Full opinion text
OPINION OF THE COURT JAMES HUNTER, III, Circuit Judge: This is a petition brought by the American Iron and Steel Institute and several individual steel companies to review regulations promulgated by the Administrator of the Environmental Protection Agency on June 28, 1974. In these regulations, entitled “Effluent Guidelines and Standards, Iron and Steel Manufacturing Point Source Category,” the Administrator established nationwide single number effluent limitations for point sources in the iron and steel industry engaged in “primary” (or basic manufacturing) operations. Contending that the Administrator’s regulations do not conform to the requirements of the Federal Water Pollution Control Act Amendments of 1972 (hereinafter, the “Act”), the petitioners seek judicial review of the Administrator’s actions under section 509(b)(1) of the Act. Two other steel companies — Youngstown Sheet and Tube Company, and CF&I Steel Corporation— filed similar petitions in the Sixth and Tenth Circuits, respectively. By order of each Circuit, the cases were transferred, to this Court, and they have been consolidated with the petitions filed here. In addition, the Natural Resources Defense Council, Inc., has filed a brief as Amicus Curiae. I Petitioners’ first, and most basic, challenge is to the Administrator’s very power to promulgate nationwide single number effluent limitations for existing point sources. Petitioners contend that the limitations which are to be binding on them can only be established by the permit-granting authorities (principally, the States), which are to follow guidelines promulgated by the Administrator. The Administrator contends that he is not merely empowered to promulgate guidelines, but may establish limitations which are binding throughout the country and which must be incorporated into any permit issued to any individual point source. The answer to this dispute, which goes to the very heart of the administration of the Act, depends upon our resolution of the interrelationship of three key sections of the Act — 301, 304 and 402. Petitioners rely heavily on the fact that there is no section in the Act which explicitly authorizes the Administrator to establish single number effluent limitations for existing point sources. They contend that the lack of such explicit authorization cannot have been an oversight, since the Act expressly authorizes the Administrator to promulgate other types of regulations. For example, the Administrator has the explicit authority to set standards for new point sources under section 306(b)(1)(B) and for toxic discharges under section 307(a)(2), and to establish pretreatment standards under section 307(b) and water quality standards under section 303(b). Many of these sections not only explicitly authorize the Administrator to promulgate regulations establishing such standards, but also specify in some detail the times and procedures to be followed. In contrast to these sections, the only section explicitly authorizing the Administrator to establish any regulations pertaining to effluent standards for existing point sources — section 304(b) — merely authorizes the promulgation of “guidelines” rather than precise standards or single number limitations. While section 301(b) refers to “effluent limitations” for existing point sources, that section does not explicitly authorize the Administrator (or anyone else) to promulgate regulations establishing such limitations. Rather, using the passive voice, that section merely states that effluent limitations for such point sources “shall be achieved” by July 1, 1977 through the application of the “best practicable control technology currently available” (hereinafter, “BPCTCA”), and by July 1, 1983 through the application of the “best available technology economically achievable” (hereinafter, “BATEA”). Under petitioners’ construction of the Act, these effluent limitations are to be “achieved” through the permit process. They contend that the permit issuing authorities, under section 402, are to determine the effluent limitations to be achieved by applying, to individual point sources, the factors enumerated in the guidelines previously promulgated by the Administrator under section 304(b). The two consolidated cases present graphic examples of the differing consequences of the two interpretations. Both Youngstown and CF&I contend that they have local problems which can be fully appreciated only by a local permit-issuing authority. Youngstown claims that its plants in the Mahoning Valley provide one third of the direct employment in that area and that they are indirectly responsible for a significant percentage of the remaining jobs. It also claims that its plants are very old (several were built before World War I), that many will be forced to close if the limitations promulgated by the Administrator are enforced, and that because of the heavy concentration of steel plants along the shores, the river is unavailable for recreational uses anyway. Youngstown claims that a local authority, in appreciation of these factors, might have required somewhat less stringent controls. CF&I points to a different problem. It contends that the installation of anti-pollution devices in its Colorado plants would cause a significant net loss of water through evaporation, which would have serious consequences in a state where water is a scarce and valuable resource. Contending that only a local authority would fully appreciate the impact of the anti-pollution devices on scarce water resources, CF&I argues that its case is another illustration of the necessary for flexibility at the local level. We acknowledge that these arguments are not without force, but we believe that the Administrator does have the authority to promulgate effluent limitations under section 301. While we admit that Congress did not express its intent on this point with particular clarity, we conclude, after examining the entire statutory scheme and the legislative history, that the Administrator’s power to promulgate effluent limitations under section 301 can be inferred. We thus respectfully disagree with the contrary conclusion reached by the Eighth Circuit in CPC International, Inc. v. Train, 515 F.2d 1032 (8th Cir. 1975). With respect to the peculiarly local problems of some point sources, as illustrated by Youngstown and CF&I, we believe that our holding that the Administrator does have power under section 301 does not preclude some flexibility at the local level. This point will be addressed in detail in part II of this opinion. Perhaps the strongest indication in the Act that the Administrator has the power under section 301 to promulgate effluent limitations can be found in section 509(b)(1). This section provides for judicial review “of the Administrator’s action . . . (E) in approving or promulgating any effluent limitation or other limitation under section SOI . and (F) in issuing or denying any permit under section 402.” Not only does this section explicitly refer to the Administrator’s action in promulgating a section 301 effluent limitation, but its separate references, in subsections (E) and (F), to section 301 effluent limitations and section 402 permits indicates that the limitations and permits have independent significance. The Eighth Circuit in CPC International discounted the importance of section 509(b)(1)(E) by contending that the section 301 limitation referred to must be one promulgated under section 301(c), which empowers the Administrator to “modify the requirements of subsection (b)(2)(A) of this section” with respect to an individual plant upon a stringent showing of hardship and good faith efforts at compliance. We cannot accept the Eighth Circuit’s reasoning on this point. In the first place, section 301(c) does pot authorize the Administrator to promulgate any effluent limitations. Rather, it merely authorizes him to relax the requirements of section 301(b) with respect to 1983 “BATEA” standards, and thus it cannot be read to be within the scope of section 509(b)(1)(E). Furthermore, the legislative history shows quite clearly that section 301(c), which was only added during the House-Senate Conference, did not even exist at the time section 509(b)(1)(E) was originally drafted. Further support for the Administrator’s power can be found in section 505. which permits citizen suits to enforce compliance with the Act. Section 505(a)(1) permits such suits against any person “who is alleged to be in violation of (A) an effluent standard or limitation under this Act. . . . ” Section 505(f) defines “effluent standard or limitation,” for purposes of section 505(a), as “. . . (2) an effluent limitation or other limitation under section 301 . ; or (6) a permit or condition thereof issued under section 402 of this Act . . . .” If petitioners’ view were correct — that effluent limitations required under section 301 can only be established through the permit process under section 402 — then subsections (2) and (6) of section 505(f) would ^appear redundant. The fact that Congress felt it necessary to include separate references to section 301 limitations and to section 402 permits indicates that they have independent significance and that a person could be in violation of section 301(b) independently of section 402. Petitioners attempt to explain away this redundancy by arguing that the separate references merely indicate that “the legislative draftsmen appreciated that a § 402 permit condition is not always an effluent limitation — it might be a monitoring or reporting requirement.” This contention fails to eliminate the redundancy, and it fails to explain why Congress did not consider it sufficient to include only a reference to a violation of a section 402 permit. While a permit may contain conditions other than section 301 limitations, petitioners’ position is that a section 301 limitation can only be established through the permit process. Thus, given the reference to violations of section 402 permits, any reference to violations of section 301 limitations would still be superfluous under petitioners’ view. The Eighth Circuit in CPC International attempted to explain away the redundancy on a different ground. It viewed the independent reference to section 301 as being necessary to encompass section 301(f), which prohibits the discharge of radiological, chemical, or biological warfare agents and high-level radioactive wastes into navigable waters. However, section 301(f) is not an “effluent limitation,” but rather a flat prohibition on all such discharges. Section 301(b) is the part of section 301 which requires compliance with effluent limitations, and section 301(f) would seem to be a type of “other limitation” referred to in section 505(f)(2). Section 401(a)(1) also lends support to the Administration’s position. This section requires an applicant for a permit to provide the permit-issuing agency with “a certification from the State that any such discharge will comply with the applicable provisions of sections 301, 302, 306, and 307 of this Act.” The section then proceeds to state: “In the case of any such activity for which there is not an applicable effluent limitation or other limitation under sections 301(b) and 302, and there is not an applicable standard under sections 306 and 307, the States shall so certify . . . .” This section, by requiring the State to certify the absence of a section 301 effluent limitation when someone has applied for a permit, seems to presuppose that section 301 limitations have an existence independent of the section 402 permits, and that they are generally expected to have been promulgated prior to the issuance of such permits. There are a number of other sections in the Act which refer to the establishment of effluent limitations “under” or “pursuant to” section 301. While it is true that none of these sections specify how these limitations are to be established or who is expected to establish them, the repeated references to such limitations, when coupled with the complete absence of any qualification that such limitations are to be established through the permit process, is further support for the position that Congress intended the section 301(b) limitations to have an independent existence. If they are to have such an independent existence, it is reasonable for the Administrator to conclude that the promulgation of regulations establishing such limitations is within his inherent rule-making power. Those sections containing references to effluent limitations include section 301(e) [“Effluent limitations established pursuant to this section”]; section 302(c) [“The establishment of effluent limitations under this section shall not operate to delay the application of any effluent limitation established under section 301 or section 306 of this Act”]; section 316(b) [“any standard established pursuant to section 301”]; and section 316(c) [“effluent limitations established under section 301”]. Finally, we believe that a joint reading of sections 304 and' 301 lends support for the Administrator’s position and undercuts that of the petitioners. Sections 304(b)(1)(A) and 304(b)(2)(A) require that the Administrator’s regulations, which are to be applicable by the years 1977 and 1983, respectively, “identify the degree of effluent reduction attainable” through application of the best technology practicable or available. This meshes in very neatly with section 301(b), which requires the “achievement” of effluent limitations by those dates applying the same respective levels of technology. Reading the two sections together, it would seem inconsistent to require, on the one hand, both the achievement of effluent limitations applying certain levels of technology [section 301(b)] and the promulgation of regulations by the Administrator which “identify . . . the degree of effluent reduction attainable” through those levels of technology [section 304(b)], while at the same time allowing permit grantors to determine for themselves, bound only by section 304(b) guidelines, what levels of effluent limitation are to be achieved. The Eighth Circuit in CPC International relied in part on the permit provisions of the Act, especially section 402(d)(2), which provides: “No permit shall issue ... if the Administrator within ninety days of the date of transmittal of the proposed permit by the State objects in writing to the issuance of such permits as being outside the guidelines and requirements of this Act.” The Eighth Circuit, emphasizing the phrase “as being outside the guidelines,” concluded that “[i]t is hard to imagine a clearer indication that the permit-issuing authority is to follow the guidelines promulgated under § 304(b), and is not to refer to independent regulations under § 301.” We respectfully disagree, and believe that the Eighth Circuit read too much into section 402(d)(2). Even if the word “guidelines” is a specific reference to the section 304(b) guidelines, the Eighth Circuit’s emphasis on that word ignores the fact that the permit grantors must also comply with the “requirements” of the Act. In section 402 (b)(1)(A), the Act requires that any permits issued “apply, and insure compliance with, any applicable requirements of sections 301, 302, 306, 307 and 403.” Section 304, unlike section 301, is never explicitly mentioned in section 402. We thus read the whole of section 402 as requiring compliance both with section 301 limitations and with any guidelines promulgated by the Administrator. We also believe that a close reading of the legislative history supports the Administrator’s position. The Senate Report accompanying the bill stated: “It is the Committee’s intention that pursuant to subsection 301(b)(1)(A), and Section 304(b) the Administrator will interpret the term ‘best practicable’ when applied to various categories of industries as a basis for specifying clear and precise effluent limitations to be implemented by January 1, 1976.” Virtually identical language was used by Senator Muskie, the principal author of the bill, in explaining the Report of the House-Senate Conference Committee. Senator Bentsen, a member of the Senate Public Works Committee which drafted the bill, gave perhaps the clearest indication that Congress contemplated that the Administrator would establish section 301 limitations: “In phase I, for point sources of pollutants, effluent limits shall be established not later than January 1, 1977 [now July 1, 1977], which comply with specifically defined levels of effluent controls and treatment. As defined in section 301(b)(1) of the bill, and as elaborated in the regulations which we anticipate the Administrator shall issue pursuant to section 301 and section 304, these 1976 [now 1977] goals shall be at least . . . the ‘best practicable control technology currently available’ for [industrial] point sources. ” The petitioners, as did the Eighth Circuit in CPC International, also rely on portions of the legislative history as support for their interpretation, but we believe that a close reading of the parts on which they rely does not contradict our analysis. Many of the statements indicating the need for flexibility, precise federal guidelines and a meaningful state role do not prove that the Administrator does not have the power to establish limitations under section 301. As we shall explain in greater detail in part II of this opinion, they merely indicate that the Administrator was empowered to promulgate guidelines in addition to limitations. Furthermore, some of the legislative history relied on by petitioners is ambiguous. One example is the following exchange between Senator Mathias and Senator Muskie: “Mr. Mathias. Does section 301(b) (2)(A) on page 76 comtemplate that a State, or the Administrator if appropriate, might be able to set the 1981 [now 1983] effluent limitations almost on an individual point source by point source basis? “Mr. Muskie. Section 301(b)(2)(A) as well as section 301(b)(1) anticipate individual application on point sources through the procedures under the permit program established under section 402.” Unlike the petitioners, we do not see Senator Muskie’s answer as support for the view that section 301 effluent limitations can only be established through the permit process. Rather, it is fully consistent with the Administrator’s position that previously promulgated nationwide effluent limitations are to be applied to individual plants through the permit process. Given the question asked, his answer must be taken primarily as a denial that the Administrator was to set limitations on a plant-by-plant basis, not that he lacked the authority to establish nationwide limitations. He merely indicated that those limitations were to be applied by local authorities on a local basis. The strongest statement in support of the view that effluent limitations were to be established, as well as applied, through the permit process appears in a letter from EPA Administrator William Ruckelshaus to Chairman Blatnik of the House Public Works Committee: “Effluent limitations required by Section 301 would be established and applied to all point sources of discharges covered by the Act by means of permits issued under Title IV.” We believe that this statement should not be given great weight. This Act was predominantly a congressional product, with the Administration playing a relatively insignificant rule in its drafting, at least with respect to those parts of the Act at issue here. Consequently, we believe that the comments by the EPA Administrator reflecting his interpretation of then pending legislation should be given less weight in determining the meaning of the Act than statements by Senators such as Muskie and Bentsen who had a far greater responsibility for the Act’s drafting. The Eighth Circuit in CPC International also relied heavily on the fact that several Representatives, principally Abzug and Rangel, were concerned about the lack of a federal veto power over the permit grantors in the early House version of the bill. The Eighth Circuit concluded that “the creation of the veto power would make no sense if the EPA was already empowered to promulgate regulations under § 301.” We disagree. We believe that a veto power could have been considered just as necessary to ensure compliance by the permit grantors with section 301 limitations as with section 304 guidelines. As a final point, we note that the Supreme Court has consistently stated that, where an Act of Congress is fairly susceptible of differing constructions, the interpretation made of it by the agency charged with its administration should be given considerable deference. Udall v. Tallman, 380 U.S. 1, 16-18, 85 S.Ct. 792, 13 L.Ed.2d 616 (1965); Power Reaction Co. v. Electricians, 367 U.S. 396, 408, 81 S.Ct. 1529, 6 L.Ed.2d 924 (1961); McLaren v. Fleischer, 256 U.S. 477, 480-81, 41 S.Ct. 577, 65 L.Ed. 1052 (1921). This position was most recently reiterated in Train v. NRDC, 421 U.S. 60, 87, 95 S.Ct. 1470, 43 L.Ed.2d 731 (1975), where the Supreme Court, in upholding the Administrator’s interpretation of section 110 of the Clean Air Act, stated: “We therefore conclude that the Agency’s interpretation of §§ 110(a)(3) and 110(f) was ‘correct’ to the extent that it can be said that any particular interpretation of a complex statute such as this is the ‘correct’ one. Given this conclusion, as well as the facts that the Agency is charged with administration of the Act, and that there has undoubtedly been reliance upon its interpretation by the States and other parties affected by the Act, we have no doubt whatever that its construction was sufficiently reasonable to preclude the Court of Appeals from substituting its judgment for that of the Agency.” Similarly, the Supreme Court also stated: “Without going so far as to hold that the Agency’s construction of the Act was the only one it permissibly could have adopted, we conclude that it was at the very least sufficiently reasonable that it should have been acóepted by the reviewing courts.” For the reasons expressed earlier, we believe that the Administrator’s interpretation of this very complex Act — at least with respect to his power to promulgate effluent limitations under section 301 — is the more reasonable one, and should be given appropriate deference. II Having concluded that the Administrator does have the power to promulgate effluent limitations under section 301, we are now faced with the question of interpreting the nature of the Administrator’s powers and duties under section 304 and of reconciling our construction of section 304 with that of section 301. We begin by observing that the Administrator not only has the power, but the explicit obligation to promulgate “guidelines” under section 304. In these guidelines, the Administrator is specifically required (1) to “identify the degree of effluent reduction attainable through the application” of the relevant standard of technology [sections 301(b)(1)(A) and 301(b)(2)(A)], and (2) to “specify factors to be taken in to account in determining the control measures and practices to be applicable to point sources . . . within such categories or classes ” [section 304(b) (1) (B)] (emphasis added). This second requirement clearly contemplates that the guidelines promulgated by the Administrator on the basis of broad categories or classes of industries are to provide guidance to those authorities (presumably, the permit grantors) which determine the precise degree of effluent control required of any individual point source. The only other interpretation possible would be that the Administrator is required to promulgate guidelines for broad categories or classes which are to guide himself in setting precise limitations for specific point sources — clearly an illogical interpretation, and, as noted below, one which is contrary to the clearly expressed congressional intent that the Administrator is not to consider the particular circumstances of individual point sources. Sections 304(b)(1)(B) and 304(b) (2) (B) also specify a number of factors to be taken into account in determining the control measures required for compliance with the respective 1977 and 1983 technological standards. These factors include cost, the age of the equipment and facilities involved, engineering aspects of application of various types of control techniques, and non-water quality environmental impact. The first question to address is who is to consider these factors — the Administrator in promulgating the guidelines, or the permit grantors in applying the guidelines to individual point sources? The statute is somewhat ambiguous on this point. By listing the various “factors” in sections 304(b)(1)(B) and 304(b)(2)(B) immediately after stating that the guidelines are to “specify factors to be taken into account,” Congress would appear to have intended that these were factors to be considered by those authorities applying the guidelines to individual point sources within the categories outlined by the Administrator. On the other hand, the statute states that these are factors “relating to the assessment” of the relevant technology, and sections 304(b)(1)(A) and 304(b)(2)(A) state that the Administrator is required to “identify . . . the degree of effluent reduction attainable through the application” of the relevant level of technology. This suggests that the factors are ones to be considered by the Administrator in “assessing” the proper level of technology. The legislative history, however, indicates quite clearly that a full consideration of these factors at the permit-issuing stage was not intended. In Senator Muskie’s statement in support of the bill as approved by the House-Senate Conference Committee, he explained: “The Conferees intend that the factors described in Section 304(b) be considered only within classes or categories of point sources and that such factors not be considered at the time of the application of an effluent limitation to an individual point source within such category or class.” Similarly, Representative Dingell, in discussing one of the factors (“cost”) enumerated in section 304(b), stated: “The conference report emphasizes on page 121 a very important point. The report states: ‘The conferees intend that the Administrator or the State, as the case may be, will make the determination of the economic impact of an effluent limitation on the basis of classes and categories of point sources, as distinguished from a plant by plant determination.’ “Thus, a plant-by-plant determination of the economic impact of an effluent limitation is neither expected, nor desired, and, in fact, it should be avoided.” Nevertheless, it does appear that Congress contemplated that some degree of consideration of the enumerated factors was to be made by the permit grantors on a plant-by-plant basis. The Senate Report, after referring to the section 304(b) factors, stated: “In applying effluent limitations to any individual plant, the factors cited above should be applied to that specific plant.” At first glance, this statement seems hard to reconcile with the previously quoted statements indicating that the factors were not to be considered on a plant-by-plant basis. We note, however, that the Senate Report just quoted said that the factors were to be “applied,” not that they were to be “considered.” We therefore believe that the statements quoted can be reconciled if we determine that any “consideration” to be given the specified factors by the permit grantors should be considerably less than de novo. Thus, it seems that the Administrator is to conduct the primary consideration of the enumerated factors for classes and categories and is to specify to the permit grantors how some variation in the standards should be made in light of those factors. In other words, the permit grantors are to have a limited and carefully circumscribed discretion to take into account factors as specified by the Administrator. To hold that they have no discretion does not make sense in light of the clear command in section 304(b)(1)(B) that the “guidelines” promulgated by the Administrator “specify factors to be taken into account in determining the control measures and practices to be applicable to point sources within such categories and classes.” Nor would it make any sense in light of the very fact that the permit grantors are to be guided by “guidelines.” Finally, it would be inconsistent with Congress’ concern that the guidelines provide precise guidance. As Senator Muskie stated: “Except as provided in Section 301(c) of the Act, the intent is that effluent limitations applicable to individual point sources be as uniform as possible. The Administrator is expected to be precise in his guidelines so as to assure that similar characteristics, regardless of their location or the nature of the water into which the discharge is made, will meet similar effluent limitations.” This indicates to us that, while uniformity was clearly a major congressional concern, some amount of local variation, carefully circumscribed by precise guidelines, was contemplated. In short, uniformity was to be achieved by effluent standards within a given category which were similar, rather than identical or unitary. This conclusion is further buttressed by several indications in the legislative history that the guidelines were intended to provide permissible “ranges” of effluent limitations. With respect to the 1977 “BPCTCA” standards the Senate Report, for example, stated: “The Administrator should establish the range of best practicable levels based upon the average of the best existing performance by plants of various sizes, ages, and unit processes within each industrial category.” With respect to the 1983 “BATEA” standards, the same Report continued: “. . . [T]he Committee intends that effluent limitations be based upon application of best available technology as defined by the Administrator. In making the determination of ‘best available’ the Committee expects the Administrator to apply the same principles involved in making the determination of best practicable as outlined above except that rather than the range of levels established in reference to the average of the best performers in an industrial category the range should at a minimum be referenced to the best performer in any industrial category.” Similarly, Senator Muskie, in summarizing the bill as reported out of the House-Senate Conference, stated: “The Administrator should establish the range of best practicable levels based upon the average of the best existing performance by plants of various sizes, ages, and unit processes within each industrial category.” Having concluded that the local permit-issuing authorities have a carefully circumscribed degree of discretion, as evidenced by Congress’ concern that the Administrator’s guidelines “specify factors to be taken into account” and that they provide “ranges,” we are then faced with the problem of reconciling this limited amount of discretion with out previous conclusion that the Administrator has the power to promulgate effluent limitations under section 301. The answer, in our view, lies in the fact that Congress clearly contemplated that there was to be a uniform “ceiling” which no polluter would be permitted to exceed. The key to understanding the interrelationship between sections 301 and 304 can, we believe, be found in the following excerpt from the Senate Report: “In defining best practicable for any given industrial category, the Committee expects the Administrator to take a number of factors into account. These factors should include the age of plants, their size and the unit processes involved and the cost of applying such controls. In effect, for any industrial category, the Committee expects the Administrator to define a range of discharge levels, above a certain base level applicable to all plants within that category. In applying effluent limitations to any individual plant, the factors cited above should be applied to that specific plant. In no case, however, should any plant be allowed to discharge more pollutants per unit of production than is defined by that base level.” In our view, the section 301(b) limitations represent a single number effluent limitation which prescribes the minimum amount of control (the “base level”), or conversely, the maximum amount of effluent discharge (a “ceiling”) that is permissible. In determining this “base level,” and concomitant pollutant ceiling, the Administrator is to consider the numerous differences in processes and capabilities of point sources. Having determined the “base level,” and the “ceiling,” he must then promulgate guidelines which are to guide the permit-issuing authorities in deciding whether, and by how much, the limitation to be applied to any individual point source is more stringent than the base level (in terms of requiring more effective technology), and more stringent than the ceiling (in requiring a lower amount of effluent discharge). Thus, we reconcile sections 301 and 304 in the following manner: the section 301 limitations represent both the base level or minimum degree of effluent control permissible and the ceiling (or maximum* amount of effluent discharge) permissible nationwide within a . given category, and the section 304 guidelines are intended to provide precise guidance to the permit-issuing authorities in establishing a permissible level of discharge that is more stringent than the ceiling. Ill We now turn to the regulations being challenged to see if they comply with our interpretation of the statutory scheme. It is immediately apparent that the regulations establish only single number effluent limitations for each of twelve subcategories within the iron and steel industry. While the regulations comply with the requirement in sections 304(b)(1)(A) and 304(b)(2)(A) that they “identify . . . the degree of effluent reduction attainable” through application of the relevant standard of technology (“BPCTCA” or “BATEA”), they do not specify any factors to be taken into account in determining the control measures to be applied to individual point sources within the categories and classes, as required by sections 304(b)(1)(B) and 304(b)(2)(B). Furthermore, they do not specify permissible “ranges” of limitations below the ceiling. The Administrator, however, contends that he satisfied these requirements. In the regulations themselves, he states: “The Agency considers that the limitations already represent ranges, taking into account differences in processes used and other factors. Subcategorization has been used to take these factors into account with different limitations for each subcategory. Within subcategories, exceptions to the limitations have been provided where appropriate, thus constituting a range. Each numerical limitation represents a maximum value over a given period of time. This, in effect, represents a range from zero up to the specific limitation.” 39 Fed.Reg. 24117 (comment # 25). ' We believe that this response is insufficient. We do not agree that the “exceptions” referred to satisfy the “range” requirement, since only one of the twelve subcategories — By-Product Coke —contains any such exception. The statement that a range exists from zero up to the limitation is unsupportable since for all of the subcategories the Administrator concluded, with respect to the new source standards, that a “zero discharge” level was not feasible at this time. Thus, the “range” admittedly encompasses standards that cannot be met. More important, the regulations provide absolutely no guidance to the permit-issuing authorities as to what factors to consider or how to set the particular discharge level within a feasible range below the ceiling established under Section 301(b). Finally, we disagree with the Administrator’s contention that “sub-categorization” provides a range. The Administrator’s subeategorization merely divided the entire iron and steel making industry by means of the types of processes employed, and it does not reflect any of the innumerable differences within the particular subcategories. No guidance is given with respect to the remaining section 304(b) factors, such as age, costs and engineering aspects, which we previously concluded must be “specified” in order to guide the permit grantors in exercising their carefully circumscribed discretion in setting precise standards for individual point sources. The Administrator contends that sufficient flexibility in the regulations is provided through the “variance” procedure, which allows individual discharges to obtain variances from the limitations upon a showing that the factors relevant to a particular point source are “fundamentally different from the factors considered in the establishment of the guidelines.” Our responsibility, however, is not to determine whether the Administrator has provided for flexibility, but whether he has followed the statutory scheme established by Congress. Regardless of whether the establishment of a variance procedure is within the Administrator’s discretion, we do not believe that the Administrator can ignore his obligation to promulgate guidelines specifying factors* to be considered and ranges above a base level. We also note that the variance procedure provides for less flexibility than we believe Congress contemplated, since it permits deviations from otherwise rigid and unitary limitations only where the circumstances of the particular plant are “fundamentally different” than those from which the effluent limitation was derived. Having concluded that the regulations failed to constitute valid “guidelines” (despite their title), since they failed to provide meaningful ranges or guidance in considering individual factors, the question then arises as to the scope of our remand. Should we merely remand for promulgation of guidelines, leaving the single number limitations intact, or should we also remand for reconsideration of the limitations? We conclude that the latter approach is preferable. The guidelines can serve a meaningful purpose under our analysis only if they are coupled with limitations which represent the “ceiling” or maximum degree of effluent discharge permissible. In promulgating these regulations, the Administrator appears to have considered it his duty to establish uniform nationwide limitations, rather than establishing a ceiling with detailed guidelines to guide the grantors within a feasible range below the ceiling. Consequently, although we do not suggest that the present limitations cannot perforce represent ceilings, it is possible that those limitations might be more stringent than they would have been if they had merely been intended to represent such levels. Thus, we believe that the Administrator, in addition to promulgating guidelines, should reconsider the limitations with the base level and ceiling concepts in mind. IV Petitioners raise several other contentions with respect to the process by which the Administrator determined the 1977 “BPCTCA” and 1983 “BATEA” standards in his regulations. As noted in part II of this opinion, we believe it was clearly Congress’ intent, despite the ambiguous wording in section 304(b), that the Administrator consider the enumerated factors in setting standards for classes and categories of industries. Petitioners contend that the Administrator failed to give sufficient consideration to these factors or to articulate his reasoning with sufficient clarity. At the outset, it is important for us to articulate our standard of review. As in other cases involving review of an administrative agency’s rule-making actions we are governed by an “abuse of discretion” standard — in other words, we must not substitute our judgment for that of the agency, but must determine whether the Administrator’s actions were “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” Administrative Procedure Act § 10(e), 5 U.S.C. § 706(2); Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971); Delaware Citizens for Clean Air, Inc. v. Administrator, 480 F.2d 972, 975-76 (3d Cir. 1973). In order to facilitate meaningful judicial review, we should require administrative agencies to “articulate the standards and principles that govern their discretionary decisions in as much detail as possible.” Environmental Defense Fund, Inc. v. Ruckelshaus, 142 U.S.App.D.C. 74, 439 F.2d 584, 598 (1971). See also Greater Boston Television Corp. v. FCC, 143 U.S.App.D.C. 383, 444 F.2d 841, 851 (1970), cert. denied, 403 U.S. 923, 91 S.Ct. 2229, 29 L.Ed.2d 701 (1971). However, we should not reverse an agency’s decision that is not fully articulated where we can reasonably discern the basis for the agency’s action. Bowman Transportation, Inc. v. Arkansas — Best Freight System, Inc., 419 U.S. 281, 286, 95 S.Ct. 438, 42 L.Ed.2d 447 (1974). Furthermore, the Administrator’s conduct is entitled to a “presumption of regularity.” Citizens to Preserve Overton Park, supra, 401 U.S. at 415, 91 S.Ct. 814. A. “Age.” — The first factor enumerated in section 304(b) which petitioners contend the Administrator did not fully consider was that of “age.” Petitioners contend that the Administrator, faced with the statutory mandate to consider age, concluded that age was not relevant. However, this inaccurately states what the Administrator did. The Administrator did in fact consider age, but concluded, after studying the plants sampled, that “the processes and treatment systems are similar regardless of the age and size of the plant. Furthermore, as the regulations themselves state, the division of the industry by process was also, generally speaking a division by age: “The Agency has subdivided the steel making segment primarily along operational lines because the waste water volumes and pollutant parameters vary with the type of operation being conducted. In addition, the processes reflect the age of the technology employed. Subcategorization of coke making by the older beehive and the newer by-product operations and steel making by the older open hearth and the newer basic oxygen and electric arc furnace operations is indirectly subcategorization by age. “The treatment technology to be applied is primarily a function of the' pollutants present and hence is a function of the type of operation conducted. The type of pollutants present is not a function of the size or age of the operating facilities. Land availability for application of the treatment technology is not a function of size or age since many new as well as old mills are limited on the area available for installation of treatment facilities and vice versa . . . Many of the older mills have better treatment than some of the newer ones and vice versa.” 39 Fed.Reg. at 24116, comment # 16 (emphasis added). We believe that this quotation shows that the Administrator did consider age and did sufficiently articulate his reasoning, to the extent that he concluded that age was not by itself relevant to the type of treatment technology or processes to be installed, and that.'separate and explicit categorization by age was not warranted. We further believe, after examining the data in the Final Development Document, that the Administrator’s conclusions in this respect are fully supportable. Where we believe the Administrator erred, however, was in his failure to consider age as it had a bearing on the cost or feasibility of retrofitting plants. While all the plants in a certain older subcategory (e. g., beehive coke) may require the same technological processes to reduce effluent discharges, the fact that all the plants within that subcategory were built long before plants in another subcategory may present special problems in installing anti-pollution devices. Similarly, in a subcategory where there is considerable variation in age, the fact that the processes are similar may mean that the same type of control technology can be installed, but it does not necessarily mean that the ease with which that technology can be installed, or the ability to comply with effluent limitations once it has been installed, is not affected by age. Since we see no evidence that the Administrator considered age in this light, remand is appropriate. If he concludes on remand that age is not relevant to the cost or feasibility of retrofitting plans, as well as to the processes and treatment system, that conclusion will be valid to the extent it is supported by the record. B. “Engineering Aspects. ” — Petitioners raise several points here in contending that the Administrator did not sufficiently consider this factor. They contend, inter alia, that for several subcategories the Administrator has suggested technology which is not currently available; that virtually all of the limitations are based on reduced flow rates which have not been demonstrated; and that many of the limitations are based on transfer technologies but lack guidelines on the engineering requirements for such transfers. Instead of considering these points as they affect the regulations in their entirety, we will defer any consideration of them until part VII of this opinion, in which we consider the specific challenges to individual limitations. The only contention we shall consider here is that in each of two subcategories — vacuum degassing and continuous casting — the Administrator surveyed only two plants which were not representative of the subcategories as a whole. In particular, they contend that the plants surveyed were built on “green field sites,” which enabled them to construct large lagoons for the treatment of waste materials, and that the treatment systems were built simultaneously with the production facilities, thereby making it easier for them to meet effluent limitations than plants which must be retrofitted. The Administrator’s response to petitioner’s challenge is that given the extraordinary complexity of promulgating uniform guidelines and limitations, he cannot be expected to consider all the engineering factors in specific situations. He further contends that, to the extent that a particular plant’s inability to comply with an effluent limitation is attributable to the fact that it is operating under conditions “fundamentally different” than the surveyed plants, it could obtain a variance. Petitioners in support of their challenge have failed to document their assertions that reliance on studies of plants with “unrepresentative” engineering characteristics has resulted in unrealistic limitations. They refer to no record deficiency that would overcome a presumption of regularity in the Administrator’s conduct. Accordingly we are convinced that the regulations cannot be declared arbitrary or capricious on this ground. C. “Non-water quality environmental impact (including energy requirements).” — Petitioners contend that there is no evidence in the record to indicate that the Administrator considered these factors. We disagree, however, and believe that the Final Development Doeument accompanying the regulations (especially section VIII of the Document, found in Appendix at 1522a — 1581a) clearly shows that he did consider them in some detail. For each category, and for each level of technology (1977 “BPCTCA” and 1983 “BATEA” standards), the Administrator considered both the problems of air pollution and solid waste disposal, as well as the problem of additional energy requirements caused by installation of the necessary anti-pollution devices. In the by-product coke subcategory, for example, the Document identified the types of air pollution emissions likely to occur and concluded that “[i]f a vapor recirculation or solvent extraction facility for dephenolization is added to the system, significant reductions in both parameters are achieved.” (Appendix at 1534a). Similarly, the Document identified the solid wastes likely to be generated by treatment systems, and concluded that they could either be internally consumed through reuse in the mill, incinerated, or used as landfill (Id.). With respect to additional energy requirements, the Document estimated the additional annual operating costs in terms of power for each of the two treatment alternatives — $2,175.00 for alternative I, and $31,500.00 for alternative II (Appendix at 1523a). The Administrator then evaluated the overall impact of the limitations on air quality, solid waste problems and energy requirements for all subcategories (Appendix at 1583a — 1584a, 1629a and 1630a), and concluded that “[t]he enhancement of water quality management provided by these proposed effluent limitations substantially outweighs the impact on air, solid waste, and energy requirements” (Appendix at 1584a). Given the failure of petitioners to present contrary evidence as to the effects of the limitations in these areas, we cannot say that the Administrator’s consideration of these factors was insufficient or that his conclusions were arbitrary and capricious. A special problem is presented by CF&I Steel Corporation. As noted at the outset of this opinion, CF&I objects to the effluent limitations on the ground that they did not take into account the problem of water loss caused by recycle systems, a problem which it claims is serious in the arid and semi-arid western states. It appears, however, that CF&I never raised this contention at the agency level, when these regulations were being formulated with extensive participation on the part of the rest of the petitioners. Normally, this would preclude our consideration of this contention for the first time at this stage. See Unemployment Compensation Commission of the Territory of Alaska v. Aragon, 329 U.S. 143, 155, 67 S.Ct. 245, 91 L.Ed. 136 (1946); Portland Cement Association v. Ruckelshaus, 158 U.S.App.D.C. 308, 486 F.2d 375, 394 (1973). Petitioners claim that the rule barring de novo consideration of challenges to agency action in a reviewing court should not apply here since the agency proceedings were rule-making rather than adjudicatory and since the Administrator failed to follow his mandatory and affirmative responsibility to consider certain statutory factors. Whatever merit there may be to this contention in the abstract, we do not believe that this is an appropriate case to deviate from the general rule. Labelling these proceedings “rule-making” cannot obscure the fact that CF&I’s challenge is to actions of the Administrator which affect it alone. Only three other steel mills in the entire country are located in arid or semi-arid regions, and CF&I is the only company which has alleged that installation of anti-pollution devices would have an adverse effect on the water supply. Thus, this contention is one of particular and localized concern, and was not a problem which we believe was so obvious to the Administrator that it was arbitrary and capricious for him to have failed to consider it sua sponte during the rule-making stage. We believe it would be highly disruptive to the administrative process to allow a company to sit back and wait until the regulations were published in final form before coming forth and contending that the agency had failed to consider the peculiarly local impact of its regulations in a certain region. However, while we do not believe that we can invalidate the regulations on the ground that the Administrator failed to consider this problem, we have already indicated that the matter should be remanded to the Administrator for the promulgation of guidelines and for the reconsideration of the limitations in light of the “base level” and “ceiling” concepts. We do not mean to preclude CF&I from raising its particular concerns about water scarcity in the context of the proceedings on remand. Cf. Portland Cement Association, supra at 394 — 95. D. “Costs. ” — The final factor which petitioners contend the Administrator did not sufficiently consider was that of “costs.” We have already concluded that the Administrator did not sufficiently consider the problem of age as it pertained to the cost and feasibility of retrofitting existing plants, and our present discussion of the cost factor will not repeat what was said previously. Our first task here is to clarify what Congress intended when it directed the Administrator to take costs into account. It is immediately apparent that Congress contemplated that the Administrator should give greater consideration to the cost of compliance when defining 1977 “BPCTCA” technology levels than he should when defining the 1983 “BATEA” levels. Section 304(b)(1)(B), states that the factors relating to an assessment of “BPCTA” “shall include consideration of the total cost of application of technology in relation to the effluent reduction benefits to be achieved from such application.” This is in contrast to section 304(b)(2)(B), which merely requires that the assessment of the factors “shall take into account . . the cost of achieving such effluent reduction.” Nevertheless, while costs were intended to be given greater weight in defining “BPCTCA”, it is clear that even with that 1977 standard, the cost of compliance was not a factor to be given primary importance. Furthermore, Congress clearly intended that the Administrator consider costs on a class or category basis, rather than as a plant-by-plant basis. As Senator Muskie stated in support of the House-Senate Conference Committee Report: “The modificátion of subsection 304(b)(1) is intended to clarify what is meant by the term ‘practicable.’ The balancing test between total cost and effluent reduction benefits is intended to limit the application of technology only where the additional degree of effluent reduction is wholly out of proportion to the costs of achieving such marginal level of reduction for any class or category of sources. “The Conferees agreed upon this limited cost-benefit analysis in order to maintain uniformity within a class and category of point sources subject to effluent limitations and to avoid imposing on the Administrator any requirement to consider the location of sources within a category or to ascertain water quality impact of effluent controls, or to determine the economic impact of controls on any individual plant in a single community.” With respect to the 1983 “BATEA” standards, Senator Muskie intended that the type of assessment should be basically the same, except that there should be no cost-benefit analysis. Although there is some ambiguity as to precisely how much weight should be given to cost, it seems that the Administrator would be governed by a standard of “reasonableness”: “In making the determination of ‘best available’ for a category or class, the Administrator is intended to apply the same principles involved in making the determination of ‘best practicable’ (outlined above), except as to cost-benefit analysis . “While cost should be a factor in the Administrator’s judgment, no balancing test will be required. The Administrator will be bound by a test of reasonableness. In this case, the reasonableness of what is ‘economically achievable’ should reflect an evaluation of what needs to be done to move toward the elimination of the discharge of pollutants and what is achievable through the application of available technology — without regard to cost.” With respect to the overall impact of the legislation, Congress clearly contemplated that cleaning up the nation’s waters might necessitate the closing of some marginal plants. As Senator Bentsen stated: “There is no doubt that we will suffer some disruption in our economy because of our efforts; many marginal plants may be forced to close.” In sum, while it is clear that the Administrator must consider cost, some amount of economic disruption was contemplated as a necessary price to pay in the effort to clean up the nation’s waters, and the Administrator was given considerable discretion in weighing costs. With these expressions of legislative, intent in mind, we turn to consider the specific points raised by petitioners. Their contentions are essentially twofold: (1) that the Administrator’s consideration of costs was insufficient and incomplete, in part because he excluded certain cost factors and because his promulgation of regulations for only “primary” (or “phase I”) operations precluded him from evaluating the full impact of pollution control devices; and (2) that the Administrator acted arbitrarily and capriciously in promulgating regulations in the face of evidence that many plants cannot raise the necessary capital to finance the installation of anti-pollution devices and would be forced to close. With respect to the first contention, we believe it is clear that the Administrator did consider the cost impact of the regulations. The EPA contractor developed operating and capital cost estimates for each level of technology within each subcategory, and cost-benefit diagrams were prepared for each subcategory. The estimated costs were based on an “average size plant,” as determined from the surveyed plants, and were then multiplied by the total number of facilities within each subcategory to obtain a figure representing the industry-wide costs per subcategory. The EPA also estimated the total cost (including amortization) of complying with both the “BPCTCA” and “BATEA” standards to be $82.3 million, a figure which represented 0.37% of the 1972 gross revenue of the steel industry. It further estimated that the total annual costs for both air and water pollution controls after 1983 would be $1.24 billion, or 5.54% of the industry’s gross revenue in 1972. On the basis of these figures, and after referring to the section of the Final Development Document discussing the deleterious effects of the pollutants sought to be controlled, the Administrator concluded: “The Agency believes that the benefits of thus reducing the pollutants discharged justify the associated costs, which though substantial in absolute terms, represent a relatively small percentage of the total capital investment in the industry.” Given our standard of review of agency action and the relative weights we believe Congress intended the Administrator to assign as between the need for pollution abatement and costs, we conclude that this assessment, when coupled with the economic impact analysis to be discussed below, was neither arbitrary nor capricious. A troublesome question is presented by petitioners’ contention that the costs estimated by the Administrator were artificially low because certain factors were excluded. Petitioners point to the fact that the Final Development Document listed several factors which admittedly affected costs, but which were excluded from the cost analysis. Petitioners contend that these factors, which include such things as land acquisition and site clearance costs, generate costs which are as large as those which the Administrator included in his estimates. However, petitioners have pointed to no data which would support their allegation as to the magnitude of these excluded costs. factors. Furthermore, many of these factors were excluded because they were inherently site-specific or because they could not be evaluated. For these reasons, we do not believe we can conclude that the exclusion of these factors from the costs analysis was arbitrary or capricious. Petitioners’ chief objection here is that since regulations for only primary (or phase I) operations were promulgated, no cost consideration was given to the impact of the yet-to-be-promulgated regulations for secondary operations. The “Kearney Report” estimated that those costs would be very substantial, and that the total water pollution-control costs would be three to five times as high as those for