Opinion for the Court filed by Circuit Judge ROBB.
ROBB, Circuit Judge:These consolidated cases present challenges to regulations and actions of the Environmental Protection Agency and its Administrator pursuant to the Federal Water Pollution Control Act Amendments of 1972 (FWPCA), 33 U.S.C. § 1251 et seq. (Supp. IV). In No. 74-1967 American Paper Institute (API), joined by eight companies engaged in the production of pulp, paper and paperboard,1 challenged the EPA regulations by way of a suit in the United States District Court for the District of Columbia. The court held that under Section 509(b)(1) of the FWPCA jurisdiction to review these regulations was in a United States Court of Appeals and not in the District Court; accordingly the court granted EPA’s motion to dismiss for lack of subject matter jurisdiction. American Paper Institute v. Train, 381 F.Supp. 553 (D.D.C.1974). The plaintiffs appeal.
The appeal from this dismissal has been consolidated here with petitions for review filed by API and the eight paper manufacturing companies. For convenience we shall refer to the appellants and the petitioners collectively as API. The petitions for review challenge both the regulations and the implementation of the FWPCA by EPA. These regulations were promulgated May 29, 1974 and establish effluent limitations guidelines and new source perform*186anee standards for five subeategories of the pulp, paper and paperboard industries.2 We affirm the District Court’s dismissal, uphold as valid the Agency’s promulgated regulations and dismiss the various petitions for review.
I. THE STATUTE
The FWPCA establishes a comprehensive scheme for federal regulation of water pollution. The ultimate objective of the Act is to eliminate completely the discharge of pollution into navigable waters by 1985. The Act seeks to accomplish this by means of phased reductions in the amounts of pollutants which are discharged from municipal and industrial sources. Recently, in a lengthy and detailed opinion by another panel, this court reviewed and interpreted certain provisions of the Act which play a key role in this process. American Frozen Food Institute v. Train, 176 U.S.App.D.C. 105, 539 F.2d 107, No. 74-1464 (1976). Because the present case involves a discussion of these same statutory provisions, we dispense with an elaborate analysis of what is indeed a “complicated and lengthy statute,” AFFI v. Train, supra, 539 F.2d at 115, and simply sketch, in abbreviated fashion, those sections of the Act with which we must deal.
Section 301, entitled “Effluent Limitations”, requires all existing point source3 dischargers of pollution to meet two levels of effluent standards: one level by July 1, 1977, and the other by July 1, 1983. 33 U.S.C. § 1311. The 1977 level requires the application of the “best practicable control technology currently available” for point sources other than publicly owned treatment works. (BPCTCA or “best practicable”) Section 301(b)(1)(A). The 1983 level requires the application of the “best available technology economically achievable” for these same sources. (BATEA or “best available”) Section 301(b)(2)(A). These same provisions of Section 301 also state that the defined levels of technology are to be given more specific meaning in regulations issued by the EPA pursuant to Section 304.
Section 304, entitled “Information and Guidelines”, indicates the method by which precise definitions for the two standards are to be achieved. 33 U.S.C. § 1314. Under Section 304(b)(1), which applies to the 1977 BPCTCA standard, the Administrator is to establish guidelines for effluent limitations. These regulations are to (1) identify the degree of effluent reduction attainable through best practicable control technology currently available for classes and categories of point sources, other than publicly owned treatment works, and (2) specify factors to be taken into account in determining the control measures and practices applicable to such point sources. Section 304(b)(1)(A) & (B). Under Section 304(b)(2), which applies to the 1983 BATEA standard, the Administrator is to publish regulations which take into account similar considerations. Thus, the effluent limitations required under Section 301 are to be achieved pursuant to guidelines established by EPA under Section 304.
Another section of importance in this case is Section 306. 33 U.S.C. § 1316. Unlike Sections 301 and 304 which pertain to effluent standards required for existing sources, Section 306 governs the standard for new sources of pollution discharge. This section specifies twenty-seven industries for which standards for new sources must be set, including the pulp, paper and paperboard industry. The standard for new sources must reflect the greatest degree of effluent reduction achievable through application of the “best available demonstrated control technology” (BADCT), including where practicable, a standard permitting no discharge of pollutants.
Hence, under Sections 301, 304 and 306 of the FWPCA the Administrator is responsi*187ble for regulations which define three separate levels of pollution control:
(a) the 1977 “best practicable” (BPCTCA) standard for existing sources;
(b) the 1983 “best available” (BATEA) standard for existing sources;
(c) the “best available demonstrated” (BADCT) standard for new sources.
In general, the statute envisions the 1977 standard as a minimum level of control, the 1983 standard as more stringent, and the new source performance standards as the most stringent.4
Primary enforcement of these standards is to be accomplished under Section 402 of the Act. 33 U.S.C. § 1342. This section establishes a permit system as the basic mechanism for enforcing the effluent limitations established under Section 301. Permits for pollutant discharge may be issued by the Administrator or by a state which has adopted a permit program approved by the Administrator. Section 402(a) and (b). The function of the permit is to define precisely each individual discharger’s obligations under the Act. EPA retains authority to review operation of a state’s permit program and may withdraw approval of a program which is not being administered in compliance with this section. Section 402(c) and (d).
The final statutory provision that is important here is Section 509. 33 U.S.C. § 1369. This section, entitled “Administrative Procedure and Judicial Review”, establishes that the United States courts of appeal have exclusive jurisdiction to review the Administrator’s actions in promulgating effluent limitations under Section 301. Section 509(b)(1)(E).
II. EPA PROCEEDINGS
In January 1973 EPA contracted with a private engineering consultant firm, Wapora, Inc., to make an analysis of the pulp and paper industry to aid EPA in developing effluent limitations and new source performance standards. The Wapora study concentrated only on that portion of the industry involved in unbleached pulp and paper (Phase I), and the regulations at issue here are only these Phase I regulations.5
The Wapora study resulted in a 245-page draft report which was submitted to EPA in June 1973. The draft report' stated that as a result of various factors this Phase I category of the paper industry should be subdivided into five subcategories: unbleached kraft, sodium base neutral sulfite semi-chemical, ammonia base neutral sulfite semi-chemical, unbleached kraft-NSSC (cross recovery), and waste paperboard. (App. 41) The report then discussed water use and waste characterization, selection of various pollutants, control and treatment technologies, costs, energy, non-water quality aspects, and implementation requirements. (App. 44-162) This information was then evaluated to make a preliminary determination of the effluent limitations and new source performance standards for each subcategory. (App. 162 — 180)
In July 1973 copies of the contractor’s draft report were circulated and comments solicited from 160 government, state, industry, and private institutions. Twenty-five organizations, including API, responded. Their comments resulted in changes being made in the draft report and in the forthcoming proposed regulations. (App. 1335)
On August 6, 1973 EPA published an “Advance Notice of Public Review Procedures” which set out EPA’s general course in developing the regulations for the effluent limitations and new source performance standards. (App. 730) The notice explained the legal authority for the regulations, the general methodology employed in their development, and the means by which *188EPA would facilitate comment on them. (App. 730-734)
On January 15, 1974 the proposed guidelines and standards were published in the Federal Register. (App. 1331) The proposed regulations were supported by the contractor’s draft report in its proposed form — known formally as the Development Document for Proposed Effluent Limitations Guidelines and New Source Performance Standards for the Unbleached Kraft and Semichemical Pulp Segment of the Pulp, Paper and Paperboard Mills Point Source Category. [Draft Development Document] (App. 956-1330) In addition, a second document concerning economic analysis and entitled Economic Analysis of Proposed Effluent Guidelines, Pulp, Paper and Paperboard Industry was made available to the public and circulated along with the Draft Development Document. (App. 735-955) Once again comments were solicited; thirty-seven parties, including API, responded and additional changes were made in the regulations. (App. 2199, 2203)
A technical hearing on the regulations was held on April 15, 1974. This hearing was in response to API’s request, after the two comment periods, for an additional opportunity to comment on the regulations. (App. 1935-36) A transcript of this hearing may be found at App. 1937 — 2185.
On May 29, 1974 EPA promulgated final regulations setting forth “final effluent limitations guidelines for existing sources and standards of performance and pretreatment standards for new sources in the pulp, paper, and paperboard category of point sources.” (App. 2199) These are the regulations at issue on this appeal. They are supported by a 340-page final development document which bears substantially the same title as the earlier draft: Development Document for Effluent Limitations Guidelines and New Source Performance Standards for the Unbleached Kraft and Semichemical Pulp Segment of the Pulp, Paper, and Paperboard Mills Points Source Category. [Final Development Document] (App. 2210-2565)
The promulgated regulations are divided into the five subcategories previously mentioned. See p. 9 supra. In comments preceding the promulgated regulations EPA alluded to the flexibility provision contained within each of the five subcategories. Comment (7) states:
Section 304(b)(1)(B) of the Act provides for “guidelines” to implement the uniform national standards of section 301(b)(1)(A). Thus Congress recognized that some flexibility was necessary in order to take into account the complexity of the industrial world with respect to the practicability of pollution control technology. In conformity with the Congressional intent and in recognition of the possible failure of these regulations to account for all factors bearing on the practicability of control technology, it was concluded that some provision was needed to authorize flexibility in the strict application of the limitations contained in the regulation where required by special circumstances applicable to individual dischargers. Accordingly, a provision allowing flexibility in the application of the limitations representing best practicable control technology currently available has been added to each subpart, to account for special circumstances that may not have been adequately accounted for when these regulations were developed. (App. 2203) (Emphasis added.)
The flexibility provision incorporated into each of the five subcategories reads as follows:
In establishing the limitations set forth in this section, EPA took into account all information it was able to collect, develop and solicit with respect to factors (such as age and size of plant, raw materials, manufacturing processes, products produced, treatment technology available, energy requirements and costs) which can affect the industry subeategorization and effluent levels established. It is, however, possible that data which would affect these limitations have not been available and, as a result, these limitations should be adjusted for certain plants in this industry. An. individual discharger *189or other interested person may submit evidence to the Regional Administrator (or to the State, if the State has the authority to issue NPDES permits) that factors relating to the equipment or facilities involved, the process applied, or other such factors related to such discharger are fundamentally different from the factors considered in the establishment of the guidelines. On the basis of such evidence or other available information, the Regional Administrator (or the State) will make a written finding that such factors are or are not fundamentally different for that facility compared to those specified in the Development Document. If such fundamentally different factors are found to exist, the Regional Administrator or the State shall establish for the discharger effluent limitations in the NPDES permit either more or less stringent than the limitations established herein, to the extent dictated by such fundamentally different factors. Such limitations must be approved by the Administrator of the Environmental Protection Agency. The Administrator may approve or disapprove such limitations, specify other limitations, or initiate proceedings to revise these regulations. (App. 2204-8) (Emphasis added.)
These, then, are the final regulations promulgated by EPA and challenged by API in the District Court, and on these petitions for review.
The only issue before the District Court was whether it had jurisdiction to review the regulations. The court held that Section 509 of the FWPCA provides for review by a United States Court of Appeals and not by a United States District Court. American Paper Institute v. Train, 381 F.Supp. 553 (D.D.C.1974).
Before we turn to the merits of the challenged regulations, three preliminary matters must be addressed. These are the questions of our jurisdiction under Section 509, the Administrator’s implementation of the Act under Sections 301 and 304, and the “range” of effluent limitations for this industry. All three are issues raised by API and recently resolved by this court in American Frozen Food Institute v. Train, 176 U.S.App.D.C. 105, 539 F.2d 107, No. 74-1464 (1976). Accordingly, our discussion will be brief.
III. RECENTLY RESOLVED ISSUES
A. Jurisdiction
API contends that the regulations promulgated by EPA are Section 304 guidelines reviewable by the District Court under the Administrative Procedure Act, 5 U.S.C. § 551 et seq. (1970). EPA argues that the Section 304 guidelines are inextricably intertwined with the Section 301 effluent limitations and that the regulations are therefore effluent limitations guidelines promulgated pursuant to both Sections 301 and 304. Thus, argues EPA, review is lodged exclusively in the courts of appeal under Section 509 which states in pertinent part:
(b)(1) Review of the Administrator’s action
******
(E) in approving or promulgating any effluent limitation or other limitation under section 301, 302, or 306 . may be had by any interested person in the Circuit Court of Appeals of the United States .
Because Section 509 makes no mention of judicial review of Section 304 guidelines, Section 509 will not apply and review is not lodged exclusively in this court unless we determine that the regulations published by EPA are effluent limitations guidelines under Sections 301 and 304.
In AFFI v. Train, supra, this court analyzed the language and legislative history of the FWPCA in detail and resolved this question in favor of the Agency. We note that the Eighth Circuit takes a contrary view, CPC International, Inc. v. Train, 515 F.2d 1032 (8th Cir. 1975), but that the Second, Third, Fourth, Seventh and Tenth Circuits are in agreement with us. Hooker Chemicals & Plastics Corp. v. Train, 537 F.2d 620 (2d Cir. 1976); American Iron & Steel Institute v. EPA, 526 F.2d 1027 (3rd Cir. 1975); E. I. duPont de Nemours & Co. *190v. Train, 528 F.2d 1136 (4th Cir. 1975), cert. granted, 425 U.S. 933, 96 S.Ct. 1662, 48 L.Ed.2d 174 (1976) (No. 75-978) (duPont I); American Meat Institute v. EPA, 526 F.2d 442 (7th Cir. 1975); American Petroleum Institute v. EPA, 526 F.2d 1343 (10th Cir. 1975). Accordingly, we reaffirm the view expressed in AFFI v. Train, supra, 539 F.2d at 129-30, that this court has jurisdiction under Section 509 of the FWPCA to review the effluent limitations guidelines which have been promulgated by the Administrator pursuant to Sections 301 and 304 of the Act.
B. Implementation
The question presented here is whether the effluent limitations guidelines published by EPA pursuant to Sections 301 and 304 reflect a proper implementation of the FWPCA. API contends that EPA should merely publish guidelines to assist the Section 402 permit-granting authorities, which in turn will establish effluent limitations on an individual, plant-by-plant basis. EPA says it must promulgate effluent limitations which shall be nationally uniform.
In AFFI v. Train, supra, this court examined the language of the Act and its legislative history and again resolved this issue in favor of the Agency. The court said:
As we have pointed out in Parts I and II, both the legislative history of the Act and its statutory structure and language mandate our rejecting the argument that the Act required the Administrator to set “guidelines” and “effluent limitations” for individual plants. We believe that Congress intended individual plant considerations to be taken into account within the nationally set effluent limitations in the granting of state permits under § 402(b) and (e). ... at 131.
Although once again at odds with the view expressed by the Eighth Circuit in CPC International, supra, strong support for our position is found in similar holdings by the Second, Third, Fourth and Seventh Circuits. Hooker Chemicals & Plastics Corp. v. Train, supra; American Iron & Steel Institute v. EPA, supra, E. I. duPont de Nemours & Co. v. Train, 541 F.2d 1018, No. 74-1261 (4th Cir. 1976), cert. granted, - U.S. -, 96 S.Ct. 3165, 49 L.Ed.2d 1183 (1976) (Nos. 75-1473 & 75-1705) (duPont II); American Meat Institute v. EPA, supra. See also EPA v. State Water Resources Control Board, 426 U.S. 200, 96 S.Ct. 2022, 48 L.Ed.2d 578 (1976).
Contrary to API’s fears, this interpretation will not reduce the function of the permit-granting authorities to that of a “clerical rubber stamping exercise.” As noted earlier in our discussion, there is a flexibility provision incorporated into the regulations for establishing effluent limitations guidelines for each of the five subcategories. This provision has recently been upheld by this circuit and two others. See AFFI v. Train, supra, E. I. duPont de Nemours & Co. v. Train (duPont II), supra, and Natural Resources Defense Council v. EPA, 537 F.2d 642 (2d Cir. 1976). Under the provision, the permit-granting authorities have a critical role to play: to define precisely the discharger’s obligations under the Act where “fundamentally different factors” require an adjustment in the effluent limitations established for the particular subcategory. As stated by the Second Circuit:
Not all of the thousands of plants in operation can be expected to fit into prefabricated molds or templates. By specifying a permit procedure, Congress implicitly conferred on the permit-grantor the privilege of construing the broader regulations in light of the specific type of plant applying for the permit. Without variance flexibility, the program might well founder on the rocks of illegality.
Natural Resources Defense Council v. EPA, supra, at 647. Accordingly, we reaffirm the view expressed in AFFI v. Train, supra, 539 F.2d at 131, that the EPA is to promulgate effluent limitations guidelines of a nationally uniform nature.
C. Range
API argues that the entire set of regulations is invalid because the Agency has established single number effluent limi*191tations for each subcategory instead of a range of numeric values. In support of its “range” argument petitioners refer to the following statement in the Senate Report on the 1972 FWPCA amendments:
In effect, for any industrial category, the Committee expects the Administrator to define a range pf discharge levels, above a certain base level applicable to all plants within that category.6
EPA’s response is that it has fulfilled the Committee’s expectation. The Administrator has taken the pulp, paper and paperboard industrial category, divided it into two phases, and established five subcategories for the first phase.7 He has then assigned separate effluent limitations for each subcategory in Phase I. Thus, the individual numerical limitations for the five subeategories provide the “range” for Phase I of the pulp, paper and paperboard industrial category.
In AFFI v. Train, supra, this court considered and rejected API’s argument. We approved EPA’s method of establishing single number effluent limitations because “the permit issuing authority under § 402 (state or federal) will clearly be able to employ any limitation it finds appropriate for a specific plant which falls between a ‘range’ of zero pollutant discharge and the nationally set effluent limitations.” 539 F.2d at 140. On the facts presented to us, we add only that the Administrator has exercised reasonable discretion in subcategorizing the industry and establishing separate effluent limitations for the five subcategories of this phase of the industry. Although the Third Circuit has disapproved of this approach in American Iron & Steel Institute v. Train, supra, the Second and Fourth Circuits are in agreement with us. Hooker Chemicals & Plastics Corp. v. Train, supra, and E. I. duPont de Nemours & Co. v. Train (duPont II), supra. We affirm the Agency’s method of establishing single number discharge limitations for each subcategory of this industry.
IV. SCOPE OF REVIEW
This court’s authority to set aside agency action is governed by Section 10(e)(2)(A) of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A) (1970). We may not interfere unless the agency’s action is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law”. We may not substitute our judgment for that of the agency. Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971). We must ensure that a rational basis exists for the agency’s decision. Bowman Transportation, Inc. v. Arkansas-Best Freight Systems, Inc., 419 U.S. 281, 95 S.Ct. 438, 42 L.Ed.2d 447 (1974). This does not mean however that we must rubber-stamp agency decision-making. It is our duty to ensure that the agency’s decision is based on a consideration of all relevant factors. Our inquiry must be searching and careful, especially in highly technical cases such as the present one. Ethyl Corp. v. EPA, 176 U.S.App.D.C. 373, 541 F.2d 1 (1976) (en banc), cert. denied, 426 U.S. 941, 96 S.Ct. 2663, 49 L.Ed.2d 394 (1976). On this basis we proceed to an examination of the specific regulations being challenged.
V. GENERAL VALIDITY OF REGULATIONS
A. Cost
Cost is one of the factors which the Administrator must consider in setting standards under the Act. Section 304(b)(1)(B) requires that he consider “the total cost of application of technology in relation to the effluent reduction benefits to be achieved” by the 1977 BPCTCA standard. This requires a cost-benefit analysis. Sections 304(b)(2)(B) and 306 mandate no such balancing. Under these sections the Administrator must consider only “the cost of *192achieving such effluent reduction” in establishing the 1983 effluent limitations guidelines and the new source performance standards.
API contends that EPA’s assessment of the costs involved in setting these standards is inadequate in two respects: first, that the cost analysis undertaken by EPA is perfunctory and ignores economic considerations; and second, that the Agency has not engaged in a cost-benefit analysis for the 1977 standard as required by Section 304(b)(1)(B). We perceive no merit in either charge.
The bulk of EPA’s cost analysis is contained in Section VIII and Appendix III of the Final Development Document. (App. 2409-2426; 2519-2548) Section VIII summarizes the costs of internal and external effluent treatment associated with the three technological levels prescribed by the Act—the 1977 BPCTCA, the 1983 BATEA and the new source performance standards. For each technology the cost of effluent treatment has been determined for the five subcategories established in this phase of the paper industry. (App. 2448, 2469, 2477) Tables showing internal and external controls for the five subcategories also indicate the investment costs, total annual costs, interest cost plus depreciation cost at 15% per year, and operating and maintenance costs. (App. 2419-2423) Appendix III of the Final Development Document presents a detailed description of the basis of the costs presented in the Section VIII tables. (App. 2519-2548) In addition, the Appendix delineates the internal and external technologies identified by the three levels of technology in terms of design limitations, unit processes and costs. Our study of these portions of the Final Development Document leads us to conclude that the Administrator’s analysis is neither arbitrary nor capricious.
API’s second contention is that the Administrator has not engaged in a cost-benefit analysis for the 1977 BPCTCA as required by Section 304(b)(1)(B). Our examination of the record however negates this charge. In a 240-page report entitled Economic Analysis of Proposed Effluent Guidelines, Pulp, Paper and Paperboard Industry, the Agency, assisted by Arthur D. Little, Inc., engaged in a careful analysis of the total cost effects of the regulations and their impact on prices, production, employment, community, balance of payments and growth. (App. 735-955) As noted in the preamble to the published regulations, the total projected impact of 1977 BPCTCA technology includes 3%-6% price increases, 7—10 potential closures out of 188 mills, and 810-1,250 potential unemployed persons (representing 1.1%—1.6% of total employment for these mills). (App. 2203) In that portion of the preamble entitled “Cost-benefit analysis”, the Administrator concludes with this statement:
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. (App. 2203)
Our review of the costs identified in Section VIII of the Final Development Document and the discussion of their impact in the Economic Analysis convinces us that there is firm record support for the Administrator’s conclusions.
B. Non-water Environmental Impact and Energy Requirements
Sections 304 and 306 of the Act require the Administrator to consider non-water quality environmental impact and energy requirements in establishing effluent limitations guidelines and new source performance standards. API argues that EPA has failed to fulfill this statutory obligation because it did not analyze energy costs and omitted consideration of certain environmental impacts.
The Final Development Document shows that EPA did analyze the energy costs and requirements of the industry. (App. 2424— 2426, 2448, 2469) After establishing both the power costs and energy requirements for all five subcategories under the 1977 BPCTCA standard, the 1983 BATEA standard, and the new source performance stan*193dard, (App. 2425-2426) the Administrator concluded that these energy considerations “are not substantial (less than 1 per cent [in the case of BPCTCA]) and should not be great enough to warrant concern on either a national or regional basis.” (App. 2448, 2469) In view of this assessment, the Administrator cannot be charged with failing to perform his statutory duty.
We reach the same conclusion regarding the Agency’s consideration of the non-water environmental impact of these regulations. We note in the Final Development Document that the Administrator discusses air pollution, noise potential, solid waste disposal and by-product recovery. (App. 2427— 2432) API takes the Administrator to task for failing to consider the impact of increased lime treatment and the implications of reuse of condensate streams. Our review of the Administrator’s discussion of non-water environmental impact does not convince us that his treatment of these issues constitutes arbitrary or capricious action. There is no indication that increased lime treatment is a significant factor or that the Administrator’s discussion of the air pollution problems does not encompass the problems inherent in reuse of condensate streams. Accordingly, we decline to find invalid this portion of his analysis.
VI. THE 1977 EFFLUENT LIMITATIONS GUIDELINES
EPA must establish effluent limitations guidelines for 1977 which reflect the application of “the best practicable control technology currently available (BPCTCA).”8 In the establishment of BPCTCA standards, Section 304(b)(1)(B) of the Act requires the Administrator to consider the following factors: (1) the total cost of application of technology in relation to the effluent reduction benefits to be achieved from such application; (2) the age of equipment and facilities involved; (3) the processes employed; (4) the engineering aspects of the application of various types of control techniques; (5) process changes; (6) non-water quality environmental impact (including energy requirements); and (7) such other factors as the Administrator deems appropriate. BPCTCA limitations will normally be established on the basis of the average performance of the best existing plants in each subcategory. Although the BPCTCA effluent standards will rely primarily on external (or “end of pipe”) treatment, internal control measures may be required if they are considered normal practice within the industry.
As noted earlier, this unbleached phase of the pulp, paper and paperboard industry has been subdivided into five subcategories: unbleached kraft, sodium base neutral sulfite semi-chemical, ammonia base neutral sulfite semi-chemical, unbleached kraft— NSSC (cross recovery), and waste paperboard. For existing sources in each subeategory, the 1977 effluent limitations guidelines limit the discharge of three pollutants: “BOD5”, “TSS”, and “pH”. These terms require explanation:
(a) BOD5. The initials “BOD” stand for “biochemical oxygen demand”. This is a measure of the oxygen consuming capabilities of organic matter. Dissolved oxygen in the water is consumed during decomposition of organic pollutants. Biochemical oxygen demand is commonly termed “BOD5” because the analytical techniques utilized for measurement involve a five-day period.
(b) TSS. The initials “TSS” stand for “total suspended solids”. This is a measure of the materials suspended in a sample of water. Suspended solids in pulp and paper mill effluents are generally fibrous materials lost in the pulping of wood and the production of paper.
(c) pH. This is a logarithmic expression of the concentration of hydrogen ions in a sample of waste water. It essentially describes the acidity or alkalinity of a water sample. At a pH of 7, the hydrogen and hydroxyl ion concentrations are approximately equal and the water is neutral. *194Lower pH values indicate acidity while higher values indicate alkalinity. Final Development Document at 83-86. (App. 2309-2312)
The 1977 effluent limitations guidelines permit the discharge of certain amounts of BOD5 and TSS expressed in pounds per ton of product (and kilograms per 1,000 kilograms of product) and the presence of a certain amount of pH for each subcategory. These limitations are set out in the following table.
BPCTCA Effluent Limitations Values in kg/kkg (Ibs/ton)
BOD5 TSS
Subcategory 30 Day Daily Max 30 Day Daily Max
Unbleached Kraft 2.8 (5.6) 5.6(11.2) 6.0 (12.0) 12 0(24.0)
NSSC-Ammonia 4.0 (8.0) 8.0(16.0) 5.0 (10.0) 10.0(20.0)
NSSC-Sodium 4.35(87) 8.7(17.4) 5.5 (11.0) 11.0(22.0)
Unbleached Kraft-NSSC 4.0 (8.0) 8.0(16.0) 6.25(12.5) 12.5(25.0)
Paperboard from WastePaper 1.5 (3.0) 3.0( 6.0) 2.5 ( 5.0) 5.0(10.0)
pH for all subcategories shall be within the range of 6.0 to 9.0.
Final Development Document at 216 (App. 2442).
A. Internal Control Measures
API’s first challenge to the 1977 effluent limitations is that in order to meet the BOD5 limitations, the majority of mills in the industry will be required to make extensive internal changes in their equipment. This, API argues, contravenes the Act, which limits the 1977 standards to external treatment measures.
API errs in believing that the FWPCA precludes in-plant control measures as part of the 1977 BPCTCA standard. The statute and the legislative history make clear that although BPCTCA standards are to rely principally on end-of-pipe treatment of waste water, internal control measures may be included if these measures are considered normal practice within the industry.9 The Final Development Document establishes that the in-plant changes which have been identified as available in 1977 are practices which are in common use in the pulp and paper industry. (App. 2443) Fully 60-100% of the mills within the five subcategories use some or all of the identified internal controls. Id. Moreover, the record discloses that at a technical meeting on April 4, 1974, API informed EPA that various in-plant control measures specifically identified by the Agency as BPCTCA were then common practices in the industry. (App. 2019-2025) Finally, we note that the Second and Fourth Circuits are in agreement with our position that internal control measures are permissible under the 1977 standard if they are considered normal practices within the industry. See Hooker Chemicals & Plastics Corp. v. Train, supra, 537 F.2d 620, as modified by Memorandum of June 14, 1976; duPont II, supra, 541 F.2d 1018; Tanners’ Council v. Train, 540 F.2d 1188, at 1191 (4th Cir. 1976); FMC Corp. v. Train (4th Cir. 1976) 539 F.2d 973, 981.
B. Basis for 1977 Effluent Limitations Guidelines
API’s second challenge to the 1977 effluent limitations is that they may not be based upon the average of the best performing mills in each subcategory, but must reflect only the average of the industry’s current performance. This argument is unpersuasive in light of the following statement by Senator Muskie in his discussion of the Conference 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 the plants, their size, the unit processes involved, and the cost of applying such controls.
The Administrator should establish the range of “best practicable” levels based upon the average of the best existing performance by plants of various sizes, *195ages, and unit processes within each industrial category. In those industrial categories where present practices are uniformly inadequate, the Administrator should interpret “best practicable” to require higher levels of control than any currently in place if he determines that the technology to achieve those higher levels can be practicably applied.
“Best practicable” can be interpreted as the equivalent of secondary treatment for industry, but this interpretation should not be construed to limit the authority of the Administrator. Leg.Hist. at 169-70. (Emphasis added.)
Accord, American Meat Institute v. EPA, 526 F.2d 442, 453 (7th Cir. 1975); Hooker Chemicals & Plastics Corp. v. Train, supra, 537 F.2d at 633; Tanners’ Council v. Train, supra, 540 F.2d at 1191.
We find no merit in API’s remaining contentions that EPA improperly used data showing maximum calendar month discharges to set effluent limitations and that unexplained discrepancies between the Draft and Final Development Document render the regulations invalid. On the first point, there is no inconsistency between the data and the regulations because both use twelve fixed-period daily averages per year. On the second point, because the Draft Development Document represents only a stage in the rulemaking process, subject to comment and change, there is no requirement that the Agency explain minor changes and refinements which inevitably result during such a process.
C. Identification of Effluent Reduction
API’s third challenge to the 1977 effluent limitations is that, contrary to the requirement of Section 304, EPA has failed to identify the degree of effluent reduction attainable through application of the best practicable control technology currently available. EPA contends that it has fully complied with Section 304 by establishing for each subcategory the quantity or quality of the various pollutants which may be discharged after application of the BPCTCA standard. We believe petitioners’ argument is more a matter of semantics than one of substance.10
Section 304(b)(1)(A) requires the Administrator to publish regulations providing guidelines for effluent limitations which
identify, in terms of amounts of constituents and chemical, physical, and biological characteristics of pollutants, the degree of effluent reduction attainable through the application of the best practicable control technology currently available for classes and categories of point sources , . (Emphasis added.)
This is precisely what EPA has done. It has (1) identified the three types of pollutants whose discharge must be limited under the 1977 standard (BOD5, TSS and pH), and (2) established in the regulations themselves the amounts of these pollutants which may be discharged in the five subcategories of this phase of the paper industry. We think this is a reasonable response to the statutory requirements and we reject the contention that EPA is also required to furnish data as to the pollution reduction which can be expected from the technologies which comprise the 1977 BPCTCA standard. If the Agency chooses to promulgate effluent limitations in the form of percentage removals, as it has done in the case of two subcategories of the 1983 color limitations, we would find this alternate method of establishing effluent limitations guidelines acceptable under the language of Section 304. However we do not believe that the above-quoted portion of Section 304(b)(1)(A) dictates to the Administrator that in adopting one or the other of these two methods he must also estimate how much reduction in pollution will be achieved as a result of the enunciated standard.
D. 1977 BOD5 and TSS Limitations
API’s fourth and final challenge to the 1977 effluent limitations is directed at the *196permissible amounts of BOD5 and TSS which may be discharged by mills under the BPCTCA standard. API contends that the 1977 BOD5 and TSS limitations cannot be met by the majority of mills performing at the “best practicable” level and therefore these limitations are invalid. In weighing petitioner’s argument, we must examine the method employed by EPA to establish these limitations.
EPA’s technique for developing effluent limitations guidelines and new source performance standards is set out in the Agency’s notice of August 6, 1973. (App. 730)
[T]he development of regulations for effluent limitations guidelines and standards of performance [is] undertaken in the following manner. The point source category is first studied for the purpose of determining whether separate limitations and standards are appropriate for . different segments within the category. This analysis includes a determination of whether differences in raw material used, product produced, manufacturing process employed, age and size of plants, waste water constituents and other factors require development of separate limitations and standards for different segments of the point source category. The raw waste characteristics for each such segment are then identified. This includes an analysis of (1) the source, flow and volume of water used in the process employed and the sources of waste and waste waters in the plant; and (2) the constituents of waste waters. The constituents of the waste waters which should be subject to effluent limitations guidelines and standards of performance are then identified.
Next, the control and treatment technologies existing within each segment are identified. This includes an identification of each distinct control and treatment technology, including both in-plant and end-of-process technologies, which exists or is capable of being designed for each segment. It also includes an identification of the effluent level resulting from the application of each of the treatment and control technologies, in terms of the amount of constituents and the chemical, physical, and biological characteristics of pollutants. The problems, limitations and reliability of each treatment and control technology are also identified. In addition, any non-water quality environmental impact, such as the effects of the application of such technologies upon other pollution problems, including air, solid waste, noise and radiation is examined. Finally, the energy requirements of each control and treatment technology are determined, as well as the cost of the application of such technologies.
This information is then evaluated in order to determine what levels of technology constitute the “best practicable control technology currently available”, “best available technology economically achievable” and the “best available demonstrated control technology, processes, operating methods, or other alternatives.” In identifying such technologies, various factors are considered including the total cost of the application of technology in relation to the effluent reduction benefits to be achieved from such application, the age of equipment and facilities involved, the process employed, the engineering aspects of the application of various types of control techniques, process changes and non-water quality environmental impact (including energy requirements). (App. 731)
The accumulation of information from this process resulted in a large data base. As shown in the following table, the total number of mills included in the subcategories under study was 218 of which 84 discharged to municipal systems.11 Of the 134 mills discharging to receiving waters, raw waste data from mill records was available *197for 64 mills. Final effluent data from mill records was available for 30 mills. Of these 30, 18 mills were finally chosen as best performers.
NUMBER OF MILLS PER SOURCE OF INFORMATION
Subcategory m , , . . . Mill Records, NCASI 12 Total Discharging Discharging----Number To To Receiving Raw Final Best of Mills Municipal Waters Waste Effluent(a) Performers
Unbleached Kraft 27 0 27 24 12 6
NSSCSodium 14 1 13 6 3 1
NSSCAmmonia 2 0 2 1 1 0
Kraft-NSSC 10 0 10 10 4 3
Paperboard-Waste Paper 165 83 82 23 9(b) 8
Totals 218 84 134 64 30 18
(a) Effluents from biological treatment
(b)One of the nine mills only had primary treatment
(App. 2242)
EPA’s method of choosing “best performers” for each subcategory is based on data contained in Sections III, V, VII and VIII and in the Appendices of the Final Development Document. Following the accumulation and analysis of information and data from every available source, best performing mills were chosen on the basis of those mills in each subeategory with the lowest current levels of pollution discharge. This is best illustrated by the following example.
In the unbleached kraft subcategory, raw waste data were available for 24 of the 27 mills. Final effluent data were available for 12 of these 27. After these data were examined six of the seven unbleached kraft mills achieving the best final effluent qualities in the country were selected as best performers. (App. 2450-2451) (The mill achieving the best final effluent quality in this subcategory, mill UK-3, was not included as a best performer because its external treatment system was judged to be atypical of the subcategory.) (App. 2449) The best performers in this subcategory are found in Tables 61 and 62 of the Final Development Document. (App. 2450-2451)
Once the Agency had selected its “best performers” it was in position to determine the effluent limitations guidelines for the 1977 BPCTCA standard. As previously mentioned, the legislative history demonstrates that Congress required EPA to establish 1977 effluent limitations based on “the average of the best existing performance by plants of various sizes, ages, and unit processes within each industrial category.” Leg.Hist. 169, 1468. (Emphasis added.) To continue our example of the unbleached kraft subcategory, Tables 61 and 62 provide us with the following information about the six best performers:
Final Effluent Data Showing Maximum Monthly Discharges of BOD5 and TSS in kg/kkg (Ibs/ton)
Mill BOD5 TSS
UK-1 2.3 (4.6) 6.1 (12.1)
UK-2 1.3 (2.7) 3.8 ( 7.6)
UK-5 3.1 (6.1) 7.1 (14.2)
UK-6 3.7 (7.4) 3.9 ( 7.9)
UK-7 3.1 (6.2) 4.6 { 9.3)
UK-8 3.3 (6.6) 3.5 ( 7.0)
Average 2.8 "(5.6) 4.8"~( 9/7)
(App. 2450-2451)
The 1977 BOD5 limitation finally established by EPA for the unbleached kraft subcategory was the figure derived from *198averaging the BOD5 figures from the six best performers: 2.8 (5.6). Ordinarily the 1977 TSS limitations would have been the average of the TSS discharge figures from these same six dischargers, but in this particular subcategory a less stringent limitation was finally selected because of practical considerations.13
(1) BOD5
API’s attack on the 1977 BOD5 limitations is based on a discussion of “percentage removals”. API equates the 1977 BPCTCA for removal of BOD5 at 85% and then presents calculations to show that in reality EPA is requiring treatment systems which must be capable of removing 92% of the annual average BOD5. This, API argues, exceeds what EPA has identified as BPCTCA.
The flaw in petitioners’ argument is that the 85% figure has not been used by the Agency to define the 1977 BOD5 limitation. It is not in the regulation establishing the effluent limitations guidelines for BOD5 and does not appear in that section of the Final Development Document discussing the BPCTCA standard.14 Although there is one reference to an 85% BOD, removal figure in a discussion of secondary treatment systems in Appendix III of the Final Development Document, four district biological units are analyzed there and in the analysis of only one—natural oxidation ponds—does this figure appear. (App. 2536) It appears that API has taken this figure out of context, misapplied it, and now criticizes the Agency for requiring limitations which exceed it. We therefore regard petitioners’ discussion of percentage removals as irrelevant in the context of the 1977 BOD5 limitations and uphold these regulations as promulgated.
(2) TSS
API says the 1977 TSS limitations are invalid because the Agency has failed to specify technology capable of removing “non-settleable biological solids” and has not stated what “percentage of reduction” of this material can be achieved by the operations EPA describes.15 EPA responds that these allegations are irrelevant: because the TSS limitations are based on actual waste water treatment performance data, the settleability of biological solids and the reduction capabilities of the identified technologies are included in the final effluent TSS levels.
Although EPA’s response is satisfactory to us on both points, we are unpersuaded by API’s arguments for additional reasons. First, we find unsound petitioners’ method of framing an argument on the basis of definitions not employed by the Agency and not found in the record. The absence of any discussion of the “non-settle-able” nature of biological solids reinforces our view of the validity of the TSS limitations as promulgated. Second, we have already considered and rejected API’s “percentage reduction” argument. As we said in discussing API’s third challenge to these effluent limitations guidelines, EPA is in full compliance with Section 304(b)(1)(A): it has (1) identified the pollutants whose discharge must be limited under the 1977 standard (BOD5, TSS and pH), and (2) established in the regulations themselves the amounts of these pollutants which may be discharged in the five subcategories of this phase of the paper industry. EPA is not required to furnish data as to the pollution reduction or “percentage reduction” which can be expected from the technologies which comprise the 1977 BPCTCA standard.
*199Finding no merit in any of petitioners’ challenges to the 1977 effluent limitations guidelines, we uphold these limitations.
VII. THE 1983 EFFLUENT LIMITATIONS GUIDELINES
EPA must establish 1983 effluent limitations guidelines which reflect application of “the best available technology economically achievable (BATEA).”16 In establishing BATEA standards, Section 304(b)(2)(B) of the Act, 33 U.S.C. § 1314(b)(2)(B), requires the administrator to consider the following factors: (1) the age of equipment and facilities involved; (2) the process employed; (3) the engineering aspects of the application of various types of control techniques; (4) process changes; (5) the cost of achieving such effluent reduction; (6) non-water quality environmental impact (including energy requirements); and (7) such other factors as the Administrator deems appropriate. The distinction between the 1977 BPCTCA standard and the 1983 BATEA standard is explained succinctly by Senator Muskie in his discussion of the Conference Report:
In making the determination of “best available” for a category or class, the Administrator is expected to apply the same principles involved in making the determination of “best practicable” (outlined above), except as to cost-benefit analysis. Also, rather than establishing the range of levels in reference to the average of the best performers in an industrial category, the range should, at a minimum, be established with reference to the best performer in any industrial category.
The distinction between “best practicable” and “best available” is intended to reflect the need to press toward increasingly higher levels of control in six-year stages. Through the research and development of new processes, modifications, replacement of obsolete plans and processes, and other improvements in .technology, it is anticipated that it should be possible, taking into account the cost of controls, to achieve by 1983 levels of control which approach and achieve the elimination of the discharge of pollutants.
As to the cost of “best available” technology, the Conferees agreed upon the language of the Senate Bill in Section 304(b)(2). 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.
Leg.Hist. at 170.
To a certain extent, the 1983 limitations must be anticipatory because no one can possibly know with certainty what will be “the best available technology economically achievable” in 1983. But we do know that the 1983 level of pollution control must be stringent enough to make possible the attainment of the ultimate goal of the Act: the complete elimination of pollution discharges to the Nation’s waters by 1985. In light of this, EPA’s 1983 effluent limitations guidelines must be upheld if the Agency can demonstrate that by 1983 there will be a technology which, if implemented, may reasonably be expected to achieve the standard.
For existing sources in each subcategory of this phase of the pulp and paper industry the 1983 effluent limitations guidelines seek to regulate the discharge of four pollutants: BOD5, TSS, pH and “color”. The “color” pollutant results from the various chemical pulping processes of the industry. Spent cooking liquors containing highly colored compounds are removed from the pulp in a washing sequence following the cooking process. The wash water is highly colored and large amounts of color are ultimately discharged into receiving streams. This color ranges from brown to black. EPA *200seeks to regulate the discharge of this pollutant for four of the five subcategories. For two subcategories (unbleached kraft and unbleached kraft — NSSC (cross recovery)) specific color controls are set in terms of pounds of color per ton of production (and its metric counterpart). For the other two subcategories (ammonia base neutral sulfite semi-chemical and sodium base neutral sulfite semi-chemical) color control is expressed in terms of percentage removal rather than maximum allowed mass units.
The 1983 point source discharge limitations for each of the four pollutants for this unbleached phase of the pulp and paper industry are set out in the following table, found in the Final Development Document. (App. 2464)
BATEA Effluent Limitations Values in kg/kkg (lbs/ton)
BOD5 TSS
30 Day Daily Max 30 Day Daily Max Subcategory
Unbleached Kraft 1.35(2.7) 2.7 ( 5.4) 1.85 ( 3.7) 3.7 ( 7.4)
NSSC-Ammonia 3.2 (6.4) 6.4(12.8) 2.6 ( 5.2) 5.2(10.4)
NSSC-Sodium 2.25(4.5) 4.5 ( 9.0) 2.5 ( 5.0) 5.0(10.0)
Unbleached Kraft-NSSC 1.6 (3.2) 3.2 ( 6.4) 2.1 ( 4.2) 4.2 ( 8.4)
Paperboard from Waste Paper 0.65(1.3) 1.3 ( 2.6) 0.8 ( 1.6) 1.6 ( 3.2)
Color
Subcategory 30 Day Daily Max
Unbleached Kraft 10 (20) 15 (30)
NSSC-Ammonia 75% removal
NSSC-Sodium 75% removal
Unbleached Kraft-NSSC 12.5 (25) 25.0 (37.5)17
Paperboard from Waste Paper
pH for all subcategories shall be within the range of 6.0 to 9.0
A. 1983 BOD5 Limitations
API’s first challenge is directed at the 1983 BOD5 limitations. These limitations are based on a specific achievable raw waste load for each subcategory. Final limitations are derived by applying to the raw waste load for each subcategory percentage reductions currently achievable by identified external treatment technologies. API’s challenge to these limitations is that there is no rational basis to support the raw waste loads for each subcategory.
Table 71 of the Final Development Document (App. 2471) provides the figures being challenged by API. The raw waste loads for the five subcategories are as follows:
Raw Waste
Subcategory kg/kkg "(lbs/ton)
Unbleached Kraft 12.5 (25)
Sodium-NSSC 17.5 (35)
Ammonia-NSSC 25 (50)
Kraft-NSSC 15 (30)
Paperboard-Waste Paper 10 (20)
(App. 2471)
API takes the unbleached kraft subcategory as an example and argues that there is no support in the Final Development Document for the BOD5 waste load figure of 25 pounds per ton of production. In this we think they err.
EPA projected the raw waste loads, upon which the 1983 limitations are based, on actual raw waste data for mills operating today. As shown in Table 13 of the Final Development Document (App. 2282) the raw waste load used as a basis for the 1983 limitation for the unbleached kraft subcategory is that load currently being attained by Mill UK — 2. The raw waste BOD for this mill is 24.5 pounds per ton, which EPA rounded up to 25. The raw waste loads for the other four subcategories were derived from similar tables showing mills presently achieving comparable or in some cases lower amounts of raw waste loads. See Tables 14, 18, 19 and 21, Final Development Document. (App. 2288, 2296, 2299-2302) Thus contrary to petitioners’ assertions, the Agency evidences a rational basis for selecting raw waste loads for each subcategory. In light of this we conclude that API’s challenge to the 1983 BOD5 limitations is without merit and we uphold these regulations as promulgated.
B. 1983 TSS Limitations
API challenges the 1983 TSS limitations on two grounds. First, says API, the tech*201nology of mixed- media filtration has not been demonstrated in actual paper mill waste treatment systems and therefore should not be included as technology for the 1983 standard. Second, API says the Agency has failed to explain its method of calculating final TSS limitations and thus its action is arbitrary and capricious. In light of our examination of the record we reject both contentions.
As explained in Section X of the Final Development Document the 1983 limitations guidelines were determined by using the 1977 TSS limitations as a base and then applying a reduction figure of 60%. This figure was derived from Agency estimates that application of coagulation and filtration technology, including mixed media filtration, would remove 60% of the TSS from biological treatment effluent. Final Development Document at 244. (App. 2470) Although petitioners now contend that mixed media filtration technology has not been actually demonstrated, statements by industry representatives and reports from industry sources contradict this assertion.
The first item in the record which contradicts API is a technical bulletin by the National Council of the Paper Industry for Air and Stream Improvement (NCASI) dated May 1973. (App. 1728) This industry bulletin contains the results of a NCASI study of mixed media filtration at three different mills. The study indicates that the three different types of mills were successful in removing significant amounts of TSS effluents with mixed media filtration. With the use of chemical additives, two of the three systems were able to remove as much as 96% and 97% of the TSS material present. (App. 1740)
The. second record item refuting petitioners’ contention that mixed media filtration is not a demonstrated technology is a statement by Peter Wrist, Vice President of the Mead Corporation (one of the petitioners) and Chairman of the Board of NCASI, acknowledging that mixed media filtration has shown the potential of removing the TSS material in the amounts estimated by EPA. At an EPA technical hearing on April 4, 1974, following publication of the proposed limitations, Mr. Wrist stated:
I would like to comment now on the advanced biological solids treatment techniques. The document indicates that in the proposed standards for BATEA and NSPS that there will be the use of mixed media filters for post-clarification of biological solids.
It is further indicated that with the use of this technique a further reduction of sixty percent of the level leaving the secondary basis is achievable.
The industry has been experimenting with this technique in a number of locations on a pilot-plant scale. We would agree that based on the very limited testing that we have seen, that this level of percentage reduction is achievable. (App. 1971-1972)
This evidence effectively rebuts API’s argument that mixed media filtration is not a demonstrated technology. We believe the Administrator, in selecting this particular technology, has shown that mixed media filtration may reasonably be expected to achieve the 1983 TSS standard. We turn to API’s second argument on the invalidity of these TSS limitations.
API argues that the Agency has failed to explain its method of calculating final TSS limitations. Once again, taking the unbleached kraft subcategory as an example, API contends that it is not apparent how the Agency arrived at 3.7 pounds of TSS per ton.
EPA’s calculations, as set out in the record (App. 3545 et seq.) and as explained in its brief, (pp. 86 et seq.) are as follows: First, in the unbleached kraft subcategory, reduce the 1977 TSS limitations (12.0) by 60% to account for the predictable effect of filtration and coagulation. This yields 4.8 pounds per ton. Second, determine from the Final Development Document the 1977 BOD5 raw waste load (34 pounds per ton) and the 1983 BOD5 raw waste load (25 pounds per ton). (App. 2452 and 2471) These calculations are necessary because of the direct relationship between reductions in raw waste BOD5 and reductions in ef*202fluent TSS levels. As the Final Development Document explains:
Generally, the TSS levels are directly affected by the BOD5 raw waste load, as the TSS in biological treatment system effluents generally are biological organisms generated in the removal of BOD5. Final Development Document at 244. (App. 2470)
Third, multiply the 4.8 lbs. by 25/34 to arrive at the 1983 TSS limitations for the unbleached kraft subcategory. The result is 3.5. The actual limitation set by EPA is slightly less stringent: 3.7 lbs. of TSS per ton of production.
We think the Agency’s method of calculating TSS limitations for 1983 is not so cloaked in secrecy or so inexplicable and unreasonable that we must hold the regulations invalid. Accordingly we reject API’s argument and uphold as valid EPA’s 1983 TSS limitations.
C. 1983 Color Limitations
.API’s third challenge focuses on the 1983 color limitations. API presents three arguments in support of its contention: First, EPA has erred by failing to give adequate consideration to petitioners’ comments concerning color regulations. Second, the Agency has set effluent limitations for col- or removal which cannot be met by the prescribed technology. Third, EPA’s cost analysis for color removal is deficient. We will examine these arguments in order.
(1) Comments
Petitioners argue that the Agency failed to consider two comments concerning color removal. The first comment is that although color may be objectionable from an aesthetic point of view, it has no harmful effects on receiving waters. The second comment is that the EPA-designated technologies for color removal are unproven and cannot meet the proposed 1983 color limitations.
At the outset we note that during the rulemaking proceedings the Agency afforded all interested parties, including API, two separate comment periods. API responded on both occasions. (App. 1335, 2199) In addition, EPA held a technical hearing in direct response to API’s request for an additional opportunity to comment on the proposed regulations. (App. 1935-1936) The transcript of this hearing is part of the record. (App. 1937-2185) Finally, in the preamble to the final regulations the Agency has set out a summary of the significant comments it has received together with the Agency’s response. (App. 2199-2204)
We must disagree with API’s assertion that the Agency has failed to consider its two comments on color removal. The very questions which API now says were ignored by EPA are specifically addressed by the Agency in the preamble. In preamble comment No. 24 EPA responds to API’s suggestion that color has merely an aesthetic impact and is not a proper subject for regulation.18 In comments Nos. 9 and 22 the *203Agency responds to petitioners’ argument that the technologies designated by EPA as capable of color removal could not meet the proposed limitations. In comment No. 9, the Agency addresses industry comments concerning color removal by the EPA-designated technology of reverse osmosis.19 In comment No. 22 EPA discusses the technology of lime treatment for removal of color.20
We reject API’s argument that the 1983 color limitations are invalid for failure to consider API’s comments on color removal.
2. Technologies
In their second argument petitioners charge that EPA has set effluent limitations for color removal which cannot be met by the prescribed technologies of lime treatment and reverse osmosis. According to the Final Development Document, Section X, (App. 2463-2472) color reduction for four of the five subcategories regulated by the 1983 BATEA standard is to be achieved by these two external treatment processes. For unbleached kraft and unbleached kraftNSSC (cross recovery) mills, color reduction can be achieved by minimum lime treatment. For NSSC-Sodium and NSSC-Ammonia mills, color reduction can be achieved by reverse osmosis. These two treatment technologies are explained at length in Section VII of the Final Development Document, (App. 2317-2408) but for purposes of our discussion a simplified explanation will suffice:
Minimum Lime — this technology involves the addition of lime to the pulping waste water at a stage of the treatment process prior to biological treatment for reduction of the level of oxygen-demanding substance. The lime combines with the color bodies to form a sludge. The sludge is then settled out of the treated waste. The lime is then either disposed of or recycled and reused. Final Development Document 119-125. (App. 2345-2351) Reverse Osmosis — this technology is generally employed to remove solutes from mill effluents. Basically, osmosis is the flow of a solvent from a solution of higher solvent concentration to a solution of lower solvent concentration across a semipermeable membrane. The driving force for solvent flow is the osmotic pressure difference which is proportional to the difference in concentration between the two solutions. If external pressure in excess of the osmotic pressure difference is applied, a reversal of solvent flow occurs — hence the term “reverse osmosis”. By these means, water containing impurities can be forced through a suitable membrane, leaving the impurities behind in concentrated solution. Final Development Document 157 — 163. (App. 2383— 2389; see also App. 374-382)
It is important to understand EPA’s methodology for setting color limitations. Color is discharged from pulp and paper mills in the form of organic “color bodies”. *204These are derived from the various chemical pulping processes involving raw wood. For purposes of the effluent limitations guidelines, color is measured in terms of color units (CU’s). Mass units of color are derived by multiplying the number of CU’s by the flow of the waste water stream in which they are found. EPA is then able mathematically to translate these calculations into limitations expressed in terms of pounds of color per ton of production and its metric counterpart.21 EPA’s 1983 color limitations for the four subcategories of this phase of the pulp and paper industry are set out below.
BATEA Effluent Limitations Values in kg/kkg (lbs/ton)
Color
Subcategory 30 Day Daily Max
Unbleached kraft 10 (20) 15 (30)
NSSC-Ammonia 75% removal -
NSSC-Sodium 75% removal -
Unbleached kraft-NSSC (cross recovery) 12.5 (25) 25.0 (37.5)
Paperboard from Waste Paper
With this in mind we proceed to examine API’s arguments that the Agency has set effluent limitations which cannot be met by the prescribed technologies of minimum lime treatment and reverse osmosis.
(a) Minimum Lime Technology
API’s challenge to this technology is based on results obtained by EPA from two full-scale demonstration projects, the Interstate Paper Company at Riceboro, Georgia, and the Continental Can Company at Hodge, Louisiana.
The Interstate Paper Company began operation of its minimum lime technology in March 1968. In December 1971 the company published a 117-page report describing the performance and operating results of color removal by this technology for the unbleached kraft subcategory. (App. 3895) The Continental Can Company began operation of its color removal technology in August 1971, and its 124-page report was published in February 1974. (App. 3762) The Hodge mill is both an unbleached kraft and an unbleached kraft-NSSC (cross recovery) mill and its report dealt with the results obtained in both subcategories.
Both reports from the demonstration projects concluded that minimum lime technology was a highly successful method for achieving significant color removal from mill effluent. For example, at the Hodge mill, average color reduction was 80% for all-kraft effluent with an 85-93% removal result at an upper range of lime dosage. (App. 3766; see also App. 3906)
API asserts that EPA’s analysis of these two demonstration projects was deficient in two respects: First, EPA failed to consider the point at which it measured color units in mill waste. Second, EPA ignored the additional complications for unbleached kraft-NSSC (cross recovery) which result from the presence of spent semi-chemical pulping liquor in the waste. We are not persuaded however that EPA omitted consideration of either factor.
API says the Agency failed to recognize that, because lime treatment must precede the biological treatment for successful color removal, a substantial regain in color units exceeding the established color limitations results during the secondary treatment. EPA responds to this contention in comment No. 22 of the preamble to the regulations:
The Agency believes that the lime treatment process has been satisfactorily demonstrated and can be applied to new mills. It is the Agency’s judgment that the operating problems being experienced by the two full-scale installations are inherent to the specific mills involved. For *205example, the problem of increases in color levels through the biological treatment systems are [sic] not indicative of problems of the lime treatment technology. These are a result of the specific biological systems involved, as color is being leached into the waste water from the ground. In addition, the Agency feels that many of the operating problems can be solved by new mills by being able to design the color removal system into the total mill design rather than adding on the unit operation to an existing mill. (App. 2201)
In a letter dated March 12, 1974 and made a part of the record (App. 1824-1827), Continental Can Company commented on the proposed 1983 color limitations and the problem of color removal reversion. Continental stated:
The color of the effluent must be in the 100-150 cu to satisfy the guidelines. Continental’s Hodge installation can achieve these levels in the effluent from the color system. However, during passage through the biologic system following color removal reversion takes place. The color of the effluent from the system to the waterway is closer to 350 cu. (App. 1826)
As finally promulgated by EPA, and as explained in the Final Development Document, the color effluent limitations for the unbleached kraft-NSSC (cross recovery) subcategory were based on 380 CU’s. (App. 2472) Thus, the EPA appears to have made a reasonable accommodation in the regulations for the phenomenon of color reversion.
API says EPA ignored another complication which prevents the lime treatment from achieving the limitations established by EPA for unbleached kraft-NSSC (cross recovery). According to API the presence of spent semi-chemical pulping liquor in the waste has the effect of doubling the percentage of the original color remaining in the treated waste. Thus, argues API, higher color levels are necessarily found in the unbleached kraft-NSSC (cross recovery) subcategory than in the unbleached kraft subcategory effluent.
It does not appear that EPA has ignored this doubling factor in establishing the color limitations for the unbleached kraft-NSSC (cross recovery) subcategory. As API acknowledges, these limitations were based on actual operating data from the color removal system at the Continental Can Company, Hodge, Louisiana. The Hodge mill is both an unbleached kraft and an unbleached kraft-NSSC (cross recovery) mill. The limitations for both subcategories were based on actual effluent data and did take into account the higher color levels and lower efficiency of removal. As Table 31 of the Final Development Document indicates, the average color in the unbleached kraft effluent at the Riceboro mill was 120 CU’s with a range of monthly averages from a low of 83 CU’s to a high of 170 CU’s. (App. 2349) Although the maximum monthly discharge was 170 CU’s the established limitations for the unbleached kraft subcategory were based on a conservative 250 CU’s, apparently to provide a substantial margin of safety. (App. 2472) As discussed earlier, the limitations for the unbleached kraftNSSC (cross recovery) subcategory were based on 380 CU’s — the average level achieved during the maximum month at Hodge. Id. This figure is slightly more than double the 170 CU’s maximum monthly discharge for unbleached kraft. Because limitations for both subcategories were based on actual mill data which reflected this differential, we reject petitioners’ contention that EPA established color limitations which ignore this fact. We thus conclude that API’s dual attack on the minimum lime technology for color removal is without merit. Accordingly, in the context of this argument, we uphold the 1983 color limitations for the unbleached kraft and unbleached kraft-NSSC (cross recovery) subcategories of these Phase I regulations.
(b) Reverse Osmosis
For the subcategories of NS SC-Ammonia and NSSC-Sodium, 1983 color limitations are set at 75% removal and are based on the reverse osmosis technology. The Final Development Document explains that al*206though this technology has not been demonstrated at “full mill scale”, pilot scale studies have indicated that at least 75% reduction of color should be achievable. (App. 2472) The Final Development Document also informs us that
color limitations of 75% removal for both sodium and ammonia base NSSC subcategories will be changed to kilograms of color per metric ton of production (pounds of color per short ton of production) at a later date when the technology has been proven through further development. (App. 2465)
Petitioners’ main contention is that the 1983 color limitations for these two subcategories are prescribed pursuant to technology which is non-existent. The Agency has dealt with this contention in the preamble to the final regulations. Comment No. 9 states:
(9) A large number of comments were made which stated that color removal by reverse osmosis, which is recommended in the proposed regulations as the color removal technology for NSSC mills, has not been demonstrated. It was stated that the mill referenced in the Development Document as demonstrating reverse osmosis for color removal is an atypical mill and is not using reverse osmosis specifically for color removal. It was pointed out that the reduction in color is actually an additional benefit of the reverse osmosis operations for water reuse. Thus, commenters suggested that the limitations for color for the NSSC subcategories should be removed from the final regulations.
The Agency acknowledged in the Development Document that the technology of reverse osmosis for color removal has not been fully demonstrated. Thus, color removal was not included for new sources for the NSSC subcategories. However, the Agency believes that reverse osmosis will be further demonstrated and will be an available technology for color removal by 1983. In addition, the technologies of ion exchange-resin adsorption and ultra-filtration as discussed in the Development Document are projected to be available to reduce color in pulp and paper mill effluents by 1983. Thus, the Agency is recommending reverse osmosis for NSSC mills for color removal for 1983, but other color removal technologies are also projected to be available for implementation in 1983. (App. 2200).
We have examined the basis for the Agency’s belief that reverse osmosis will be an available technology for color removal by 1983. From what is set out in both the Final Development Document at 157 — 163 (App. 2383-2398) and the various technical papers discussing the capabilities of this technology (App. 374, 378, 383, 3562, 3587), we believe there is adequate record support for the Agency’s conclusion that reverse osmosis will be available for color removal by 1983. Accordingly we decline the invitation to find the 1983 color limitations for NSSC-Ammonia and NSSC-Sodium invalid and uphold these regulations.
In summary, we reject petitioners’ argument that there is no support in the record for the Agency’s conclusion that the 1983 effluent limitations for color removal can be met by the prescribed technologies of lime treatment and reverse osmosis. We now address petitioners’ remaining argument.
(3) Cost
API’s third and final contention regarding the 1983 color limitations is that EPA has failed to “balance the benefits of color removal against its costs.” A subsidiary argument advanced by petitioners is that color is a matter of aesthetics and there is no evidence that it is harmful to receiving waters. Our response to these allegations will be brief.
First, in part V of this opinion we have considered petitioners’ general allegations concerning the Agency’s cost analysis. Our examination of the pertinent portions of the Final Development Document, as outlined in part V, confirms our conclusion *207that the Agency has fully considered the costs of color removal. Moreover, API errs in believing that the Agency must perform a cost-benefit analysis of the 1983 color limitations. Section 304(b)(2)(B) mandates no such balancing for the 1983 limitations, nor does Section 306 require it for the new source performance standards. All that is required under these two sections is that the Administrator consider “the cost of achieving such effluent reduction” in establishing the respective standards. This the Agency has done.
Second, we reject, as did the Agency, the assertion by petitioners that although color may be objectionable on aesthetic grounds, it has no harmful effect on receiving waters. There is record support for the Agency’s views, as stated in comment No. 24, quoted in footnote 18, supra, that color is a major pollutant. The detrimental effects of color are amply documented in the record. (App. 310, 393, 502, 611, 634, 683, 3043) These are in addition to what the Agency has set out in Section VI of the Final Development Document, referred to in comment No. 24. We find this evidence persuasive and we reject the contention that there is no basis for the Agency's conclusions with respect to color.
Having considered all the challenges to the 1983 effluent limitations guidelines, we find no merit in any of them.
VIII. THE NEW SOURCE PERFORMANCE STANDARDS
EPA must establish new source performance standards (NSPS) which reflect the application of the “best available demonstrated control technology (BADCT)”.22 Section 306(a)(2) of the Act defines a “new source” as any facility discharging pollutants “the construction of which is commenced after the publication of proposed regulations prescribing a standard of performance under this section.” 33 U.S.C. § 1316(a)(2). Section 306(a)(1) defines a new source “standard of performance” as “a standard for the control of the discharge of pollutants which reflects the greatest degree of effluent reduction which the Administrator determines to be achievable through application of the best available demonstrated control technology, processes, operating methods, or other alternatives, including, where practicable, a standard permitting no discharge of pollutants.” 33 U.S.C. § 1316(a)(1). In establishing NSPS under Section 306(b)(1)(B), the Administrator is to 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). He is also required to revise such standards “from time to time, as technology and alternatives change" under procedures outlined in this same section.
The distinction drawn by Congress between new sources and existing sources reflected a belief that new pollution-producing facilities could attain discharge levels at less cost and with fewer difficulties than existing sources. As stated by Senator Muskie in his discussion of the Conference agreement on Section 306:
In order to assure that a reasonable cost test is met, the Conference agreement clarifies the fact that the Administrator must take into account the cost of compliance with any new source performance standards as applied to any category or class of new sources. The Conferees would expect that this cost test would be considerably more restrictive than the test which would be applied to “best available technology” because pollution control alternatives are available to a new source which are not available to existing sources.
It may be that in most instances, the technology for elimination of discharge of pollutants from new sources can be achieved on a considerably more reasonable basis than for existing sources. The Conferees intend that this alternative be examined carefully and each determination of standards applicable to any cate*208gory of new sources be periodically re-examined by the Administrator to insure that any new source constructed does the best that can be done in terms of performance.
Leg.Hist. at 172. ■
The NSPS established by the Administrator regulate the discharge of the same four pollutants controlled under the 1983 BA-TEA standard: BOD5, TSS, pH and color. The NSPS for BOD5, TSS and pH are more stringent than the 1977 BPCTCA limitations but less so than the 1983 BATEA limitations. The NSPS for color apply to only two subcategories — unbleached kraft and unbleached kraft-NSSC (cross recovery) — and the standards are the same as the 1983 limitations for these subcategories. The following table sets out the New Source Performance Standards as promulgated by EPA.
New Source Performance Standards Values in kg/kltg (lbs/ton) ,«
BOD5 TSS
Subcategory 30 Day Daily Max 30 Day Daily Max
Unbleached Kraft 1.55(3.1) 3.1 ( 6.2) 3.75 ( 7.5) 7.5(15.0)
NSSC-Ammonia 3.75(7.5) 7.5(15.0) 3.75 ( 7.5) 7.5(15.0)
NSSC-Sodium 2.6 (5.2) 5.2(10.4) 3.85( 7.7) 7.7(15.4)
Unbleached Kraft-NSSC 1.9 (3.8) 3.8 ( 7.6) 4.0 ( 8.0) 8.0(16.0)
Paperboard from Waste Paper 0.75(1.5) 1.5 ( 3.0) 2.0 ( 4.0) 4.0 ( 8.0)
Color
Subcategory 30 Day Daily Max
Unbleached Kraft 10 (20) 15 (30)
NSSC-Ammonia - -
NSSC-Sodium - -
Unbleached Kraft-NSSC 12.5 (25) 25.0 (37.5)23
Paperboard from Waste Paper - -
pH for all subcategories shall be within the range of 6.0 to 9.0
API presents three challenges to these NSPS. The first is that EPA’s consideration of the economic, energy and environmental costs of the technology prescribed for new sources has been inadequate. We have addressed this contention in part V of this opinion. No further discussion is necessary.
API’s second challenge is that in setting NSPS for color EPA has (1) failed to give adequate consideration to petitioners’ comments concerning color regulation (2) set NSPS for color removal which cannot be met by the prescribed technology of lime treatment, and (3) been deficient in its cost analysis for color removal. If these arguments have a familiar ring it is because they are the same arguments raised by API in regard to the 1983 color limitations. In fact, petitioners have reproduced verbatim in their brief challenging the new source performance standards the color arguments they made in their brief challenging the 1983 color limitations guidelines. Because the NSPS for color for the unbleached kraft and unbleached kraft-NSSC (cross recovery) subcategories are the same as the 1983 effluent limitations guidelines for these two subcategories, and because petitioners’ arguments are identical, our discussion of these contentions in part VII of this opinion applies to petitioners’ allegations here. Accordingly, we uphold the NSPS for color for the reasons set out in our discussion of the validity of the 1983 color limitations.
API’s third challenge is directed at the NSPS for TSS. Petitioners say these are invalid because the Agency has failed to specify technology capable of removing “nonsettleable biological solids” and has not stated what “percentage or reduction” of this material can be achieved by the operations it does describe. As noted in an earlier part of this opinion24 we have accepted as valid the Agency’s answer to these allegations, rejected API’s “percentage reduction” argument, and found unsound petitioners’ method of framing an argument on the basis of definitions not employed by the agency and not found in the record. We see no need to add to what we have already said, especially because petitioners have once again reproduced here almost verbatim the argument made in their brief challenging the 1977 and 1983 guidelines.
*209We believe it appropriate to make one further comment. The extent to which the NSPS for TSS are more stringent than the 1977 limitations for this pollutant is a result of the lower raw waste levels deemed achievable by new sources. The internal technologies needed to attain the raw waste loads for new sources are described in Section VII of the Final Development Document. (App. 2317-2408) The application of these technologies to the NSPS is explained in Sections X and XI at pages 240 and 249-251 of the Final Development Document. (App. 2466, 2475 — 77) Because the new source performance standards for TSS are based on current performance levels of external technology and available internal technology, the standards are reasonable and we find them valid.
Having considered petitioners’ three contentions regarding the invalidity of the NSPS and having found no merit in any of them, we uphold as valid these regulations as promulgated by EPA under the BADCT standard.
IX. CONCLUSION
This case has presented complex and intertwined legal and technical challenges to regulations promulgated by the Administrator of the Environmental Protection Agency pursuant to the Federal Water Pollution Control Act Amendments of 1972. We have sought to steep ourselves in the knowledge necessary to assure us that the Agency has exercised a reasoned discretion in establishing the regulations. On the record as a whole we believe the Agency has indeed exercised a reasoned discretion. Accordingly, we uphold the 1977 and 1983 effluent limitations guidelines for existing sources and the standards of performance for new sources for this phase of the Pulp, Paper and Paperboard Point Source Category.
The judgment of the District Court in No. 74-1967 is affirmed. The petitions for review are dismissed.
So Ordered.
. Boise Cascade Corporation, Georgia-Pacific Corporation, International Paper Company, Mead Corporation, Owens-Illinois Incorporated, Packaging Corporation of America, Westvaco, and Weyerhaeuser Company. API is a non-profit trade association.
. 40 C.F.R. § 430 (1975).
. A “point source” is defined as “any discernible, confined and discrete conveyance, including but not limited to any pipe, ditch, [etc.], from which pollutants are or may be discharged.” Section 502(14), 33 U.S.C. § 1362(14).
. For an introduction to the Federal Water Pollution Control Act as seen by the present General Counsel of EPA see Zener, The Federal Law of Water Pollution Control in Federal Environmental Law 683 (E. Dolgin and T. Guilbert eds. 1974).
. The Phase II regulations concern the bleached segments of the pulp and paper industry. They are not at issue in this litigation.
. Congressional Research Service, A Legislative History of the Water Pollution Control Act Amendments of 1972, at 1468 (1973) (hereinafter "Leg. Hist.”)
. According to EPA, approximately 16 additional subcategories will be included in the Phase II segment of the pulp and paper industry. Brief for Respondent at 37.
. 33 U.S.C. § 1311(b)(1)(A)
. See Section 304(b)(1) (“control measures and practices”). (Emphasis added.) The Senate Consideration of the Report of the Conference Committee contains the following statement by Senator Muskie: “ ‘[b]est practicable’ can be interpreted as the equivalent of secondary treatment for industry, but this interpretation should not be construed to limit the authority of the Administrator.” Leg.Hist, at 170. (Emphasis added.) See also Leg.Hist. at 1468.
. Petitioners also raise this same argument regarding the 1983 effluent limitations. Our discussion here is a response to both challenges.
. Plants which discharge to municipal treat-ment systems are not subject to effluent limitations under Sections 301 and 304.
. “NCASI” stands for the National Council of the Paper Industry for Air and Stream Improvement. (App. 1728)
.As noted in the Final Development Document, the TSS limitations were based on Mill UK-1 (6.1 (12.1)) and not on the average of the best performers’ TSS levels (4.8 (9.7)). The less stringent limitation was chosen because a portion of the treatment system used by some of the best performers required very large land areas not always available at all unbleached kraft mills. (App. 2450-2452)
. Section X of the Final Development Document. (App. 2463-2472)
. Petitioners raise this same argument regarding the new source performance standards. Our discussion here is a response to both challenges.
. 33 U.S.C. § 1311(b)(2)(A).
. This figure appears in the regulations as promulgated. In the Final Development Document the figure is 18.75 (37.5). There is no explanation for the discrepancy.
. (24) It was suggested by several commenters that the color limitation should be removed from the regulations as color has not been established as a pollutant and its inclusion for BATEA and NSPS is not justified.
As discussed in Section VI of the Development Document, the Agency believes that color is a major pollutant parameter and has the following detrimental effects: (1) Color in receiving waters retards sunlight transmission and interferes with photosynthesis thereby reducing productivity of the aquatic community; (2) color alters the natural stream color and is thereby aesthetically displeasing; (3) color has a detrimental effect upon downstream municipal and industrial water users, as color, even when not visually apparent (i. e. turbid streams), must be removed before use in municipal and industrial water supplies; (4) color bodies complex with metal ions, such as iron or copper, forming tar-like residues which remove metals from the stock available to stream organisms for normal metabolism, and the complexes can have direct inhibitory effects on some of the lower scale of organisms in the aquatic community; (5) color is an indicator of potentially inhibitory compounds discharged to the aquatic environment; and (6) color in receiving waters affects fish productivity and fish movements. Therefore, the limitations for color for BATEA and NSPS were not removed from the regulations as the Agency believes that color is a major pollutant parameter.
. Comment No. 9 is reproduced in the text infra at pages 50-51.
. (22) Many comments were received that stated that the technology of lime treatment for removal of color has not been fully demonstrated. The commenters stated that a large number of operating problems continue to give difficulty in achieving adequate consistent color removal levels and that specifically, increases of color following color removal during biological treatment are being experienced at two mills employing color removal systems. It was suggested that since the lime treatment technology is still developing, the color removal limitations for new sources should be removed.
The Agency believes that the lime treatment process has been satisfactorily demonstrated and can be applied to new mills. It is the Agency’s judgment that the operating problems being experienced by the two full-scale installations are inherent to the specific mills involved. For example, the problem of increases in color levels through the biological treatment systems are not indicative of problems of the lime treatment technology. These are a result of the specific biological systems involved, as color is being leached into the waste water from the ground. In addition, the Agency feels that many of the operating problems can be solved by new mills by being able to design the color removal system into the total mill design rather than adding on the unit operation to an existing mill.
. For two subcategories (NSSC-Ammonia and NSSC-Sodium) the 1983 color limitations are expressed in the form of percentage removals. As explained in the Final Development Document, the Agency selected this particular method because of the stage of the development of the identified technology. These particular limitations “will be changed to kilograms of color per metric ton of production (pounds of color per short ton of production) at a later date when the technology has been proven through further development.” Final Development Document at 239. (App. 2465)
. 33 U.S.C. § 1316(b)(1)(B)
. See n. 17, supra.
. See n. 15, supra.