g., § 1311(b)(2)(A) (calling for achievement of BAT “as determined in accordance with regulations issued by the Administrator pursuant to section 1314(b)(2)”); § 1311(b)(1)(A) (requiring achievement of BPT “as defined by the Administrator pursuant to section 1314(b)”); § 1311(b)(2)(E) (speaking of BCT “as determined in accordance with regulations issued by the Administrator pursuant to section 1314(b)(4)”), evidences Congress’ intent that the Act’s technology-based standards remain dormant until activated by promulgation of EPA guidelines. See Industry Brief at 139-40.
Although it is consistent with some fragments of the statutory language, Industry’s interpretation ultimately proves too much, and thus must be rejected as inconsistent with the structure of the Act. Section 1342(a)(1) requires EPA, in approving permits in the absence of formally promulgated effluent limitations guidelines, to exercise its best professional judgment (BPJ) as to proper effluent limits. When issuing permits according to its BPJ, EPA is required to adhere to the technology-based standards set out in § 1311(b). NRDC v. Costle, 568 F.2d 1369, 1380-81 (D.C.Cir.1977).86 This is so despite the fact that EPA has not yet rigorously defined these standards through notice-and-comment rulemaking.
States issuing permits pursuant to § 1342(b) stand in the shoes of the agency, and thus must similarly pay heed to § 1311(b)’s technology-based standards when exercising their BPJ. Thus, notwithstanding Industry’s contrary assertions, States are required to compel adherence to the Act’s technology-based standards regardless of whether EPA has specified their content pursuant to § 1314(b). Accordingly, EPA contends that it may veto state permits that, in its estimation, fail to comply with § 1311’s dictates.
In response, Industry asserts that even if a State must use its best professional judgment to comply with the § 1311(b) standards, its substantive decisions as to what constitutes BAT, BPT, et al., are subject to EPA veto only if inconsistent with nationally promulgated guidelines. In other words, EPA’s interpretation of what § 1311(b) requires becomes an independent requirement of the Act only when determined pursuant to § 1314(b). Industry argues that its interpretation preserves Congress’ intentional balance of federalism and uniformity: While state-issued permits must comply with formally promulgated national guidelines, in the absence of such guidelines ad hoc federal judgments cannot trump their ad hoc state counterparts. Under a contrary view, Industry contends, EPA would be able to roam freely, unconstrained by standards and at liberty to reject state-issued permits.
(b)
The battle lines are thus sharply drawn. Unfortunately, the express terms of § 1342(d) do not provide a ready answer to the interpretive question before us. The crucial language, “guidelines and requirements,” fails conclusively to support either side’s view. Following the Supreme *208Court’s mandate to bring to bear the “traditional tools of statutory construction,” we therefore proceed to employ other interpretive tools by exploring whether either the legislative history or, more importantly, the structure of the statute sheds light on the question at hand.
(c)
Industry’s reading of the legislative history focuses broadly on Congress’ intent that States play the primary role in administering the Act. As to that general proposition there can be no reasonable doubt. But the general, pro-federalism thrust of the statutory regime does not manifest itself in the legislative history in helpfully specific ways as to the issue at hand: whether in the absence of formally promulgated effluent limitations guidelines, EPA can veto state permits on the basis of inadequate effluent limits. That being so, Industry understandably relies rather generally on the asserted narrowness of the veto authority and Congressional intent that federal superintendence of the States be limited.
Now to the specific of the history of the Act’s veto authority. The Senate version of the Clean Water Act contained a provision, characterized by Industry as a broad veto power, that stated as follows: “No permit shall issue until the Administrator is satisfied that the conditions to be imposed by the state meet the requirements of this Act.” See S. 2770, 92d Cong., 1st Sess. § 402(d)(2) (1971), reprinted in A Legislative History of the Water Pollution Control Act Amendments of 1972 at 1690 (1973) (hereinafter cited as “1972 Legislative History”). The House version, in contrast, did not contain a veto provision at all. Instead, the House preferred to rely on the Administrator’s more general power to withdraw approval of state programs in their entirety. See 1972 Legislative History at 814; 33 U.S.C. § 1342(c)(3). The Conference Committee settled on the language ultimately enacted, which as we have seen grants EPA the power to reject a particular state permit if it is “outside the guidelines and requirements” of the Act.
Industry argues that this sequence of events on Capitol Hill shows that the veto power, as it emerged in the final bill, was narrower than that contained in the initial Senate version. Industry also quotes liberally from passages of the legislative history indicating that Members of Congress intended federal veto authority to be exercised sparingly. See, e.g., 1972 Legislative History at 262 (“The managers [of the bill] expect the Administrator to use this authority judiciously”). Industry argues, ultimately, that the broad power asserted by EPA in the veto regulations is inconsistent with Congress' intent that the veto authority be exercised in a restrained manner.
In our view, Industry’s parsing of the legislative history fails to identify the requisite indications of Congress’ intent on the issue at hand. No one denies that federal overview of the States is meant to be limited. But the fundamental fact remains that nowhere in the legislative history did Congress evince its intent on the specific question before us.87
As to the scope of the veto authority, it is ultimately unhelpful to fall back to general principles of federalism, for those principles, however important to our polity, do not answer the specific question at hand. Contrary to Industry’s unflattering characterizations, EPA’s oversight of state permits is not ad hoc, nor is the agency’s discretion unfettered. The requirement articulated in section 1342(a)(1) that the factors listed in sections 1311 and 1314 be considered operates as a significant check on the agency’s discretion.
In sum, the Act envisions a significant role for the States in its administration, but nothing in the legislative history indicates Congress’ intent on the issue at hand. Contrary to Industry’s assertions, EPA’s *209discretion is constrained under the Act; so long as the agency exercises its veto power “judiciously,” the veto regulations are not at odds with Congress’ view of abiding principles of federalism. And it goes without saying that we are not confronted, in this global challenge to EPA’s regulations, with a specific action by federal authorities that is seen by Industry as riding roughshod over the felt interests of a particular State. Finally, we cannot but observe that Industry is in the odd role of seeking to carry on a reverse sort of parens patriae role, attending to the interests of the several States, when the States themselves have not seen fit (at least in litigation) to call the veto regulation into question. Like Congress itself, the States have been silent on the point.
(d)
Next, we consider the structure of the Act. See United Savings Ass’n of Texas v. Timbers of Inwood Forest Assocs., Ltd., 484 U.S. 365, 108 S.Ct. 626, 630, 98 L.Ed.2d 740 (1988) (“Statutory construction ... is a holistic endeavor”); American Mining Congress v. EPA, 824 F.2d 1177 (D.C.Cir.1987). Industry points out that the Act is premised on shared state and federal authority, and that the veto regulations, in authorizing ad hoc federal rejection of state permits, gut this carefully crafted structure. As Industry puts it, “[wjhere judgments have to be made in the absence of statutory requirements or regulations, the Act envisions that the states with delegated NPDES authority will exercise their good judgment free of second-guessing by EPA.” Industry Brief at 141.
But, again, the fact that the structure of the Act envisions that States will play a primary role under the Act does not mean that the agency’s veto regulations go beyond what the statute permits. For, as we have seen, the Act undeniably provides some veto authority; the question is how much (or by what standard the authority is to be exercised). Industry can point to nothing in the Act that would reasonably lend itself to delimiting the bounds of the veto authority (other than the requirements, already described, which cabin EPA’s discretion). Absent such a showing, relying on the pro-federalism bent of the Act’s structure does not, for reasons already stated, get us very far in resolving a highly specific question of statutory interpretation. Moreover, the veto regulations are not, as Industry would have it, “totally at odds with the principles of federalism embodied in the statute.” Reply Brief for Industry Petitioners at 89. Congress could not have intended, as Industry prefers, in all cases for state judgments about the proper conditions for permits to take precedence; the very presence of veto authority over individual permits belies this rather hard view of federalism principles.
EPA, on the other hand, mounts what seem to us more persuasive arguments based on the statute’s structure. First, nowhere in the Act is there an indication that Congress intended that promulgation of section 1314 effluent limitations guidelines would be a precondition for vetoing state permits. On the contrary, where Congress did intend to establish a precondition to EPA action, it provided for it, namely in section 1342(b). That provision requires EPA to issue state program guidelines under 1314(i)(2) before permitting authority may be transferred to the States. Since a State must have permitting authority before it may submit permits to EPA for approval (and possible rejection), section 1314(i)(2)’s state program guidelines are, in effect, a prerequisite to the agency’s veto authority taking effect.
In addition, EPA argues that Industry’s position may lead to results that would violate the Act. To return to basics, the statute sets forth an absolute prohibition on the discharge of any pollutants, unless in compliance with certain sections of the statute. See 33 U.S.C. § 1311(a). The Act then fixes certain technology-based requirements to be achieved by certain dates. See 33 U.S.C. § 1311(b). Under section 1342(a)(1), until promulgation of section 1314 national effluent limitations guidelines, the Administrator may approve permits if in his best professional judgment they come within the Act’s technology-based requirements. Under Industry’s *210view, however, until EPA has promulgated national guidelines setting forth its view of § 1311(b)'s technology-based requirements, the agency is powerless to supplant a state permitting authority’s judgment of whether a particular permit meets the technology-based standards. From this, as EPA emphasizes, States would be able (under Industry’s view) to approve permits that plainly violate section 1311(a) — the Act’s bedrock prohibition of pollutant discharges —and the federal authority must nonetheless stand helplessly aside, awaiting the uncertain coming of national effluent limitations guidelines. Cf S. Beckett, Waiting for Godot.
This is the sort of situation (although admittedly stated in the extreme) that the Act’s veto provision, by its very nature, was designed to address. EPA persuasively argues that, although the States are to play a large role in administering the Act, section 1342(d)’s veto authority provides considerable evidence that Congress intended federal minima (even if crafted on the basis of BPJ) to take precedence. In short, the several States are to be centrally involved in the Act’s administration, but their involvement is to be in the achievement of federal goals. By virtue of Congress’ policy decision to allow EPA to approve federal permits based on the Administrator’s best professional judgment, it is consonant with that policy to allow a similar power with respect to vetoing state permits. Rejection of a state permit in the latter situation represents no more an ad hoc judgment than approval of a federal permit in the former.
We are persuaded that the agency’s interpretation is more in keeping with the Act’s overall structure. But when the plain language of the Act is not clearly dispositive, and when respectable interpretive arguments are advanced in contradiction to the agency’s view, we proceed out of an abundance of caution to Chevron’s second step. In short, we cannot say with certainty that Congress’ intent requires the interpretation EPA has adopted. To recap: The Act is, to be sure, pro-federalism in its operation; but the very existence of a veto authority cuts against Industry's rather absolutist view that all interpretive calls are to be made in favor of state authority. Still, we cannot say with confidence that Congress’ intent is clear on the issue at hand.
(e)
Inasmuch as Congress’ intent has eluded us, we proceed to the second step of the familiar Chevron inquiry. See Continental Airlines, Inc. v. Department of Transportation, 843 F.2d 1444 (D.C.Cir.1988); Rettig v. Pension Benefit Guaranty Corp., 744 F.2d 133 (D.C.Cir.1984). In our view, EPA’s interpretation is sufficiently reasonable to pass muster under Step Two. For one thing, as we have just seen, the statute’s structure cuts more in favor of the agency’s reading. For another, EPA’s interpretation strikes a reasonable balance between state and federal interests which enjoy recognition under the Act. Moreover, the goal of limited federal supervision, which was undeniably a pivotal part of Congress’ intent, is by no means rendered nugatory by the regulations. By its terms, section 1342(d) envisions that EPA will supervise the States in their issuance of permits. Subsection 6, in seeking to preserve federal oversight while maintaining to a large degree the primary roles of States as permitting authorities, strikes a reasonable balance between state and federal interests. Finally, the fundamental purpose of the Clean Water Act is, of course, to “restore and maintain” the purity of our country’s waters. See 33 U.S.C. § 1251. Although this is a joint state-federal venture, it is EPA that has been charged by Congress with oversight of progress toward the achievement of these legislatively ordained goals. It seems reasonable, therefore, for EPA to retain a check to ensure that the several States do not, even if unintentionally, impede progress toward achievement of the Act’s basic purpose.
But there is another, less obvious, indicator of reasonableness. In view of EPA’s apparent inability to issue national effluent limitations guidelines as Congress had envisioned, the agency's rule at least retains a *211check on state permitting authority that the national effluent limitations guidelines were themselves meant to serve. Indeed, to accept Industry’s view means that an important weapon placed by Congress in EPA’s arsenal would have remained useless indefinitely. EPA’s interpretation strikes us as a reasonable accommodation of competing interests achieved in a set of circumstances apparently unforeseen by Congress. Cf. NRDC v. Costle, 568 F.2d 1369, 1378-83 (D.C.Cir.1977).
We acknowledge that early interpretations of the agency’s veto authority by two of our sister circuits look in a somewhat different direction. Specifically, the Sixth Circuit, in Ford Motor Co. v. EPA, 567 F.2d 661 (6th Cir.1977), and the Ninth Circuit, in State of Washington v. EPA, 573 F.2d 583 (9th Cir.1978), refused to uphold EPA’s rejection of state-approved permits on the basis of perceived infirmities not grounded in the specific language of the Act or regulations duly promulgated under section 1314. In short, both rejected EPA’s contention that its interpretation of the requirements of the Act ought to bind the States in the absence of federally promulgated guidelines. Accordingly to our sister circuits, “ ‘[t]he promulgation of standards, issuance of permits, and conformance with effluent limitations bespeaks a rational implementation strategy anticipating a discrete sequence of events____’” State of Washington, 573 F.2d at 592 (quoting Republic Steel Corp. v. Train, 557 F.2d 91, 95 (6th Cir.1977)). The inference drawn a decade ago by those two courts was that Congress must have intended federal veto authority (on the basis of inadequate permit conditions) not to take effect until this “sequence of events’’ played itself out.
The fundamental difficulty with this view is that, for reasons already stated, we see nothing in the language or history of the statute to indicate that Congress intended this to be the case. On the contrary, section 1342 placed temporal limits on EPA’s granting permitting authority to the States (the issuance of state program guidelines under section 1342(i)(2)). Moreover, section 1342(a) allows EPA to approve permits in the absence of effluent limitations guidelines based on a best professional judgment standard. These factors (which were not addressed by our sister circuits) support the inference that Congress would have intended the federal agency to supervise state permits in the absence of formally promulgated effluent limitations guidelines. At a minimum, it would seem that if Congress had harbored the opposite intent, it would have so stated. In addition, we are now confronted, ten years later, with the actual “sequence of events” having unfolded, for whatever reason, in quite a different way than originally envisioned by Congress.
Finally, we cannot overlook the fact that the doctrinal framework for review of agency interpretations of their governing statutes has been helpfully clarified since these cases were decided. Chevron mandated a framework for- analysis that was less clear in 1977 and 1978, when State of Washington and Ford Motor Co. were decided. Cf. INS v. Cardoza-Fonseca, 480 U.S. 421, 107 S.Ct. 1207, 1225, 94 L.Ed.2d 434 (1987) (Scalia, J., concurring in the judgment). In Ford Motor Co. and State of Washington, the courts, understandably, did not inquire into the clarity of Congressional intent in the manner ordained by Chevron and its progeny; rather, in that pre-Chevron era, the two courts simply interpreted the Clean Water Act and gave effect to their interpretations, without the benefit of the C%ewow-articulated framework. Now, of course, it is well established that unless Congress’ intent is clear, courts are bound to defer to a reasonable agency interpretation. In this post-Chevron era, it is quite likely that our sister circuits would find the Act susceptible to differing interpretations and thus deem themselves bound to uphold the agency’s view.
2
Next, we consider Industry’s challenge to 40 C.F.R. § 123.44(c)(5) (subsection 5), which provides that the Regional Administrator may reject proposed state permits if in his judgment the permit’s provi*212sions relating to “maintenance of records, reporting, monitoring, sampling, or the provision of any other information,” 44 C.F.R. § 123.44(c)(5), are inadequate to assure compliance with the Act or regulations issued under it. Industry’s challenge to subsection 5 is essentially the same as that mounted against subsection 6. In Industry’s view, the Act does not allow EPA to reject proposed permits on the basis of ad hoc judgments about the adequacy of a particular permit’s conditions. Rather, rejection of state permits must be based either on the explicit requirements of the Act or guidelines issued under it.
This challenge misses the mark. The EPA has established guidelines for determining whether the information provided by the federal permittees is adequate. To enforce compliance with permit conditions, EPA has enacted a number of monitoring and record-keeping requirements. See, e.g., 40 C.F.R. §§ 122.41©, 122.44(9(1), and 122.48(b) (“[r]equired monitoring includes gathering information on] type, intervals, and frequency sufficient to yield data which are representative of the monitored activity ...”); 40 C.F.R. § 122.21(p) (“[applicants shall keep records of all data used to complete permit applications and any supplemental information ... for a period of at least 3 years ... ”). These regulations apply directly to permit issuance by the EPA itself. For states, EPA has incorporated the same provisions by reference, with the qualification that they may impose more stringent ones. 40 C.F.R. § 123.25(a). Accordingly, it seems altogether premature to consider whether the Administrator could veto state permits on the basis of deficiencies in record-keeping requirements not based on guidelines.
3
Industry's final argument, which relates to both subsection 5 and 6, is that EPA violated the fundamental tenet of administrative law that it must respond to comments submitted on proposed rules. See 5 U.S.C. § 553(e). Specifically, Industry maintains that the agency erred in retaining the proposed veto regulations, without responding to Industry’s substantial legal objections to the validity of subsections 5 and 6.
EPA replies that there was no need to respond to Industry’s objections. The Notice of Proposed Rulemaking dealt fully, EPA contends, with the issues raised by the veto regulations, including the bulk of Industry’s legal objections. See 43 Fed. Reg. 37,078, 37,087 (1978). The Federal Register notice discussed the Ford Motor Co. and State of Washington decisions, noting EPA’s view that the veto regulations, although perhaps in tension with dicta in those two cases, were consistent with the cases’ holdings. See id. at 37,087-88. The agency’s view, as set forth in the NOPR, was that the veto regulations themselves constitute “guidelines” under the Act and therefore address the concern that federal supervision be pursuant to formally promulgated standards.
In EPA’s view, none of the specific objections voiced during the comment period required response. The reason for this seemingly harsh attitude was simple; any response would, in the main, have been only a restatement of what had already been set forth in the preamble to the proposed rules. EPA contends that under these circumstances no response was necessary.
We agree. The APA requirement of agency responsiveness to comments is subject to the common-sense rule that a response be necessary. Failure to respond is not grounds for APA invalidation unless the points raised in the comments were sufficiently central that agency silence would demonstrate the rulemaking to be arbitrary and capricious. See, e.g., ACLU v. FCC, 823 F.2d 1554, 1581 (D.C.Cir.1987), cert. denied, — U.S.-, 108 S.Ct. 1220, 99 L.Ed.2d 421 (1988); Baltimore Gas & Electric Co. v. United States, 817 F.2d 108, 116 (D.C.Cir.1987); Home Box Office, Inc. v. FCC, 567 F.2d 9, 35 n. 58 (D.C.Cir.1977). The fundamental purpose of the response requirement is, of course, to show that the agency has indeed considered all significant points articulated by the public; in addition, agency responsiveness aids in *213the Congressionally sanctioned process of judicial review of agency action. See Citizens to Save Spencer County v. EPA, 600 F.2d 844, 883-84 (D.C.Cir.1979).
In the circumstances before us, EPA, in effect, responded in advance; as a result, there was no error in failing to respond to legal objections that were thereafter raised during the comment period. The agency had, for all practical purposes, already noted and addressed the objections. EPA’s legal position (which we have sustained) was clear to the public at all times, and the court is thus not faced on review with uncertainty as to whether the agency overlooked substantive objections. Accordingly, EPA did not run afoul of section 553(c) of the APA.
IV. Toxicity Limitations
EPA has adopted the view that permit writers may express technology-based or water-quality-based limits in terms of “toxicity.” 40 C.F.R. § 125.3(c)(4). Such a limit would typically state how great a concentration of a particular effluent would be necessary to kill a certain percentage of a particular type of marine life in a specified time period. For instance, the toxicity of an effluent may be expressed in terms of its “LC50,” i.e., the concentration of that effluent that would kill 50% of a sample of a particular biological species within a predetermined time period. To say that an effluent has an LC50 of 20% would be to say that a 20% concentration of that effluent instream would kill 50% of the test sample within a specified time period. Obviously, the less toxic an effluent, the greater the concentration necessary to cause 50% fatalities, and the higher the LC50. EPA Brief at 81 n. 1.
The agency explains that the ability to so frame effluent limits allows the permit writer to regulate effluents that include combinations of chemicals that are innocuous individually but cause harm in the aggregate. 49 Fed.Reg. 38,009/3.
Industry raises four objections. We find its flat-out attack on the agency’s statutory authority ripe but without merit. We find unripe its claims that the agency has failed to demonstrate the existence of a reliable methodology for implementing such limits and that the provision entails an impermissible trespass on the state’s rights to set their preferred water-quality standards. Finally, we find certain procedural attacks hybrid, meriting some — but not a definitive —legal treatment at this point. We address the challenges in that order.
A. Statutory Authority.
As did the regulation on imposing conditions unrelated to effluent quality, see section II.A above, 40 C.F.R. § 125.3(c)(4) contains a qualification that might seem to insulate it from legal attack: it calls for use of toxicity-based limits only where it is shown that those limits “reflect the appropriate requirements ... of the [Clean Water] Act.” But as in the prior instance we read the regulation to reflect a definitive agency view that some such situations will arise, and to encourage permit writers to act accordingly. The purely legal character of the issue, the agency’s crystalized position and its asserted and understandable interest in review now, EPA Brief on Ripeness at 19-21, and the improbability that the issue will prove null, all point toward ripeness in light of Congress’s expressed preference.
Industry’s assertion of complete want of statutory authority rests on a twofold argument that (1) “toxicity” is not a pollutant and that (2) therefore the agency lacks the statutory authority to regulate it within the NPDES. It is the second phase that misfires. While “toxicity” appears to be an attribute of pollutants rather than a pollutant itself, we see no reason why this should preclude the agency from using it as a measure to regulate effluents that are pollutants. Under § 502(6) of the Act, 33 U.S.C. § 1362(6), the term “pollutant” includes “industrial, municipal, and agricultural waste.” Any discharge to which a toxicity limit could be applied would seem to fall within this broad definition.
B. Technical Feasibility.
Industry next argues that promulgation of § 125.3(c)(4) was arbitrary and *214capricious because, it says, the agency failed to demonstrate that where technology-based limits are framed in terms of toxicity, the limits could be properly related to treatment technology. This issue appears to be not only unripe, but virtually conceded. The agency admits that it cannot lawfully use toxicity as a technologically-based limitation until it is able to develop “a supporting record,” EPA Brief at 103, which it implicitly acknowledges having not yet done. Clearly this issue must await the day when a permit writer uses toxicity and tries to build the necessary record.
C. Intrusion on State Authority.
Industry attacks the authority of a federal permit writer to use toxicity to set water-quality-based limitations to meet “narrative” state water quality standards. The agency evidently claims the authority to set toxicity limits when a state fails to establish specific standards to supplement its general commitment to limiting the toxicity of discharges, or fails to promulgate adequate standards at all. EPA Brief at 99-102. Industry believes that such action would intrude on substantive decisions that the statute reserves for the states.
As we see it, the parties are asking this court to (1) hypothesize a narrative state water-quality standard or envision a hypothétical situation in which a state’s standards fail to measure up to the agency’s desires, (2) imagine an agency response in the form of a toxicity-based limit, and (3) imagine the arguments that the agency and parties disputing its action might then make. This is far too amorphous for us to get ahold of at this stage.
D. Procedural Claims.
Industry charges that the regulation suffers from a fatal case of APA neglect. Grasping the claims requires a brief review of the regulation’s origins.
EPA initially promulgated § 125.3(c)(4) in May 1980 as part of its consolidated permit regulations. 45 Fed.Reg. 33,290, 33,512 (May 19, 1980). To implement the regulation, it prepared and informally circulated a draft toxicity testing document and toxicity reduction manual. In 1982, partly in response to serious feasibility questions raised by industry, the agency proposed to shelve the regulation until it could study its draft testing manuals in light of the various comments and reach a conclusion “on whether and how toxicity testing should be accomplished.” 47 Fed.Reg. 52,072, 52,079 (Nov. 18, 1982). In the next two years, it prepared a draft policy for the development of water-quality-based permit limits, in which it discussed the uses of limits framed in terms of toxicity, “Notice — Development of Water Quality-Based Permit Limitations for Toxic Pollutants; National Policy,” 49 Fed.Reg. 9016, 9017 (March 9, 1984), and a draft technical support document. It circulated these informally. It received comments not only on them but also on its 1982 decision to put toxicity limits on hold. Commenters split on the soundness of using toxicity. 49 Fed.Reg. at 38,009-10. In its final regulations issued in 1984, the agency adopted its original position, rehabilitating § 125.3(c)(4).
Industry challenges the agency’s decisionmaking process on two fronts. First, it asserts that the agency should have exposed its 1984 policy statement and its technical support document to notice and comment under § 4 of the APA, 5 U.S.C. § 553. Second, it claims that the agency violated § 553 by failing to respond adequately to industry’s attacks on the contested provision.
In part industry appears to claim that the agency’s 1984 policy statement and its draft technical support document were “rules” requiring notice and comment. The claim seems ripe but wrong. So framed, the issue is purely legal; the agency invites its resolution now, hoping to clear a problem out of the way. We can readily accommodate the parties.
The two documents do not appear to have independent legal significance. We do not believe that while an informal rule-making is pending (as was true here between 1980 and 1984), an agency is barred from any information-gathering process other than issuances of new or revised *215notices of proposed rulemaking. Cf. NRDC v. Thomas, 838 F.2d 1224, 1242 (D.C.Cir.1988) (upholding regulations altered from original proposal and noting agency's informal efforts, on eve of final promulgation, to solicit feedback from affected parties).
But there is a subtler issue that we believe unripe. When, as and if a permit writer concludes that it is technically feasible to relate treatment technology to levels of toxicity, the basis of that finding will be at issue. So long as the minds of permit writers are open on the issue — as they should be in light of EPA’s positions here— no harm will have arisen from any incompleteness in EPA’s circulation of documents. Cf. McLouth Steel Products Corp. v. Thomas, 838 F.2d 1317, 1324-25, (D.C.1988) (defects in procedure in adopting scientific model do not prevent agency use thereof where in individual applications it is completely open to scientific attacks). On the other hand, to the extent that the permit writer assumes away the technical issues on the basis of 40 C.F.R. § 125.3(c)(4), an objecting permittee seems likely to prevail. First, the permit writer would appear to have read the regulation as saying more than it does — to have disregarded its qualification. See section IV.A. above. Second, if the permit writer relied on the technical document without allowing an adequate opportunity for rebuttal, the permittee might successfully object that the technical conclusion was arbitrary and capricious. Clearly we cannot now resolve the legal issues that might arise from all these speculations. The same is equally true of industry’s objections that the agency did not respond well enough to its technical critiques of toxicity limits.
V. Non-Adversary Panel Procedures
In 1979 the EPA promulgated, as one aspect of its 1979 final NPDES regulations, rules establishing some short-cut procedures for hearings on initial permit grants and initial grants of variances. These were known as non-adversary panel procedures (“NAPP” or “NAP procedures”). 44 Fed. Reg. 32,854, 32,944-47 (June 7, 1979). The EPA viewed the affected decisions as involving “initial licensing,” which the APA explicitly exempts from some elements of formal adjudication. See, e.g., 5 U.S.C. § 556(d) (allowing agency to provide for admission of evidence in written form “where a party will not be prejudiced thereby”); id. § 554(d) (setting forth restrictions on contracts by persons making recommended decisions, but exempting initial licensings). But see id. § 557(d)(1) (barring communications between any “interested person outside the agency” and any person expected to be involved in the decisional process). As the EPA later explained, it believed that “[hjearings on initial licensing [were] exempt from formal evidentiary hearing requirements for a number of reasons, most importantly because the complex policy decisions in initial licensing decisions [were] more akin to rulemaking than adjudication.” 49 Fed.Reg. at 38,040/2-3. Thus it limited application of the procedures to “initial licensings,” defined (in the current version) to include both “the first decision on an NPDES permit applied for by a discharger that has not previously held one and the first decision on any variance requested by a discharger.” 40 C.F. R. § 124.111(a)(l)(ii). In 1980, it repromulgated these regulations without substantial modifications. 45 Fed.Reg. at 33,504-08. In the sort of proceedings covered, the regulations allowed the Regional Administrator to decide whether to use the procedures. Id.
The NAP procedures came under heavy fire from industry litigants, who claimed that the Clean Water Act and APA required more formal adjudicatory procedures. The EPA, evidently concerned that “the panel hearings might not prove useful if invoked upon unwilling participants,” 49 Fed.Reg. at 38,041/1, made a limited retreat. As part of a 1982 settlement of various court challenges brought by industry, the agency issued proposed rules under which the NAP procedures would only be used with the consent of the permit applicant. 47 Fed.Reg. at 52,092. The settlement provided that if the agency adopted final rules substantially the same as those *216proposed, the litigants would withdraw their challenges. Id. at 52,072/3.
In 1984 the agency had a change of heart. After reconsidering the issue, it concluded that the NAP procedures did not impermissibly restrict the rights of applicants, that Regional Administrators had not in the past invoked the procedures in inappropriate situations, and that they weren’t likely to in the future. 49 Fed.Reg. at 38,041/1. Believing that the Regional Administrators were better placed than permit applicants to decide whether the NAP procedures would “aid decisionmaking or expedite permit issuance,” id., the agency restored the Regional Administrators’ power to adopt NAP procedures. Id.
The sole challenger of the NAP procedures is the American Petroleum Institute (“API”). It objects first to some of the procedures on their merits, as violations of the APA or Clean Water Act. We will review these claims seriatim, disposing on the merits of what we find ripe. That doesn’t amount to much. Like the evidentiary precepts considered in subsection II.D., these regulations have an amorphous character. It is certainly conceivable that their application (or misapplication) might infringe procedural rights of applicants, but we find no such infringement on the face of the regulations. The API also contends that the regulations violate the settlement agreement. We find this ripe and completely without merit.
A. The Merits of the Procedures.
API's attacks fall into two categories. First, it claims the regulations unduly restrict oral presentation of evidence and use of cross-examination. Second, it claims that the provisions for composition of the hearing panels violate APA provisions on who may make recommended decisions and on ex parte communications with those who do.
Before addressing these, we should note what is not disputed. The API argues that under § 402(a)(1), 33 U.S.C. § 1342(a)(1), the hearings in question are ones “required by statute to be determined on the record after opportunity for an agency hearing” within the meaning of 5 U.S.C. § 554, and therefore governed by the procedural requirements of 5 U.S.C. §§ 554, 556, and 557; the agency agrees. The agency claims that the decisions at issue are initial licensings and thus enjoy the APA-provided exemptions; API does not dispute this, but argues that the Clean Water Act imposes some procedural requirements more stringent than those mandated by the APA.
We further note that while use of the NAP procedures is at the discretion of the Regional Administrator, 40 C.F.R. § 124.111(a)(l)(i), the agency has stated that it does not intend them to use the NAP procedures “if the factual issues involved make the decisionmaking more akin to adjudication than to rulemaking.” 49 Fed.Reg. at 38,040/3. The amount of discretion vested in the Regional Administrators and their presiding officers concerning if, when and how to apply the NAP procedures weighs heavily against our addressing arguments based on how API believes the regulations will be applied. We are not soothsayers and at this point can only respond to the API’s facial attacks.
1. Oral testimony and cross-examination. The regulations provide for each party to file in writing “all of its comments on the draft permit.” § 124.118. This sounds quite draconian, but the same section softens it. Oral commentary is permissible for “points that could not have been made in written comments, or to emphasize points which are made in the [written] comments, but which the party believes can more effectively be argued in the hearing context.” Id.
As to cross-examination, the regulations’ initial, bark again is worse than their bite. § 124.120(d) states that only the presiding officer and panel members may cross-examine, id., but then permits others to do so when the presiding officer “determines, after consultation with the panel, that the cross-examination would expedite consideration of the issues.” Id. Also, participants may submit written questions which the officer may ask “at his or her sole discretion.” Id. Also, after the hearing any party may request the opportunity to *217conduct cross-examination at a supplementary hearing before the AU. § 124.121(a).
Insofar as API claims that applicable law gives permit applicants an absolute right to present testimony orally and to cross-examine, we find its claim ripe. The issue is purely one of law and the agency’s position—denying any such absolute right —is final. Given the congressional interest in speed and the agency’s desire to press ahead, EPA Brief on Ripeness at 21-23, we see no institutional obstacle to an immediate answer.
On the merits, however, the answer is plainly no. The law simply does not afford applicants the sort of absolute right asserted. As to agency insistence on written evidence, 5 U.S.C. § 556(d) makes clear that “when a party will not be prejudiced thereby,” agencies conducting initial licensing proceedings may “adopt, procedures for the submission of all or part of the evidence in written form.”
API responds that under Seacoast AntiPollution League v. Costle, 572 F.2d 872 (1st Cir.), cert. denied, 439 U.S. 824, 99 S.Ct. 94, 58 L.Ed.2d 117 (1978), the public hearing requirements of the Clean Water Act preempt § 556(d) and require “more than an opportunity to submit evidence in written form.” API Brief at 24. But Seacoast (which this Circuit has not adopted) falls far short of finding an absolute right to present all evidence in oral form. There the court explicitly noted the suitability of receiving in writing “the actual baseline data (which presumably all parties will be happy to have submitted in written form).” Id. at 880. It appeared to believe that oral presentation was suitable where expert credibility in the interpretation of data was at stake. Id. This sort of concern would seem to be accommodated handily by § 124.118, which, after all, explicitly allows for oral testimony “to emphasize points ... which the party believes can more effectively be argued in the hearing context.” Cf. 49 Fed.Reg. at 38,040/3 (agency does not intend for NAPP to be used in hearings that are likely to be dominated by disputed factual concerns). At any rate, we cannot say in advance that the regulations on their face call for any violation of Seacoast’s message.
The same is true for cross-examination. Seacoast itself recognized that a party to an adjudicatory hearing did “not have an absolute right to cross-examine witnesses,” id., and that the “plain language of 5 U.S. C. § 556(d) limits the right to instances where cross-examination is ‘required for a full and true disclosure of the facts.’ ” Id. (This applies even where the exemption for initial licensing is not applicable.) We cannot say that the regulations’ provisions for discretionary allowance of cross-examination are bound—or even likely—to violate this standard.
2. The panels’composition. NAP proceedings are held before a Presiding Officer (who may be someone other than an AU, but only if all parties consent), assisted by a panel made up of three or more specialists, who need not be agency employees. 40 C.F.R. §§ 124.119(a), 124.-120(a). At the conclusion of the process, the Presiding Officer or a specifically designated member of the panel, possibly in consultation with other panel members, drafts a recommended decision. The Presiding Officer or the Regional Administrator forwards a copy of the draft to the Administrator and to each of the parties, id. §§ 124.116 & 124.124, presumably only when satisfied that it is correct. After all internal appeals have been exhausted, the Administrator, who is free in the interim to consult with panel members and other agency personnel, hands down the final agency decision. Id. at § 124.126.
API raises two challenges. First, it contends that by allowing the participation of outsiders the regulations permit non-EPA employees to decide the outcome of agency hearings in violation of 5 U.S.C. § 557(b). This provides that only the presiding employee, or an employee qualified to preside at hearings pursuant to § 556(b) (in this case in AU), shall recommend decisions to the agency where the agency is to make a decision without having itself presided at the reception of the evidence. As we have noted, § 124.119(a)(1) allows someone other than an AU to preside at NAPPs *218only on consent of the parties, so there is no possibility that (without that consent) the deciding officer would be anyone other than an ALJ or (under § 124.116) the Regional Administrator. API’s claim that the regulations violate § 557(b) by delegating “the initial decision to non-agency employee[s],” API Brief at 25, seems to be demonstrably false.
API’s second claim is that the participation of non-EPA employees will inevitably taint hearings with extra-record evidence. Again they point to Seacoast, and again, even if Seacoast were law in this Circuit, it would not establish the illegality of the procedures. In fact, that case approved EPA’s use of the panel format in NAPP hearings, and reversed the panel’s decision (on this point) only because its report explicitly relied on scientific literature not in the record. 572 F.2d at 881; cf. United Steelworkers v. Marshall, 647 F.2d 1189, 1218 (D.C.Cir.1980) (construing § 556(e), in the context of a formal rule-making, to require remand of an agency decision only where an agency has substantially relied on “hard data or new legal arguments” not appearing in the record). We note that there is no reason to suppose that panel members who are not EPA employees would be interested persons within the meaning of § 557(d) (forbidding ex parte communications with such persons). See also United Steelworkers, 647 F.2d at 1218-20 (finding no substantive difference between use of agency staff and use of outside consultants to analyse and evaluate record evidence). Thus, we find no facial defect in the agency’s plan for non-EPA employees to play a role on NAPP panels.
B. Defects in the Mode of Adoption.
We find these claims ripe on the same grounds as the parallel attacks on the promulgation of the toxicity limits, see section IV.D., but also without merit.
API claims that by promulgating a final rule giving the decision on use of NAP procedures to the Regional Administrator, EPA “reneged” on its obligations under the 1982 settlement agreement. But in that agreement EPA bound itself only to propose regulations under which the consent of all parties would be required — as in fact it did. It never bound itself as to the content of the final regulations. Industry clearly understood that, as it reserved for itself the ability to challenge the final regulations to the extent that they differed from the 1982 proposals. Further, a binding promise to promulgate in the proposed form would seem to defeat Congress’s evident intention that agencies proceeding by informal rulemaking should maintain minds open to whatever insights the comments produced by notice under § 553 may generate. Cf. Association of National Advertisers v. FTC, 627 F.2d 1151, 1165-70 (D.C.Cir.1979), cert. denied, 447 U.S. 921, 100 S.Ct. 3011, 65 L.Ed.2d 1113 (1980) (seeming to be ready to find a due process violation in rulemaking where agency member has unalterably closed mind on critical issue, despite Bi-Metallic Inv. Co. v. Colorado, 239 U.S. 441, 36 S.Ct. 141, 60 L.Ed. 372 (1915) (due process restrictions not applicable to legislative activities of administrative agency)); Jeffrey M. Gaba, Informal Rule-making by Settlement Agreement, 73 Geo. L.Rev. 1241, 1256-61 (1985).
API acknowledges, as it must, that the agreement left the agency free to alter the regulatory proposals. But it argues that the agency's elimination of the requirement of all parties' consent must, as it drew no hostile comments, evidence agency bad faith. A variation of this theme is that we can infer from the switch that EPA intended all along to “renege,” so that its 1982 proposed regulations were not adequate public notice of its true intentions. We see no basis for inferring bad faith. API’s implied premise seems to be that in the absence of comments critical of the 1982 proposals, the agency could not lawfully alter them. We think this premise — which would bar an agency from adjusting proposed rules on the basis of second thoughts originating in the agency — inconsistent with the flexibility intended by the APA. A commentator on negotiated rulemaking urges agencies engaged in such a process not to make changes in the proposed rules except in response to a meritorious com*219ment, arguing that a practice of making sua sponte changes may lead the negotiation process to unravel, see Philip J. Harter, Negotiating Regulations: A Cure for the Malaise, 71 Geo.L.Rev. 1, 102 (1982), presumably because it reduces participants’ benefits from the negotiation. But the soundness of that advice to the agency, however great, cannot impair its legal right to make such changes.
Finally, API argues that the agency failed to explain adequately its decision to place the use of the NAP procedures at the discretion of the Regional Administrators, rather than conditioning their use on party consent. The agency said simply that it believed Regional Administrators to be “in a better position” to make the choice than applicants. This is hardly an earthshaking point, but it surely is enough to override EPA’s 1982 conclusion — itself hardly the product of compelling analysis— that it was unwise to impose NAP procedures on unwilling parties. Such a judgment call requires no more.
VI. The Antibacksliding Controversy
Although non-experts such as ourselves may picture water pollution controls becoming steadily more stringent over time, this is apparently not the case. In some instances new restrictions are more lax than old ones covering equivalent effluents. This raises the question of what happens when a permit holder asks, at the time of renewal or reissuance, to get the benefit of the new rule — to “backslide” as the parties call it.
In its original 1979 regulations, 44 Fed. Reg. 32,864, and in its 1984 promulgation, 40 C.F.R. § 122.44(1), EPA prohibited backsliding except under relatively narrow circumstances. In the initial phase of this suit, industry challenged two aspects of this rule. First, it objected to the ban on backsliding from “BPJ” limits (technology-based limitations set, in the absence of a national guideline, according to a permit writer’s best professional judgment, pursuant to § 402(a)(2) of the CWA, 33 U.S.C. § 1342(a)(2)) when a more lenient federal guideline had taken effect. It attacked this rule on the grounds that the agency lacked statutory authority and had failed adequately to explain its decision to retain BPJ antibacksliding after proposing to shelve the rule in 1982. Second, industry challenged EPA’s authority to prohibit permit-tees from backsliding from their original “new source performance standards” (“NSPS”) under § 306 of the CWA, 33 U.S.C. § 1316.88
On February 4, 1987, Congress passed the Water Quality Act of 1987, Public Law 100-4, 101 Stat. 7 et seq. (“WQA” or “1987 amendments”). In § 404 of the WQA, it restricted backsliding, both with respect to existing BPJ limits and also water-quality-based limits. This obviously moots much of industry’s attack. But there are several residual issues. First, as the WQA does not cover backsliding from NSPS, industry renews its contention that the EPA has lacked statutory authority to impose anti-backsliding.89 Second, industry claims that the WQA antibacksliding restrictions can be applied only on a prospective basis.
*220A. The NSPS Issue.
There are a variety of scenarios under which NSPSs may evolve, and the ripeness issue in fact turns on getting them straight. First, EPA may simply issue a new NSPS. As NSPSs are by definition applicable only to new sources, one would not expect such an NSPS to apply to a source built under an earlier one, and in fact industry, apparently content to rely on its ability to petition for a rulemaking to modify the old NSPS, does not call for a right to backslide in this situation. Second, EPA may revise an NSPS. The parties evidently assume that a revision would by definition cover a source to which the prerevision standard had applied (except to the extent that, since NSPSs are limited to new sources, the agency would explicitly preclude application of stricter terms). Thus there is no dispute here. The final scenario is the case of the agency simply withdrawing an NSPS, leaving nothing in its' stead. Industry’s concern appears to be exclusively with this possibility. Industry Petitioners’ Response to February 3, 1988 Order requiring Clarification of Statements Made at Oral Argument at 3; Reply to EPA’s Post-Argument Supplemental Statement at 2. In this situation, industry contends that the permittee ought to be able to backslide to the effluent levels prescribed by the applicable, technology-based guidelines for existing sources. Industry Post-Amendment Brief at 19.
The parties now dispute whether any such withdrawal ever takes place, the agency saying never, industry saying it has happened, pointing to what it argues is an actual instance. EPA Post-Argument Supplemental Statement at 3; Industry Brief at 85-86 n. 3. On the record before us, we are unsure of the proper characterization of this prior episode. But we are left with a situation where the agency forswears any intention to create the scenario that industry fears, EPA Post-Argument Supplemental Statement at 3, and there is only at most one past occurrence. The institutional interests in avoiding the waste of judicial resources on speculative claims clearly militate against our addressing the substantive claim. No hardship is asserted. We therefore find the issue unripe. Obviously this disposition can have no bearing on any claims dischargers may have arising out of the alleged past episode.
B. Attack on the BPJ Regulation — Ripeness.
Industry’s second contention is that the antibacksliding measures of the WQA may only be applied prospectively. This means that where a discharger holds a current BPJ permit containing stricter limits than would have applied in the absence of EPA’s antibacksliding rules from 1979 to 1987, the baseline for application of the WQA’s antibacksliding rules should be the level to which the permit limits would have slid in the absence of those rules. This is so because prior to the 1987 amendments, in Industry’s view, all aspects of EPA’s antibacksliding rules were unlawful and thus any limits affected by the unlawful regulations are invalid.
Industry’s challenges the pre-WQA anti-backsliding. Industry contends, first, that EPA’s antibacksliding regulation was invalid because it was unauthorized by the unamended Clean Water Act and inconsistent with the specific language and general purposes of the statute. Second, Industry claims that the agency acted arbitrarily by retaining its prohibition on backsliding from BPJ permit limits in its 1984 modified rules without adequately explaining its decision to do so.
We believe both claims are ripe. Each is based on purely legal contentions and neither requires further factual development. The agency asserts an interest in prompt resolution of the controversy. EPA Brief on Ripeness at 25-26. And, significantly, delay here may cause significant hardship. If industry’s contentions are correct, postponing our consideration of the merits could force permittees into continued compliance with unlawfully strict effluent limitations. We have some concern that permittees should have litigated these issues when their permits were renewed or reissued, but conclude that that should not be a barrier: permittees may have refrained from such litigation in the belief that the *221matter was to be resolved in this lawsuit or that 33 U.S.C. § 1369(b) precluded them from raising the claim more than 90 days after issuance of the rules.
C. Attack on the BPJ Regulation — Merits.
1
Industry’s first attack on EPA’s anti-backsliding regulation90 (that the pre1987 CWA did not authorize such a restriction) raises a question of statutory interpretation, to be resolved under the analytical framework provided by Chevron U.S.A., Inc. v. NRDC, 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), and its progeny. See supra p. 167. Reviewing the language, legislative history, and policies of the statute, we conclude that Congress’s intent on the question of EPA’s pre-1987 anti-backsliding authority is not clear; because EPA’s interpretation of the statute is reasonable, we uphold the challenged regulation.
(a)
In mounting its first attack on the BPJ regulation, Industry relies on the statute’s language and legislative history.91 Industry’s textual argument focuses first on section 402(a)(1), 33 U.S.C. § 1342(a)(1), which authorizes the issuance of BPJ permits. That section provides:
[T]he Administrator may ... issue a permit for the discharge of any pollutant, or combination of pollutants ... upon condition that such discharge will meet either all applicable requirements under sections [301 ... 306] fie., national effluent guidelines], or prior to the taking of necessary implementing actions relating to all such requirements, such conditions as the Administrator determines are necessary to carry out the provisions of this [Act].
Id. § 1342(a)(1).
Under Industry’s reading, the statute is clear on its face: BPJ limits are simply temporary measures designed to fill the gap until nationally uniform effluent limitations guidelines have been developed. At that point, those uniform limitations are to supercede BPJ limits. As Industry parses the statutory language, EPA may issue permits containing applicable guidelines under section 301 and EPA may issue permits containing BPJ limits — but only so long as effluent limitations guidelines have not yet been adopted. Under this reading of the statutes, EPA’s refusal to modify permits containing BPJ limitations on pollutants for which effluent limitations guidelines have been adopted exceeds its statutory authority.
Industry also relies on legislative history to bolster its restrictive interpretation. In Industry’s view, that history confirms that Congress did not intend to allow EPA to perpetuate BPJ limits once effluent guidelines were adopted. Industry relies on a passage from the House Committee Report discussing BPJ permits:
The Committee further recognizes that the requirements under section [] 301 ... will not all be promulgated immediately upon enactment of this bill. Nevertheless, it would be unreasonable to delay issuing of permits until all the implementing steps are necessary [sic]. Therefore, subsection (a)(2) provides that prior to the taking of the necessary implementing actions relating to all such requirements, the Administrator may issue permits during this interim period with such conditions as he determines are necessary to carry out the provisions of this Act. Thus, the new permit pro*222gram may be initiated without undue delay upon enactment of this Act.
H.R.Rep. No. 911, 92d Cong., 2d Sess. 126 (1972), reprinted in 1972 Legislative History at 813.92 This passage, Industry contends, clearly evinces Congress’s intent that BPJ limits should do no more than allow the permit program to get underway. After EPA’s promulgation of national effluent limitations guidelines, Industry argues, Congress intended that the temporary, stop-gap BPJ limits adopted during the interim period would give way.
Industry’s second textual argument focuses on sections 301 and 304 of the Act. 33 U.S.C. §§ 1311, 1314. Read in conjunction with section 402(a)(1), id. § 1342(a)(1) (requiring imposition of “applicable” requirements under section 301), these provisions require adoption and implementation of nationally uniform effluent limitations guidelines for industrial categories and classes of point sources.93 Industry argues that the language of these sections evinces Congress’s intent that effluent limitations be applied uniformly to all point sources within industry categories and classes. EPA flouts this intent, Industry contends, when it requires BPJ permittees to abide by BPJ permit limits after the applicable effluent guidelines have been promulgated.
Industry sees support for its textual interpretation in the legislative history of the CWA and judicial opinions construing the statute. As to the former, Industry quotes a statement by Senator Muskie in presenting the Conference Committee Report to the Senate. There, Senator Muskie reiterated that similar point sources should be treated the same.94 Industry then points to judicial opinions that it claims recognize Congress’s desire for uniformity in application of section 301 limitations and section 304 guidelines.95 This emphasis on uniformity, Industry observes, is not gratuitous; in the absence of equal treatment, permittees subject to more stringent limitations may be placed at a significant competitive disadvantage in relation to others in the same industry who are subject to less stringent guidelines.
EPA responds, first, by emphasizing that its backsliding regulation furthers the overall purposes of the CWA. The agency observes that Congress’s primary goal in enacting the CWA was the complete elimination of pollutant discharges. Id. § 125(a)(1); see also NRDC v. Costle, 568 F.2d 1369, 1371 (D.C.Cir.1977). Pursuant to this ambitious goal, the Act instructs EPA to establish technology-based limita*223tions, which are to become progressively more stringent. 33 U.S.C. § 1311. The increasingly strict standards are, in the Act’s express terms, to result in “reasonable further progress” toward this national goal. Id. § 1311(b)(2)(A)(i). The anti-backsliding regulation provides a means, EPA argues, to implement these mandates. Specifically, the prohibition on backsliding ensures that permittees will meet the most stringent limits required by the statute, thereby ensuring “further progress” toward elimination of all pollutant discharges.96
EPA contrasts its implementation of the Act’s far-reaching purposes with Industry’s position, which EPA disparages as undermining the CWA’s goals. As we have seen, Industry interprets the CWA to require the agency to allow BPJ permittees to increase the amount of pollutants discharged, even when the permittee can and has been meeting more stringent BPJ requirements. This position, EPA argues, cannot be reconciled with the Act’s directive to promote reasonable further progress toward eliminating the discharge of pollutants. Id. § 1311(b)(2)(A). At the same time, EPA acknowledges that requiring permittees to retain BPJ limits may conflict with the Act’s goal of uniformity; nonetheless, prohibiting backsliding indisputably furthers the overriding statutory goal — the elimination of all pollutant discharges. Id. § 1251(a)(1).97 Since it falls to the agency to reconcile competing legislative goals, EPA’s determination to promote further progress toward pollutant reduction by prohibiting backsliding is, the agency contends, entitled to substantial deference from the Article III branch. National Wildlife Federation v. Gorsuch, 693 F.2d 156, 166-67 (D.C.Cir.1982).
Responding to Industry’s textual arguments, EPA maintains that Industry’s interpretation fundamentally mischaracterizes the import of a BPJ determination under section 402(a)(1). In EPA’s view, BPJ limits are not merely temporary measures, rendered obsolete by the advent of national effluent guidelines. Rather, BPJ limits constitute case-specific determinations of the appropriate technology-based limitations for a particular point source. In what EPA characterizes as a “mini-guideline” process, the permit writer, after full consideration of the factors set forth in section 304(b), 33 U.S.C. § 1314(b), (which are the same factors used in establishing effluent guidelines), establishes the permit conditions “necessary to carry out the provisions of [the CWA].” § 1342(a)(1). These conditions include the appropriate BPT or BAT effluent limitations for the particular point source. In EPA’s view, the resultant BPJ limitations are as correct and as statutorily supported as permit limits based upon an effluent limitations guideline. Moreover, EPA urges, nothing in the statute indicates that Congress intended general guidelines for broad categories to invalidate these point source-specific limitations.
(b)
With the battle lines thus drawn, we turn to the first step of the Chevron analysis: *224the search for Congress’s intent with respect to the propriety of EPA’s prohibition on backsliding. Employing the traditional tools of statutory construction, INS v. Cardoza-Fonseca, 480 U.S. 421, 107 S.Ct. 1207, 1221, 94 L.Ed.2d 434 (1987), we find that they yield no definitive answer to the question before us.
First, the text of the statute fails to resolve the specific issue presented. As we read it, section 402(a)(1) acts only to preclude the establishment of BPJ permit limits once applicable effluent guidelines are in place. Id. § 1342(a)(1). It provides a broad framework for the ÑPDES program; in orderly fashion, the statute in effect creates a two-phase process (BPJ permit issuance, then national effluent guidelines). The statute thus makes clear that BPJ permits are no longer to be created once national guidelines are in place. But the provision tells us nothing about the fate of BPJ permit limits established prior to the guidelines’ promulgation. As EPA emphasizes, these permits were (presumably) validly issued and reflect the appropriate level of technology control mandated by the Act. The statute simply does not say, as Industry would have it, that promulgation of national standards invalidates these heretofore valid BPJ limits. Nor, however, does it indicate that BPJ limits must be retained. Again, the statute provides that EPA can no longer establish new BPJ limits for categories of point sources subject to a national guideline; but it overreads the statute’s language to maintain that EPA must invalidate existing BPJ limits once national guidelines are in place.
Nor does the CWA’s legislative history help resolve the issue. Like the statutory language itself, the passage referred to by Industry, see supra p. 74, tells us that EPA is to issue permits containing BPJ limits only until national guidelines are in place. H.R.Rep. No. 911, 92d Cong., 2d Sess. 126 (1972), reprinted in 1972 Legislative History at 813. It further instructs that the purpose of these ad hoc limits is to bridge the gap between promulgation of the Act and adoption of effluent guidelines. Id. But having established when EPA may properly establish BPJ permit limits, the legislative history then falls silent; at no point does it address the treatment of extant BPJ limits once the pertinent industry-wide standards are in place.98 There is, in short, an implicit gap in the statutory scheme. See Chevron, 467 U.S. at 851, 859-64, 104 S.Ct. at 2786, 2790-92.
Similarly, the other statutory provisions relied upon by Industry, namely sections 301 and 304, evince no controlling congressional intent. Those two sections do indeed require adoption and implementation of nationally uniform effluent limitations guidelines for industrial categories and classes of point sources. 33 U.S.C. §§ 1311, 1314. And, one congressional purpose in this respect was clear: the Article I branch sought to maximize horizontal equity, or, in the words of Senator Muskie, “to assure that similar point sources with similar characteristics ... meet similar effluent limitations.” Legislative History at 172. Contrary to Industry’s assertion, however, nothing in all this specifies that EPA must apply these uniform guidelines *225uniformly to all point sources within industry categories, no matter what. To the contrary, the statute does not in fact require uniformity in all circumstances; it explicitly provides for exceptions to the general egalitarian rule.99 Moreover, EPA has long asserted authority, which has specifically been upheld by the Supreme Court, to grant variances from effluent limitations guidelines.100 Thus, although exalting the value of uniformity, the statute simply does not require uniformity in all circumstances.
Finding no clear indication of Congress’s intent with respect to the precise issue at hand, we undertake the second step in the Chevron analysis — to determine whether the agency’s interpretation of the statute is reasonable or permissible. Chevron, 467 U.S. at 843, 104 S.Ct. at 2782. We think it is.
First, EPA’s backsliding prohibition represents a logical (although by no means inevitable) reading of the statute’s various provisions. As we have seen, section 402(a)(1) accords EPA the authority to establish BPJ limitations. Nothing in that grant of authority or any other provision in the statute limits the duration of those BPJ limits. Thus, there is no statutory basis for the far-reaching proposition that promulgation of guidelines automatically revokes (or otherwise invalidates) these limits. Moreover, it bears reemphasizing that in establishing BPJ limits, EPA considers the same statutory factors used to establish national effluent guidelines. BPJ limits thus represent the level of technology control mandated by the CWA for the particular point source. That being the case, it is not unreasonable for EPA to conclude that the “necessarily rough-hewn” industry-wide limitations do not reflect the appropriate discharge limits for a particular plant, whose specific characteristics have been taken into account by the permit writer. See Chemical Manufacturers Ass’n v. NRDC, 470 U.S. 116, 120, 105 S.Ct. 1102, 1106, 84 L.Ed.2d 90 (1985).
Finally, EPA’s anti-backsliding approach obviously results in the discharge of fewer pollutants, in conformity with the overriding goal of the CWA. American Frozen Food Inst. v. Train, 539 F.2d 107, 124 (D.C.Cir.1976) (“The principal purpose of the Act is to achieve the complete elimination of all discharges of pollutants into the nation’s waters____”) Although this position may conflict with the competing goal of uniformity, it is, as we have had occasion elsewhere to observe, the agency’s function to reconcile such competing legislative aims. EPA’s preference in favor of the overriding goal of the Act at the expense of the lesser value of uniformity represents, we believe, a reasonable exercise of the agency’s judgment.
Nor do Industry’s various attacks persuade us that EPA’s interpretation is unreasonable. Industry argues, first, that BPJ limitations are based on less data than subsequent categorical standards, which results in inaccurate BPJ limits. Industry provides no support for this contention, however. It may well be true that national guideline writers have the benefit of performance information from other members of the industry category; but BPJ permit writers have the not inconsiderable benefit of information specific to the particular point source. We are scarcely in a position to evaluate which of these represents the “better” determination. In the absence of evidence to the contrary, it is reasonable to conclude that an individualized, point source-specific determination represents an accurate assessment of the statutory requirements applicable to that source.101
*226Industry next attacks the agency’s emphasis on the accuracy of BPJ limits as “disingenuous.” EPA only adheres to these “finely tuned” limits, Industry points out, when subsequently promulgated national guidelines are less stringent; when more stringent guidelines are established, the agency “summarily casts aside BPJ limits” and requires permits to reflect the stricter limitations. Industry Petitioners’ Reply Brief at 49; see 40 C.F.R. § 122.44(c)(1) (1987). This is surely odd, as one-way ratchets tend to be, but it does not strike us as “disingenuous”; indeed, EPA’s policy is, we are convinced, fully consistent with the goals and structure of the Act. As we observed above, the primary purpose of the CWA is the elimination of all pollutant discharges. See supra p. 92-93. The central mechanism for achieving this goal is promulgation and imposition of increasingly stringent effluent limits. In light of this, it is reasonable for the EPA to require conformance with the most stringent limit applicable to a particular source, whether that be BPJ or a national guideline.
Industry’s third point calls into question the agency’s admittedly heavy reliance on the general goals of the statute. Industry contends that in relying on the CWA’s general water quality goals to perpetuate BPJ permit limits, see id. § 1251(a)(1), EPA is permitting general statutory goals to override substantive statutory provisions. EPA’s perpetuation of BPJ permit limits, Industry argues, ignores the specific mechanism established by Congress to implement the CWA’s goals, namely the system of increasingly stringent national effluent guidelines.
In our view, Industry has mischaracterized the nature of EPA’s action. The anti-backsliding regulation does not override any of the Act’s substantive provisions; rather, the program represents the agency’s “fitting together” of one substantive provision (the BPJ permitting program) with another (subsequent national guidelines) in the absence of an express direction from Congress as to how the two should work together. As the agency charged with interpreting the complicated statutory provisions that comprise the CWA, EPA is entitled to considerable deference in the interpretive process of making the regulatory machinery work. See, e.g., Chevron, 467 U.S. at 865-66, 104 S.Ct. at 2793; Chemical Manufacturers Ass’n, 470 U.S. at 125-33, 105 S.Ct. at 1107-11; Nat’l Wildlife Federation, 693 F.2d at 166.
In reaching this conclusion, we note that EPA has provided for the modification of a BPJ permit where experience shows that operation and maintenance costs under the permit “are totally disproportionate from [those costs] considered in the development of a subsequently promulgated effluent limitations guideline.” 40 C.F.R. § 122.62(a)(15) (1987). Nothing we say here either interprets or endorses the “totally disproportionate” standard or any particular view of what it may mean. We have no doubt that in considering applications for modification of a BPJ permit under this section EPA will be sensitive to Congress’s very serious concerns for equal treatment of point sources (particularly of course those in competition with each other), concerns reflected in the statutory structure ánd legislative history which we recited above.
2
In its second attack, Industry claims that EPA acted arbitrarily and capriciously by inadequately explaining its decision in 1984 to retain the anti-backsliding rule after having proposed its elimination.
*227(a)
An understanding of Industry’s contention requires a brief restatement of the challenged regulation’s history. The concept of backsliding first appeared in the Code of Federal Regulations in 1979 when EPA adopted rules limiting the circumstances under which NPDES permits can be reissued or modified to incorporate less stringent limitations. Under these rules, permit limits based on BPJ could not be replaced by less stringent effluent limitations guidelines established in national rulemakings. See 40 C.F.R. § 122.15(i) (1979).
In 1982, EPA proposed revisions to these NPDES rules. The revised rule would have allowed BPJ permittees, either upon permit reissuance or through modification during the permit term, to obtain revised permit limits based upon subsequently promulgated, less stringent effluent limitations guidelines. EPA Consolidated Permit Regulations, 47 Fed.Reg. at 52,072, 52,084-86, 52,089 (1982). EPA’s rationale for the revisions was two-fold: first, the agency sought to apply effluent guidelines in a nationally consistent manner so that companies operating under the less stringent, subsequently promulgated guidelines would not have an unfair advantage over companies with more stringent BPJ permits. Id. at 52,084. Second, the agency feared that the limit on backsliding might encourage dischargers to challenge second-round permits containing BPJ limits in order to avoid being locked into more stringent limits than might ultimately be contained in effluent guidelines. Such challenges, the agency contemplated, would result in a drain on EPA’s resources. Id.
In its final rule, EPA retained the anti-backsliding prohibition, concluding that sufficient justification did not exist to change the policy. EPA NPDES Permit Regulations, 49 Fed.Reg. 37,998, 38,021 (1984). Industry claims that EPA failed adequately to explain this policy “reversal.” This is, Industry contends, not surprising, inasmuch as there was no evidence to support the agency’s rejection of its own proposal.
(b)
As this background illustrates, Industry’s position suffers from a fundamental defect: EPA's 1984 rulemaking did not alter the agency’s position on backsliding; rather, the agency in 1984 refused to alter its prior stand. The 1982 proposal was just that — a proposal. EPA’s “policy” was, and is, embodied in the 1979 regulations. It would be odd indeed to demand fresh evidence in support of a decision to continue an existing policy. Yet this is what Industry would have us require.
The real point is, for EPA’s decision to be sustainable, the agency must have engaged in reasoned decisionmaking. This means, of course, that EPA must have considered the issue and provided a reasoned basis for its action. Here, EPA clearly identified its reasons for declining to jettison its anti-backsliding rule. The proposed change, EPA explained, was prompted not by any perceived illegality in the existing regulation, but rather in response to equity concerns. Id. at 38,019. These alleged inequities were not, EPA found, supported in the record; specifically, the record failed to support Industry’s claim that the anti-backsliding policy imposed significant economic and competitive disadvantages on BPJ permittees. Id. at 38,020. With respect to the second concern articulated in the proposal, EPA determined that the possibility of numerous challenges to second-found BPJ permits was insufficient, standing alone, to justify a policy reversal. The agency explained that it would attempt to forestall such challenges to the extent possible by deferring BPJ permit reissuance when promulgation of the applicable guideline is expected. Id.
Having reconsidered its concerns, EPA concluded that since adopting the proposal would damage the integrity of BPJ permits and do nothing to further the statutory goal of eliminating the discharge of pollutants, it would reembrace its longstanding policy. Id. at 38,019-20. This judgment appears to us to be both reasonable and fully explained.
*228D. Application of the New WQA Rules.
Finally, industry argues that even if we uphold the agency’s old anti-backsliding rules, the new WQA rules must be applied only prospectively. It points out that the agency's regulation had more exceptions and was therefore less strict than the WQA. It argues that permittees may have accepted certain limitations before the WQA in reliance on their ability to backslide in the event of more lenient guidelines, within the terms of then-applicable agency antibacksliding rules. Had the WQA controlled, with its greater rigidity as to backsliding, these firms might have fought harder against the imposition of disputed limits in their initial permits. Industry argues that it would be an inequitable reading of the WQA to lock permit-tees into limits based on now-invalid premises.
This theory attacks only a likely future agency interpretation of the WQA. So far EPA has adopted no regulations under the WQA; there is no agency action to attack. The claim is quintessentially unripe.
VII. Net/Gross Limits
Should a source be able to take a credit for pollutants in its intake water? The EPA has ruled that as a general matter it may not — that pollution will be measured “gross.” 40 C.F.R. § 122.45(g). Its supporting reasoning seems internally contradictory. On the one hand it says that “within a broad range of influent pollutant concentrations, treatment systems typically reduce pollutants to a certain level.” 49 Fed.Reg. at 38,026/1. This suggests that denial of credit is suitable because it will usually cost the discharger no more to attain that “certain level,” regardless of intake pollution. But EPA goes on to explain that the gross evaluation is correct because otherwise sources on polluted streams will get a kind of unfair competitive edge — the ability to comply with effluent limitations, by virtue of intake credits, “by utilizing a lower level of treatment than their competitors on cleaner streams.” Id.
In any event, the agency has provided for a credit where either the applicable effluent limitation guidelines direct net treatment or where the discharger shows that pollutants in the intake water prevent compliance with the applicable technology-based limitations. 40 C.F.R. § 122.45(g)(1). For credit under the second theory, the discharger must show that the control system used or proposed for use would meet applicable limitations and standards in the absence of intake water pollutants if “properly” installed and operated. § 122.45(g)(1)(h). (The discharger can enjoy the credit even if the control system could meet the standards on a gross basis, but at “significant additional expense.” 49 Fed.Reg. at 38,027/3.) Even then, the source will enjoy a credit only to the extent necessary to meet the technology-based limits. § 122.45(g)(3).
EPA’s regulations impose three further limitations. First, credit for generic pollutants (such as biochemical oxygen demand (“BOD”) or total suspended solids (“TSS”)) will generally not be available unless the permittee can demonstrate that the constituents of the generic measure in the effluent are substantially similar to the constituents of the generic measure in the intake water. § 122.45(g)(2). Second, credit for pollutants in intake water that has been drawn from a body of water different from that into which the discharge flows will be granted only if this will lead to no environmental degradation. § 122.45(g)(4). Third, EPA made the exception provisions inapplicable to discharges of “raw water clarifier sludge” and “filter backwash,” § 122.45(g)(5), i.e., pollutants filtered from the intake water by dischargers, such as electric generating plants, that need clean water for their internal operations. It preferred to leave this matter to the discretion of its permit writers. 49 Fed.Reg. at 38,027/1.
Industry attacks the regulations along two lines. First, it urges that they directly violate the Clean Water Act, which regulates only “discharges of pollutants,” 33 U.S.C. § 1311(a), which it in turn defines as “any addition of any pollutant to navigable water.” Id. § 1362(12) (emphasis add*229ed). Even if the net/gross regulations are not simply barred by the statute, industry argues that the specific criteria for denial of credit are arbitrary and capricious.
The first issue appears purely legal in character and the agency can well be said to have crystalized its rejection of the industry viewpoint. As in our approach to the agency’s use of limits framed in terms of toxicity, see section C above, we might well carve this issue out from the more fact-bound claims and find it ripe. The EPA, however, invokes our decision in Diamond Shamrock Corp. v. Costle, 580 F.2d 670 (D.C.Cir.1978), where we held a previous and similar industry challenge to the net/gross regulations unripe. The authority of Diamond Shamrock as a guide to review of regulations under the Clean Water Act, however, has been somewhat sapped. The case came before us on appeal from the district court, which had exercised jurisdiction under 28 U.S.C. § 1331. This was four years before we held, in NRDC v. EPA, 673 F.2d 400 (D.C.Cir.), cert. denied, 459 U.S. 879, 103 S.Ct. 175, 74 L.Ed.2d 143 (1982), that the NPDES regulations were reviewable under 33 U.S.C. § 1369(b). Accordingly, the court did not consider Congress’s preference for early review, a preference that under EaglePicher Industries, Inc. v. EPA, 759 F.2d 905 (D.C.Cir.1985), we read as tending to eliminate or at least radically weaken the requirement of a showing of hardship.
We find, however, that it makes sense to persevere with Diamond Shamrock. EPA asserts that the claims are unripe and thus evidently believes its institutional interests favor deferral of review. It also notes that some firms may have refrained from participation in this attack, on the supposition that Diamond Shamrock barred review. We certainly do not feel we may infer such absence with any certainty, and we have no assurance that such absent parties could have added much to the legal dispute. On the other hand, and noting that the broad legal issue may have links to the concededly fact-dependent ones that we cannot now fully perceive, we think these factors enough to tilt the balance in favor of finding not only the fact-dependent claims, but also the more purely legal one, unripe.
VIII. Upset Defense
A. Ripeness.
The agency’s regulations provide an affirmative defense to an action for noncompliance with technology-based permit limitations where the source shows that the non-compliance was due to an “upset.” Industry here objects to the agency’s failure to extend this defense to noncompliance with water-quality-based limits.
The agency defines an “upset” as “an exceptional incident in which there is a temporary and unintentional noncompliance with permit effluent limitations because of factors beyond the reasonable control of the permittee.” 49 Fed.Reg. at 38,038/1. Upsets may be caused by external events, such as power failures or storms, or by unpreventable failures of effluent treatment equipment; noncompliance due to operational error, improperly designed treatment facilities, inadequate treatment facilities, lack of preventive maintenance, or careless or improper operation is excluded. 40 C.F.R. § 122.41(n)(l). To qualify for the defense, a discharger must also perform any required remedial measures. § 122.41(n)(3)(iv).
In 1982 the agency proposed to extend the upset defense to violations of water-quality-based limits. 47 Fed.Reg. at 52,089/1. The defense would be available only to permittees who could demonstrate that despite the upset, instream water quality standards were maintained in all stream segments and for all parameters that could have been affected by the discharge. Id. Although EPA did not regard this extension as legally required, id. at 52,079/2, it offered the proposal on the view that there was no reason to punish a permittee for an upset if it could prove the absence of injury to water quality standards. Id.
In 1984, after reevaluating its proposal in light of various criticisms, the agency concluded that it would be impractical to extend the upset defense to violations of wa*230ter quality-based effluent limitations. 49 Fed.Reg. at 33,038/2. It reasoned that “[although the proposal would seemingly allow permittees to claim an upset defense, the costs, burdens, and technical difficulty of establishing that water quality standards were not violated would make the defense nearly impossible to establish.” Id. at col. 3. Rather than leave in place an affirmative defense it believed “illusory,” it decided to deny extension of the defense and to rely instead on case-by-case prosecutorial discretion. Id.
Industry’s objections are two-fold. It contends first that the agency is legally required to provide for such a defense, at least where the discharge does not result in the violation of a water quality standard. Second, it maintains that the agency’s decision to scrap its 1982 proposal was arbitrary and capricious because, even if the defense (as proposed) could not be met and was therefore “illusory,” the agency failed to evaluate potential alternatives. It points to the curiosity of the agency’s rejecting the defense on the ground that it was not good enough for the very interests that sought it.
We find the challenge ripe. The agency’s position is clearly final. The issues raised may fairly be viewed as ones of law: whether the statute requires the proposed extension of the defense and, if not, whether the agency gave a reasoned explanation of its decision.
The agency’s proposed exercise of prosecutorial discretion is not a close enough substitute for the defense to justify deferral of review. The agency would in no way be bound to exercise that discretion favorably. See 47 Fed.Reg. at 52,089/1. Moreover, even a firm agency commitment to exercise prosecutorial discretion would leave permittees without an affirmative defense in citizens’ enforcement suits under § 505 of the Act, 33 U.S.C. § 1365. Further, unlike the net/gross dispute, the issue cannot be resolved at the time permits are issued; accordingly failure to resolve it now seems more likely to have an impact on dischargers’ primary conduct. Even though such matters as “careless or improper operation” will disqualify a source for the defense, it seems probable that its absence will somewhat increase compliance expenditures.
Finally, the agency’s institutional interests, as reflected in its brief on ripeness, at 29, evidently favor immediate review. We will address the claims on the merits.
B. Merits.
Lacking infallibility, no pollution control technology works perfectly all of the time. Occasionally, through no fault of the operator, the technology will fail, and pollution levels in the effluent will correspondingly rise. Current EPA regulations provide that when permit effluent limitations based on technological capabilities are briefly exceeded as the result of such an incident, the offending plant will nevertheless be deemed to be in compliance with the Act.102 This is the so-called “upset defense.” 103
EPA, however, has refused to make this defense available to plants exceeding permit limitations predicated upon water quality standards.104 In the agency’s view, the Act makes an important distinction between technology-based permit limitations and water quality-based permit limitations by requiring water quality standards to be observed at all times in all situations. On the other hand, because the technology used to satisfy water quality-based permit limitations is no more foolproof than that employed to meet technology-based permit limitations, industry petitioners contend that the rationale for the upset defense extends to water quality-based limitations as well. These petitioners further argue *231that a violation of a water quality-based permit limitation caused by an upset does not translate automatically into a violation of water quality standards in the body of water receiving the effluent. We disagree with industry petitioners that an upset defense for water quality-based permit limitations is required by the Act. We find meritorious, however, their claim that EPA acted arbitrarily when it declined to provide a limited form of the defense.
The need for an upset defense in connection with technology-based permit limitations was first recognized in 1977 by the Ninth Circuit in Marathon Oil Co. v. EPA,105 in connection with offshore oil and gas platforms. Prior to that decision, EPA refused to recognize such a defense under any circumstances. Typically, EPA’s practice in formulating an effluent discharge standard for an industry was to collect data from plants in the industry applying the technology at an appropriate level in an exemplary fashion.106 EPA would then derive from the average of these data a standard for effluent discharge which these plants would be able to meet most of the time.107 The standard was purposefully not set at a level low enough to accommodate occasional failures of control technology at these plants, as this in EPA’s view would have resulted in too lax a standard.108 This meant that a plant utilizing controls at a suitable level of technology in a commendable manner would invariably be in violation of the standard some of the time.109
The court recognized that EPA was in effect holding dischargers to a higher standard of technology than that demanded by the Act:
The Federal Water Pollution Control Act requires point sources of pollution to utilize the “best practicable control technology currently available” [BPCTCA] prior to 1983. The EPA cannot impose a higher standard without violating the Control Act. And yet the permits as currently written do exactly that. The permits on their face require petitioners to meet the standards 100 percent of the time. But platforms using BPCTCA can only be expected to achieve the effluent standards 97.5 percent of the time in the case of deck drainage and 99 percent of the time in the case of produced water. We, therefore, remand to the EPA with instructions to insert upset provisions into the permits.110
Industry petitioners insist that a similar result should obtain in this case. A reasonable reading of the Act, they say, reveals no meaningful distinction between technology-based and water quality-based limitations for this purpose;111 although limitations of the latter type may be more stringent than those grounded on technology, both are achieved through use of pollution control technology, and therefore are equally subject to upsets. They maintain that once limitations have been set and plant owners deploy appropriate pollution control equipment, they can be held responsible only for violations that could have been prevented.
In a second argument, industry petitioners contend that EPA’s regulations obscure an important dissimilarity between permit limitations applicable to an individual plant and standards applicable to a body of water. An upset exceeding permit limitations at a single plant, they submit, does not *232always lead to infringement of water quality standards for the body of water receiving that plant’s effluent. Because EPA has authority to enforce only those water quality permit limitations on effluents that are necessary to preserve water quality standards, they assert that EPA cannot penalize a plant for an upset that does not contravene these standards.
Industry petitioners’ argument, however, overlooks the fundamental differences between the two types of permit limitations clearly envisioned by Congress. While Section 301(b)(1)(A) of the Act mandates merely the “application of the best practicable technology currently available,” Section 301(b)(1)(C) requires satisfaction of “any more stringent limitation ... necessary to meet water quality standards.” 112 The import of this language is made clear by the legislative history:
[A]ny point source will be required to have effluent limitations which are consistent with the water quality standards and concomitant load limits for the receiving waters. Thus, if the best practicable control technology is inadequate to meet the water quality standards and load limits, a more stringent requirement would be imposed upon the given point source.113
And
[t]he requirements of H.R. 11896, will assure that water quality standards are met, and that even if such “best practicable control technology currently available” is not sufficient to meet water quality standards, each point source will still be required to be so equipped to further enhance the quality of our waters. In other words we require the upgrading of our waters to a much greater degree than does S. 2770.114
This evidence strongly supports EPA’s position that Congress did not intend to tie compliance with water quality-based limitations to the capabilities of any given level of technology. A technology-based standard discards its fundamental premise when it ignores the limits inherent in the technology.115 By contrast, a water quality-based permit limit begins with the premise that a certain level of water quality will be maintained, come what may, and places upon the permittee the responsibility for realizing that goal.116
In our view, the Act’s differing treatment of water quality and technology-based standards provides persuasive evidence that the broad argument advanced by industry petitioners must fail. Nothing in the Act indicates suitably that individual technological failures may excuse violations of water quality-based limitations. Nor, contrary to industry petitioners’ second argument, does the Act mandate that permit limitations properly established in accordance with water quality standards be ignored in favor of a direct measurement of water quality standards instream. The Act does not expressly provide for an upset defense, and, looking to the Act’s structure and legislative history, we discern no clear congressional intent to confer one. While at first blush it seems odd to expect dis-chargers to go beyond the limits of extant technology, it becomes apparent that Congress had a deep respect for the sanctity of water quality standards and a firm convic*233tion of need for technology-forcing measures.
In circumstances where, as here, congressional intent is unclear, it is our duty to uphold an agency’s decision representing a permissible interpretation of its governing statute.117 We are satisfied that it does, and that the EPA’s outcome reflects a reasonable accommodation of competing considerations. Indubitably, the Act requires overall water quality to be maintained, and even though an individual permit violation may not always produce an infringement of water quality standards for a particular body of water, the agency is at liberty to make the regulatory judgment not to allow permittees to assert an upset defense. EPA could reasonably view the impact to water quality under such a forgiving regime as potentially dangerous.
This conclusion does not, however, mark the end of the inquiry that we have been summoned to conduct. Although EPA is not statutorily compelled to provide an upset defense, we nevertheless must ask whether EPA acted arbitrarily in declining to allow it. Industry petitioners charge that it did when first it suggested the desirability of an upset defense for water quality permit limitations and later omitted it in the final regulations.
In its 1982 proposed regulations, EPA stated its belief that there is “no reason to penalize a discharger [able to] prove that an upset occurred and that water quality standards were met despite its non-complying [sic ] discharge.”118 EPA thus envisioned an extension of the upset defense to water quality-based permit limitations, with the added requirement that the permittee show that water quality standards were observed in all stream segments for all pollutants which the upset could have affected.119
EPA never altered its interpretation of the statute that it had the statutory authority to establish an upset defense for water quality-based permit limitations.120 In its final regulations, however, EPA declined to extend the upset defense to water quality-based limitations, concluding that practical obstacles confronting a permittee attempting to establish the defense would be so great as to render the defense illusory.121 Our task is to ascertain whether this decision withstands scrutiny.
The scope of judicial review of agency decisionmaking under the arbitrary-and-capricious standard is narrow.122 Nonetheless, the agency must “examine the relevant data and articulate a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made.’ ” 123 An agency rule normally must be overturned if the agency has “entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.” 124 We cannot of course substitute our judgment for that of the agency.125 “What we do require is that the [agency] come to grips with the obvious ramifications of its approach and address *234them in a reasoned fashion.” 126
EPA’s explanation of its decision to deny an upset defense for water quality permit limitations falls short of these standards. In its final rule, EPA listed the elements it deemed satisfactory proof that water quality standards were maintained during an upset.127 But then, without discussing any record evidence, EPA noted several problems that would arise naturally during the course of an effort to demonstrate that fact, and summarily dismissed the defense as impracticable.128. There is no indication in the record before us that EPA made any serious effort to measure the intensity of these problems or to consider possible solutions to them.
EPA’s decision rests upon its estimate that the defense would be virtually impossible to establish to its satisfaction. On this account, it felt that even to hold out the defense as a hope would be “illusory.” 129 EPA’s only argument, in essence, is that the interests of the industry petitioners would be ill-served by the defense because EPA cannot envision circumstances under which petitioners would be able to make it out. Yet EPA offers no more than mere speculation to support its conclusion. These are not adequate grounds upon which to sustain an agency’s action. Because it is elementary that agency action may be upheld only on grounds articulated by the agency,130 EPA must consider its position.131
In so ruling, we do not mean to imply that EPA must permit the defense. EPA may decide to disallow it if, for example, it reasonably foresees a waste of agency resources in dealing with a complicated defense that can hope to succeed only rarely. Nor do we dispute EPA’s position that it has authority to determine in the first instance what constitutes sufficient proof that water quality was maintained during an upset. Applying its technical expertise, EPA may, for instance, require monitoring schemes after an upset, establish standards for the promptness with which such monitoring must begin, and lay out procedures by which evidence derived from mathematical models may be used to supplement monitoring data. The agency’s experience with water quality monitoring programs utilized for other purposes 132 likely would be valuable in application of these techniques. Yet EPA, in delineating its final rule, failed even to mention efforts of this nature.
In summary, the Act, by virtue of the important differences between technology-based and water quality-based standards, does not require EPA to provide for an upset defense to water qüality-based permit limitations violations. Because the agency reasonably could decide not to provide the defense, and the Act does not require it, the agency is at liberty to decide in its expert judgment not to allow the defense. Nevertheless, having included water quality-based permit violations in the proposed regulations, we find that EPA acted arbitrarily in eliminating the defense solely on the ground that it would be of little use to dischargers. EPA must therefore conduct further proceedings to determine whether to extend the upset defense to violations of water quality-based permit limitations. The existing upset regulation pertaining to noncompliance with technology-based permit limitations is to remain in force and effect.
*235IX. APA Continuance
A. Ripeness.
The Clean Water Act prohibits the discharge of pollutants into navigable waters without a permit. 33 U.S.C. § 1311(a). NPDES permits, whether issued by a state or by the agency, are good for only five years. Id. at § 1342(a)(3). We deal here with the status of sources whose permits have expired but who have not received renewals due to the great backlog of unprocessed renewal applications. At oral argument the agency estimated that these numbered more than 60,000. Absent some mechanism for continuance, each source owner operating under an expired permit would be in violation of the Act and subject to enforcement actions and penalties. §§ 1319, 1365.
The risk that such delays might halt the operations of innocent permittees in violation of Congress’s intention is not unique to environmental regulation and was in fact recognized by the writers of the APA. Its § 9(b) provides a general solution:
When the licensee has made timely and sufficient application for a renewal or a new license in accordance with agency rules, a license with reference to an activity of a continuing nature does not expire until the application has been finally determined by the agency.
5 U.S.C. § 558(c). EPA has promulgated a regulation aimed at structuring the application of § 9(b) to NPDES permits. 40 C.F. R. § 122.6(a) provides that:
When EPA is the permit-issuing authority, the conditions of an expired permit continue in force under 5 U.S.C. § 558(c) until the effective date of a new permit ... if:
(1) The permittee has submitted a timely application ... which is a complete ... application for a new permit; and
(2) The Regional Administrator, through no fault of the permittee does not issue a new permit with an effective date ... on or before the expiration date of the previous permit (for example, when issuance is impracticable due to time or resource constraints).
NRDC attacks the agency’s authority to approve continuances of out-of-date permits under the CWA and § 9(b) of the APA. The agency argues that neither challenge is ripe for review. We disagree.
The environmentalists begin with the undisputed fact that the Act utilizes, in part, a “technology-forcing” approach, mandating the implementation of increasingly stringent restrictions over time. Primarily they point to 33 U.S.C. § 1311(b)(2)(A), which requires use of “best available technology” (“BAT”) by July 1, 1984. They claim that allowing sources whose permits have expired to continue discharging at higher than BAT levels allows these operators to violate § 1311 and flouts the Act’s explicit timetable.
The agency here argues — rather confusingly — that the contention is not ripe. It alludes to two possibilities that might require case-by-case consideration. First, it notes that for some permittees the BAT requirement may be no more stringent than BPT. EPA Brief at 266. But at the same time it argues strenuously that continuance is permissible even when BAT is more stringent and indicates clearly that this view is its final position. Moreover, the agency does not contend that all permittees would be in compliance with BAT. Thus we do not see how the possibility of some license continuances being in accordance with NRDC’s view of the law militates against immediate review of the rule as to the many that will violate that view. Second, the agency alludes to the possibility that, even though 40 C.F.R. § 122.6(a) be legal, some invalid continuances might be issued under it. Again, we fail to see how that militates against resolving now the basic legal question whether the Act allows EPA to apply § 9(b) to permittees not yet in compliance with BAT. Finally, assuming merit in the environmentalists’ challenge, they would suffer hardship were we to postpone resolution of their claim: prolonged periods over which thousands of operators would unlawfully discharge pol*236lutants into the nation’s waters. We thus find the issue ripe for review.
NRDC's alternative claims that the regulation violates § 9(b) of the APA are also ripe. First, it argues that to be consistent with § 9(b)’s requirement that the permittee’s operations be “of a continuing nature,” the agency's regulation must be explicitly limited to cover only continuing commercial operations undertaken on a daily basis. NRDC Brief at 54-55. It specifically objects that seasonal operations such as those of oil and gas drilling cannot be classified as continuing. Second, it argues that an application for permit renewal cannot be “sufficient,” as required by § 9(b), if the permit sought to be renewed contains provisions that the agency could not allow in the renewed permit — here, provisions short of BAT. Id. at 56.
The agency’s position has generally crystalized and the disagreement between the parties is, once again, over a matter of statutory interpretation. Of course our disposition may fail to resolve the lawfulness of applying the regulation to some episodic discharges, the character of which we cannot now know. But those potential gaps — which may be more theoretical than real — are not enough to destroy the generally legal character of the issue. Resolution of the issues is likely to eliminate dispute over a great number of outstanding applications, and delay would (if petitioners be correct) materially injure their interests in unpolluted (or less polluted) water. We find the claims ripe.
B. Merits.
Under the Act, the term of a permit may not exceed five years.133 EPA developed a backlog of permit renewal applications; according to the agency, forty percent of the 60.000 outstanding permits had expired without reissuance,134 including more than 5.000 since June 30, 1984, alone.135 Because discharge of a pollutant is illegal without a permit,136 every source owner operating under an expired permit would be in violation of the Act, and subject to enforcement actions and penalties, absent some mechanism for continuance of the permit.137
To avoid such “irreparable injuries” to parties subject to licensing requirements,138 Section 558(c) of the Administrative Procedure Act (APA) provides that
[w]hen the licensee has made timely and sufficient application for a renewal or a new license in accordance with agency rules, a license with reference to an activity of a continuing nature does not expire until the application has been finally determined by the agency.139
A regulation implementing the NPDES explicitly applies Section 558(c) to the permit program.140 NRDC attacks this provision on grounds of its alleged inconsistency with the basic scheme of the Act and its asserted failure to ensure that continued permits meet the prerequisites of Section 558(c). We disagree with both premises of this challenge, and accordingly uphold the *237regulation.141
1. Consistency With Clean Water Act Scheme. The primary objection to the regulation as continuance is that it allows sources meeting only best practicable technology requirements to keep on discharging at current levels after the date for achievement of best available technology has passed.142 Therefore, it is contended, the regulation implicitly extends the statutory deadline for best available technology.143
We believe, however, that this case is governed by the Supreme Court’s decision in Pan-Atlantic Steamship Corp. v. Atlantic Coast Line Railroad Co.144 There, as here, the substantive standards to be met by the renewal application for a permanent license differed drastically from those governing issuance of the initial temporary permit,145 but the Court allowed the permit to continue in force:
[W]e see no reason why [Section 558(c) ] may not be invoked to protect a person with a license from the damage he would suffer by being compelled to discontinue a business of a continuing nature, only to start it anew after the administrative hearing is concluded____ [0]nce the conditions of [Section 558(c)] are satisfied, an extension in the interests of economy and efficiency is authorized.
... Unless the authority is vested in the Commission by [Section 558(c)], the operation, no matter how essential or necessary, must be discontinued at the end of 180 days. We think such a reading of the law would mutilate the administrative system which Congress created by the two Acts. Where the remedy for an evil is clear, the remedial provisions of the Administrative Procedure Act should be given full effect.146
Correspondingly, without a similar extension of authority, any industrial operation requiring a discharge permit, no matter how important or essential that operation may be, would be brought to a halt on June 30, 1984, or at the later expiration date of its current permit, regardless of “ ‘the serious hardships occasioned both to [it] and to the public by expiration of [its] license before the agency finds time to pass upon its renewal.’ ”147
In allowing continuance of an expired broadcast license pursuant to a statutory provision similar to Section 558(c), we ourselves recognized that the congressional purpose was “protection of licensees from the uncertainties stemming from protracted administrative consideration of applications for license renewals,” and that “there is no indication in the language or history ... that this safeguard is to subsist only *238for a limited interval.” 148 Neither is there any hint that the safeguard afforded by Section 558(c) exists only while the permit remains unchanged. As in Pan-Atlantic Steamship, the standard that EPA must here apply in renewing discharge permits differs from that controlling issuance of earlier permits, and the agency cannot renew unless a more stringent test is satisfied. In both cases, however, the agency’s lack of independent statutory power to extend the permit is over-balanced by Section 558(c); in each instance, the expired permit is continued, not by affirmative agency action, but by operation of law. We conclude that EPA’s recognition of this result is entirely proper.149
2. Consistency With Section 558(c). NRDC also asserts that the regulation temporarily extending expired permits is deficient because it does not incorporate all the criteria of Section 558(c). NRDC focuses on the section’s requirements that the application for a renewal permit be “sufficient,” and that the permitted activity be "of a continuing nature.” 150
We begin by noting that, as EPA emphasizes,151 the regulation does not purport independently to extend permits. Rather, it simply delineates the circumstances under which “the conditions of an expired permit continue in force under 5 U.S.C. 558(c).” 152 Thus, because the expiring permit is continued only by authority of Section 558(c), a requirement that the permit application meet that section’s standards as well as those set by the agency is implicit in the regulation itself.153
Additionally, NRDC’s contention that the regulation as written may apply to noncontinuing activities seriously distorts the statutory language. NRDC would limit the words “of a continuing nature” to “continuing commercial operations undertaken on a daily basis,” and would thereby exclude such activities as offshore drilling.154 Yet, as NRDC concedes, the APA does not define “continuing” at all, much less in the narrow manner advocated.155 Nothing in the APA indicates that “continuing” activities may not be weekly, monthly, seasonal or even intermittent in nature.156 Indeed, as respondent indicates, NRDC’s restrictive definition would have astonishing results for some climate-based activities, such as exploratory drilling, that can operate year-round in some areas and only seasonally in others.157 We see no reason to interpret Section 558(c) so narrowly.
NRDC’s second claim holds out at least a modicum of merit. Citing Section 558(c)’s requirement that applications be “sufficient,” NRDC contends that any renewal application for a permit that does not meet best available technology requirements is by definition insufficient, and thus may not *239be continued in force. EPA’s response is that “sufficient” means only that the application must be procedurally proper.158
At least one court has accepted NRDC’s argument. In an alternative holding, the District of Alaska opined:
[A]ny application must be a “sufficient” application. In this case, ARCO and Exxon applied to continue operation under [best practicable technology] standards. Since EPA could not issue any permit for after June 30 that did not provide for [best available technology] standards, EPA could not legally grant the applications and the applications therefore weren’t “sufficient.” 159
We find this reasoning unconvincing. As we have pointed out, the extension follows automatically from operation of Section 558(c), and not from action of the agency.160 The terms of the expiring permit are unrelated to those of the anticipated new permit, and the agency can legally grant permit renewal applications after June 30, 1984, provided they satisfy statutory and administrative standards. In almost any renewal scheme, the licensing authority has broad power, and sometimes the duty, to change permit requirements, even by adopting new standards. Were a renewal permit application to be deemed insufficient merely because the standards therefor differ from those controlling issuance of the earlier permit, the operation of Section 558(c) would be seriously curtailed.161 On the contrary, by its terms, Section 558(c) tests, not the sufficiency of the expiring permit, but the sufficiency of the renewal permit “application ... in accordance with agency rules.”162 EPA’s construction of this language as referring only to procedural standards set by the agency comports well with the plain language of the statute,163 and we will not upset it here.
* * * * * *
In sum, we hold that the following issues are not ripe for review: effluent-related permit conditions; admission of evidence in permit proceedings; three of the four challenges to EPA’s regulations permitting the establishment of toxicity-based limits; backsliding from NSPS permit limits; and net/gross limits. On the merits, we find both the state program requirements established by EPA and the veto authority regulations to be within the permissible bounds of the agency’s discretion. We reject Industry’s argument that EPA lacks statu*240tory authority to frame effluent limits in terms of toxicity; similarly without merit are those procedural challenges to EPA’s toxicity regulations that we have concluded are ripe. We also hold that the non-adversary panel procedures permitted by EPA’s regulations do not, on their face, infringe permit applicants’ procedural rights, nor do the regulations violate the parties’ 1982 settlement agreement. We further hold that EPA possessed authority under the Clean Water Act to prohibit backsliding from BPJ permit limits and that its decision to continue to do so in its 1984 regulations was fully explained and supported by the record. We also uphold the agency’s regulation providing for continuances of out-of-date permits. We conclude, however, that EPA lacks authority to impose conditions in permits that are unrelated to effluent limits. Finally, we hold that although the CWA does not require EPA to provide for an upset defense to water quality-based permit limitations' violations, the agency acted arbitrarily in eliminating that defense from its final regulations. Accordingly, the petitions for review are denied in part and granted in part.
Judgment accordingly.
. The environmental petitioners are the Natural Resources Defense Council, Inc. and Citizens For a Better Environment. The industry petitioners include the following: the American Iron and Steel Institute, the American Paper Institute, the National Forest Products Association, the American Petroleum Institute, the American Textile Manufacturers Institute, the American Wood Preservers Institute, the Chemical Manufacturers Association, the National Coal Association, Burlington Industries, Ford Motor Company, Kennecott, United Technologies Corporation, and Virginia Electric and Power Company.
. These NEPA issues include: (1) EPA’s imposition of permit conditions unrelated to effluent limits based on its NEPA review; (2) the agency1 s refusal to adjust effluent limits as a result of this review; and (3) the admissibility of evidence in permit proceedings.
. This group of procedural issues includes challenges to the standards for (1) state allowance of public participation; (2) maximum penalties under state programs; and (3) EPA’s veto of permits approved by states.
. This antibacksliding controversy actually involves two separate issues. First, Industry challenges the propriety of prohibiting backsliding from permit limits based on new source performance standards ("NSPS”). Second, Industry contests the ban on backsliding from "BPJ” permit limits, which are those established, in the absence of a national guideline, according to the permit writer’s best professional judgment.
. § 509(b) of the Act provides:
(1) Review of the Administrator’s action (A) in promulgating any standard of performance under section 1316 of this title, (B) in making any determination pursuant to section 1316(b)(1)(C) of this title, (C) in promulgating any effluent standard, prohibition, or pretreatment standard under section 1317 of this title, (D) in making any determination as to a State permit program submitted under section 1342(b) of this title, (E) in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, or 1316 of this title, and (F) in issuing or denying any permit under section 1342 of this title, may be had by any interested person in the Circuit Court of Appeals of the United States for the Federal judicial district in which such person resides or transacts such business upon application by such person. Any such application shall be made within ninety days from the date of such determination, approval, promulgation, issuance or denial, or after such date only if such application is based solely on grounds which arose after such ninetieth day. (2) Action of the Administrator with respect to which review could have been obtained under paragraph (1) of this subsection shall not be subject to judicial review in any civil or criminal proceeding for enforcement.
33 U.S.C. § 1369(b) (emphasis added).
. It is important to keep in mind that the NEPA obligations set forth in § 511(c) apply only to those situations in which the EPA is the direct permitting authority. At this point most NPDES permits are issued by the states, 39 of which are now authorized to implement the federal program. EPA Brief on Ripeness at 12 n. 9.
. Specifically, EPA points to section 105 of NEPA, which provides that "[t]he policies and goals set forth in this [Act] are supplementary to those set forth in existing authorizations of Federal agencies.” 42 U.S.C. § 4335. According to the agency, the '"action forcing’ requirements of section 102(2) of NEPA, in particular the requirement for an environmental impact statement (EIS), were intended to give effect to the 'goal expressed in section 102(1) that agencies interpret and administer their statutes in accordance with the policies of NEPA set forth in section 101(b).'” EPA Brief at 30-31. To do this in the CWA context, EPA argues, it must "evaluate all the environmental impacts and alternatives concerning the planning, design, construction and location of new sources” and be able to impose any condition necessary to account for any adverse environmental effects. EPA Brief at 32-33.
. The CWA also authorizes EPA to take certain other actions relating to the regulation of pollution discharges, such as promulgation of effluent limitations guidelines, 33 U.S.C. §§ 1311, 1316, and collection of pertinent information, id. § 1318. EPA does not contend that these other functions authorize the challenged regulations. But even considering the whole of EPA’s authority under the CWA, the imposition of non-discharge related conditions in new point source permits does not fall within its scope.
. 40 C.F.R. § 124.85(e) provides:
If a hearing is granted under this subpart for a new source subject to NEPA, the Presiding *196Officer may admit evidence relevant to any environmental impacts of the permitted facility if the evidence would be relevant to the Agency's obligation under § 122.29(c)(3). If the source holds a final EPA-issued RCRA, PSD, or UIC permit, or an ocean dumping permit under the Marine Protection, Research, and Sanctuaries Act (MPRSA), no such evidence shall be admitted nor shall cross-examination be allowed relating to: (1) Effects on air quality, (2) effects attributable to underground waste management practices, or (3) effects of ocean dumping subject to the MPRSA, which were considered or could have .been considered in the PSD, RCRA, UIC, or MPRSA permit issuance proceedings. However, the presiding officer may admit without cross-examination or any supporting witness relevant portions of the record of PSD, RCRA, UIC, or MPRSA permit issuance proceedings.
. The agency may, in fact, have pulled some of the teeth from its evidentiary restriction when, in its Brief at p. 61 n. 63, it asserted that
the bar on cross-examination is not absolute, as NRDC appears to believe. The rule provides that "the presiding officer may admit without cross-examination or any supporting witness relevant portions of the record [of the prior proceeding].” 40 C.F.R. § 124.85(e) (emphasis added). The permissive "may" gives the presiding officer limited discretion to allow cross-examination or a supporting witness if, for example, new evidence had become available or circumstances had changed since the prior hearing.
. 33 Ü.S.C. § 1251(b) (1982).
. Id. § 1342(b).
. Environmental Policy Division, Congressional Research Service, Library of Congress, A Legislative History of the Water Pollution Control Act Amendments of 1972 at 274, 860, 1366, 1482, 1489 (Comm. Print 1973) [hereinafter 1972 Legislative History]; District of Columbia v. Schramm, 203 U.S. App.D.C. 272, 278, 631 F.2d 854, 860 (1980). By 1985, 37 states had assumed the permit program. Brief for Respondent at 9.
. 33 U.S.C. § 1342(b) (1982).
. The Administrator is authorized "to prescribe such regulations as are necessary to carry out his functions under this chapter.” Id. § 1361(a). The Act set a deadline within which he had to "promulgate guidelines establishing the minimum procedural and other elements of any State program,” explicitly including enforcement provisions. Id. § 1314(i)(2).
. Id. § 1342(b); see Citizens for a Better Environment v. EPA, 596 F.2d 720, 722 (7th Cir.1979). The Administrator’s determinations on certification of state programs are subject to judicial review. 33 U.S.C. § 1369(b)(1)(D) (1982).
. 1972 Legislative History, supra note 13, at 356.
. Id. at 263, 452-53, 473, 577.
. We note that a statement heavily relied upon by CBE, Brief for Petitioner CBE at 12, 17, is taken somewhat out of context. The Senate Report states that ”[i]t is intended that the program ... will, even after incorporation of any or all States, be uniform in as many procedural requirements as possible,” 1972 Legislative History, supra note 13, at 1472. This is not, however, quite the smoking gun CBE would have us believe, for it refers simply to "the minimum requirements for the acquisition of information from owners and operators of point sources applying for permits," and more particularly to "application forms and the actual permit forms." Id.
. See, e.g., 1972 Legislative History, supra note 13, at 172, 309, 378; see also Weyerhaeuser Co. v. Costle, 191 U.S.App.D.C. 309, 340, 590 F.2d 1011, 1042 (1978); NRDC v. Train, 166 U.S.App.D.C. 312, 329-30, 510 F.2d 692, 709-10 (1975).
. 1972 Legislative History, supra note 13, at 262, 355, 359, 814.
. 33 U.S.C. § 1251(b) (1982) (quoted in text supra at note 11).
. Id. § 1342 (1982). In particular, the veto provision, id. § 1342(d)(2), was a carefully crafted compromise between the Senate and the House. See District of Columbia v. Schramm, supra note 13, 203 U.S.App.D.C. at 278, 631 F.2d at 860.
. 1972 Legislative History, supra note 13, at 262.
. See Consolidated Permit Regulations, 45 Fed. Reg. 33,290 (aim is to "ensure the benefits of public participation while intruding less into the States’ management of their judicial and administrative systems") [hereinafter 1980 Regulations]; id. at 33,383 (regulations "set minimum levels of fines and penalties which States must have the authority to recover in order to ensure *199effective State enforcement programs,” but "reduce[ ] the levels below those available to EPA based on the large volume of comments from States requesting such relief’).
. Chevron U.S.A. v. NRDC, 467 U.S. 837, 845, 104 S.Ct. 2778, 2783, 81 L.Ed.2d 694, 704 (1984) (quoting United States v. Shimer, 367 U.S. 374, 382-83, 81 S.Ct. 1554, 1560-61, 6 L.Ed.2d 908, 915 (1961)); see also Capital Cities Cable v. Crisp, 467 U.S. 691, 699-700, 104 S.Ct. 2694, 2700-01, 81 L.Ed.2d 580, 589 (1984).
. 33 U.S.C. § 1251(e) (1982). Section 1342(b)(3), (7) makes this requirement a prerequisite for state assumption of the permit program.
. Promotion of public participation in the process of issuing permits, in the form of notice and opportunity for hearing, is required on both federal and state levels. Id. § 1342(a), (b)(3). Public notice and opportunity to comment in rulemaking proceedings is, of course, mandated by the APA. 5 U.S.C. § 553 (1982).
. 1972 Legislative History, supra note 13, at 249; see id. at 819, 1430, 1490.
. As characterized by CBE, the participation requisite at the state level would allow citizens to "(1) initiate an enforcement action against polluters in state court; (2) intervene as of right in any enforcement action brought by the state; (3) initiate an action in state court against an enforcement agency for failure to perform a non-discretionary duty; and (4) recover litigation costs for participating in enforcement.” Brief for Petitioner CBE at 25. Compare 33 U.S.C. § 1365 (1982). CBE demands as well access to information and judicial review of state-issued permits. Brief for Petitioner CBE at 25.
. 1972 Legislative History, supra note 13, at 328, 1497; see Brief for Petitioner CBE at 20-21.
. 1972 Legislative History, supra note 13, at 1497; NRDC v. Train, supra note 20, 166 U.S.App.D.C. at 320-21, 323, 510 F.2d at 700-01, 703.
. 1972 Legislative History, supra note 13, at 217, 221; American Frozen Food Inst. v. Train, 176 U.S.App.D.C. 105, 124, 539 F.2d 107, 126 (1978).
. 1972 Legislative History, supra note 13, at 467; see also id. at 663, 673-74, 820-21, 1499. This anxiety was assuaged by limitations on standing, inability to recover damages, and potential assessment of costs against guilty parties. Id. at 217, 663, 821, 1499; see 33 U.S.C. § 1365(a), (d), (g) (1982).
. 33 U.S.C. § 1365(b)(1)(B) (1982). The EPA Administrator opposed even this provision. See 1972 Legislative History, supra note 13, at 856; see also Hudson River Sloop Clearwater v. Consolidated Rail Corp., 591 F.Supp. 345, 349-50 (N.D.N.Y.1984), aff'd in part and rev’d in part, 768 F.2d 57 (2d Cir.1985) (declining to find lack of diligent state prosecution under § 1365 on account of denial of intervention as of right in state proceeding because that section is limited by its terms to intervention in federal courts).
. 10 Envtl.L.Rep. (Envt.L.Inst.) 20,208 (D.D.C.1980), aff’d en banc, 209 U.S.App.D.C. 259, 653 F.2d 514, cert. denied, 454 U.S. 822, 102 S.Ct. 106, 70 L.Ed.2d 93 (1981).
. Id. at 20,211.
. Id.
. 33 U.S.C. § 1342(a)(3) (1982); see Brief for Petitioner CBE at 25-26.
. Citizens for a Better Environment v. EPA, supra note 16, 596 F.2d at 725 n. 8.
. 33 U.S.C. § 1342(b)(1)(A), (B) (1982).
. Id. § 1342(b)(3), (7).
. 1980 Regulations, supra note 25, 45 Fed.Reg. at 33,383.
. Chevron U.S.A. v. NRDC, supra note 26, 467 U.S. at 843, 104 S.Ct. at 2782, 81 L.Ed.2d at 703.
. While imposition of a citizen suit requirement might be within the Administrator’s discretion, we alone may not imply this additional enforcement remedy absent any statutory support or legislative history. Middlesex County Sewerage Auth. v. National Sea Clammers Ass’n, 453 U.S. 1, 17, 101 S.Ct. 2615, 2625-26, 69 L.Ed.2d 435, 448-49 (1981); District of Columbia v. Schramm, supra note 13, 203 U.S.App. D.C. at 281, 631 F.2d at 863.
. Brief for Petitioner CBE at 28.
. 40 C.F.R. § 123.27(d) (1987). EPA notes the availability of an additional remedy — citizen petitions to the agency to withdraw the state's certification to administer the permit program. See 1980 Regulations, supra note 25, 45 Fed. Reg. 33,383. We accord this privilege little weight in the present context; both the legislative history of the Act and the caselaw evolving indicate that this authority ought not to be invoked except in circumstances of egregious state flouting of the Act’s demands. 1972 Legislative History, supra note 13, at 452; Save the Bay v. EPA, 556 F.2d 1282, 1290 (5th Cir.1977).
. 1980 Regulations, supra note 25, 45 Fed.Reg. at 33,383.
. 33 U.S.C. § 1251(e) (1982).
. 1972 Legislative History, supra note 13, at 1490.
. Id. at 819.
. United States v. Ketchikan Pulp Co., 74 F.R.D. 104, 108 (D.Alaska 1977) (citing Friends of the Earth v. Carey, 535 F.2d 165, 172 (2d Cir.1976).
. We, of course, defer to the agency’s reasonable construction of its regulation. Udall v. Tallman, 380 U.S. 1, 16-17, 85 S.Ct. 792, 801, 13 L.Ed.2d 616, 625 (1965); Ashland Exploration, Inc. v. FERC, 203 U.S.App.D.C. 436, 444-45, 631 F.2d 1018, 1021-22 (1980), cert. denied, 450 U.S. 915, 101 S.Ct. 1538, 67 L.Ed.2d 340 (1981).
. 1980 Regulations, supra note 25, 45 Fed.Reg. 33,383; Brief for Respondent at 248-50; see Fed.R.Civ.P. 24(a).
. 33 U.S.C. § 1365(b)(1)(B) (1982); see also Hudson River Sloop Clearwater v. Consolidated Rail Corp., supra note 35, 591 F.Supp. at 349-350.
. See text supra at note 47; see also 28 C.F.R. § 50.7 (1986) (establishing policy of Department of Justice respecting consent judgments in actions to enjoin pollutant discharge); 1980 Regulations, supra note 25, 45 Fed.Reg. at 33,383.
. Citizens for a Better Environment v. EPA, supra note 16, 596 F.2d at 726.
. We ground our holding on EPA’s present construction that the regulations foreclose use of the second option by states lacking a mechanism for permissive intervention.
. 1980 Regulations, supra note 25, 45 Fed.Reg. 33,382.
. Water Quality Act of 1987, Pub.L.No. 100-4, § 312, 101 Stat. 7, 43-44 (1987) (codified at 33 U.S.C.A. § 1319(c)(1)(B) (1982)).
. Id. (codified at 33 U.S.C.A. § 1319(c)(2)(B)).
. Id. (codified at 33 U.S.C.A. § 1319(c)(3)(A)).
. Id. (codified at 33 U.S.C.A. § 1319(c)(4)).
. Id. at § 313, 101 Stat. at 45 (codified at 33 U.S.C.A. § 1319(d)).
. Id. § 1342(b)(7).
. Id. § 1314(i)(2)(C).
. 40 C.F.R. § 123.27(3) (1987).
. Brief for Petitioner NRDC at 91-92.
. See 1972 Legislative History, supra note 13, at 163, 802, 1482.
. Id. at 1482-1483.
. Id at 802, 1482.
. Brief for Petitioner NRDC at 92-94.
. 33 U.S.C. § 1370(1)(B) (1982).
. Brief for Petitioner NRDC at 96.
. Moreover, the notion that states may adopt standards more stringent than those imposed on the federal level, 1972 Legislative History, supra note 13, at 331, conflicts with NRDC’s proposal. If the Act’s penalty provisions were binding on the states, they would restrict at the upper as well as the lower level of enforcement. Ironically, if NRDC were correct, the penalties states are empowered to impose would implicitly be limited to the statutory maxima. That result would be at odds with the precept of state freedom to impose more rigorous requirements.
. Water Quality Act of 1987, supra note 60, § 313(b)(2), 101 Stat. at 45 (codified at 33 U.S.C. A. § 1319 note (1982)).
. Brief for Respondent EPA at 257-258.
. See text supra at notes 35-45.
. See text supra at notes 35-36.
. Water Quality Act of 1987, supra note 60, § 313(b)(2), 101 Stat. at 45 (codified at 33 U.S.C.A. § 1319 note (1982)).
. See Post-Argument Reply Brief for Petitioner NRDC at 2-4.
. National Pollutant Discharge Elimination System Revision of Existing Regulations, 43 Fed.Reg. 37,078, 37,108 (1978).
. EPA has indicated that the enforcement provisions generated more comments than any other aspect of the proposed state permit-program regulations. 1980 Regulations, supra note 25, 45 Fed.Reg. at 33,381.
. Id. at 33,382.
. Chevron U.S.A. v. NRDC, supra note 26, 467 U.S. at 865, 104 S.Ct. at 2792-93, 81 L.Ed.2d at 716.
. Cf. 43 Fed.Reg. 37,078, 37,087 (1978) (preamble to proposed veto regulations):
Technology-based requirements may be imposed in permits through the application of an EPA promulgated effluent guideline or on a case-by-case basis under section [1342(a)(1) ] of the Act. Case-by-case determinations must consider factors listed in section [1314(b) ] of the Act for development of EPA effluent guidelines.
. Industry quotes from a passage in the legislative history which purportedly indicates that formally issued "guidelines” would have to be the basis for rejection of individual permits. See 1972 Legislative History at 727. But the "guidelines" referred to in the quoted passage were the state program guidelines required by section 1314(i)(2), not the effluent limitations guidelines required by section 1314(b).
. NRDC also raised a challenge to the anti-backsliding regulations, asserting that 40 C.F.R. § 122.44(l)(2)(iii), which allows a source operating under a BPJ permit to backslide to a subsequently promulgated, less stringent "best conventional technology” (“BCT”) guideline, violates the CWA. The agency informs the court that it is in general agreement with NRDC’s position and that its decision not to rescind the offending provision was based on a false factual assumption. EPA Brief at 144. It asks the court to remand the issue to it for reconsideration. Industry raises no objection to this suggestion, Industry Intervenor-Respondents Brief at 21, and we see no reason to deny the agency’s request.
. While this argument is a potentially broad one, here we address only industry’s argument that the agency has no authority to restrict backsliding from NSPS limitations. In its post-argument brief, Industry raised, for the first time, challenges to the agency's antibacksliding rules as applied to effluent guidelines, best management practices, and other permit limits and conditions. As these challenges should have been and could have been, but were not raised in Industry’s original petition for review, we decline to address them.
. The NPDES rules (those adopted in 1979 and 1984) include both a general "anti-backsliding” rule governing when less stringent limits can be imposed in reissued permits, and permit modification rules, governing when permits can be modified during their terms to impose less stringent limits. The two sets of provisions are closely interrelated and the parties have referred to them collectively. We thus continue the practice of referring to the regulations collectively as the “anti-backsliding rule.”
. All references to the Clean Water Act in this section are to the unamended, pre-1987 Clean Water Act.
. The Committee’s reference to subsection (a)(2), rather than (a)(1), appears to be an error; subsection (a)(2) does not address EPA’s authority to impose interim permit limits, which is the subject of the Committee's discussion.
. Section 301(b)(2)(A), for example, requires dischargers to comply with BAT effluent limits for "categories and classes of point sources.” Id. § 1311(b)(2)(A). Similarly, section 304(b)(1)(A) instructs EPA when adopting or revising BPT effluent limits to identify the "best practicable control technology currently available for classes and categories of point sources.” Id. § 1314(b)(1)(A). Section 304(b)(2)(A) echoes this language, requiring EPA to identify BAT effluent reductions attainable for "classes and categories” of point sources. Id. § 1314(b)(2)(A). Finally, section 301(e) provides that "[ejffluent limitations established pursuant to this section ... shall be applied to all point sources of discharge of pollutants in accordance with the provisions of this [Act]." Id. § 1311(e).
. Specifically, Senator Muskie stated the following: “The Administrator is expected to be precise in his guidelines so as to assure that similar point sources with familiar characteristics, regardless of their location or the nature of the water into which the discharge is made, will meet similar effluent limitations." Senate Consideration of the Report of the Conference Committee (Oct. 4, 1972), Remarks of Senator Muskie, Exhibit 1, reprinted in 1972 Legislative History at 172.
.The opinions cited by Industry for support include, inter alia, E.I. du Pont de Nemours & Co. v. Train, 430 U.S. 112, 129, 97 S.Ct. 965, 975, 51 L.Ed.2d 204 (1977) ("the language of the statute supports the view that § 301 limitations ... are to be based primarily on classes and categories’’); United States Steel Corp. v. Train, 556 F.2d 822, 844 (7th Cir.1977) (citing du Pont for the proposition that § 301 effluent limitations and § 304 guidelines are "uniform national standards and are not to vary from plant to plant”); and American Frozen Food Inst. v. Train, 539 F.2d 107, 127 (D.C.Cir.1976) (express language of Act and its legislative history demonstrate that guidelines and effluent limitations were intended to serve as “controlling standards” under the Act).
. EPA observes that its regulation does provide for equitable considerations, permitting the agency to relieve BPJ permittees of any undue burden in two situations: first, where a permit holder can show that the BPJ limits are unattainable despite installation and proper operation of the necessary treatment equipment; and second, where compliance with permit limits would result in costs wholly disproportionate to those considered in the subsequently promulgated effluent limitations guidelines. 40 C.F.R. §§ 122.44(l)(2)(i), (iv), 122.62(a)(15) (1987).
. EPA proffers its own selection from the legislative history of the CWA to counter Industry's claim that uniformity was Congress’s paramount goal: “ ‘The sole purpose of the Act is the enhancement of environmental quality.’ ’’ EPA Brief at 152 (quoting House Consideration of the Report of the Conference Committee (Oct. 4, 1972), Remarks of Rep. Jones, reprinted in 1972 Legislative History at 232). Although we were frankly unable to locate the quoted passage in Representative Jones’ remarks, we did find a statement by Senator Muskie to the same effect: "The purpose of the Act is to establish a detailed regulatory mechanism for restoring and maintaining the ... integrity of the Nation's waters. The goal of the Act is to eliminate the discharge of pollutants into the Nation's waters by 1985.” Senate Consideration of the Report of the Conference Committee (Oct. 4, 1972), Remarks of Senator Muskie, Exhibit 1, reprinted in 1972 Legislative History at 181.
. EPA contends that the legislative history of the 1987 Water Quality Act specifically addresses the treatment of BPJ limits once applicable guidelines have been promulgated. Specifically, EPA points to the Senate Report, which states that the WQA’s express prohibition on backsliding was intended to "clarify the Clean Water Act’s prohibition of backsliding on effluent limitations.” H.R.Rep. No. 1004, 99th Cong., 2d Sess. 153 (1986); see also S.Rep.No. 50, 99th Cong., 1st Sess. 45 (1985) (Conference Committee Report’s description of Senate Bills: "The Senate Bill adds new subsections to section 303 and 402 which clarify the Clean Water Act’s prohibition of backsliding on effluent limitations.’’) Although they may indicate more current congressional approbation of EPA’s position, post-enactment statements are of highly limited relevance in determining the scope of EPA’s pre-1987 authority. As the Supreme Court has cautioned, "the views of a subsequent Congress form a hazardous basis for inferring the intent of an earlier one.” United States v. Philadelphia Natl Bank, 374 U.S. 321, 348-49, 83 S.Ct. 1715, 1733, 10 L.Ed.2d 915 (1963) (citation omitted); see also United Sav. Ass’n v. Timbers of Inwood Forest Assocs., Ltd., 484 U.S. 365, 108 S.Ct. 626, 634, 98 L.Ed.2d 740 (1988) (suggesting irrelevance of legislative history); United States v. Taylor, — U.S. -, -, 108 S.Ct. 2413, 2424-25 (1988) (Scalia, J., concurring).
. See, e.g., 33 U.S.C. § 1311(c) (economic hardship variances from BPT guidelines); id. § 1311(g) (water-quality based variances from BAT limitations guidelines for non-toxic, non-conventional pollutants).
. EPA will establish alternative effluent limitations for permittees who can demonstrate that they are "fundamentally different,” in terms of the relevant statutory factors, from plants considered in developing the guidelines. The Supreme Court upheld these "FDF variances" in Chemical Manufacturers Ass’n v. NRDC, 470 U.S. 116, 105 S.Ct. 1102, 84 L.Ed.2d 90 (1985).
.Industry cites the two major exceptions to EPA’s antibacksliding rule as proof that even EPA recognizes the inaccuracy of BPJ limits. See supra note 96 (backsliding from BPJ is permitted when the limits are unachievable or if *226circumstances have materially and substantially changed). In our view, however, the exceptions reinforce the rule: when BPJ fails to produce achievable limits or disproportionately burdens a permittee, EPA will allow backsliding to the general standard.
Industry also argues that the exceptions unfairly put the burden of proof on the discharger to show that it meets the exception, ie., that it should be allowed to meet the national standard. This, argues Industry, reverses the usual "presumption" that the national standard applies and one seeking a variance bears the burden of proof. We see no such reversal. The permittee always bears the burden of demonstrating that validly issued permit requirements should be revised or eliminated. 40 C.F.R. § 122.62 (1987).
. 40 C.F.R. § 122.41(n)(2) (1987).
. Id. § 122.41(n)(l).
. Water quality standards apply to particular bodies of water and are designed to protect the designated uses of a body of water. 33 U.S.C. § 1313(c) (1982). A set of water quality standards serves as the basis for determining the water quality-based permit limitations applicable to each plant discharging its effluent into that body of water. Scott v. City of Hammond, 530 F.Supp. 288, 289 (N.D.Ill.1981).
. 564 F.2d 1253 (9th Cir.1977).
. Id. at 1266.
. Id
. See id
. Id at 1272.
. Id. at 1272-73 (footnote omitted).
. Both the technology-focused provisions of § 301(b)(1)(A) and the water quality-focused language of § 301(b)(1)(C) require satisfaction of "limitations.” Industry petitioners reason that because both types of limitations are to be met in the same manner, both types should be subject to the same exceptions. EPA counters with the argument that the water quality provision of § 301(b)(1)(C) calls for observance of "any more stringent limitation necessary to meet ... water quality standards," 33 U.S.C. § 1311(b)(1)(A) & (C) (1982) (emphasis added), implying that water quality-standards are stricter and less compromising.
. 33 U.S.C. § 1311(b)(1)(A), (C) (1982).
. House Debate on H.R. 11896 (Mar. 27, 1972), Legislative History, supra note 13, at 379 (statement of Rep. Clausen). The House version of § 301(b)(1)(C) is virtually identical to the enacted version. Compare H.R. 11896, 92d Cong., 2d Sess. § 301(b)(1)(C) (1972), reprinted in Legislative History, supra note 13, at 963, with 33 U.S.C. § 1311(b)(1)(C) (1982).
. House Debate on H.R. 11896 (Mar. 27, 1972), Legislative History, supra note 13, at 353 (statement of Rep. Blatnik).
. See Marathon Oil Co. v. EPA, supra note 105, 564 F.2d at 1273.
. Industry petitioners themselves acknowledge the fundamental division between technology-based and water quality-based approaches to pollution control embodied in the Act. Brief for Joint Industry Petitioners at 117-126, 166-167.
. Chevron U.S.A. v. NRDC, supra note 26, 467 U.S. at 843, 104 S.Ct. at 2782, 81 L.Ed.2d at 703.
. 47 Fed.Reg. 52,072, 52,079 (1982).
. Id. at 52,089.
. EPA’s brief maintains that the decision to disallow an upset defense was based purely on "legal grounds.” Brief for Respondent at 207. An examination of the final regulations, however, reveals that EPA based its decision solely on practical grounds. 49 Fed.Reg. 37,998, 38,-038 (1984). Nor do EPA’s arguments in the brief point to any statutory obstacles standing in the way of such a defense. Brief for Respondent at 201-206.
. 49 Fed.Reg. 37,998, 38,038 (1984).
. E.g., Motor Vehicles Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443, 458 (1983).
. Id. (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168, 83 S.Ct. 239, 246, 9 L.Ed.2d 207, 216 (1962)).
. 463 U.S. at 43, 103 S.Ct. at 2866, 77 L.Ed.2d at 458.
. Id.
. Clark-Cowlitz Joint Operating Agency v. FERC, 264 U.S.App.D.C. 58, 76, 826 F.2d 1074, 1092 (1987) (en banc).
. 49 Fed.Reg. 37,988, 38,038 (1984).
. Id.
. Id.
. E.g., Burlington Truck Lines v. United States, supra note 123, 371 U.S. at 168-169, 83 S.Ct. at 246, 9 L.Ed.2d at 216; SEC v. Chenery Corp., 332 U.S. 194, 196, 67 S.Ct. 1575, 1577, 91 L.Ed. 1995, 1999 (1947).
. We need not dwell on EPA’s promise to exercise prosecutorial discretion to protect plants from unfair punishment in the absence of an upset defense. That would be of little or no value to plants that EPA chose to prosecute, and that could have availed themselves of the defense were it available. Moreover, a decision not to prosecute would not preclude a private enforcement action under 33 U.S.C. § 1365(a)(1) (1982).
. See, e.g., 40 C.F.R. § 130.4 (1987).
. 33 U.S.C. § 1342(a)(3), (b)(1)(B) (1982); 40 C.F.R. § 122.46(a) (1987).
. Brief for Petitioner NRDC at 44 n. 19 (purported agency estimate).
. Brief for Respondent at 276. This was the deadline for implementation of best available technology and best conventional pollution control technology, the second stage of effluent limitations under the Act. See 33 U.S.C. § 1311(b) (1982).
. 33 U.S.C. § 1311(a) (1982).
. Id. §§ 1319, 1365.
. 92 Cong.Rec. 5654 (1972) (statement of Rep. Walter).
. 5 U.S.C. § 558(c) (1982).
. The regulation provides:
When EPA is the permit-issuing authority, the conditions of an expired permit continue in force under 5 U.S.C. 588(c) until the effective date of a new permit ... if:
(1) The permittee has submitted a timely application ... which is a complete ... application for a new permit; and
(2) The Regional Administrator, through no fault of the permittee does not issue a new permit with an effective date ... on or before the expiration date of the previous permit (for example, when the issuance is impracticable due to time or resource constraints).
40 C.F.R. § 122.6(a) (1987) (cross-references omitted).
. As a preliminary matter, we are aware of the statement in Costle v. Pacific Legal Found., 445 U.S. 198, 210-11 n. 10, 100 S.Ct. 1095, 1103 n. 10, 63 L.Ed.2d 329, 341 n. 10 (1980), that because EPA had not acted on the NPDES permit application there at issue, the permit was continued by operation of law — Section 558(c) —after its expiration. That continuance apparently was unchallenged. Reply Brief for Petitioner NRDC at 22 n. 13; see also Student Pub. Interest Research Group v. Fritzscke, Dodge & Olcott, Inc., 579 F.Supp. 1528, 1538 (D.N.J.1984), aff'd, 759 F.2d 1131 (3d Cir.1985) (citing 40 C.F.R. § 122.6 continuance with apparent approval).
. The Act established July 1, 1984, as the deadline, 33 U.S.C. § 1311 (1982), but compliance with any valid permit shields the discharger from enforcement action. Id. § 1342(k).
. Brief for Petitioner NRDC at 41-46.
. 353 U.S. 436, 77 S.Ct. 999, 1 L.Ed.2d 963 (1957).
. See id. at 440-441, 444-445, 77 S.Ct. at 1002, 1004-1005, 1 L.Ed.2d at 967-968, 969-970 (dissenting opinion) (contending that § 558(c) ought not apply "since the permanent license sought is not of the same type and class as the old license"; prerequisites to the permanent license included notice and hearing, a finding of public convenience and necessity, and judicial review if in order, while the temporary permit was obtainable ex parte, based on "need[ ] on an emergency basis to meet specific transportation problems.”).
. Id. at 439-40, 77 S.Ct. at 1002-03, 1 L.Ed.2d at 966-67.
. County of Sullivan v. CAB, 436 F.2d 1096, 1099 (2d Cir.1971) (quoting Pan Atlantic S.S. Corp. v. Atlantic Coast Line R.R., supra note 144, 353 U.S. at 444-45, 77 S.Ct. at 1005, 1 L.Ed.2d at 967-70) (dissenting opinion).
. Committee for Open Media v. FCC, 177 U.S. App.D.C. 376, 382-83, 543 F.2d 861, 867-68 (1976) (footnote omitted).
. This does not mean that EPA is at liberty to extend permits indefinitely, as NRDC contends. Brief for Petitioner NRDC at 49. By virtue of 5 U.S.C. § 706(1) (1982), which authorizes a reviewing court to “compel agency action unlawfully withheld or unreasonably delayed," a party may seek relief from prolonged agency inaction with respect to pending permit renewal applications. Costle v. Pacific Legal Found., supra note 141, 445 U.S. at 220 n. 14, 100 S.Ct. at 1108 n. 14, 63 L.Ed.2d at 347 n. 14.
. Brief for Petitioner NRDC at 54-58; see 5 U.S.C. § 558(c) (1982).
. Brief for Respondent at 277.
. 40 C.F.R. § 122.6(a) (1986) (emphasis added).
. While the regulation is relatively clear in this regard, we would in any event defer to EPA’s reasonable interpretation of its own rule. Udall v. Tollman, supra note 53, 380 U.S. at 16-17, 85 S.Ct. at 801, 13 L.Ed.2d at 625; Ashland Exploration, Inc. v. FERC, supra note 53, 203 U.S.App.D.C. at 439-40, 631 F.2d at 1021-22.
. Brief for Petitioner NRDC at 55. Even oil and gas drilling, the sole example NRDC cites, involves ongoing planning and drilling, albeit at different sites. Brief for Respondent at 277-278.
. Brief for Petitioner NRDC at 55.
. The bare fact that a source operator subject to the regulation is applying for renewal of a five-year operating permit is itself some evidence of a continuing activity.
. Brief for Respondent at 277-278 n. 10.
. Id. at 278-279. The regulation cross-references 40 C.F.R. § 122.21(e) (1987), which defines a "complete” application.
. Kitlutsisti v. ARCO Alaska, Inc., 592 F.Supp. 832, 843 (D.Alaska 1984). The court also concluded that § 558(c) did not extend to general permits such as that in issue.
Other statements, however, tending superficially in the same direction may easily be explained. Lac Courte Oreilles Band v. RPC, 166 U.S.App.D.C. 245, 510 F.2d 198 (1975), indicated that § 558(c) might be inapplicable in the situation there presented because if the .agency lacked power to renew the permit under any circumstances, the application was by definition not “sufficient." Id. at 251 n. 21, 510 F.2d at 204 n. 21. In Bankers Life & Casualty Co. v. Callaway, 530 F.2d 625, 633-34 (5th Cir.1976), cert. denied, 429 U.S. 1073, 97 S.Ct. 811, 50 L.Ed.2d 791 (1977), the parties argued that “section 558(c) does not apply unless the application is sufficient," and that “[a]t the time the renewal application was filed, it was insufficient because it lacked the required local consents," but that "[a]s time passed, it became more incomplete with the addition of new laws.” The court found § 558(c) inapplicable because nonrenewal resulted from a conscious decision of the agency not to renew, rather than from the temporal exigencies at which the section was aimed. Thus, both Lac Courte Oreilles Band and Bankers Life may readily be distinguished as turning on the agency’s inability or disinclination to renew a permit, as opposed to the present situation, where the agency could issue a new permit upon an application meeting governing standards but for delays caused by resource constraints.
. See text supra at notes 148-149.
. See, e.g., Pan-Atlantic S.S. Corp. v. Atlantic Coast Line R.R., supra note 144 (temporary permit continued despite different standards for permanent authority).
. 5 U.S.C. § 558(c) (1982) (emphasis added).
. This interpretation also parallels that of 33 U.S.C. § 1342(k) (1982), which immunized polluters with pending permit applications “containing the necessary information for processing of the application" from prosecution until December 31, 1974. NRDC v. Train, supra note 20, 166 U.S.App.D.C. at 316, 510 F.2d at 696.