Solite Corporation v. U.S. Environmental Protection Agency, and William K. Reilly, Administrator, Epa, the Aluminum Association, Intervenors

952 F.2d 473, 293 U.S. App. D.C. 117

• Court: Court of Appeals for the D.C. Circuit
• Decided: February 26, 1992
• Docket: 89-1629, 89-1665, 89-1696, 89-1722, 89-1724, 89-1727 to 89-1729, 89-1731, 89-1732, 90-1029, 90-1086, 90-1125, 90-1149, 90-1195, 90-1198, 90-1200, 90-1203, 90-1206, 90-1207 and 90-1216
• Status: Published

Opinion

PER CURIAM:

CONTENTS

Page

I. Background. 477

II. Discussion. 482

A. Standard of Review. 482

B. High Volume Criteria. 483

1. Congressional Intent. 483

2. APA Compliance. 484

3. Methodological Objections. 485

(a) Consideration of Other Special Wastes. 485

(b) Overly Stringent High Volume Measurement. 486

C. Low Hazard Criterion. 488

D. The 50 Percent Rule. 490

E. Future Waste Streams. 491

F. Application of Section 3004(x) to Non-BeviU Wastes. 492

G. The Mixture Rule. 493

H. The Treatment Permit Requirement. 494

I. Du Pont’s Chloride-Ilmenite Process. 494

J. Chrome Tailings. 495

1. Background. 495

Page

2. Analysis. 496

K. Lead Process Wastewater. 497

1. Background. 497

2. Analysis. 498

L. Lightweight Aggregate Air Pollution Control Dust/Sludge. 498

1. Background. 498

2. Analysis. 499

III. Conclusion. 500

In Environmental Defense Fund v. EPA, 852 F.2d 1316 (D.C.Cir.1988), cert. denied, 489 U.S. 1011, 109 S.Ct. 1120, 103 L.Ed.2d 183 (1989) {“EDF II”), this court ordered the Environmental Protection Agency (“EPA”) to determine the ore and mineral processing wastes that qualify under the Bevill Amendment “mining waste exclusion,” 42 U.S.C. § 6921(b)(3)(A)(ii), for exemption, at least temporarily, from the hazardous waste management regime of Subtitle C of the Resource Conservation and Recovery Act (“RCRA”), 42 U.S.C. §§ 6921-6939b. 1 This case is the follow up to that 1988 decision. Upon the EPA’s completion of the rulemaking we ordered in EDF II, members of the mineral and chemical processing industry, urging broader interpretation of the Bevill Amendment mining waste exemption from regulation, petitioned for our review.

We uphold the EPA’s determinations in principal part. On two matters, however, we conclude that reconsideration by the Agency is necessary: (1) we vacate EPA’s determination that lightweight aggregate residuals do not meet the Agency’s high volume criterion, for the record shows that EPA decided that matter without providing due notice and opportunity to comment; (2) we remand for EPA to provide a more adequately reasoned explanation of its denial of the Bevill Amendment exclusion for the wastes from Du Pont’s chloride-ilmenite process. Finally, we remand without opinion two aspects of this rulemaking that EPA based upon the Subtitle C mixture rule, which we vacated recently in another case.

I. Background

Subtitle C of RCRA, enacted in 1976, requires EPA to develop a comprehensive regulatory scheme for the treatment, storage, and disposal of hazardous wastes. See 42 U.S.C. §§ 6921-6931. Section 8002 of RCRA, 42 U.S.C. § 6982, directs EPA to conduct special studies and research on certain categories of waste, including mining waste. See 42 U.S.C. § 6982(f) (requiring EPA to conduct “a detailed and comprehensive study” of adverse environmental effects of “solid wastes from active and abandoned surface and underground mines,” addressing adequacy of current means of disposal and utilization of such wastes). In its first proposal for regulations governing hazardous waste management under Subtitle C, EPA noted a category of “special waste,” for which “special standards” might be appropriate. 43 Fed. Reg. 58,946, 58,992 (1978). EPA’s “limited information” indicated that “such waste occurs in very large volumes, that the potential hazards posed by the waste are relatively low, and that the waste generally is not amenable [to the Subtitle C controls developed for industrial and manufacturing wastes].” Id. at 58,991-92. EPA identified, as one category of special waste, hazardous wastes “from the extraction, benefi-ciation, and processing of ores and minerals.” Id. at 59,016. 2

In final regulations released in May 1980, EPA identified the characteristics of hazardous waste and listed as subject to Subtitle C regulation specific hazardous wastes. 45 Fed.Reg. 33,084 (1980). EPA concluded that the less inclusive definition of hazard and more relaxed regulatory requirements adopted in the final rule “accomplish[ed] the objective of, and eliminate[d] the need for, a special solid waste category.” Id. at 33,174. Consequently, EPA’s final regulatory scheme abandoned the concept of special waste, and would have subjected all wastes meeting EPA’s modified hazard criteria to Subtitle C regulation. Id. at 33,-175.

A month before the Subtitle C regulations were to take effect, Congress altered EPA’s course by enacting the Bevill Amendment as part of the Solid Waste Disposal Act Amendments of 1980. In pertinent part, the Bevill Amendment directed EPA to study any adverse health and environmental effects “of solid waste from the extraction, beneficiation, and processing of ores and minerals,” and to submit a report to Congress by October 21, 1983. 42 U.S.C. § 6982(p). 3 The Amendment further required EPA, within six months of reporting to Congress, and after public hearings and an opportunity for comment, either to “determine to promulgate regulations” governing the subject wastes, or to “determine that such regulations are unwarranted.” 42 U.S.C. § 6921(b)(3)(C). Pending pursuit and completion of the mining waste studies and, thereafter, until EPA’s final regulatory determination, the Bevill Amendment prohibited the Agency from regulating mining and mineral processing wastes as hazardous wastes within the compass of Subtitle C. 42 U.S.C. § 6921(b)(3)(A)(ii).

In November 1980, EPA amended its hazardous waste regulations to incorporate the Bevill Amendment’s exemption of mining wastes from Subtitle C controls. 45 Fed.Reg. 76,618 (1980). The regulatory amendment tracked the statutory language, exempting from Subtitle C regulation “[sjolid waste from the extraction, be-neficiation and processing of ores and minerals.” Id. at 76,620; 42 U.S.C. § 6921(b)(3)(A)(ii). As “an immediate but temporary accommodation” of requests from the mining industry for “clear guidance” on the scope of the exemption, EPA announced an interpretation of the Bevill Amendment “to include solid waste from the exploration, mining, milling, smelting and refining of ores and minerals.” 45 Fed.Reg. at 76,618-19. The Agency questioned whether Congress intended so broad an exclusion from regulation, however, and announced that it would review the legislative history of the Bevill Amendment and consider public comments to determine definitively the proper scope of the exemption. Id. at 76,618-19.

EPA did not meet the October 1983 statutory deadline to complete the Bevill study and report to Congress. Consequently, all mining wastes remained exempt from Subtitle C regulation. In September 1984, the Environmental Defense Fund and two citizens groups sued the Agency for failure to complete the studies and make the regulatory determinations as required by law. See Concerned Citizens of Adamstown v. EPA, No. 84-3041 (D.D.C. Aug. 21, 1985) (discussed in EDF I, 852 F.2d at 1311, and EDF II, 852 F.2d at 1321). EPA, in response to the litigation, indicated its intention “to reinterpret the mining waste exclusion to remove certain smelting and refining {i.e., processing] wastes ... from its scope.” See EDF II, 852 F.2d at 1321. The district court granted summary judgment for the Adamstown plaintiffs, and imposed on EPA a timetable for the Agency’s reinterpretation of the mineral pro cessing waste category, completion of the required study, and rendition of the regulatory determination for processing wastes. See EDF II, 852 F.2d at 1321. 4

In compliance with the court order in Adamstown, EPA proposed to limit its definition of ore and mineral processing wastes, for Bevill Amendment purposes, to those wastes meeting the “special waste” criteria first mentioned in EPA’s 1978 proposed Subtitle C regulations. 50 Fed.Reg. 40,292, 40,294 (1985). EPA later withdrew its proposed reinterpretation, however, because the proposal “did not set out practically-applicable criteria for distinguishing processing from non-processing wastes.” 51 Fed.Reg. 36,233, 36,235 (1986).

In EDF II, this court, again on the petition of the Environmental Defense Fund, held that “EPA’s decision to withdraw its proposed [processing wastes] reinterpretation in its entirety was arbitrary and capricious and contrary to law because it ... reaffirmed an impermissibly over-broad interpretation of the Bevill Amendment.” EDF II, 852 F.2d at 1326. Reviewing EPA’s interpretation of the Bevill Amendment under the analytical framework of Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), the court first determined that “the statutory term ‘processing’ does not on its face admit of a standard definition, and ... the precise meaning of that term is not fully apparent from the structure of the statute.” EDF II, 852 F.2d at 1327. The court therefore turned to the legislative history of the Be-vill Amendment “for more precise guidance on Congressional intent.” Id.

Upon reviewing the legislative history of the Bevill exclusion from regulation, the court found the guidance it sought. The EDF II panel concluded from the legislative history that “the key to understanding Congress’s intent is the concept of ‘special waste’ articulated in the regulations proposed by EPA on December 18, 1978 following the enactment of RCRA.” Id. (citing 43 Fed.Reg. 58,991-92 (1978); 50 Fed. Reg. 40,293-94 (1985)). From the remarks of Congressman Bevill and other legislators on the House floor when the Bevill Amendment was introduced, and the Conference Committee Report accompanying the 1980 amendments to RCRA, 5 the court found it

clear that Congress did not intend the mining waste exclusion to encompass all wastes from primary smelting and refining. On the contrary, Congress intended the term “processing” in the Bevill Amendment to include only those wastes from processing ores or minerals that meet the “special waste” criteria, that is, “high volume, low hazard” wastes.

Id. at 1328-29.

In granting relief, the court imposed on EPA a schedule “for fulfilling its statutory obligations with respect to processing wastes under the Bevill Amendment.” Id. at 1331. The court ordered EPA to propose, and, after notice and comment, determine “which processing wastes remain within the Bevill exclusion as ‘high volume, low hazard’ special wastes.” Id. As to the wastes so identified, the court further directed EPA to complete the study required by the Bevill Amendment, report to Congress, and make a regulatory determination, within six months of the submission of EPA’s report to Congress, with respect to those wastes. Id.

Following EDFII, EPA issued three notices of proposed rulemaking (“NPRMs”), culminating in two final rules. In the first NPRM, published October 20, 1988 (“10/88 NPRM”), EPA proposed definitional criteria to identify mineral “processing” operations, and quantitative measures to identify “high volume” wastes. 53 Fed.Reg. 41,-288, 41,290-94 (1988). To develop a high volume criterion, EPA evaluated the annual generation rates of individual mineral processing waste streams in metric tons on both a plant-specific and an industry-wide basis. Id. at 41,292-94. The Agency quantified the high volume measure by examining the generation rates of mineral extraction and beneficiation wastes, oil and gas wastes, and wastes currently managed under Subtitle C. Id. at 41,293-94. The proposed volumetric threshold of 50,000 metric tons per year (“mt/yr”) per facility was calculated to retain within Bevill coverage mineral processing wastes generated in volumes comparable to the largest volume wastes managed at the top ten percent of Subtitle C facilities. Id. at 41,294.

EPA declined at that stage to establish a “low hazard” standard. Id. at 41,294-95. Applying the proposed criteria, EPA identified 15 wastes meeting the “processing” definition and “high volume” measure, and solicited püblic comment on its proposed approach. Id. at 41,288, 41,296.

In a second NPRM, published April 17, 1989 (“4/89 NPRM”), EPA revised the “special waste” criteria and modified the list of mineral processing wastes proposed for retention within the Bevill mining waste exclusion. 54 Fed.Reg. 15,316 (1989). The Agency proposed to establish separate high volume thresholds for “solid and acqueous liquid waste streams,” to account for large differences noted by EPA, i.e., Subtitle C facilities, the Agency observed, typically managed liquids in far larger volumes than solids. Id. at 15,331. EPA retained the 50,000 mt/yr standard for solid wastes and proposed a 1.5 million mt/yr standard for liquid wastes. Id.

Further augmenting the initial notice, EPA’s 4/89 NPRM proposed a “low hazard” criterion to screen out from Bevill Amendment coverage mineral processing wastes that are “clearly not low hazard.” Id. at 15,331. The low hazard standard was intended to “remove from the [Bevill] exclusion wastes that may pose risk or hazard to such a great extent or magnitude that they cannot be considered ‘special wastes,’ irrespective of volume.” Id. at 15,322. EPA proposed to test all wastes meeting the “processing” definition and “high volume” criteria for corrosivity, using a pH test, and for toxicity and mobility, using a newly developed procedure called Method 1312. Id. at 15,339-40. Any high volume processing waste that failed either proposed test would be ineligible for Bevill Amendment exemption from Subtitle C regulation. Id. at 15,340.

Because of limited data, EPA was then unable to apply the new hazard criteria to most high volume processing wastes. Id. The Agency planned to promulgate a final rule in September 1989 to establish the special waste criteria and take final action on six wastes proposed for Bevill coverage and three wastes proposed for elimination from that category. Id. at 15,317. The Agency conditionally retained an additional 33 wastes that satisfied the “processing” and “high volume” criteria, but for which EPA lacked sufficient data to test for hazard. Id. at 15,342. EPA planned to test those wastes and projected that in January 1990 it would finally determine their Bevill Amendment status. Id.

In a final rule published in September 1989 (“9/89 Rule”), EPA settled on the “special waste” criteria for identifying Be- vill mineral processing wastes. 54 Fed. Reg. 36,592 (1989). As proposed in the 4/89 NPRM, EPA established separate high volume criteria for liquid and non-liquid wastes. Id. at 36,607. In quantifying the high volume cut-offs, EPA replaced as its source of data on Subtitle C facilities its 1985 National Biennial Report of Hazardous Waste Generators and Treatment, Storage, and Disposal Facilities Regulated Under RCRA (Draft) (“1985 Biennial Report”), used in both the 10/88 and 4/89 NPRMs, with its newly available National Survey of Hazardous Waste Treatment, Storage, Disposal, and Recycling Facilities (“TSDR Survey”). Id. at 36,608. The TSDR Survey, containing “detailed information about volumes and specific types of wastes generated and managed at Subtitle C facilities during calendar year 1986[,] ... allowed EPA to conduct a waste stream-level analysis of current management practices and hazardous waste volumes managed at facilities regulated under [S]ubtitle C of RCRA.” Id. Using this more recent and complete data, EPA

examined individual waste-code data for waste streams entering Subtitle C landfills to develop a revised criterion for solid/sludge materials, and for waste streams entering wastewater treatment processes, surface impoundments, and injection wells to develop a cut-off value for liquid waste streams. The final criterion values reflect the largest single waste code managed at the 95th percentile of the Subtitle C facilities employing these hazardous waste management techniques.

Id.; see also Development of the High Volume Criterion for Mineral Processing Wastes (Aug. 18, 1989), reproduced in Joint Appendix (“JA”) at 171-73. This methodology produced a final high volume criterion of 45,000 mt/yr per facility for non-liquid wastes and one million mt/yr per facility for liquid wastes. 54 Fed.Reg. at 36,594.

The final low hazard screening prescription established a maximum level of mobility and toxicity for both liquid and non-liquid wastes, and a range of pH corrosivity, applicable only to liquid wastes, as the outer limits beyond which a waste was considered “clearly not low hazard” and thus ineligible for Subtitle C exclusion under the Bevill Amendment. Id. at 36,596, 36,630-31. The final toxicity/mobility standard, which employed EPA’s newly developed Method 1312, excluded from Bevill coverage any waste containing hazardous constituents at a level greater than 100 times the maximum contaminant level (“MCL”) in EPA’s primary drinking water standards. Id. at 36,630. The final corro-sivity test placed outside Bevill coverage any waste with a pH of less than one or greater than 13.5. Id. at 36,631.

EPA applied these final “special waste” criteria to determine the Bevill status of 23 mineral processing wastes: five wastes were retained within the Bevill Amendment and 18 were ruled outside Bevill’s scope. Twenty wastes that appeared to satisfy the “processing” definition and high volume standard were conditionally retained within the Bevill exclusion, pending accumulation of further data necessary to apply the low hazard criterion. Id. at 36,631-32.

In a third NPRM issued in mid-September 1989 (“9/89 NPRM”), EPA applied the final “special waste” criteria to the 20 mineral processing wastes conditionally retained for Bevill coverage in the 9/89 Rule, and proposed their final Bevill status. 54 Fed.Reg. 39,298 (1989). In a rule published in January 1990 (“1/90 Rule”), EPA made final the Bevill status of the 20 wastes, removing five from Bevill coverage and retaining the remaining 15. 55 Fed.Reg. 2322, 2342 (1990). Thus, EPA designated a total of 20 mineral processing wastes — five identified in the 9/89 Rule and 15 identified in the 1/90 Rule — for continued temporary exemption from Subtitle C regulation, pending the Agency’s detailed study, report to Congress, and ultimate regulatory determination under the Bevill Amendment. Id. 6 Any wastes ruled outside Bevill’s scope would be subject to Subtitle C regulation if found to exhibit one or more of the hazardous characteristics identified by EPA in 1980 pursuant to RCRA. See 54 Fed.Reg. at 36,595.

In these consolidated petitions for review, a number of mineral and chemical production companies, several associations representing mining and chemical industry interests, and the Mayor and City Council of Baltimore (collectively “petitioners”) challenge on numerous grounds EPA’s 9/89 and 1/90 Rules. Essentially, petitioners argue that EPA’s “special waste” criteria subvert congressional intent by excluding all but a few mineral processing wastes from the waste-specific study and regulatory determination required under the Be-vill Amendment. Petitioners charge that EPA has violated Congress’ statutory directive by paring down the scope of the Bevill study, thereby assigning RCRA regulatory status to allegedly Bevill wastes from mineral processing. See Joint Brief of Consolidated Petitioners and Intervenors (“Jt. Brief”) at 11 (summary of argument).

In the main, we find that EPA’s rulemak-ing is consistent with this court’s decision in EDF II and reflects a permissible interpretation of the Bevill Amendment. We therefore affirm EPA’s determinations in principal part.

II. Discussion

A. Standard of Review

In line with this court’s opinion in EDF II, EPA reviewed the administrative development of the “special waste” concept during the late 1970s. See 54 Fed.Reg. at 15,318-19. The Agency examined closely a key background document, completed in the fall of 1979, that (1) identified the criteria EPA used to develop the original list of “special wastes” in the 1978 proposed RCRA regulations, and (2) refined those criteria in response to comments on the 1978 proposal. Id.; see Draft Background Document, Introduction and Criteria for Special Waste (Nov. 2, 1979) (“Special Waste Background Document”), portions reproduced in JA at 200-07, JA Supp. at 593. EPA found it clear that, in 1978, the Agency had in view “several specific criteria (principally low hazard, high volume, and infeasibility of Subtitle C technical requirements) ... to evaluate potential special wastes”; moreover, the Agency then anticipated that “the group of wastes [eligible for] the temporary exemption from full Subtitle C regulation [would] be both finite and relatively small.” 54 Fed.Reg. at 15,-319.

Reiterating that “high volume and low hazard” are “the key elements in identifying special wastes,” EPA recalled the function these factors play in determining which wastes are appropriate for temporary, and possibly permanent, exclusion from Subtitle C regulation under the Bevill Amendment:

High volume is the principal indicator of whether a particular waste is amenable to management under Subtitle C of RCRA. A consideration of hazard is necessary to identify and remove from the exclusion wastes that may pose risk or hazard to such a great extent or magnitude that they cannot be considered “special wastes,” irrespective of volume.

Id. at 15,322. EPA observed, however, that neither the legislative history of the Bevill Amendment, nor the regulatory background of the “special waste” concept, precisely defines the terms “high volume” or “low hazard.” See id.; see also 54 Fed.Reg. at 36,595. Nor do these terms “convey a self-evident, accepted meaning.” See EDF II, 852 F.2d at 1316.

Because Congress has not “directly addressed the precise question at issue,” that is, the appropriate delineation of the “special waste” concept, and the attendant “high volume” and “low hazard” criteria, the question on review is “whether the agency’s answer is based on a permissible construction of the statute.” Chevron, 467 U.S. at 843, 104 S.Ct. at 2782. In EDF II, we held that “Congress intended the Bevill exclusion to encapsulate the ‘special waste’ concept articulated by the EPA in 1978,” 852 F.2d at 1329, and we review EPA’s rules, with the deference mandated by Chevron, for conformity to the statute so understood.

B. High Volume Criteria

EPA understood that Congress intended the Bevill exclusion to encompass “only those waste streams that are generated in such quantities as to be potentially unmanageable under [SJubtitle C regulations.” 54 Fed.Reg. at 36,611. Consistent with Agency documents prepared in conjunction with EPA’s 1978 proposal on special wastes, the Agency viewed volume as “the most relevant and objective measure” of the technical feasibility of subjecting waste to Subtitle C controls. See 54 Fed. Reg. at 15,326; see also id. at 15,318 (quoting Special Waste Background Document (“apparent technological difficulty in applying [proposed Subtitle C] regulations to the waste because of volumes involved at typical facilities” was a “driving force” in developing the original list of “special wastes”)). EPA logically concluded that “comparison of mineral processing waste volumes with those of wastes managed under Subtitle C controls” was the most “appropriate analytical basis for developing the high volume criterion.” 54 Fed.Reg. at 36,611.

The Agency consequently devised a methodology to calculate “volumetric cutoffs ... reflecting] some of the largest quantities of individual and identifiable waste streams managed at facilities that are currently in the Subtitle C regulatory system.” Id. at 36,629. EPA examined the generation rates of Subtitle C wastes managed at facilities disposing of solid/sludge materials in landfills, or employing wastewater treatment processes, surface impoundments, and injection wells to manage hazardous liquid wastes — disposal techniques that would be used to manage mineral processing wastes under Subtitle C. Id. at 36,608, 36,629. The Agency identified the largest individual waste stream managed by such techniques at each facility, and “computed univariate statistics on the resulting distribution.” Id. at 36,629. The final high volume criteria — 45,000 mt/yr per facility for solids, one million mt/yr per facility for liquids — represent “the largest individual waste stream managed by the facility at the 95th percentile of the relevant distribution.” Id.

Beyond question, a high volume standard for Bevill mineral processing wastes is in order. Petitioners do not argue otherwise; rather, they challenge the particular high volume criteria established by EPA on three grounds. First, petitioners charge that the selected criteria frustrate the congressional intent underlying the Bevill Amendment. Next, they complain that, in violation of the Administrative Procedure Act (“APA”), 5 U.S.C. § 553, EPA failed to provide adequate notice and opportunity to comment on the data and methodology used in developing the high volume standards. Finally, petitioners assert that EPA’s method of calculating the quantitative volumetric thresholds was improperly biased toward overly stringent standards and was therefore arbitrary and capricious.

We conclude that EPA’s high volume criteria are consistent with a permissible interpretation of the Bevill Amendment, were developed in accord with the APA, and reflect methodological choices properly within the Agency’s discretion.

1. Congressional Intent

Petitioners argue that EPA’s “unreasonably high volumetric thresholds ... transformed what was intended by Congress to be a screening analysis into a final method for determining the regulatory status of mineral processing wastes.” Jt. Brief at 31. Petitioners challenge EPA’s “false equation of ‘high volume’ and ‘non-amenability’ ” as without “basis in the language or legislative history of the Bevill Amendment.” Jt. Reply at 5-6 & n. 8. Congress intended EPA to determine the amenability of mineral processing wastes to Subtitle C regulation through the multi-factored analysis required by the Bevill Amendment, petitioners assert, not through the application of a rigid high volume screening criterion. Id. at 7-8.

For a waste within its scope, the Bevill Amendment no doubt requires an in-depth, multi-factored, context-specific study and regulatory determination. See 42 U.S.C. §§ 6921(b)(3), 6982(p). This court found it equally clear, however, that Congress did not intend to sweep all mineral processing wastes within Bevill Amendment coverage, but only those that satisfied threshold high volume, low hazard criteria to be determined by EPA. See EDF II, 852 F.2d at 1329. 7 We discern no departure from Congress’ instruction, or abuse of discretion, in EPA’s use of precise quantitative measures to make the threshold classifications, while reserving for wastes thus classed “Bevill” the closer study antecedent to determining the regime for their regulation. Petitioners apparently would have the Agency make a multi-factored analysis twice over. We see no legislative command for the scheme petitioners ask us to order.

2. APA Compliance

Petitioners accuse EPA of having “performed a classic ‘bait-and-switch’ ” by “substituting at the last moment one set of data and analysis for another” when the Agency replaced the 1985 Biennial Report with the TSDR Survey as the source of data on which the final quantitative volumetric measures were based. Jt. Brief at 13. In petitioners’ view, EPA thus deprived the public of the notice and opportunity to comment guaranteed by the APA. See 5 U.S.C. § 553. We conclude that the EPA’s high volume criteria do not fail on notice and comment grounds.

The APA requires that a notice of proposed rulemaking include “either the terms or substance of the proposed rule or a description of the subjects and issues involved,” 5 U.S.C. § 553(b), and that the agency “give interested persons an opportunity to participate in the rulemaking through submission of written data, views, or arguments.” 5 U.S.C. § 553(c); see Air Transport Ass’n v. CAB, 732 F.2d 219, 224 (D.C.Cir.1984). Integral to the notice requirement is the agency’s duty “to identify and make available technical studies and data that it has employed in reaching the decisions to propose particular rules.... An agency commits serious procedural error when it fails to reveal portions of the technical basis for a proposed rule in time to allow for meaningful commentary.” Connecticut Light and Power Co. v. NRC, 673 F.2d 525, 530-31 (D.C.Cir.), cert. denied, 459 U.S. 835, 103 S.Ct. 79, 74 L.Ed.2d 76 (1982).

At the same time, consistent with the APA, an agency may use “supplementary” data, unavailable during the notice and comment period, that “expand[s] on and confirm[s]” information contained in the proposed rulemaking and addresses “alleged deficiencies” in the pre-existing data, so long as no prejudice is shown. Community Nutrition Institute v. Block, 749 F.2d 50, 57-58 (D.C.Cir.1984); see also Air Transport Ass’n, 732 F.2d at 224 (APA requirements met despite agency’s reliance on internal staff studies unavailable to public before rule’s adoption, because “critical elements” of proposal did not change and “final rule was a ‘logical outgrowth’ of the proposed rule”) (quoting Small Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d 506, 547 (D.C.Cir.1983)).

Petitioners do not point to inaccuracies in the data contained in the TSDR Survey. See Community Nutrition Institute, 749 F.2d at 58 (no prejudice from agency’s response to comments in form of “new scientific studies” where petitioners did “not even suggest that the new studies were defective”). Nor does the record suggest that EPA hid or disguised the information it used, or otherwise conducted the rulemaking in bad faith. Cf Connecticut Light and Power Co., 673 F.2d at 531 (“To allow an agency to play hunt the peanut with technical information, hiding or disguising the information that it employs, is to condone a practice in which the agency treats what should be a genuine interchange as mere bureaucratic sport.”).

Rather, petitioners’ complaint is with the way EPA used the data to calculate volumetric thresholds for Bevill status qualification of solid and liquid processing wastes. EPA’s methodology in selecting the high volume criteria, however, did not change significantly from the proposed notices to the final rule, and petitioners had ample opportunity to criticize EPA’s approach. In fact, the updated and expanded data in the TSDR Survey enabled EPA to respond to several industry objections to use of the data in the 1985 Biennial Report; by providing a more precise quantitative measure based on more complete information, the TSDR Survey confirmed the prior estimate of the upper bounds of waste volumes currently managed under Subtitle C controls. See Community Nutrition Institute, 749 F.2d at 58 (“It is impossible to perceive why correction of an asserted deficiency in earlier studies— which correction confirms the accuracy of those studies — should give rise to an additional opportunity to comment.”); cf. American Iron & Steel Inst. v. OSHA, 939 F.2d 975, 1009-1010 (D.C.Cir.1991) (OSHA violated APA by basing finding of economic feasibility on post-comment data apparently less reliable than, and inconsistent with, “all other important evidence in the record”).

In sum, because EPA’s methodology remained constant, and because the added data was used to check or confirm prior assessments, we hold that the Agency did not violate the notice and comment provisions of the APA by using the TSDR data.

3. Methodological Objections

This court’s review of EPA’s methodology in developing the high volume criteria is governed by the APA’s “arbitrary or capricious,” “abuse of discretion” standard. 5 U.S.C. § 706(2)(A). Under this familiar standard, we inquire “ ‘whether the [Agency’s] decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment.’ ” Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Automobile Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443 (1983) (citation omitted). The record need only demonstrate that the Agency met its obligation to “examine the relevant data and articulate a satisfactory explanation for its action.” Id.

(a) Consideration of Other Special Wastes

Petitioners complain that in developing the high volume criteria for mineral processing wastes, EPA neglected to evaluate the volumes of wastes clearly within the Bevill Amendment’s scope and thereby ignored “the only direct evidence of Congressional intent relating to the special waste concept.” Jt. Brief at 16. EPA did indeed base its high volume criteria principally on a comparison of volume data from regulated Subtitle C facilities, but the Agency did not ignore the volumes of wastes within Bevill’s compass. The rulemaking record documents the consideration given to the generation rates of such wastes. See 53 Fed.Reg. at 41,293-94; 54 Fed.Reg. at 15,-330; Internal EPA Memorandum dated Oct. 19, 1988 re: Quantitative Basis for High Volume Mineral Processing Waste Criteria, reproduced in JA at 241-48. Quantitative comparisons showed that mineral processing wastes proposed for continued Bevill status in the 10/88 NPRM

are generated in quantities that are comparable to many of the other special wastes, and that the lower bounds of the quantitative criteria approximate the lower limit of the quantities of waste generated by the other industry types contained within the Bevill exclusion.

EPA Mem., Oct. 19, 1988 at 1, JA at 241.

True, EPA rejected the suggestion to “use the lowest of extraction and beneficiation waste generation rates to establish the high volume threshold.” 54 Fed.Reg. at 36,611-12; see also 54 Fed.Reg. 15,330. As EPA reasoned, “Congress intended the Bevill exclusion to cover only those waste streams that are generated in such quantities as to be potentially unmanageable under [S]ubtitle C regulations”; therefore, the volumes of wastes currently managed under Subtitle C controls formed the “appropriate analytical basis for developing the high volume criterion.” 54 Fed.Reg. at 36,611; see also 54 Fed.Reg. 15,329. In EPA’s view,

comparisons with Subtitle C wastes are not only reasonable and appropriate, but necessary. Comparisons with other Be-vill wastes, on the other hand, do not provide conclusive evidence but do suggest boundaries on what might be considered a high volume special waste.

Id. at 15,329; see also id. at 15,330; 54 Fed.Reg. at 36,611. 8

Nothing in the legislative history of the Bevill Amendment indicates that Congress contemplated a specific volumetric threshold, or intended the generation rate of any particular Bevill waste to serve as the cutoff for mineral processing wastes to qualify for exemption from Subtitle C regulation. Rather, the legislative record, as the EDF II panel concluded, demonstrates only that “Congress intended the Bevill exclusion to encapsulate the ‘special waste’ concept articulated by the EPA in 1978.” EDF II, 852 F.2d at 1329. Given the absence of more precise instructions from Congress, the EDF II court properly left to EPA the task of setting criteria for determining which mineral processing wastes are special wastes. Id. at 1331.

EPA, we are satisfied, provided a reasonable explanation for its decision to base the high volume criteria on the volumes of waste generated and managed at Subtitle C-regulated facilities, and to use generation rates of other Bevill wastes as a “reality check” on the volume thresholds selected. See 54 Fed.Reg. at 15,330; 54 Fed.Reg. at 36,611.

(b) Overly Stringent High Volume Measurement

Petitioners charge that EPA’s data selection from the TSDR Survey and methodological choices “piled one conservative assumption atop another, resulting in high volume criteria that are so unreasonably high as to be arbitrary and capricious.” Jt. Brief at 18-19. Specifically, petitioners argue that by examining only facilities that use land disposal to manage solid wastes or treat hazardous wastewaters, ignoring small volume waste generators, and inelud-ing commercial waste management facilities, EPA improperly narrowed the comparative data base to exclude the majority of Subtitle C wastes. Id. at 19-22. Petitioners also contest EPA's decisions to consider only the largest waste stream managed at each facility in its data base, id. at 23, to use data that aggregated the volumes of individual Subtitle C waste streams, id. at 24-26, to establish separate high volume thresholds for liquid and solid wastes, id. at 27-28, and to select the 95th percentile as the statistical cutoff to define the numerical high volume standards. Id. at 28-30.

EPA, we conclude, reasonably interpreted the “special waste” concept when the Agency decided to quantify the “high volume” criterion through an analysis of the technical feasibility of Subtitle C controls. EPA adequately explained, as logically within this permissible interpretation, each of the methodological decisions petitioners challenge. To the extent petitioners quarrel with EPA’s methodology or selection of a precise numerical measure, they have failed to demonstrate that the Agency acted in an arbitrary and capricious manner. See State Farm, 463 U.S. at 43, 103 S.Ct. at 2866.

Petitioners fault EPA for not basing its high volume criteria on “the full universe of industrial Subtitle C facilities.” Jt. Brief at 19. But EPA reasonably limited the comparative data base to facilities using those Subtitle C management techniques that would most likely be employed to manage hazardous mineral processing wastes — landfills for solid wastes, and wastewater treatment processes, surface impoundments, and injection wells for liquid wastes. 54 Fed.Reg. at 36,608, 36,629. Small quantity waste generators, and those that do not employ these waste management techniques, could supply no data relevant to EPA’s inquiry into the technical feasibility of managing large volumes of mineral processing waste under Subtitle C.

EPA similarly rejected the suggestion that “a relevant comparative analysis should reflect ‘typical’ quantities of hazardous waste generated.” 54 Fed.Reg. at 15,-329. Instead, endeavoring to identify mineral processing wastes generated in volumes so high as to resist Subtitle C controls, EPA appropriately concentrated its examination on the largest volume wastes currently managed under Subtitle C. Id. at 15,329; 54 Fed.Reg. at 36,611.

The Agency likewise reasonably decided to include data from commercial Subtitle C waste management facilities in its analysis. EPA reasoned that information on waste volumes managed by commercial facilities was relevant to the Agency’s determination of “technical feasibility^] ... the fundamental issue addressed by the volume criterion.” 54 Fed.Reg. at 36,612. “[Cjonsider-ations of differential economic incentives facing operators of commercial and private hazardous waste management facilities,” EPA said, “are not relevant in resolving this [technical feasibility] issue.” Id. at 36,630.

EPA’s decision to establish separate high volume thresholds for liquid and solid mineral processing wastes is similarly rooted in the Agency’s legitimate focus on technical feasibility. EPA proposed setting separate volumetric standards in the 4/89 NPRM, noting that Subtitle C facilities typically manage liquid wastes in far larger volumes than solid wastes. 54 Fed.Reg. at 15,331. The Agency finalized this approach in the 9/89 Rule. EPA thus properly recognized the “highly significant differences” in treatment processes and management options that rendered management of large volumes of wastewater “more technically feasible” than management of large volumes of solids. 54 Fed.Reg. at 36,630.

EPA acknowledged the analytical inconsistency of using data on the generation rates of aggregated Subtitle C waste streams to calculate a waste stream-specific high volume threshold for mineral processing wastes. 53 Fed.Reg. at 41,293-94; 54 Fed.Reg. at 15,329; 54 Fed.Reg. at 36,-609-10. The TSDR Survey data, however, enabled EPA substantially to disaggregate the data on Subtitle C waste generation rates. Id. at 36,610; see also Development of the High Volume Criterion for Mineral Processing Wastes (Aug. 8, 1989) at 1-2, JA at 171-72 (describing methodology for developing high volume criterion based on TSDR Survey data, including formula for calculating disaggregated waste generation rates). 9

Finally, EPA properly exercised its discretion in selecting the 95th percentile as the statistical cutoff to define the numerical high volume standards, permitting a 5% overlap between the volumes of Bevill mineral processing wastes and the volumes of wastes currently managed under Subtitle C. See 54 Fed.Reg. at 36,629. Environmental watchdog groups recommended a cutoff at the 99th percentile, allowing for an overlap of only one percent, while members of the mineral processing industry favored an overlap of “at least 10 percent,” translating to a cutoff at the 90th percentile. Id.; see Comments of National Audubon Society, et al. (May 31, 1989) at 55; Comments of American Mining Congress (May 31,1989) at 44. EPA explained in the 9/89 Rule that the percentile overlap was reduced from 10%, as proposed in the 10/88 and 4/89 NPRMs, to 5% because “problems with the data used in the analysis” (i.e., the 1985 Biennial Report), which had justified a lower percentile cutoff, had been resolved when “better data” (i.e., the TSDR Survey ) became available. 54 Fed.Reg. at 36,-613. EPA determined that this “much stronger” data supported a 5% overlap. Id.

Given EPA’s logical conclusion that current management of large volume hazardous wastes indicates that Subtitle C regulation of that quantity of waste is technically feasible, the Agency reasonably selected a percentile cutoff that eliminated from Be-vill Amendment coverage those mineral processing wastes generated in volumes lower than the largest volume Subtitle C-managed wastes. In selecting a 95th percentile cutoff, resulting in a 5% overlap between the generation rates of large volume Subtitle C wastes and Bevill mineral processing wastes, EPA chose “a numerical standard ... within a ‘zone of reasonableness’ ” that warrants judicial approbation. See Small Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d 506, 525 (D.C.Cir.1983) (citations omitted).

C. Low Hazard Criterion

To qualify as a “special waste” covered by the Bevill Amendment, a mineral processing waste, EPA has ruled, must not only be high volume, it must also be low hazard. Petitioners challenge the very inclusion of any low hazard criterion. They also challenge the particular hazard screening tests EPA used. We find EPA’s low hazard criterion and tests compatible with the legislative design, consistent with our decisions in EDF I and EDF II, and reasonably explained. We therefore reject petitioners’ challenges.

According to petitioners, the “special waste” concept, as originally presented in EPA’s 1978 proposal, classified wastes as “special” because of their large volume and, consequently, the anticipated infeasibility of Subtitle C regulation. Jt. Brief at 33. While acknowledging EPA’s statement in the 1978 proposal that “hazardous ‘special wastes’ posed a ‘relatively low’ risk to human health and the environment,” id., petitioners insist that “volume, not hazard, is the key to the special waste concept.” Id. at 34 & n. 39. Volume, no doubt, is a necessary key, but EPA maintains it does not suffice to open the special processing waste category.

Interpretation of the Bevill Amendment to exclude “wastes that are clearly not low hazard,” 54 Fed.Reg. at 15,331, we continue to hold, is consistent with congressional intent and the regulatory history of the special waste concept. See EDF II, 852 F.2d at 1329 (“Congress intended the term ‘processing’ in the Bevill Amendment to include only ... ‘high volume, low hazard’ wastes.”). The Agency’s 1978 proposal lists low hazard as one of the three identifying characteristics of special waste, along with large volume and lack of amenability to Subtitle C controls. 43 Fed.Reg. at 58,992. In its November 1979 Special Waste Background Document, EPA explained that it had deferred regulation of special wastes, pending further study, in part because “[t]he various special wastes pose only a low potential hazard to human health and the environment.” Special Waste Background Document at 21, JA at 205. The Agency, at that time, expressly established “low potential hazard ... as a criterion for special waste status,” and observed that, consequently, “all the [special] waste streams will pose no more than a minimal threat to human health and the environment during the several year interim period” in which EPA planned to study the wastes to determine appropriate regulatory action. Id. at 21-22, JA at 205-06; see also 54 Fed.Reg. at 15,318-19 (recounting that Background Document identifies “low potential hazard” as a key criterion EPA would use to evaluate petitions to designate special wastes).

Petitioners further maintain that EPA’s imposition of a low hazard screening criterion conflicts with this court’s interpretation of the Bevill Amendment in EDF I. Jt. Brief at 34-35. In EDF I, the court upheld EPA’s regulatory determination under the Bevill Amendment to exempt hazardous mineral extraction and beneficiation wastes from Subtitle C controls, and to subject them instead to the less restrictive standards of RCRA Subtitle D. See EDF I, 852 F.2d at 1310. The court rejected the Environmental Defense Fund’s argument that EPA was required to regulate under Subtitle C any wastes found in the Bevill study to be “hazardous” under RCRA. “EPA’s interpretation of the Bevill Amendment as allowing the agency discretion to base its regulatory determination on a variety of factors,” the EDF I court held, “is a permissible construction of the statute.” Id. at 1313.

Petitioners argue that by applying a low hazard screening criterion, EPA has made hazard the determinative factor in establishing the regulatory status of high volume processing wastes that would otherwise qualify for Bevill coverage, contrary to EDF I. It is not inconsistent, however, to interpret the Bevill Amendment as requiring a “context-specific” determination of hazard as one of several factors to be considered in regulating wastes definitively placed within Bevill’s scope, e.g., wastes from the extraction and beneficiation stages of mining, EDF I, 852 F.2d at 1315-16, and, at the same time, to construe the Amendment’s terms to exclude from Be-vill’s scope processing wastes that do not qualify as “low hazard.” EDF II, 852 F.2d at 1329. This is indeed the very interpretation our circuit approved in EDF I and EDF II, companion cases decided the same day.

EPA adopted in the rulemaking on review low hazard tests generally less stringent than the RCRA hazardous waste characteristic tests normally used for Subtitle C wastes. Thus a waste qualifying as “low hazard” under Bevill might subsequently be found to contain a RCRA hazardous characteristic. See 54 Fed.Reg. at 36,597. 10 Consistent with EDF I, EPA assured that

the low hazard criterion is solely a preliminary screening device to determine which mineral processing wastes are special wastes, and will not be used in determining which wastes will subsequently be regulated under Subtitle C, either as a result of today’s rule or in the upcoming regulatory determination.

Id.

In sum, EDF I held only that EPA is not required to regulate high volume special wastes under Subtitle C simply because they exhibit a RCRA hazardous characteristic. That holding does not mean EPA is prohibited from using hazard as a criterion, along with volume, to determine which processing wastes fit the “special waste” category. Petitioners’ argument both overreads EDF I and is at odds with EDF II, which constantly juxtaposes “high volume, low hazard” in the context at hand. See EDF II, 852 F.2d at 1322, 1325, 1326, 1329, 1331.

Petitioners further object to the tests EPA selected. The Agency selected Method 1312 as the procedure for extracting waste constituents from solid mineral processing wastes to test for toxicity and mobility. 54 Fed.Reg. at 36,597. EPA considered Method 1312, a recently developed leaching procedure, to be “generally less aggressive” than Method 1310, the leaching procedure used to measure the hazardous toxicity characteristic under RCRA. 54 Fed.Reg. at 36,597, 36,601. Petitioners argue that EPA “failed to verify the accuracy and precision of Method 1312” in measuring “hazards that may be posed by mineral processing wastes,” and that selection of the procedure was therefore arbitrary and capricious. Jt. Brief at 38.

While EPA “ha[d] not yet completed its evaluation of Method 1312” when the procedure was first proposed in the 4/89 NPRM, the Agency reported that “work conducted to date indicates that the procedure is of acceptable precision.” 54 Fed. Reg. at 15,340. Because EPA’s evaluation of Method 1312 was incomplete, the Agency tested several mineral processing wastes using the “more aggressive extraction medium” of Method 1310, and neutral water testing, considered “only slightly less aggressive than the Method 1312 leaching medium.” Id. After retesting these wastes using Method 1312, EPA determined in the 9/89 Rule that “the new sampling and analytical data obtained using Method 1312 confirm the Agency’s earlier findings” from Method 1310 and neutral water testing. 54 Fed.Reg. at 36,601.

This court does not “demand certainty where there is none,” Small Ref. Lead Phase-Down Task Force, 705 F.2d at 525, and the Agency here may apply its “ ‘expertise to draw conclusions from ... probative preliminary data not yet certifiable as “fact,” and the like.’ ” AMC v. EPA, 907 F.2d 1179, 1187 (D.C.Cir.1990) (quoting Ethyl Corp. v. EPA, 541 F.2d 1, 28 (D.C.Cir.) (en banc), cert. denied, 426 U.S. 941, 96 S.Ct. 2662, 49 L.Ed.2d 394 (1976)). Given the respect properly accorded EPA’s judgment in this area, use of Method 1312 as a testing mechanism, we rule, withstands petitioners’ attack.

Petitioners also challenge EPA’s pH cor-rosivity test as arbitrary and capricious because “the pH standard overstates the potential hazard” of wastes containing “un-buffered” acids. Jt. Brief at 40. Petitioners state that a small amount of “unbuf-fered” acid in a waste will result in a very low pH, “even though the waste’s acidic content or potential hazard ... may be quite minimal.” Id. As EPA explained, however, “the fact that mineral acids are not appreciably buffered does not alter the fact that wastes of such low pH may pose a hazard” prior to treatment. 54 Fed.Reg. at 36,600. Petitioners have not demonstrated that EPA’s refusal to except unbuf-fered acids from the corrosivity test is arbitrary or capricious.

EPA selected a pH range of 1.0 to 13.5 “to identify which wastes clearly are so corrosive that they do not merit continued regulatory exclusion and further study.” Id. The final low hazard corrosivity standard represents a relaxation by one order of magnitude of the RCRA hazardous characteristic pH standard of 2 to 12.5. Id. EPA determined that “any further increase in the pH range may result in wastes that are clearly not low hazard remaining in the Bevill exclusion, which may in turn compromise the protection of human health and the environment.” Id. Because the agency “examined the relevant data and ... articulated a rational explanation for its action,” EPA’s predictive low hazard screening criterion warrants our respect on review. Eagle-Picher Industries v. EPA, 759 F.2d 905, 921 (D.C.Cir.1985).

D. The 50 Percent Rule

The 9/89 Rule interprets the Bevill Amendment as applying only to wastes from mineral processing operations with feedstocks containing less than 50 percent scrap materials, see 54 Fed.Reg. at 36,629; for this purpose, in-process materials, which are derived from prior mineral processing operations, are counted as scrap regardless of their mineral content. See id. at 36,619-20. The Amendment itself does not define the terms “ores [or] minerals.” Because the statute is “silent” on this issue, our role is merely to determine whether EPA has permissibly construed the statute. See Chevron, 467 U.S. at 843, 104 S.Ct. at 2781.

Petitioners assert that the 50 percent rule is not a permissible construction of the Bevill Amendment because it runs afoul of an underlying environmental objective of RCRA, namely “to conserve valuable material and energy resources by ... encouraging process substitution, materials recovery, properly conducted recycling and reuse, and treatment.” 42 U.S.C. § 6902(a), (a)(6). They point out that the limitation on in-process materials acts as a disincentive to the reuse of processed materials and thus promotes the extraction of additional virgin ores and minerals.

According to EPA, the 50 percent rule strikes a balance between, on the one hand, promoting the reuse of scrap and in-process materials and, on the other, restricting the Bevill exclusion to primary processors. See 54 Fed.Reg. at 36,620. While a rule making the Bevill exclusion unavailable to some in-process materials may to that extent discourage recycling, it is necessary, in the Agency’s view, to “maintain[] the essential upper bound on the amount of non-ore present in a feedstock in order to ensure that wastes from operations that primarily process materials other than ores and minerals are not provided with an exclusion that Congress did not intend.” Id.

We take the Agency’s point. EPA must distinguish somehow between primary mineral processing operations, which qualify as Bevill mineral processing, and secondary processing operations, which do not. The 50 percent rule supplies a reasonable line of demarcation, classifying every operation on the basis of whether it is primarily of the one or of the other sort.

Petitioners also argue that the 50 percent rule purports to regulate the production process, which the Agency acknowledges would exceed its authority. As we have just established, however, Congress delegated to EPA the authority — indeed the obligation — to determine which processes qualify for the Bevill mineral processing exclusion, and it is reasonable for the Agency to exclude from the Bevill category operations that use predominantly scrap and in-process materials. That decision may have an effect upon the economics of the production process, but it is not for that reason alone a regulation of the production process itself. Thus, we uphold the 50 percent rule against the two challenges pressed by petitioners.

E. Future Waste Streams

The 9/89 Rule denies Bevill status to any waste stream that did not exist when the Rule was promulgated, and to any extent waste stream that did not meet the Bevill criteria at the time of rulemaking. According to the Agency, it could and did make

a one-time determination of Bevill status. Wastes not yet in existence and wastes not meeting the high volume/low hazard criteria during any of the past five years would ... not be eligible for Bevill exclusion status in the future.

54 Fed.Reg. at 36,595. As a result, an after-arising waste stream that meets all of the substantive criteria for Bevill status will be regulated under either Subtitle C or Subtitle D of RCRA. Id. at 36,596.

Petitioners argue that denying Bevill status to future waste streams is contrary to the purpose of the Bevill Amendment. They point in particular to special wastes that do not currently satisfy the high volume threshold but will do so in the future.

The statutory provision directing EPA to study Bevill wastes suggests by its terms that a one-time study is sufficient. See 42 U.S.C. § 6982(p) (“The Administrator shall conduct a detailed and comprehensive study ... [and] shall publish a report of such study_”). The way in which the temporary nature of the Bevill Amendment is expressed in the statute also lends support to EPA’s interpretation. See 42 U.S.C. § 6921(b)(3)(A) (Bevill exclusion operative “until at least six months after ... submission of the applicable study”).

In any event, this court’s holding in EDF II secures EPA’s position that a one-time determination is sufficient. There we interpreted the Bevill Amendment as an exclusion specifically for “the category of wastes designated as ‘special wastesf ]’ ... in EPA’s 1978 proposed hazardous waste regulations.” EDF II, 852 F.2d at 1329 (quoting 51 Fed.Reg. 36,234 (1986)). While we did not go so far as to foreclose Bevill status for a future waste that might satisfy a pre-set criterion, we clearly enough rejected the theory that Congress intended the coverage of the Bevill exclusion to evolve with time.

In sum, while the Bevill Amendment itself is silent on the coverage of future waste streams, the statutory framework of RCRA and the precedent of EDF II both support EPA’s position. In these circumstances, EPA’s position is surely reasonable, and we must uphold it.

F. Application of Section 300^(x) to Non-Bevill Wastes

In 1984 Congress added to RCRA the Hazardous and Solid Waste Amendments, commonly known as the “Simpson Amendment,” codified at section 3004(x). This section authorizes EPA to relax Subtitle C regulations for “the extraction, beneficiation or processing of ores and minerals” — a list of operations virtually identical to that in the Bevill Amendment. Accordingly, in the course of implementing the Bevill Amendment, EPA addressed the question whether a waste that is denied Bevill status is nonetheless eligible for relaxed regulation pursuant to section 3004(x).

In the preamble to its 9/89 Rule, EPA read section 3004(x) as having the same scope as the Bevill Amendment: “the Simpson Amendment would not apply to wastes that are not special wastes and that would therefore be removed from the Be-vill exclusion by the proposed rule.... [T]he proper reading of section 3004(x) is that it applies only to special wastes as defined by today’s final rule.” 54 Fed.Reg. at 36,624-25.

In this EPA merely assigned the same meaning to the same statutory formula in the two places where Congress used it. That seems eminently reasonable, in the absence of a particular reason to do otherwise. Compare Alabama Power Co. v. Costle, 636 F.2d 323, 396 (D.C.Cir.1979) (“Given no expression of any contrary intent in the Act or in the legislative history regarding these definitions, we must assume that the meaning of a particular term is to be consistent throughout the Act.”) with National Ass’n of Casualty and Surety Agents v. Board of Governors of the Fed. Reserve Sys., 856 F.2d 282, 287 (D.C.Cir.1988) (upholding agency’s different interpretations of similar statutory phrases based upon “their different economic impact”).

Petitioners challenge the Agency’s position by appealing to a competing norm of statutory construction: “[W]here ... Congress adopts a new law incorporating sections of a prior law, Congress normally can be presumed to have had knowledge of the interpretation given to the incorporated law, at least insofar as it affects the new statute.” Lorillard v. Pons, 434 U.S. 575, 581, 98 S.Ct. 866, 870, 55 L.Ed.2d 40 (1978). Accord Lindahl v. OPM, 470 U.S. 768, 782 & n. 15, 105 S.Ct. 1620, 1628 & n. 15, 84 L.Ed.2d 674 (1985); Haig v. Agee, 453 U.S. 280, 301 & n. 50, 101 S.Ct. 2766, 2779 & n. 50, 69 L.Ed.2d 640 (1981); Dart v. United States, 848 F.2d 217, 229 (D.C.Cir.1988). Thus, when Congress enacted section 3004(x) in 1984, petitioners assert, it adopted EPA’s contemporaneous interpretation of the scope of the Bevill Amendment, which included all solid waste from the smelting and refining of ores and minerals. See 45 Fed.Reg. at 76,619.

In effect, petitioners advance a Chevron step one argument, i.e., that “Congress has directly spoken to the precise question at issue.” 467 U.S. at 842, 104 S.Ct. at 2781. They rely principally upon two points of legislative history in support of their position. First, Senator Randolph inserted into the record the following statement:

[0]f course, we recognize that under the ongoing reexamination of what wastes the mining waste suspension should apply to, some wastes — including previously listed wastes — may be regulated ... even earlier than the conclusion of the study.

130 Cong.Rec. 20,845 (July 25, 1984). This statement bespeaks an intent to include within the scope of section 3004(x) wastes that were then within the Bevill exclusion but that might subsequently be denied Be-vill status. Senator Randolph, at least, viewed the scope of section 3004(x) as potentially broader than that of the Bevill Amendment.

Second, the House bill would have made section 3004(x) applicable only to wastes subject to the Bevill study, see 129 Cong. Rec. 30,854 (Nov. 3, 1983); the Senate version, which was enacted (with modifications not relevant here), would have applied instead to the seemingly broader category of “solid waste from the extraction, beneficiation or processing of ores and minerals, ... if such solid waste is subject to regulation under this subtitle.” S. 757, 98th Cong., 1st Sess. § 6(a) (1983). The committee re ports tell a different tale, however. The Senate Committee Report indicates that even in the Senate version the authority to modify Subtitle C regulations did not cover wastes listed as hazardous prior to passage of the Bevill Amendment. S.Rep. No. 284, 98th Cong., 1st Sess. 29 (1983). And the Conference Report says of section 3004(x) — again, referring to the Senate version as modified — that “[t]his authority is intended to extend to all of the wastes required to be studied by EPA pursuant to section 8002(f), (n), (o), and (p).” H.R.Conf. Rep. No. 1133, 98th Cong., 2d Sess. 94 (1984), U.S.Code Cong. & AdmimNews 1984, 5576, 5649, 5665. With the Senate initially circumscribing the scope of section 3004(x), and the conferees then explicitly tying that section to the Bevill Amendment, these reports greatly undercut the plausibility of petitioners’ interpretation.

The statement of an individual senator carries little weight when there is other, more reliable legislative history, such as these committee reports, to the contrary. See United Mine Workers v. Federal Mine Safety & Health Review Comm’n, 671 F.2d 615, 622-23 (D.C.Cir.1982); see also Chemical Mfrs. Ass’n v. EPA, 919 F.2d 158, 164 n. 10 (D.C.Cir.1990). Taken as a whole, therefore, the legislative history does not support petitioners’ argument.

In addition, we note that EPA had expressly put its interpretation of the Bevill exclusion (as of 1984) on a provisional basis. As we recounted in EDF II:

EPA ... announced [in 1980] that as a “temporary accommodation of the requests” of the mining industry it would interpret the mining waste exclusion as applying to solid waste from the “exploration, mining, milling, smelting and refining of ores and minerals.” In so doing, EPA acknowledged that such a broad interpretation of the exclusion might not be consistent with Congress’s intent in enacting the Bevill Amendment.

EDF II, 852 F.2d at 1320 (quoting 45 Fed. Reg. 76,618, 76,619 (1980), emphases omitted). In these circumstances, we hesitate to attribute to Congress an intent to adopt as its own an interpretation upon which the administrative Agency itself had cast doubt.

In sum, contrary to petitioners, we do not think that Congress can be said to have spoken directly, or if it did, to have spoken clearly enough to be discerned, to the question whether section 3004(x) incorporates EPA’s 1984 interpretation of the scope of the Bevill Amendment. The relevant legislative history is, for petitioners’ purposes, at best ambiguous. We therefore review the Agency’s interpretation of the statute under the deferential standard of Chevron step two, 467 U.S. at 843,104 S.Ct. at 2781, and uphold as reasonable EPA’s reading of section 3004(x) as coterminous with the Be-vill Amendment, for the reason assigned at the outset of this section.

G. The Mixture Rule

The 9/89 Rule provides that EPA will, “under almost all circumstances,” 54 Fed.Reg. at 36,622, apply to waste streams consisting of both Bevill and non-Bevill wastes the mixture rule it had earlier promulgated pursuant to Subtitle C, see 40 C.F.R. § 261.3. If the non-Bevill waste is a listed hazardous waste, then the resultant mixture “would be considered a hazardous waste unless and until the mixture is delisted”; if the non-Bevill waste is a hazardous “characteristic” waste, then “the mixture would be considered a hazardous waste if it exhibits one or more of the same hazardous characteristics that are exhibited by the non-excluded waste.” Id. Only if any hazardous characteristic that the mixture exhibits is attributable solely to the Bevill waste component would the mixture be deemed not to be a hazardous waste, and thus qualify for the Bevill exclusion.

In extending the Subtitle C mixture rule to the Bevill context, EPA assumed the validity of that rule, which we have since vacated on a procedural ground and remanded to the Agency for reconsideration. See Shell Oil Co. v. EPA, 950 F.2d 741 (D.C.Cir.1991). Were the Subtitle C mixture rule still in place, the Bevill mixture rule might well constitute a reasonable extension of it. Following our decision in Shell Oil, we express no opinion on this matter. 950 F.2d at 765. If the EPA desires to and successfully does repromul- gate the Subtitle C rule, it will similarly be able to repromulgate the Bevill rule, and attempt to justify the latter by reference to the former. Alternatively, the Agency may wish to justify the Bevill rule on independent grounds. Because we vacated the Subtitle C rule and remanded it to the Agency, however, see Shell Oil, the Bevill mixture rule must now also be remanded. Cf. SEC v. Chenery Corp., 318 U.S. 80, 95, 63 S.Ct. 454, 462, 87 L.Ed. 626 (1943).

H. The Treatment Permit Requirement

As a complement to the mixture rule, EPA observes in the preamble to the 9/89 Rule that a processing operation must have a treatment permit in order to perform the mixing itself:

[MJixing a characteristic hazardous waste with a Bevill waste would constitute treatment of a hazardous waste, and would be subject to the appropriate regulation for the treatment, storage, or disposal of hazardous wastes, including obtaining a permit.

54 Fed.Reg. at 36,622. Petitioners object that this conclusion is arbitrary and capricious for want of “any explanation for how [it] ... was reached, including any discussion of the statutory or regulatory definitions of ‘treatment.’ ”

Having remanded the Bevill mixture rule, we have no occasion today to pass upon the validity of this corollary rule. If the EPA reinstates the underlying mixture rule, or otherwise decides to adhere to this permit rule, petitioners may seek review of that decision in due course.

I. Du Pont’s Chloride-ilmenite Process

In the course of rulemaking, EPA twice discussed application of the Bevill Amendment to wastes generated by titanium dioxide production using a chloride process. See 54 Fed.Reg. 39,298, 39.304 (9/89 NPRM); 54 Fed.Reg. 36,592, 36,621 (9/89 Rule). On both occasions, petitioner Du Pont commented, asserting that its proprietary “chloride-ilmenite” process is different from the chloride process described by the Agency and that the waste it produces comes within the Bevill exclusion. Responding to these comments in the preamble to its 1/90 Rule, EPA acknowledged that Du Pont’s proprietary process is somewhat different but denied the Bevill exclusion to the waste it yields.

There are four sequential steps in Du Pont’s chloride-ilmenite process, the first two of which occur within the same vessel: (1) iron is leached from ilmenite ore to form iron chloride gas; (2) titanium is chlorinated from the ilmenite ore to form titanium tetrachloride gas; (3) the iron chloride is condensed and separated to form iron chloride acid; and finally (4) the titanium tetrachloride is condensed and processed to form titanium oxide pigment, the saleable product. Step (1) is a beneficiation, step (2) a processing step. In EPA’s view, the iron chloride acid waste, which is produced in gaseous form at step (1) but removed from the vessel at step (3), is a mineral processing waste that does not qualify for the Bevill exclusion. In Du Pont’s view, it is waste from the beneficiation step, and thus is covered by the Bevill exclusion. See Comments of Du Pont on EPA RCRA Proposal at 5 (Nov. 8, 1989) (“In our process, we generate first, the gaseous [form of the waste]. This step, we submit, is our ‘bene-ficiation’ process. After this conversion, the [mineral processing occurs].”).

EPA argues that because steps (1) and (2) occur within the same vessel, and were occasionally described by Du Pont as occurring “in the same process step,” see id. at 3, it may reasonably classify the operation as mineral processing. While “EPA agrees that, in part, the chloride-ilmenite process involves beneficiation of ores or minerals,” it emphasizes the similarities between Du Pont’s process and conventional chloride processes: each destroys the identity of the mineral, produces a saleable mineral product, and generates a functionally identical waste. 55 Fed.Reg. at 2329.

Because the Agency never explains why those similarities are in any way significant, it seems that the critical reason for the Agency’s treating Du Pont’s operation as mineral processing rather than benefici-ation is that “the ‘beneficiation’ wastes and the ‘processing’ wastes generated by the chloride-ilmenite process are inseparable.” Id. This is true in the sense that the wastes are physically in the same vessel; temporally, however, they are separate and distinct in that the beneficiation waste preexists the processing waste. EPA has offered no reason for thinking that the former point is more significant than the latter.

The Du Pont operation may not fit neatly into either one of the somewhat artificial categories — beneficiation and mineral processing — into which EPA has divided the universe of wastes for regulatory purposes. Compare EDF I, 852 F.2d 1309 (upholding Bevill status of all beneficiation wastes), with EDF II, 852 F.2d 1316 (EPA to determine which mineral processing wastes qualify for Bevill exclusion). That would hardly be surprising. The regulatory world is replete with imperfect distinctions; the Agency’s task is to apply them rationally in light of their imperfections. The Agency’s choice between two less than perfect alternatives cannot be accepted by the court uncritically, merely because the Agency acknowledges that each category is indeed an imperfect fit for the activity in question. Rather, EPA must articulate a reasonable justification for its choice on the basis of some policy ground or other. In this case, it simply has not done so. Therefore, the question whether to treat Du Pont’s proprietary chloride-ilmenite process as mineral processing or as beneficiation is remanded to EPA for an adequate explanation or, failing that, reconsideration. See State Farm, 463 U.S. at 56, 103 S.Ct. at 2873 (agency must “offer [a] rational connection between facts and judgment [in order] to pass muster under the arbitrary- and-capricious standard”).

Petitioner SCM claims that its sulfate process is sufficiently similar to Du Pont’s chloride-ilmenite process that if “this Court grants relief to Du Pont ... SCM should be accorded the same relief.” Having failed to make its case for similar treatment before the Agency, however, SCM will not now be heard in the first instance. See Portland Cement Ass’n v. Ruckelshaus, 486 F.2d 375, 394 (D.C.Cir.1973) (“To entertain comments made for the first time before this court would be destructive of a meaningful administrative process.”). Nor is SCM’s process so obviously similar to Du Pont’s that SCM can rest upon Du Pont’s having put its argument before the Agency. C f. NRDC v. EPA, 824 F.2d 1146, 1151 (D.C.Cir.1987) (en banc) (exhaustion requirement waived where agency considered “the identical issues” or “the very argument pressed” by petitioner) (citations omitted). Accordingly SCM’s petition is denied.

J. Chrome Tailings

1. Background

Chrome tailings are waste from the processing of chrome ore. The ore is pulverized, then mixed with lime and soda ash. That mixture is roasted, making the chromium water-soluble. Water is poured over the mixture, leaching out the chromium. The resulting liquid, sodium chromate liquor, is then drained, and constitutes the valuable product from the process. The remaining solids (chrome tailings) contain residual amounts of chrome capable of leaching out over time. At some plants tailings are released directly into the environment (untreated tailings). At other plants, the mixture is chemically treated to reduce the level of residual chrome, and only then released into the environment (treated tailings).

In the 9/89 NPRM, EPA proposed denying chrome tailings Bevill status because the tailings failed the low hazard criterion. 54 Fed.Reg. at 39,309. In response, American Chrome and Chemicals Inc. submitted a comment bringing to EPA’s attention the distinction between treated and untreated tailings. American Chrome pointed out that EPA had relied on tests of untreated tailings for its conclusions in the 9/89 NPRM. The company asserted that its (treated) tailings met the low hazard criterion and were the only tailings released into the environment at its plant. Accordingly, the company reasoned, waste from its facilities should qualify for the Bevill exclusion.

EPA agreed. In the 1/90 Rule, the Agency stated that treated tailings qualified for Bevill coverage. 55 Fed.Reg. 2353, 40 C.F.R. § 261.4(b)(7)(xiv). EPA stressed, however, that untreated tailings did not qualify, so that the on-site activity of treating the tailings was itself subject to Subtitle C. 55 Fed.Reg. at 2330.

While all this was happening, other, related issues were working their way through the rulemaking process. Petition ers Allied Signal and the City of Baltimore (collectively “Allied”) possess landfills that contain, among other things, untreated tail-ings previously deposited by now-defunct chrome processing plants. In comments to all three NPRMs, Allied directed EPA’s attention to the issue of landfill tailings. Allied argued that if currently-generated tailings received Bevill status, landfill tail-ings should too. Allied also argued, however, that if current tailings were denied Bevill status, landfill tailings should qualify for Bevill coverage as a separate category. In the 4/89 NPRM, the Agency declined to “broaden[ ]” the definition of chrome tailings “in any way.” 54 Fed.Reg. at 15,336. In the 1/90 Rule, EPA again declined to treat landfill tailings as a “separate, discrete mineral processing waste,” stating that, like other untreated tailings, this material would not be classified as waste covered by the Bevill Amendment. 55 Fed.Reg. at 2335. EPA explained that such landfill tailings were probably combined with non-mineral waste and other materials, and that untreated tailings were not low hazard. Id.

Additionally, in the 4/89 NPRM and the 9/89 Rule, the Agency clarified the application of its “regeneration rule” to this rule-making. EPA stated that if, as a result of this rulemaking, a waste stream became subjected to Subtitle C of RCRA for the first time, wastes disposed of prior to the effective date of the Rule would not be subject to “direct Subtitle C controls.” 54 Fed.Reg. at 15,338. Pursuant to the regeneration rule, this safe harbor from Subtitle C would last unless and until such predisposed waste was “actively managed.” Id. “Active management” was defined as “physically disturbing the accumulated wastes ... or disposing additional non-Be-vill hazardous wastes” into the landfill after the effective date of the Rule. Id. at 36,597. Against this backdrop, petitioners seek judicial review of EPA’s determination that chrome wastes in landfill did not qualify for the Bevill exemption.

2. Analysis

Petitioners first challenge EPA’s decision not to include landfill waste within the Bevill exclusion on procedural grounds. They argue that EPA’s conclusion that treated chrome wastes satisfied the high volume/low hazard criteria “established” in the 1/90 Rule “a new, narrowed category” of wastes. Jt. Brief at 73. The result, petitioners charge, is that landfill wastes “were in effect retroactively dropped from the scope of the September 1989 Rule.” Id. at 73-74. This, they allege, violated the Agency’s notice and comment obligations under the APA.

It generally is not a violation of notice and comment requirements to amend a proposed rule in response to a comment. NRDC v. Thomas, 838 F.2d 1224, 1242 (D.C.Cir.1988) (citing International Harvester Co. v. Ruckelshaus, 478 F.2d 615, 632 n. 51 (D.C.Cir.1973)). EPA determined that treated chrome waste satisfied the Be-vill requirements in response to submitted comments. 55 Fed.Reg. at 2353. Petitioners had the opportunity to argue to the Agency that their waste similarly satisfied the Bevill requirements; their arguments were unsuccessful. EPA’s refusal to provide further opportunity to comment on the fact that others’ arguments were successful and petitioners’ were not does not violate the APA.

Petitioners’ substantive arguments on this issue are no more persuasive. They argue first that Congress intended the “inherent qualities” of a waste to dictate its Bevill status. Jt. Brief at 78. The difference in hazardousness between treated and untreated chrome tailings, which guided EPA’s decision here, seems to be such an “inherent quality.” Petitioners next argue that EPA should have measured the volume of the waste stream when it was produced in this industry sector. Id. at 79. As we explained supra in Part B, we give great deference to EPA’s formulation and application of its high volume criteria. Since we conclude infra in Part K, that EPA reasonably decided to evaluate waste streams on an “as generated” basis, we see no reason to disturb its judgment here.

Finally, petitioners complain that EPA has “persistently” excluded chrome ore wastes in landfill from the “relevant universe of chrome ore residues,” and that this exclusion is arbitrary and capricious. Jt. Brief at 80. We conclude that while the former may be true, the latter is not. First, there was no contention that the chrome wastes in landfill were “treated” wastes; therefore, EPA reasonably concluded that landfill wastes would not be low hazard. 55 Fed.Reg. at 2335. Even petitioners admit, in comments to the 9/89 NPRM, that “certain of the fill may test EP toxic for chromium.” Comments of Baltimore City (Nov. 9, 1989), reproduced in JA at 438. Second, no chrome ore wastes in landfills had been generated during the years studied by EPA; petitioners’ challenge concerns waste that might be “regenerated” at some point in the future. Combining the regeneration rule with EPA’s interpretation of the Bevill Amendment as a one-time, “snapshot” inquiry, see supra, pp. 491-92, EPA reasonably maintains that landfill wastes do not satisfy the high volume criteria. The regeneration rule has been upheld. Chemical Waste Management, Inc. v. EPA, 869 F.2d 1526, 1536-37 (D.C.Cir.1989). EPA’s determination that chrome ore wastes in landfill did not qualify for the Bevill exemption was thus the result of the Agency’s interpretation and application of its own rules, and the interpretation was far from “plainly wrong.” Id. at 1539.

K. Lead Process Wastewater

Large amounts of water come into contact with lead during the refining process, for instance, when molten lead is cooled by spraying water on it. This “process waste-water” is one waste stream from the lead industry. Lead processing also generates hazardous air emissions, which contain (among other stuff) sulfur oxides. Some lead plants process these emissions at on-site acid plants, which produce marketable sulfuric acid from the lead plant emissions. These acid plants, in turn, produce waste-water, referred to in the industry as acid plant blowdown. See Preamble, Nonferrous Metals Manufacturing Point Source Category; Effluent Limitations Guidelines, Pretreatment Standards, and New Source Performance Standards, 49 Fed.Reg. at 8742, 8748.

In the 4/89 NPRM, EPA indicated that it would treat process wastewater and acid plant blowdown as separate waste streams. 54 Fed.Reg. at 15,327. EPA proposed retaining process wastewater for further study, but removing acid plant blowdown from the Bevill exclusion. Id. at 15,342. The 9/89 Rule removed acid plant blow-down from the Bevill exemption based on EPA’s conclusion that the waste did not satisfy the high volume criterion. Id. at 36,631. Process wastewater, however, remained a candidate for Bevill coverage. Id.

In the 9/89 NPRM, EPA revised its volume estimate for lead process wastewater. New figures showed that the waste did not satisfy the high volume criterion. 54 Fed. Reg. at 39,305. EPA also stated that its hazard data showed process wastewater contained arsenic, cadmium and lead at unacceptable levels. Id. at 39,308. Those findings did not change in the 1/90 Rule. 55 Fed.Reg. at 2341. The Agency did revise its volume figures upward to correct for one facility that had operated on only an intermittent basis. However, process wastewater still did not satisfy the volume criterion. Id. at 2340. Having met neither of the two threshold criteria, lead process wastewater did not qualify for Bevill coverage.

Petitioners’ volume objections reduce to the complaint that EPA did not aggregate the volumes of process waste-water and acid plant blowdown into one waste stream. They assert that because these two wastewaters are “generated and managed at the same facilities,” Jt. Brief at 67, it would be “more logical” to measure them as “one generic category.” Id. at 65 n. 85. Perhaps so, but as noted in Part II.A, supra, the question we must ask in reviewing the Agency’s choice is not whether we can posit a “more logical” alternative. See Udall v. Tollman, 380 U.S. 1, 16-17, 85 S.Ct. 792, 801-802, 13 L.Ed.2d 616 (1965). Indeed, when “Congress has not addressed the precise question at issue,” and that question is whether to adopt a “plantwide” or more narrow definition of a waste source, the Supreme Court has expressly instructed us to defer to permis sible agency definitions. Chevron, 467 U.S. at 843, 858, 104 S.Ct. at 2781, 2789. In this case the Agency chose to measure waste streams on an “as generated” rather than an “as managed” basis. 54 Fed.Reg. at 15,327. It did so to ensure that the high volume criteria would include those wastes not amenable to Subtitle C regulation under “current [or] alternative management” strategies. Id. (emphasis added). In light of our decision sustaining the high volume criteria, which reflect Congress’ central concern with technical feasibility, the choice to measure waste streams as generated is reasonable.

All but one of petitioners’ low hazard objections do nothing but restate their general objections to the low hazard criteria. In Part II.C, supra, we upheld those criteria. The only hazard objection specific to lead is that EPA applied its low hazard criteria to data from only two lead processing facilities. Jt. Brief at 69. There are only five lead processing facilities in the industry, however, and so EPA’s data demonstrate that at least 40 percent of plants in the industry failed to meet the low hazard criteria. We cannot say that such a determination is unreasonable. Cf. American Iron & Steel Inst. v. OSHA, 939 F.2d 975, 984-85 (D.C.Cir.1991) (per curiam).

L. Lightweight Aggregate Air Pollution Control Dust/Sludge

Lightweight aggregate (“LWA”) is a combination of materials that is itself an ingredient in certain types of concrete and asphalt. LWA is produced in a rotary kiln by heating certain minerals to very high temperatures. Some kiln furnaces burn hazardous wastes as fuels for this heating process. The furnaces employ either “wet” or “dry” air pollution control (“APC”) devices. The dry devices produce dust, the wet devices, sludge.

In the 9/89 NPRM, EPA announced that it estimated the average, per facility volume of LWA APC dust/sludge as 134,000 metric tons/year — nearly three times the 45,000 mt/yr needed to qualify as high volume. 54 Fed.Reg. at 39,305. EPA based its estimate on industry responses to questions 2.11 and 2.14 on the 1989 National Survey of Solid Wastes from Mineral Processing Facilities (“SWMPF Survey”). 11 Based on this volume estimate, and the fact that the waste met the low hazard criteria, EPA proposed retaining LWA APC dust/ sludge within the Bevill amendment.

In comments about the 9/89 NPRM, submitted on November 9, 1989, the Audubon Society pointed out that while 28 facilities generated LWA APC dust/sludge, only six submitted usable data in response to questions 2.11 and 2.14 of the SWMPF Survey. Urging that six facilities out of 28 provided an inadequate statistical base for EPA’s conclusions, Audubon suggested that the Agency “undertake additional data collection activities.” Comments of National Audubon Society (Nov. 9, 1989), reproduced in JA at 503. EPA did not attempt to gather more data in response to this comment. EPA instead developed a new methodology. Using responses to other questions on the SWMPF Survey EPA recalculated, making a series of inferences, the total per facility volume. 12 Based on the new data and new calculations, the average per-facility volume of LWA APC dust/ sludge was 15,813 mt/yr, well below the high volume cutoff. 55 Fed.Reg. at 2340.

The Agency attempted to apprise industry of these changes before promulgating its 1/90 Rule. It sent petitioners a memo detailing the new calculations on January 9, 1990. Solite received the memorandum on January 11, 1990, only one day before the Administrator signed the Final Rule on January 12, 1990. Petitioners complain that EPA did not provide the public with adequate notice of, or opportunity to comment on, its revised methodology for calculating the volume of LWA APC dust/ sludge. 13

Section 553 of the APA requires an agency to give interested parties “an opportunity to participate in the rulemaking through submission of ... arguments.” 5 U.S.C. § 553(c). An agency contemplating a change between its proposed and final rules cannot deny “commenters ... their first occasion to offer new and different criticisms which the Agency might find convincing.” BASF Wyandotte Cory. v. Costle, 598 F.2d 637 (1st Cir.1979), cited in United Steelworkers v. Marshall, 647 F.2d 1189, 1225 (D.C.Cir.1980) and Fertilizer Inst. v. EPA, 935 F.2d 1303 (D.C.Cir.1991). We find that EPA’s change in the volume determination for LWA APC dust/sludge denied petitioners just such an opportunity here.

The focus of the 9/89 NPRM, and of the 1/90 Rule, was to be application of the low hazard criteria, not the high volume criteria. See 54 Fed.Reg. at 36,593 (9/89 Rule, indicating that Agency lacked hazard information for wastes whose status was not resolved in that rule). The 9/89 NPRM itself solicited comments on “the data used to make the proposed Bevill waste exclusion decisions outlined below.” Id. at 39,-300 (emphasis added). The Agency listed LWA APC dust/sludge as a candidate for retention, id. at 39,303, and not among the wastes proposed for exclusion. Id. at 39,-309. Nothing, in short, indicated that EPA was going to reconsider its determination of the volume of LWA APC dust/sludge. Based on responses to SWMPF Survey questions not previously considered, and using methodology not previously explained, however, the Agency did just that. 55 Fed.Reg. at 2327. EPA asserts that the changes in the January 1990 Rule were a “logical outgrowth” of the 9/89 NPRM, EPA Brief at 109, and that the results it reached were substantively reasonable. Our resolution of the first point precludes our reaching the second.

A strikingly similar agency action has previously been considered and reversed by this court. Weyerhaeuser Co. v. Costle, 590 F.2d 1011, 1031 (D.C.Cir.1978). There, EPA had to calculate, based on limited data, the amount of waste produced by sulfite mills. In its penultimate pronouncement, the Agency estimated the weight to be 487 pounds per ton of product. Before issuing its final rule, however, EPA secured new data and incorporated calculations of two new factors it had not previously considered. Because neither the new data nor the calculations of the new factors had been opened for public comment, the court remanded the regulation to the Agency. Id. at 1028-31.

EPA asserts that its informal notification of petitioners by memorandum, prior to publication of the 1/90 Rule, provided “actual notice,” EPA Brief at 109-10 & n. 77, 14 and cites NRDC v. Thomas, 838 F.2d 1224, 1243 (D.C.Cir.1988), for support. That decision, in which the opinion admits “stretchpng]” notice and comment law “to its limits,” id., admits of one key distinction. EPA’s letter there was sent “two weeks before promulgation” and resulted in submission of actual comments. Id. Here, by contrast, the memorandum

reached petitioners one day before the regulation was signed; not surprisingly, comments were not forthcoming.

EPA’s determination that LWA APC dust/sludge does not satisfy the high volume criteria is based upon “a complex mix of controversial and uncommented upon data and calculations^] we cannot be sure that further and ultimately convincing public criticism of [the] changes would not have been forthcoming had it been invited.” Weyerhaeuser, 590 F.2d at 1031. We hold that EPA’s determination that lightweight aggregate air pollution control dust/sludge does not meet the high volume criteria violated notice and comment requirements. Accordingly, we direct the Agency to reconsider, after providing notice and soliciting comments, whether LWA APC dust/ sludge in fact satisfies the high volume criteria.

III. Conclusion

For the reasons stated, we affirm EPA s rulemaking in principal part, express no opinion on the Bevill mixture rule and the related requirement of a treatment permit, and return two matters to the Agency for further consideration consistent with this opinion: (1) measurement of lightweight aggregate residuals against the Agency’s high volume criterion; and (2) the Bevill status of Du Pont’s chloride-ilmenite process.

It is so ordered.

ORDER.

Feb. 26, 1992.

This matter is before the Court on respondent’s petition for rehearing. The EPA seeks to modify the terms of our remand of the mixture and treatment permit rules. In essence, the EPA asks us to uphold the Bevill rules for a mixture of a type that was not at issue in Shell Oil Co. v. EPA, Nos. 80-1532 et al. The EPA did not take such an alternative position m its briefs before this court, however. Cf. EPA Brief at 83. (“Solite’s criticism of the application of the mixture rule in the present case is more properly entertained in Shell Oil_”)• Therefore, we shall not now attempt to distinguish from the universe of Bevill mixtures those that are properly deemed independent of the decision in Shell Oil. Moreover, the EPA is free to make such subtle distinctions on remand.

The EPA’s challenge to our remand of the treatment permit rule fails for the same reason. See EPA Brief at 88-89. When the RCRA mixture rule fell, the treatment permit requirement for Bevill mixtures necessarily fell with it, except perhaps for such fine-tuning as the agency may perform on remand.

The petition for rehearing is denied.

1 . EDF II, 852 F.2d at 1318-24, recounts the complex statutory and regulatory history of the mining waste exclusion. EDF II was heard and decided in tandem with a companion case, Environmental Defense Fund v. EPA, 852 F.2d 1309 (D.C.Cir.1988) ("EDF /”), in which we upheld EPA’s decision to exempt from Subtitle C regulation, pursuant to the Bevill Amendment mining waste exclusion, all extraction and benefici-ation wastes. See infra note 4.

2 . Other "special wastes” mentioned by EPA included fly ash waste, bottom ash waste, slag waste, and cement kiln dust waste. 43 Fed.Reg. at 58,991-92.

3 . Congress directed EPA, in conducting the study, to analyze eight enumerated factors, including: "the source and volumes of such materials generated per year”; “present disposal and utilization practices”; potential health or environmental danger from disposal and reuse, and “documented cases” of such risks; “alternatives to current disposal methods”; "the impact of those alternatives on the use of phosphate rock and uranium ore and other natural resources”; and "the current and potential utilization of such materials." 42 U.S.C. § 6982(p)(l)-(8).

4 . The court in Adamstown also imposed on EPA a separate timetable in which to complete the Bevill study and regulatory determination for extraction and beneficiation wastes. See EDF I, 852 F.2d at 1311. Thus, as a result of the Adamstown case, EPA’s compliance with its obligation to study and regulate mining wastes under the Bevill Amendment proceeded on two separate tracks: one for extraction and benefici-ation wastes, firmly classified as Bevill wastes; and the other for processing wastes, still to be defined for purposes of Bevill inclusion. EPA submitted its report to Congress on extraction and beneficiation wastes on December 31, 1985, and published its regulatory determination as to those wastes on July 3, 1986. EDF I, 852 F.2d at 1312. In EDF I, we upheld EPA’s decision to regulate extraction and beneficiation wastes under RCRA Subtitle D, rather than under the stricter requirements of Subtitle C. Id. at 1310.

5 . The Conference Report, the court observed, “states clearly that the Bevill Amendment suspends regulation under Subtitle C of utility wastes as well as ‘all other wastes ... in a category designated as "special wastes” in regulations proposed by the Agency under Subtitle C on December 18, 1978.’” EDF II, 852 F.2d at 1328 (quoting H.R.Conf.Rep. No. 1444, 96th Cong., 2d Sess. 32 (1980), U.S.Code Cong. & Admin.News 1980, 5019, 5031-32).

The court also quoted Congressman Bevill’s specific reference to EPA’s 1978 proposed regulations on "special wastes” in explaining the purpose of his amendment on the House floor, id. (quoting 26 Cong.Rec. 3361 (1980)), and noted the remarks of other members of Congress referring explicitly to "special wastes,” EPA’s December 1978 proposed regulations, and "high volume, low toxicity wastes.” Id. (quoting remarks of Reps. Santini, Staggers, Findley, and Williams, reported at 26 Cong.Reg. 3348-49, 3363-65).

6 . On July 31, 1990, EPA submitted its Report to Congress on Special Wastes fr°m Mineral Processing containing the results of the Agency’s multi-factor Bevill study of each of the 20 mineral processing wastes identified in the rulemak-ing on review. See 55 Fed.Reg. 32,135 (1990). EPA published its final regulatory determination for Bevill mineral processing wastes in June 1991. See 56 Fed.Reg. 27,300 (1991); 40 C.F.R. Part 261 (1991). EPA determined that regulation under Subtitle C "is inappropriate for all 20 of the special wastes that were studied." 56 Fed.Reg. at 27,300. The Agency announced "plans to address 18 of the wastes under Subtitle D,” and to develop a program under the Toxic Substances Control Act to govern management of the remaining two wastes. Id.

7 . Petitioners incorrectly assert that failure to meet EPA's high volume standard is tantamount to determination of the regulatory status of a processing waste. See Jt. Brief at 11, 31. As earlier noted, see supra p. 481, wastes excluded from Bevill coverage through application of the high volume criterion "are not automatically subjected to [S]ubtitle C regulation; they must also exhibit one or more of the [RCRA] hazardous characteristics.” See 54 Fed.Reg. at 36,595.

8 . EPA’s comparative data indicate that only a single Bevill waste from mineral extraction or beneficiation, waste rock from zinc extraction, generated at a rate of 36,000 mt/yr, would fall outside the final 45,000 mt/yr per facility high volume threshold for solid mineral processing wastes. See Internal EPA Memorandum dated Oct. 19, 1988 re: Quantitative Basis for High Volume Mineral Processing Waste Criteria at 2, JA at 242. petitioners' objection, even assuming the validity of petitioners’ data analysis.

9 . In reply, petitioners maintain that even after EPA’s adjustment, some of the Subtitle C data continued to represent at least partially aggregated waste streams, and cite in support an analysis of EPA’s data base attached as an addendum to the joint reply brief. See Jt. Reply at 9 (citing Addendum 2 at 23-31). The discrepancy between petitioners’ analysis and EPA’s representations is troubling, particularly because petitioners had no opportunity to object to EPA’s use of the TSDR Survey as a data base, and EPA never considered or addressed petitioners’ data analysis.

That some data on Subtitle C wastes may have remained in a semi-aggregated state, however, would not seem sufficient to invalidate EPA’s methodology and the final high volume criterion as arbitrary and capricious. EPA reasonably chose to analyze mineral processing waste streams individually, consistent with the waste stream-specific evaluation apparently contemplated by the Bevill study requirement, see 42 U.S.C. § 6982(p), the regulatory structure of Subtitle C controls, and the analysis required to apply the low hazard criterion. See 54 Fed.Reg. at 36,609. EPA apparently made a conscientious effort to disaggregate the data on Subtitle C wastes to the greatest possible extent to provide the most directly comparable data base. Given the deference due EPA's interpretation of data within its area of expertise, see Ethyl Corp. v. EPA, 541 F.2d 1, 28 (D.C.Cir.) (en banc), cert. denied, 426 U.S. 941, 96 S.Ct. 2662, 49 L.Ed.2d 394 (1976), the high volume criterion survives

10 . Only the mobility/toxicity test, as applied to liquid wastes, is equivalent to the RCRA test for hazardousness under Subtitle C. See 54 Fed. Reg. at 36,602. Only two liquid processing wastes — lead process wastewater and titanium dioxide sulfate process waste acids — failed this test. Lead process wastewater also failed the high volume criterion, and titanium dioxide sulfate acids failed the relaxed Bevill low hazard corrosivity test. 55 Fed.Reg. at 2242. Because these two wastes were properly removed from Bevill coverage on other grounds, the court need not consider whether exclusion from Be-vill coverage based on use of a RCRA hazardous characteristic test would be contrary to EDF I. See 5 U.S.C. § 706 (in reviewing agency action under APA, "due account shall be taken of the rule of prejudicial error”).

11 . Question 2.11 asked: "How much of the special waste did this processing unit generate in 1988?” SWMPF Survey, reproduced in JA at 262. The answer to this question gave EPA the total volume of wastewater generated at each facility.

Question 2.14 asked: “What [was] ... the total solids content of this special waste in 1988?" This answer was either given as or converted by the Agency to a percentage figure. JA at 263.

EPA calculated the total solid waste output for each facility by multiplying the two figures: Total Volume X % solids = Total Solids. Determination of Waste Volume for Twenty Conditionally Retained Bevill Mineral Processing Wastes (Jan. 12, 1990), reproduced in JA at 163.

12 . EPA looked at responses to question 5.6, "total amount of accumulated sludge in this sur face impoundment added during 1988.” SWMPF Survey, reproduced in JA at 164. EPA divided this number by the answer to question 2.11 (total wastewater) in an attempt to calculate a "percentage of solids” figure. This derived percentage of solids was then compared to the percentage of solids directly reported in response to question 2.14. EPA then decided which of the two figures it “assumed most probable.” Id. at 258. The Agency made the same choice between the total solids figure calculated as described supra in note 11, and as derived from question 5.6.

13 . Petitioners also complain about EPA’s substantive determination of the volume of their waste. In light of our resolution of the notice and comment issue, we express no opinion on the substantive claim.

14 . EPA also implies in its brief that the very fact that "public commenters challenged the validi ty" of its volume determination put petitioners on notice. EPA Brief at 110. But "[c]omment-ing parties cannot be expected to monitor all other comments submitted to an agency.” Fertilizer Inst. v. EPA, 935 F.2d 1303, 1312 (D.C.Cir. 1991).