Opinion dissenting in part filed by Chief Judge WALD.

This case concerns the disposal of hazardous waste by "deep injection"; that is, the injection of hazardous waste into "wells" located thousands of feet beneath the surface of the earth. The Environmental Protection Agency ("EPA") has issued regulations under the Resource Conservation and Recovery Act ("RCRA") governing this method of hazardous waste disposal, and various petitioners challenge them. Industry petitioners claim the regulations are too stringent; environmental petitioners claim they are too lenient. We hold that the regulations are reasonable exercises of the EPA's authority and discretion under RCRA except insofar as they relate to the disposal of hazardous waste in "geologic repositories," that is, salt domes, salt beds, underground mines, and caves. We believe the EPA has ignored its statutory duty to promulgate standards for these repositories and has not adequately defended its permitting process against NRDC's statutory challenges, and so we remand these issues for further agency consideration. In all other respects, we deny the petitions for review.

The Resource Conservation and Recovery Act of 1976, as amended, 42 U.S.C. Secs. 6901 et seq. (1982 & Supp. III 1985), governs the land disposal of hazardous waste. As originally passed, RCRA declared that "disposal of solid waste and hazardous waste in or on the land without careful planning and management can present a danger to human health and the environment," and that "alternatives to existing methods of land disposal must be developed." 90 Stat. 2795, 2797 (1976). RCRA directed the Administrator of the EPA to promulgate regulations establishing "such performance standards, applicable to owners and operators of facilities for the treatment, storage, or disposal of hazardous waste ... as may be necessary to protect human health and the environment." 42 U.S.C. Sec. 6924.

Eight years later, a frustrated Congress, irritated at the slow pace at which the EPA was achieving RCRA's goals,2 passed the Hazardous and Solid Waste Amendments of 1984 ("HSWA"), Pub.L. No. 98-616, 98 Stat. 3221 (codified at scattered sections of 42 U.S.C. (Supp. III 1985)). As amended, RCRA pronounced that "land disposal, particularly landfill and surface impoundment, should be the least favored method for managing hazardous wastes." 42 U.S.C. Sec. 6901(b)(7). HSWA also legislated new "land ban" provisions governing the land disposal of hazardous waste, Sec. 6924(d)-(g) (set out in the margin).3 Subsection (d) governs all methods of land disposal, except disposal by deep injection, of certain, specified wastes (mostly toxic metals), popularly known as the "California list" wastes. Subsection (e) governs all methods of land disposal, except disposal by deep injection, of solvents and dioxins. Subsection (f) governs disposal by deep injection of the wastes specified in subsections (d) and (e). Finally, subsection (g) governs all methods of land disposal, including disposal by deep injection, of all hazardous wastes other than the "California list" or solvents and dioxins covered by subsections (d), (e), and (f). In toto, subsections (d) through (g) of Sec. 6924 govern all methods of land disposal of all hazardous wastes.

In 1988, the EPA promulgated final rules pursuant to subsections (f) and (g) governing the disposal of hazardous wastes by deep injection. 53 Fed.Reg. 28,118 (1988). These rules supplemented the EPA's Underground Injection Control ("UIC") program,, which governed underground injection of solid waste, to take into account injection of hazardous waste. The general scheme adopted by the EPA provided that underground injection of hazardous wastes would be prohibited unless the would-be injector obtained a permit from the EPA for a particular underground injection well. New rules located at 40 C.F.R. Part 148 identified the wastes otherwise prohibited from underground injection and specified the procedures for obtaining a permit to allow their injection. New Subpart G of 40 C.F.R. Part 146 set out technical criteria that underground injection wells must satisfy before they could be approved for hazardous waste disposal. In sum, the EPA's regulatory scheme for deep well injection of hazardous waste contained two important components: (1) the substantive standard that a hazardous waste injection well must meet, and (2) the permit procedures by which an injector must demonstrate that a well meets that standard.

The new rules also presented the EPA's interpretations of several key statutory terms. The permissibility of a land disposal method under subsections (d) through (g) of Sec. 6924 turns on whether the method will be "protective of human health and the environment for as long as the waste remains hazardous." Subsections (d), (e), and (g), but not (f), state that to meet this standard, an applicant must demonstrate to the Administrator, to a reasonable degree of certainty, that "there will be no migration of hazardous constituents from the disposal unit or injection zone for as long as the wastes remain hazardous." With regard to these safety standards, the EPA made the following interpretations: First, the EPA decided to apply the "no migration" standard to waste disposal governed by subsection (f) (deep injection of the "California list" wastes and solvents and dioxins) as well as to disposal governed by the other subsections.

Second, it decided that the term "the wastes" in the statutory no migration standard refers to the wastes that migrate out of the injection zone, and that the no migration standard is therefore satisfied if the injector demonstrates that no hazardous waste will migrate out of the injection zone.

Third, it decided that the term "migration" encompasses not only fluid migrations, but also migrations by molecular diffusion.

Fourth, it decided that in demonstrating that there will be no improper migration, an injector must show that there will be no migration for as long as the wastes remain hazardous, or for 10,000 years, whichever period is shorter.

Finally, the EPA decided that the term "injection zone," which is not defined in the statute, means any geological formation, group of formations, or part of a formation, that can meet RCRA's safety standards for containing waste.

In the preamble to its regulations, the EPA stated that the regulations might apply to the disposal of waste in geologic repositories. The EPA would, however, use individualized permit proceedings to determine whether the new regulations could appropriately govern disposal of hazardous waste in geologic repositories.

The Chemical Manufacturers Association and other industry groups (collectively referred to as "CMA" or "industry petitioners") claim that the EPA's rules impose unreasonably strict requirements that are unrelated to protection of human health and the environment. The Natural Resources Defense Council and other environmental groups (collectively referred to as "NRDC" or "environmental petitioners") claim the rules are not as protective of human health and the environment as the statute requires. The environmental petitioners also claim that the EPA cannot, consistent with the statutory directives of Sec. 6924(b),4 apply the regulations to the injection of waste into geologic repositories.

In passing on the claims of the petitioners, we are guided, as always, by Chevron U.S.A., Inc. v. NRDC, 467 U.S. 837, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984). With respect to each claim, we inquire "whether Congress has directly spoken to the precise question at issue." Id. at 842, 104 S.Ct. at 2781. In determining the intent of Congress, we must look to "the particular statutory language at issue, as well as the language and design of the statute as a whole," K Mart Corp. v. Cartier, Inc., 486 U.S. 281, 108 S. Ct. 1811, 1817, 100 L. Ed. 2d 313 (1988), and we must employ traditional tools of statutory construction, including, where appropriate, legislative history. Ohio v. United States Department of the Interior, 880 F.2d 432, 441 (D.C.Cir.1989). If the intent of Congress is clear, we must give it effect. 467 U.S. at 842-43, 104 S.Ct. at 2781. If, however, the statute is silent or ambiguous on a particular issue, we must defer to the agency's interpretation of the statute so long as it is reasonable and consistent with the statutory purpose. Id. at 844-45, 104 S.Ct. at 2782-83.

CMA claims that the EPA cannot generally ban the disposal of hazardous waste by deep injection, and require site-specific deep injection permits under subsection (f) or (g), because it has failed to make the determination, required by subsections (f)(2) and (g)(5), that absent a site-specific permitting process, deep injection of hazardous waste may not be protective of human health and the environment for as long as the wastes remain hazardous. We disagree.

The text of subsection (g) makes clear that the EPA may ban any method of land disposal of wastes governed by subsection (g) without first determining that the method is not protective of human health and the environment. Subsection (g)(1) provides that the Administrator must submit to Congress a schedule of the dates by which the agency plans to regulate the hazardous wastes not covered by subsections (d), (e) and (f). The EPA submitted this schedule to Congress in 1986. See 51 Fed.Reg. 19,300 (1986). Subsection (g) then goes on to state (and this is the language upon which CMA relies) that by the dates specified in the schedule, the Administrator "shall promulgate final regulations prohibiting one or more methods of land disposal of the hazardous wastes [on the schedule] except for methods of land disposal which the Administrator determines will be protective of human health and the environment for as long as the waste remains hazardous." Subsection (g)(5).

Nothing in this language requires the Administrator to determine that a method of land disposal is not safe before prohibiting it. Rather, the statute commands the Administrator to promulgate prohibitory regulations unless he has made an affirmative determination of safety. If by the dates specified in the schedule the Administrator has made no determinations at all (because, for example, he finds the evidence concerning safety to be in equipoise), he must promulgate regulations prohibiting one or more methods of land disposal. It seems indisputable to us that Congress did not require the EPA, before banning methods of land disposal, to determine affirmatively that they were not safe. Congress had already legislatively assumed that fact, subject only to the Administrator's decision otherwise.

Subsection (g) states further that the schedule sent to Congress "shall require that the Administrator shall promulgate regulations in accordance with paragraph (5) or make a determination under paragraph (5)" by dates certain. Subsection (g)(4) (emphasis added). The disjunctive "or" reinforces the interpretation that the EPA may promulgate the prohibitory regulations referred to in subsection (g)(5) without making any formal determinations. That is, although subsection (g) requires the Administrator to make a determination of safety before permitting a method of land disposal, it clearly does not require the Administrator to make any determination before prohibiting a method of land disposal.

CMA, however, points to the last sentence of subsection (g)(4), which states that whenever the EPA adds a new waste to its list of hazardous wastes, the agency must "determine" within six months of the new listing whether the waste shall be prohibited from disposal. CMA argues that this sentence states explicitly that the EPA must make determinations regarding the appropriate scope of prohibitions before promulgating prohibitions. This sentence, in our view, provides scant support for CMA's argument. It requires the EPA to determine only whether disposal of the newly listed waste is to be prohibited, not whether disposal is protective of human health and the environment. Since, as we have seen, the EPA can prohibit a disposal method under subsection (g) without making a safety determination, this sentence would seem to have no bearing on the issue of whether a safety determination is required. It requires only that a decision to prohibit or not to prohibit be made in the six month timeframe.

CMA also notes that the "hammer" provision of subsection (g)(6)(C), which prohibits the land disposal of any scheduled hazardous waste if the EPA fails to promulgate regulations or make determinations regarding that waste within five and a half years, provides the EPA with an unusually long leadtime if its only duty is to determine whether or not to do anything to prevent automatic prohibition. This argument misses the purpose of the hammer provision. Obviously, it would not have taken the EPA much time to issue a blanket ban on land disposal of all wastes on the schedule. But the EPA would need time to decide which, if any, waste disposal methods not to ban. The statute requires the EPA to prohibit one or more methods of land disposal except those it positively determines are safe; the 66 months were designed to afford the EPA sufficient time to affirmatively declare a safe harbor for those methods of land disposal it found worthy of rescue from the statutory hammer.

a. The EPA's Determination.

Subsection (f), which deals with deep injection of the "California list" wastes and solvents and dioxins, is worded differently from subsection (g). It says that the Administrator

shall make a determination regarding the disposal by underground injection into deep injection wells of the hazardous wastes referred to in [subsections (d) and (e) ]. The Administrator shall promulgate final regulations prohibiting the disposal of such wastes into such wells if it may reasonably be determined that such disposal may not be protective of human health and the environment for as long as the waste remains hazardous[.]

Subsection (f)(2). We find that the Administrator determined that deep well injection might not protect human health and the environment without a permit petition process.

In the preamble to the regulations, the EPA stated its belief that the regulations' new substantive standards for hazardous waste injection wells provided an appropriate level of protection for the injection of hazardous waste. 53 Fed.Reg. at 28,131 (col. 1). The EPA noted that it had identified several important ways to improve the protectiveness of the original UIC regulations. Id. For instance, the new regulations tightened the requirements concerning the choice of the well site, see id. at 28,132-33, widened the "area of review" (the geographical area within which a permit applicant must identify all existing wells and determine whether they have been properly completed or plugged and abandoned), see id. at 28,134-35, and imposed new requirements concerning the construction and operation of the well, see id. at 28,136-39. Despite these new safeguards, however, the EPA concluded that it could determine whether hazardous waste injection would protect human health and the environment only in the context of site-specific permit proceedings. Id. at 28,121 (col. 1). The agency specifically rejected the feasibility of any substantive standards that would permit a generic finding that hazardous waste injection would protect human health and the environment so long as injectors certified to the agency that their wells met the substantive standards. Id. It decided that individualized determinations, made in site-specific proceedings, would be more reliable. Id.

We think it reasonable to infer from the preamble that the EPA made the following determinations: (1) that deep well injection could, under appropriate circumstances, be protective of human health and the environment, and (2) that deep well injection might not be protective of human health and the environment unless it were governed by the agency's regulations, and particularly by the requirement that injectors obtain a site-specific permit from the agency before injecting any hazardous waste. The first determination is reflected in the EPA's decision to set up procedures for the issuance of deep injection permits; it would hardly have set up permitting procedures unless it believed that permits could be issued. As to the second determination, the agency's statement was clarion: it could not decide generically that deep injection protected human health and the environment; only a site-specific permit procedure would produce the necessary information for such a conclusion.

These determinations easily satisfy subsection (f). Indeed, insofar as CMA's claim that the EPA's determinations do not justify prohibitive regulations is concerned, the second determination, that the EPA could not generically find deep well injection to be safe, is all that is required in refutation. CMA claims that the agency failed to determine that deep injection could not be safe without a permit petition procedure. But the second sentence of subsection (f)(2) (assuming that it requires any determination at all) requires only that the EPA determine that deep well injection "may not be" protective of human health and the environment before prohibiting it. The agency's explicit rejection of the concept of deep well injection by injector certification certainly reveals an agency determination that such a process might not be protective of human health and the environment.5

b. Support for the EPA's Determination to Require

Site-Specific Permits.

CMA further claims that the EPA's decision to require individual applications for site-specific permits under subsection (f) was arbitrary and capricious. CMA asserts that the EPA's own reports show that deep well injection could safely be regulated by a rule permitting its use whenever it satisfied the substantive standards of the UIC regulations. On the basis of the record before us, we uphold the EPA's determination. That record contains ample evidence from which the EPA could rationally conclude that only site-specific inquiries by the agency into the geology of particular deep injection sites could adequately insure that deep injection would protect human health and the environment. The EPA reports cited by CMA establish, at most, that the EPA found the UIC regulations, as amended, to be acceptable substantive standards by which to judge the safety of a particular deep injection well; they do not support a conclusion that the EPA had determined that injectors could be trusted to decide for themselves whether a particular well met the standards.6

B. Application of the "No Migration" Standard to Subsection (f)

Subsections (d), (e), and (g) all state that the Administrator may not determine a method of land disposal to be protective of human health and the environment unless an interested person demonstrates that there will be no migration of hazardous constituents from the injection zone for as long as the wastes remain hazardous. Subsection (f) does not explicitly require that deep injection satisfy this "no migration" standard. However, the EPA, as a matter of policy, decided to apply the no migration standard to disposal under subsection (f) anyway. 53 Fed.Reg. at 28,120 (col. 3). The industry petitioners claim that the EPA lacked authority to inject that requirement into subsection (f).

Although subsection (f) does not contain the no migration standard, it does require the EPA to prohibit deep well injection if it determines that this method of land disposal may not be protective of human health and the environment. The phrase "protective of human health and the environment" is not defined by the statute. But surely it was within the Administrator's discretion to decide that a disposal method must meet the congressionally sanctioned no migration standard in order to protect human health and the environment. Indeed, it would have been downright surprising for him to do otherwise. Subsection (g) of the statute, which governs deep injection of all wastes not specified in subsections (d) and (e), explicitly contains the no migration standard. Had the Administrator decided that the no migration standard did not apply to subsection (f), he would have created the anomaly of legalizing injection of the wastes specified in subsections (d) and (e) without meeting the statutory standard laid down for the wastes included in subsection (g), even though there is no reason to think that the subsections (d) and (e) wastes are less hazardous than any others.7 The Administrator's decision to apply the no migration standard to subsection (f) seems almost foreordained.

CMA, however, claims that the EPA acted arbitrarily because the no migration standard, which focuses on whether any hazardous constituents cross the injection zone boundary, is totally disconnected from the protection of human health and the environment. Human health and the environment, CMA contends, are not affected until and unless hazardous waste migrates to a point of contact with something that humans actually use, such as drinking water. We find this argument unpersuasive. Congress itself decided that the no migration standard is an essential part of protecting human health and the environment so far as injection of subsection (g) wastes is concerned. How then could it be unreasonable for the EPA to apply the same standard regarding injection of subsections (d) and (e) wastes?8

C. The Content of the "No Migration" Standard--Industry Challenges

The industry petitioners claim that as a matter of statutory interpretation the no migration standard of RCRA is identical to the safety standard of the Safe Drinking Water Act ("SDWA"), 42 U.S.C. Sec. 300h(b)(1).9 We disagree.

The texts of RCRA and SDWA provide no support for the CMA's identity argument. SDWA protects sources of drinking water; RCRA protects human health and the environment. SDWA states that underground injection must not endanger drinking water sources; RCRA states that there must be no migration of hazardous constituents from the injection zone for as long as the wastes remain hazardous; it makes no reference to anything outside the injection zone that might be threatened by such a migration. The statutory texts provide no evidence whatsoever that Congress intended that the RCRA and SDWA standards be identical.

CMA's argument to the contrary rests almost entirely on some remarks made by Senator Bentsen on the Senate floor during the debates on HSWA. See Brief for Petitioner CMA at 40-41.10 These remarks do support the CMA's interpretation, but they are the remarks of a single Senator, made on the Senate floor, and without echo in the remarks of other Senators, or in the House or Senate reports. We have often noted that the comments of a single legislator cannot carry the day for a statutory interpretation contrary to that indicated by the statutory text. See, e.g., United States v. McGoff, 831 F.2d 1071, 1090 (D.C.Cir.1987); International Brotherhood of Electrical Workers v. NLRB (St. Francis Hospital), 814 F.2d 697, 713 (D.C.Cir.1987).

The only question remaining is whether the EPA's interpretation of RCRA as imposing a stricter standard on deep well injection of hazardous waste than SDWA is reasonable and consistent with RCRA's purposes. We believe that it is. RCRA's facially stronger text, requiring "no migration of hazardous constituents from the ... injection zone for as long as the wastes remain hazardous," subsection (g)(5), makes the EPA's tougher stance entirely reasonable. Its unstinting interpretation is consistent with RCRA's expressly stated purpose of minimizing the land disposal of hazardous waste.

The EPA is also accused of acting unreasonably in requiring injectors to show under subsections (f) and (g) that there would be no migration of hazardous constituents outside the injection zone via molecular diffusion.11 CMA argues that the EPA's own studies show that molecular diffusion causes hazardous constituents to migrate only a small distance from the injection zone, and poses no significant threat to underground sources of drinking water. However, CMA's characterization of the studies, even assuming it to be correct, is irrelevant. The question is not whether molecular diffusion causes hazardous waste to migrate far from the injection zone, or to migrate in a way that threatens drinking water. The EPA properly recognized the correct question to be whether molecular diffusion can cause hazardous waste to migrate outside the injection zone in violation of RCRA's "no migration" standard.

On the record before us, we uphold the EPA's finding that molecular diffusion may result in the migration of hazardous levels of hazardous constituents outside the injection zone in some cases. The EPA points to a study that shows that molecular diffusion can cause hazardous constituents to migrate dozens of feet outside the injection zone. D. Morganwalp & R. Smith, Modeling of Representative Injection Sites (undated), reprinted in J.A. 1232. Of itself this may not seem a large migration, but it must be remembered that the statute requires that there be no migration of hazardous constituents outside the injection zone. The EPA could conclude that the statute forbids even a migration of only a few feet.

The EPA therefore properly concluded from the record that molecular diffusion was a physical mechanism that could cause a forbidden migration of hazardous constituents. In designing its permitting process, the EPA certainly had the power to require that applicants, when making their "no migration" demonstrations, take into account all known mechanisms that can cause forbidden migrations. The EPA therefore properly required applicants to demonstrate that there would be no migration of hazardous constituents via molecular diffusion.

Did the EPA abuse its discretion in requiring injectors to show that there will be no migration of hazardous constituents from the injection zone for 10,000 years? CMA claims both that Congress intended a shorter period, and that 10,000 years is longer than the hazardous life of the wastes.

First, Congress never decided for itself that a shorter period than 10,000 years was adequate for keeping hazardous wastes in the injection zone. CMA can point to no decisive support for its less than 10,000 year claim in the statutory text or history. It is true that the legislative reports speak of the decomposition time of hazardous waste in terms of centuries rather than millennia. See S.Rep. No. 284, 98th Cong., 1st Sess. 15 (1983) ("Wastes chemically decompose in a land disposal facility, although often this decomposition occurs very slowly stretching over centuries."); H.R.Rep. No. 198, 98th Cong., 1st Sess. 33, reprinted in 1984 U.S.CODE CONG. & AD.NEWS 5592 ("land disposal of wastes ... might be appropriate if there is a reasonable certainty that wastes will be contained in the very long-term (i.e., at least several hundred years)."). However, the statute mandates containment "for as long as the wastes remain hazardous." Subsection (g)(5). Congress could hardly have desired the EPA to regulate waste disposal only for centuries even if scientific predictions project that the waste could remain hazardous for a much longer period. The EPA chose 10,000 years as a time limit because it would be long enough to insure that the "no migration" standard would be met, and yet short enough to come within the limitations of predictability. See 53 Fed.Reg. at 28,126. The EPA's choice was reasonable and consistent with the statutory purpose of preventing migration of hazardous constituents for as long as the wastes remain hazardous.

CMA may be correct that some injected wastes will cease to be hazardous in less than 10,000 years; in such a case the regulations specifically allow the injector to seek relief on the basis of such a showing. The regulations require the injector to show that there will be no migration so long as the wastes remain hazardous, or for 10,000 years, whichever period is shorter. The EPA's regulation properly places its emphasis on the statutory requirement of preventing migration for as long as the wastes remain hazardous. It does not require injectors to control the wastes for a longer period.

For the foregoing reasons, we deny the industry's petitions for review.III. NRDC CLAIMS

A. EPA's Interpretation of the "No Migration" Standard.

Subsections (d), (e) and (g) of Sec. 6924 ban the land disposal of hazardous wastes unless, for each of three specified methods, it has

been demonstrated to the Administrator, to a reasonable degree of certainty, that there will be no migration of hazardous constituents from the disposal unit or injection zone for as long as the wastes remain hazardous.

42 U.S.C. Sec. 6924(d), (e), (g) (emphasis added). As we noted above, EPA read this "no migration" standard into subsection (f) as well, in the interests of uniform treatment. See 53 Fed.Reg. at 28,120 (col. 3). The issue here is whether the "no migration" standard would be violated if hazardous constituents seeped out of the storage area at a time where some of the stored wastes were still hazardous, but not those actually leaving.

"Hazardous constituent" is a term of art referring to a list of chemical compounds compiled at 40 C.F.R. Part 261, Appendix VIII. Since these are defined by molecular formulae without reference to concentrations, a single molecule of such a chemical is a "hazardous constituent." A hazardous waste, by contrast, is such only if various factors, including the concentration of hazardous constituents,12 actually make it hazardous to human health or the environment. 40 C.F.R. Sec. 261.11(a)(3)(ii); 42 U.S.C. Sec. 6921(b) (requiring the Administrator to list wastes as hazardous if they contain "certain constituents (such as identified carcinogens, mutagens, or teratagens) at levels in excess of levels which endanger human health."). Thus, read literally, the "no migration" standard would seem to prohibit the migration of even a single molecule (or perhaps an appropriate de minimis amount) for the statutory time period, even though the migrating waste is itself not hazardous at all.

There is, however, an ambiguity in the statutory definition of the period--"as long as the wastes remain hazardous." To which "wastes" does the clause refer? Since the consistency of wastes may vary throughout an injection zone (or a disposal unit), wastes in one part, particularly near the perimeter, may no longer be hazardous while wastes in another part still are. Can hazardous constituents in such non-hazardous wastes migrate without violating the "no migration" standard? Or must the operator assure no migration of constituents until all wastes throughout the injection zone are no longer hazardous? Noting this textual ambiguity, the EPA interpreted "the wastes" to refer to the wastes containing the hazardous constituents that are leaving the injection zone.13 52 Fed.Reg. at 32,453 (col. 3); 53 Fed.Reg. at 28,122 (col. 2). Under this reading, hazardous constituents may migrate so long as "the wastes" immediately surrounding them at the border are no longer hazardous, or, putting it slightly differently, so long as they do not migrate in high enough concentrations to be hazardous wastes.

NRDC's most powerful argument is then that EPA has artfully transformed the standard "no migration of hazardous constituents " into a quite different standard--"no migration of hazardous wastes." We find this challenge quite close, but the ambiguity is serious enough to allow EPA under Chevron to resolve the issue in the way it has.

The EPA rests in part on a structural contention that "[o]rdinarily the term 'hazardous constituents' has no regulatory effect unless concentrations are also considered." 53 Fed.Reg. at 28,122 (col. 3). See also Brief for Respondent at 49-51; 55. This is true in the sense that RCRA "ordinarily" regulates "hazardous wastes," which of course are defined (in part) in terms of concentrations of hazardous constituents, but it is obviously not a basis for reading "constituents" as "wastes" when Congress chooses the former. Nonetheless the exceptional character of the usage may suggest a need to be on guard against facile literalism.

The parties agree that we can learn something from Sec. 6925(j), one of the rare other places in RCRA where Congress used the term "constituents" with a material regulatory effect. Section 6925(j) conditions a variance from certain technological control requirements for certain surface impoundments on the operator's proving that there will be "no migration of any hazardous constitutent [sic] into ground water or surface water at any future time." 42 U.S.C. Sec. 6925(j)(4). Even the EPA assumes that this other "no migration" standard prohibits migration of any hazardous constituent (down to a single molecule or an appropriate de minimis level). See Brief for Respondent at 54-55. Nonetheless, Sec. 6925(j) does not contain the ambiguous "for so long as the wastes remain hazardous" clause. Furthermore, it contains an additional, if not earth-shattering, textual difference. While Sec. 6924 prohibits "migration of hazardous constituents," Sec. 6925(j) prohibits "migration of any hazardous constituent." If Congress had wanted Sec. 6924's "no migration" standards to cover a single molecule, it could have expressed the point as aggressively as in Sec. 6925. And of course the context suggests a reason for a difference: Sec. 6925(j)(4) protects water resources, which are not directly implicated under the controlling language of Sec. 6924, and its "no migration" requirement works only as a condition for a variance. Overall, however, we find the statutory language and structure ambiguous and so turn to the legislative history.

Going into the Joint Conference, the House bill's precursor of Sec. 6924 did not insist on anything like "no migration." It required the EPA to ban disposal of hazardous waste into deep injection wells if it "may not be protective of human health and the environment for as long as the waste remains hazardous." H.R. 2867, 98th Cong., 1st Sess. Sec. 5(a) ("Land Disposal of Certain Hazardous Wastes") at 21 (1983). Other sections contained similar standards for prohibiting various methods of land disposal. See id. at 18-19, 23-24. The bill's only "no migration" standard was in its version of what became Sec. 6925(j). See id. Sec. 5(b) ("Interim Status Surface Impoundments") at 29. But for purposes of Sec. 6925(j) the House bill specially redefined the term "hazardous constituent" to encompass only such constituents as would migrate "in concentrations which may adversely affect human health or the environment." Id. Sec. 18 ("Hazardous Constituents") at 76. See also H.R. Conf.Rep. No. 1133, 98th Cong., 2d Sess. 96-97, reprinted in 1984 U.S.CODE CONG. & ADMIN.NEWS 5649, 5667-68 (noting House's restrictive definition). The House specifically added this because it was concerned that "there may be no facilities that meet that kind of a rigorous test," i.e., the strict "no migration" test. 129 Cong.Rec. 27,673; see also 129 Cong.Rec. 30,827-28 (adopting definition to clarify House's intention).

The Senate bill provided a "no migration" standard in its prototype of Sec. 6924 but, as even NRDC recognizes, it was clearly not as stringent as NRDC's interpretation of Sec. 6924 as enacted. See Brief for Petitioners NRDC at 22-23. It provided that

if a specified waste contains significant concentrations of one or more hazardous constituents that is highly toxic, highly mobile, or has a strong propensity to bioaccumulate, a method of land disposal may not be determined to be protective of human health and the environment for such specified hazardous waste, unless ... there will be no migration of such constituents from the disposal unit or injection zone for as long as the wastes remain hazardous.

130 Cong.Rec. 20,867-68 (setting forth final Senate Bill) (emphasis added). See also S.Rep. No. 284 at 84 (identical standard in earlier Senate draft). First, as NRDC concedes, this proposal required a showing of "no migration" only for those wastes that contained significant concentrations of certain hazardous constituents (those that are highly toxic, highly mobile, or have a strong propensity to bioaccumulate). Brief for Petitioner NRDC at 22. Moreover, although there is some ambiguity in the proposal's "no migration" standard too, the best interpretation seems to be that "such constituents" refers back to "significant concentrations" of highly toxic, etc. "hazardous constituents." That reading explains why the Senate Report uses the terms constituents and wastes interchangeably in describing the standard. See S.Rep. No. 284 at 14 (describing the standard as "no migration of [certain] constituents ") (emphasis added); id. at 15 (describing the standard as prohibiting the "migration of the wastes ") (emphasis added).

To complete the picture on the eve of conference, we note the Senate's version of Sec. 6925(j)'s "no migration" standard. This used virtually the same language as Congress enacted in Sec. 6925(j), see H.R. 2867 as amended in Senate, 98th Cong., 2d Sess. Sec. 6(c), 130 Cong.Rec. 20,866, 20,869 (col. 3),14 with no limiting definition of "hazardous constituent" equivalent to the House's. The Senate expressly noted the special ground for concern that we have already mentioned--Sec. 6925(j)'s application to migration into ground water. See 130 Cong.Rec. 20,847 (statement accompanying amendment) (expressing grave concern that hazardous contents in surface impoundments can migrate into "the surrounding soils where even dilute concentrations of toxic substances can, over time, pollute the ground water"). Thus, of the four clauses heading into conference (one from each house for each of Secs. 6924 and 6925(j)), only one, the Senate version of Sec. 6925(j), had a clause barring migration of hazardous constituents regardless of concentration.

The Conference's discussion of the various clauses gives no hint of any intention to go beyond the two houses' standards for Sec. 6924. As to Sec. 6925(j), there was a clear conflict between the House and Senate versions, because the House's "no migration" standard included a limiting definition of "hazardous constituent" and the Senate's did not. The Conference Report explicitly opted for the more stringent Senate version:

[T]he owner or operator must demonstrate no migration of any hazardous constituent into groundwater or surface water at any future time. In this formulation, the Conferees explicitly reject the provision contained in section 18 of the original House bill modifying the definition of hazardous constituent.

H.R.Conf.Rep. No. 1133 at 98, reprinted in 1984 U.S.CODE CONG. & ADMIN.NEWS at 5669. By contrast, the Conference Report is silent on Sec. 6924's "no migration" standard, even though a conflict existed here as well, between the Senate's "no migration" standard and the House's general requirement of protecting human health and the environment. Obviously the Conference followed the Senate, changing its language in the course of doing so. In place of "no migration of such constituents" ("such" referring back to "significant concentrations" of certain types of hazardous constituents), it substituted "no migration of hazardous constituents." Textually this minor change was necessary because another substantive change in the section15 had made it grammatically impossible for "such" to refer to its previous antecedent. If NRDC's interpretation were adopted, however, this small textual change would have radically altered the Senate's "no migration" standard and imposed a standard significantly more stringent than either house had passed.

Of course nothing prevents a conference from adopting a position quite different from both houses. There can be compromises like the British naval rearmament decision of 1909, when, as Churchill put it, "The Admiralty had demanded six ships; the economists offered four; and we finally compromised on eight." See Ted Morgan, Churchill: Young Man in a Hurry, 1874-1915 249 (1982). But one would expect at least a word of explanation in the Conference Report. In fact all it said was that "[t]he Conference substitute adopts provisions from both the House bill and the Senate amendment." H.R.Conf.Rep. No. 1133 at 86, reprinted in 1984 U.S.CODE CONG. & ADMIN.NEWS at 5657. Moreover, the participants seem most often to have either ignored or misunderstood the distinction between hazardous constituents and hazardous wastes, though it is central to NRDC's interpretation. See S.Rep. No. 284 at 14-15 (conflating its no migration of "such constituents" standard with a no migration of "wastes" standard); 129 Cong.Rec. 27,673 (col. 2)--74 (col. 1) (Rep. Breaux repeatedly assuring other Congressmen that "hazardous constituent" would be defined in terms of concentration levels, even before the House amended the definition of hazardous constituent). As to Sec. 6925(j), of course, Congress ultimately did focus on the distinction, made a choice, and explained it, and clearly expressed it in the statute. As to Sec. 6924, however, there is neither focus nor choice nor explanation nor clarity.

Finding no clear congressional determination in Sec. 6924 that hazardous constituents must not migrate regardless of hazard, and the reasonableness of the EPA's policy choice being unquestioned, we affirm its decision.

B. The Definition of "Injection Zone."

Though nowhere in RCRA is there a statutory definition of the term "injection zone," both EPA and NRDC agree that Congress intended to incorporate the pre-existing regulatory definition from EPA's Safe Drinking Water Act regulations, 40 C.F.R. Sec. 146.3:

Injection zone means a geological "formation", group of formations, or part of a formation receiving fluids through a well.

See S.Rep. No. 284 at 16 (citing regulations); 53 Fed.Reg. at 28,121 (col. 3); Brief for Respondent at 59; Brief for Petitioners NRDC at 27.

The EPA adopted this definition for the regulations, but it added a new spin, at least as NRDC views it. Under the new interpretation, the injection zone could extend across both the porous formations into which fluid is actually injected and the confining material, which is composed of less permeable layers of rock and clay. 53 Fed.Reg. at 28,121 (col. 3)--22 (col. 1). As a result of this view, seepage of hazardous constituents into confining material will not qualify as forbidden migration where the particular confining material has been included as part of the injection zone. NRDC argues that confining material cannot be included in an injection zone because it does not "receiv[e] fluids through a well."

First, though both parties agree that the Senate Report's reference to the SDWA definition binds EPA, we think the proper standard of review here is highly deferential. When Congress refers to a pre-existing regulatory definition, it may intend that the then-current interpretation be fro zen as part of the definition or that the agency retain at least an approximation of its ordinarily large discretion to interpret (and reinterpret) its own regulations. See United States v. Larionoff, 431 U.S. 864, 872-73, 97 S. Ct. 2150, 2155-56, 53 L. Ed. 2d 48 (1977) (agency's interpretation of its own regulations will be accepted unless "plainly inconsistent with the wording of the regulations"); General Carbon Co. v. Occupational Safety and Health Review Comm'n, 860 F.2d 479, 483 (D.C.Cir.1988) (same). Of course the interpretive discretion may not be quite as broad as where the regulation is exclusively of the agency's own choice; here (on the parties' and our assumption) the agency lacks the power to amend the regulation, and that power is the main rationale for the agency's special discretion in reading its own regulations. In this case, in any event, the Senate Report suggests the dynamic view--an expectation that as the EPA implemented the 1984 Amendments to RCRA, it would apply its experience to the definitional process:

In determining appropriate confinement from which migration shall not be allowed to occur, the terms disposal unit or injection zones should be construed ... in terms of the overall environmental integrity of the disposal practice, keeping in mind, in particular, the potential for contamination of groundwater or surface water resources.

S.Rep. No. 284 at 15.

But even if we were to ignore this passage and to treat the injection zone definition in Sec. 146.3 as if it were part of the statute, Chevron would require us to accept the agency's interpretation.

NRDC suggests that EPA's construction has created an improper overlap between "injection zone" and "confining zone," another term defined in the pre-existing SDWA regulations:

a geological formation, group of formations, or part of a formation that is capable of limiting fluid movement above an injection zone.

40 C.F.R. Sec. 146.3. NRDC argues that EPA's new reading allows the injection zone to extend into what should properly be considered only the confining zone. But the language of the definitions does not explicitly indicate that they are to be mutually exclusive. Nor does their logic or sense.

NRDC also urges that the EPA so distorted the term injection zone that it needed to create a new term, the "injection interval," defined as "that part of the injection zone ... in which the waste is ... directly emplaced," 53 Fed.Reg. at 28,148 (col. 2), codified at 40 C.F.R. Sec. 146.61(b), to serve the function Congress contemplated for the injection zone when it defined that as formations "receiving fluids through a well."

But in fact fluids from an injection well can seep into the confining material, and when they do that material will "receiv[e] fluids through a well." Moreover, because some such seeping is inevitable, see 53 Fed.Reg. at 28,122 (col. 1), NRDC's interpretation of injection zone, coupled with the "no migration" standards of Sec. 6924, would result in an absolute ban on deep well injection of wastes--a result Congress plainly did not intend. Furthermore, as EPA has argued (and NRDC offered no evidence to dispute), "there is often not a line where a large permeable strata meets relatively less permeable strata." 53 Fed.Reg. at 28,122 (col. 1). The definitional issue here is highly technical, and we can find no legal fault in EPA's solution.

C. Geological Repositories or Salt Domes.

Section 6924(b)(1) prohibits the disposal of certain wastes in salt dome formations, salt bed formations, underground mines, or caves (collectively referred to here as salt domes) until three conditions have been satisfied. Two are relevant to this case:

(B) the Administrator has promulgated performance and permitting standards for such facilities under this subchapter, and;(C) a permit has been issued under section 6925(c) of this title for the facility concerned.

Sec. 6924(b)(1)(B)-(C).

As to (B)'s requirement, EPA announced that the new regulations governing underground injection of hazardous wastes (the "amended Part 146 regulations") "could constitute performance and permitting standards for [salt domes]." 53 Fed.Reg. at 28,131 (col. 2) (emphasis added). See also Brief for Respondent at 66. But the agency hedged on the point, adding:

It is premature to discuss specific features which would make the application of Part 146 standards inappropriate. Such determinations will be made in the context of a facility's permit application.

53 Fed.Reg. at 28,131 (col. 3). NRDC charges that as caves and mines are different from the sort of porous strata for which the regulations are explicitly designed, EPA never established any basis for believing that the regulations could accomplish Congress's purpose. Comments of the Natural Resources Defense Council, Inc., et al., on EPA's Proposed Land Disposal Ban Rule for Underground Injection Wells at 66-67 (FRL-3220-3). See also 129 Cong.Rec. 27,664 (col. 3) (Rep. Breaux, sponsor of the amendment covering salt domes, stating that it banned disposal "until such time as we have a study to show that it can be done safely"). Instead of addressing this challenge, EPA said that the permit process would allow the public to challenge the applicability of Part 146 in specific geological settings. Brief for Respondent at 74; 53 Fed.Reg. at 28,131 (col. 3).

EPA says the issue is not yet ripe because it has made no final decision on when the Part 146 standards would be appropriate for salt domes; it will do so case by case as it reviews permit applications. 53 Fed.Reg. at 28,131 (col. 3); Brief for Respondent at 66. But this fails to respond to NRDC's ancillary procedural point: that subsection (B) requires advance promulgation of regulations suitable for salt domes. The point may sound persnickety, but isn't. If these regulations may be applied to salt domes, then this rulemaking is the only rulemaking format within which NRDC (or others) could challenge their suitability. NRDC naturally wants a chance to challenge generic standards rather than being reduced to prolonged guerilla warfare in the permit process. EPA's statements in both the proposed and final rules leave no doubt that it has reached the conclusion that the Part 146 regulations can (in appropriate circumstances) be used for geological formations. See 52 Fed.Reg. at 32,449 (col. 2) ("The Agency is proposing that the [amended Part 146] regulations ... constitute performance and permitting standards for such facilities as required by section [6924(b)(1)(B) ]."); 53 Fed.Reg. at 28,131 (cols. 2-3). Thus the procedural issue--must EPA have regulations in place, which it has established, in a rulemaking, as at least contingently applicable to salt domes, before it issues any permits?--is clearly ripe. Cf. Air Transport Ass'n of America v. Dep't of Transportation, 900 F.2d 369, 374 (D.C.Cir.1990).

Despite the usual broad discretion of agencies to select as between rulemaking and adjudication, SEC v. Chenery Corp., 332 U.S. 194, 203, 67 S. Ct. 1575, 1580, 91 L. Ed. 1995 (1947), NLRB v. Bell Aerospace Co., 416 U.S. 267, 294, 94 S. Ct. 1757, 1771, 40 L. Ed. 2d 134 (1974), we agree with NRDC. Section 6924(b)(1)(B) requires prior promulgation of standards "under this subchapter," surely incorporating Sec. 6924(a)'s procedural requirement that Sec. 6924 regulations be promulgated "after opportunity for public hearing." See Sec. 6924(a) and Sec. 6924(b)(1)(B) (referencing the procedures of Sec. 6924(a)). EPA has, to be sure, promulgated standards, but it has not explained their application to salt domes (which would be required substantively if they are to apply) nor defined in any way the class of salt domes to which they may apply.

We assume great flexibility for EPA in this. For example, it could presumably develop one or more sets of permit requirements for salt domes, with some sort of gateway indicating--perhaps in quite general terms16--which types fall under which set. And there is no procedural obstacle to its identifying the Part 146 regulations as such a set. But for it to do so without giving a clue as to when those regulations will be applicable (and defending the lawfulness of their application to such cases) deprives NRDC of the procedural opportunity that Sec. 6924 provided. Accordingly, we remand for the EPA either to drop the assertion of potential applicability, or to state the class of salt domes covered and justify the sense of doing so.

EPA's second contested decision was that a permit issued under the underground injection well program of the Safe Drinking Water Act (40 C.F.R. parts 144 and 145) would satisfy subsection (C)'s RCRA permit requirement ("a permit ... issued under section 6925(c)") because such a permit qualifies as a RCRA permit under EPA's permit-by-rule regulations, 40 C.F.R. Sec. 270.60. NRDC argues that because the injection well permit process does not include some of the procedural protections required by statute for RCRA permits, the regulation transforming those permits into RCRA permits is invalid. See Brief for Petitioner NRDC at 37-39, citing Secs. 6974(b) (public notice requirements), 6925(a) (requirement that permit issue before construction of facility begins), 6925(c)(3) (five-year review requirement and agency's omnibus power to add conditions to permit).

EPA's sole defense is that NRDC's statutory challenges are untimely; the original permit-by-rule regulation was promulgated in May 1980, see Consolidated Permit Regulations, 45 Fed.Reg. 33,290, 33,325-26, 33,435 (May 19, 1980), and the 90-day period for filing petitions for review, see 42 U.S.C. Sec. 6976(a), has long since expired. Brief for Respondent at 64-65. See also 53 Fed.Reg. at 28,131 (col. 2) (justifying application of permit-by-rule by reference to pre-existing regulation).

EPA's timeliness defense is dead wrong. All of NRDC's objections are based on statutory provisions added after the 90-day period for review of the 1980 permit-by-rule decision had run.17 The statutory time bar is expressly qualified to allow challenges after 90 days if "based solely on grounds arising after such ninetieth day." 42 U.S.C. Sec. 6976(a).

As EPA has offered no other defenses, we must reverse and remand this issue. On remand, EPA will, in the first instance, be able to explain more fully the procedural requirements of the various permit processes and to interpret the statutory requirements for Sec. 6925(c) permits. Nothing prevents the agency from integrating the underground injection well and RCRA procedures so as to avoid needless duplication (see, e.g., 45 Fed.Reg. at 33,325 (col. 3)), or from using other permissible techniques to alleviate administrative burdens in permitting (see NRDC v. Costle, 568 F.2d 1369, 1380-82 (D.C.Cir.1977) (Leventhal, J.) (discussing possible permitting procedures)). But before EPA can use its permit-by-rule regulations to satisfy Sec. 6924(b)(1)(C)'s requirement that a permit be "issued under section 6925(c)," it needs to demonstrate that the process meets the statutory requirements for such a permit.

D. Conclusion.

We reverse the EPA's actions on salt domes and remand for further proceedings consistent with this opinion. As to all other challenges, we deny NRDC's petition for review.

So ordered.

Congress has clearly announced that a method of land disposal of hazardous wastes cannot meet its standard of being "protective of human health and the environment" unless the disposer can prove there will be "no migration of hazardous constituents from the disposal unit or injection zone for as long as the wastes remain hazardous." 42 U.S.C. Sec. 6924(g)(5). That is a stringent requirement indeed, but one Congress intentionally imposed as the only alternative to a total ban on land disposal of hazardous waste. In its rulemaking, the EPA has, however, substituted for this requirement a far weaker requirement that there be no migration of hazardous waste from the injection zone. The difference is significant: Congress required that there be no migration of hazardous constituents; the EPA regulations allow some migration of hazardous constituents, so long as the migrating constituents are not at so high a concentration that the migrating waste is a hazardous waste. I respectfully dissent from the majority's holding that this is a permissible interpretation of the statute's "no migration" standard.1

In interpreting a statute, one begins with the words themselves. The no migration standard of subsection (g)(5) says that there must be "no migration of hazardous constituents from the disposal unit or injection zone for as long as the wastes remain hazardous." The two key terms are "hazardous constituents" and "the wastes."

1. "Hazardous Constituents "

The majority acknowledges that within the field of environmental law, "hazardous constituent" is a term of art with a very specific meaning: it refers to the substances listed in Appendix VIII to 40 C.F.R. Part 261. Any substance on this list (arsenic, for instance) is always a hazardous constituent, regardless of the concentration at which it may be present. A waste may contain hazardous constituents, yet not be a hazardous waste, if the concentration of hazardous constituents in the waste is so low that the waste poses no hazard to human health or the environment. Conversely, a waste may be a hazardous waste without containing a hazardous constituent, if it displays certain characteristics, such as corrosivity. See 40 C.F.R. Secs. 261.3 (defining "hazardous waste"), 261.22 (defining corrosivity). Thus, "hazardous constituent" and "hazardous waste," as used in RCRA, are not casual terms that vaguely describe dangerous things. They are terms of art with specific, distinct meanings.

The term used in the no migration prohibition is "hazardous constituents." By using this term, Congress expressed an unmistakable intention that there be no migration out of the injection zone of any substance listed in Appendix VIII of 40 C.F.R. Part 261 for as long as the wastes remain hazardous. Under the EPA's interpretation, however, hazardous constituents may migrate from the injection zone, so long as they do not do so at hazardous concentrations. This interpretation clashes with Congress' words: if a migrating waste contained hazardous constituents at hazardous concentrations, it would itself be a hazardous waste. The EPA is therefore reading the statute as though it simply said, "there shall be no migration of hazardous waste from the injection zone." Accepting the EPA's interpretation means accepting the notion that Congress speaks English so poorly that when it expresses a desire to ban the migration of hazardous wastes from the injection zone, it first introduces a totally different term, "hazardous constituents," and then qualifies that term in a way that robs it of its well-established meaning, rather than use the correct term in the first place.2

The EPA arrives at its reading in a convoluted manner; it interprets the term "the wastes" in the statutory phrase "for as long as the wastes remain hazardous" to refer only to the wastes that migrate from the injection zone, rather than to the wastes that remain within, and then goes on to interpret the term "hazardous constituents" in light of this strained interpretation of "the wastes." See Brief for Respondents at 44-45. Obviously the agency has gone about its interpretive task backwards. Even if the term "the wastes" were ambiguous (and the next section demonstrates that it is not), the EPA would still have to interpret it in the context of the sentence in which it appears. As we have recently said, "[w]hen a statute contains a clear term and an ambiguous term, the ambiguous term must be interpreted in light of the clear one, not vice-versa." Public Citizen v. Nuclear Regulatory Commission, 901 F.2d 147, 157 (D.C.Cir.1990). Here, the term "hazardous constituents" is a precise term of art, and the term "the wastes" is allegedly ambiguous, yet instead of interpreting "the wastes" in a way that makes sense of Congress' use in the same sentence of the term "hazardous constituents," the EPA has interpreted "the wastes" in a way that makes the term "hazardous constituents" entirely superfluous.

2. "The Wastes "

Furthermore, the EPA's interpretation of which wastes--migratory or residual--determine the permissibility of migration renders meaningless the statutory phrase, "for as long as the wastes remain hazardous." As the majority acknowledges, Maj. op. at 1159, this statutory phrase describes the time period during which migration of hazardous constituents is forbidden. Under the EPA's interpretation of "the wastes," however, the phrase has no relevance to any time period at all. As the EPA sees it, the statute does not bar migration of hazardous constituents from the injection zone for any time period at all; rather it bars migration of hazardous wastes from the injection zone, forever. That is, under the EPA's interpretation, the phrase, "for as long as the wastes remain hazardous" is only a reflection of the basic substantive restriction against migration of hazardous wastes that the statute imposes, not a temporal limitation. The majority believes that the EPA has resolved "an ambiguity in the statutory definition of the [time] period," Maj. op. at 1159, but to my mind it is peculiar to resolve such an acknowledged ambiguity in a way that makes the time period not a time period at all.

This conundrum becomes even more bizarre when one considers what the EPA's interpretation does to the word "remain." The statute bars migration of hazardous constituents "for as long as the wastes remain hazardous." In order for something to remain hazardous, it must first be hazardous. That is, the statutory term "the wastes" must refer to something that at one time is, but at some later time may not be, hazardous.

Under the EPA's interpretation, however, "the wastes" refers only to the wastes leaving the injection zone, and these wastes, according to the EPA, can never be hazardous. See Brief for Respondent at 57-59. The EPA's interpretation therefore raises pregnantly the question of why Congress would forbid migration "for as long as the wastes remain hazardous" if, as the EPA claims, Congress meant "the wastes" to refer to wastes that could under no circumstances be hazardous in the first place. Congress' use of the word "remain" is therefore yet another clear indication that "the wastes" does not refer to the wastes that migrate from the injection zone, but rather to the wastes that are deposited there in the first place.3

These interpretive difficulties can be resolved only if "the wastes" is read to refer to "the wastes injected into the injection zone." Only such an interpretation can make sense of the statutory terms Congress used in the section. We know the wastes are hazardous when injected; for as long as they remain hazardous, no hazardous constituents will be permitted to migrate from the injection zone.

The majority attempts to accommodate the statute's temporal limitation as well as its use of the word "remain" by suggesting still other meanings for the term "the wastes": the majority says it refers to "the wastes containing the hazardous constituents that are leaving the injection zone" and it "also refers to each part of the larger body of wastes in the injection zone." Maj. op. at 1159 n. 13 (emphasis added). But even overlooking the fact that the majority has simply replaced the EPA's definition with others of its own making, once the majority concedes that "the wastes" refers to wastes within the injection zone, even to "[w]astes at the very center of the injection zone that are still hazardous," id. at 1159 n. 13, it follows from the literal text of the statute that there can be no migration of hazardous constituents from the injection zone while these wastes remain hazardous. Both the majority and the EPA ultimately confront the same dilemma in their varying interpretations: if "the wastes" refers only to wastes that migrate from an injection zone, then the phrase "for as long as the wastes remain hazardous" is nonsensical; if "the wastes" includes each part of the wastes within the injection zone, then the EPA cannot allow any hazardous constituents to migrate from the injection zone for as long as the wastes within remain hazardous.4

Because Congress' use of the term "hazardous constituents" so strongly suggests that it meant to ban more than just the migration of hazardous wastes from the injection zone, the majority is forced to argue that Congress simply did not know what it was talking about. See Maj. op. at 1162 ("the participants seem most often to have either ignored or misunderstood the distinction between hazardous constituents and hazardous wastes"). In so concluding, the majority overlooks well-established tenets of statutory construction, and misreads clear indications in the legislative history that Congress was, indeed, well aware of the critical distinctions in the terms it used.

We noted recently that it is not the province of courts to "correct" Congress' drafting oversights. Rather, "we courts assume that Congress always knows the particulars whereof it speaks." We "assume that Congress intended that language which it chose to employ actually was to have meaning," and we give effect, "if possible, to every word, clause and sentence of a statute ... so that no part will be inoperative or superfluous." Consolidated Rail Corp. v. United States, 896 F.2d 574, 579 (D.C.Cir.1990) (quoting United States v. McGoff, 831 F.2d 1071, 1080 (D.C.Cir.1987) and National Association of Recycling Industries, Inc. v. ICC, 660 F.2d 795, 799 (D.C.Cir.1981)).

The only exception we recognize is the "rare case" in which "the literal application of a statute will produce a result demonstrably at odds with the intention of its drafters," United States v. Ron Pair Enterprises, 489 U.S. 235, 109 S. Ct. 1026, 1031, 103 L. Ed. 2d 290 (1989) (quoting Griffin v. Oceanic Contractors, Inc., 458 U.S. 564, 571, 102 S. Ct. 3245, 3250, 73 L. Ed. 2d 973 (1982)), and such a case "must involve, at a minimum, some clear indication of congressional intent, either in the legislative history or in the structure of the relevant statute, that informs the specific language in question." Consolidated Rail, 896 F.2d at 578. Here, since the language of the statute bars migration of hazardous constituents from the injection zone so long as the injected wastes remain hazardous, only a clear indication that Congress did not mean what it said when it chose the term "hazardous constituents" would satisfy our standard. An examination of the legislative history, however, shows the contrary. The history is replete with evidence that Congress well understood the import of using the term "hazardous constituents" rather than "hazardous wastes."

The term "hazardous constituent(s)" appears in several places in the HSWA. The history of 42 U.S.C. Sec. 6925(j), though lengthy, is quite instructive in that regard. Subsection (j) concerns surface impoundments of hazardous waste that existed on November 8, 1984, the effective date of HSWA ("existing surface impoundments"). Subsection 6924(o ) requires new surface impoundments to have double linings, but Congress was concerned that many existing surface impoundments had no lining or a single, leaky lining that allowed hazardous waste to leak into underground sources of drinking water. See 129 Cong.Rec. 27,664 (1983) (remarks of Rep. Breaux). Accordingly, subsection 6925(j) requires that four years after November 8, 1984, existing surface impoundments store no more hazardous waste unless they comply with subsection 6924(o ). However, Representative Breaux, who introduced the amendment on the House floor that later became subsection 6925(j), explained that he did not want to be unnecessarily hard on owners of existing surface impoundments. He believed that those owners should not have to alter existing impoundments if they could demonstrate their safety. Accordingly, subsection (j)(4) provides that the Administrator may modify the requirements for an existing surface impoundment if the owner can demonstrate that there will be "no migration of any hazardous constitutent [sic] into ground water or surface water at any future time."

Representative Lott questioned Representative Breaux about the meaning of his amendment. Representative Lott worried that no one could ever make a showing that there would be no migration at all from an impoundment. See 129 Cong.Rec. 27,673 (1983). Representative Breaux countered that "the definition of 'hazardous constituent' talks in terms of the concentration of that hazardous toxic chemical." Id. About a month later Representative Breaux nailed down his interpretation by introducing another floor amendment which provided, "the term 'hazardous constituent' does not include those hazardous constituents which ... will not migrate into ground or surface water in concentrations which may adversely affect human health or the environment." Id. at 30,827. His amendment passed the House.

However, after the bill went to conference, that specific amendment changing the normal definition of "hazardous constituent" was deleted, and the language "no migration of any hazardous constituent into ground or surface water at any future time" left intact. The Conference Report, commenting on this change, stated that "the Conferees explicitly reject the provision contained in [Representative Breaux's amendment] modifying the definition of hazardous constituent." H.R.Conf.Rep. No. 1133, 98th Cong., 2d Sess. 98, reprinted in 1984 U.S. CODE CONG. & AD. NEWS 5649, 5669 (emphasis added).

Thus, the history of Sec. 6925(j) shows that in formulating the final text of HSWA, Congress explicitly considered the definition of hazardous constituent as distinct from hazardous wastes. It knew what it meant when it required no migration of hazardous constituents, and it realized fully just how strict this requirement was. The Conference Report's statement strongly confirms the NRDC's interpretation that Congress used the term "hazardous constituent" as a term of art to refer to any substance listed in 40 C.F.R. Part 261 Appendix VIII regardless of concentration, and it specifically rejected EPA's more lenient interpretation.5

The term "hazardous constituents" also appears in 42 U.S.C. Sec. 6924(u), which provides that permits issued after the effective date of HSWA shall require "corrective action for all releases of hazardous waste or constituents " (emphasis added) from a disposal facility, regardless of when waste was placed there. This provision originated in the Senate Bill, and the Senate Report, commenting on it, states that "[t]he requirement for corrective action applies not just to releases of hazardous wastes, but also to releases of hazardous constituents." S.Rep. No. 284, 98th Cong., 1st Sess. 32 (1983) (emphasis added). This comment shows once again that the term "hazardous constituents" has a precise meaning that differs from the meaning of the term "hazardous wastes," and that Congress was well aware of the difference.6

The majority relies finally on the absence in Sec. 6924's history of any indication that the congressional conferees intended the final "no migration" standard to be stronger than the standard proposed by either the Senate or the House of Representatives. However, with respect to the point at issue here (whether EPA may allow hazardous constituents to migrate from an injection zone in nonhazardous concentrations), the final HSWA's "no migration" standard is not in fact any stricter than the standard contained in the Senate Bill, so that the absence of comment in the Conference Report is understandable.

As the majority notes, the Senate version of Sec. 6924's "no migration" standard read as follows:if a specified waste contains significant concentrations of one or more hazardous constituents that is highly toxic, highly mobile, or has a strong propensity to bioaccumulate, a method of land disposal may not be determined to be protective of human health and the environment for such specified hazardous waste, unless ... there will be no migration of such constituents from the disposal unit or injection zone for as long as the wastes remain hazardous.

S.Rep. No. 284, 98th Cong., 1st Sess. 84 (1983) (emphasis added). The majority contends that the term "such constituents" should be interpreted in context to refer to "significant concentrations of one or more hazardous constituents that is highly toxic," etc., and the Senate Bill therefore prohibited migration only of significant concentrations of hazardous constituents. Simply as a matter of English, this is an impossible interpretation of the term. In the Senate Bill, "such constituents" clearly means "highly toxic, mobile, or bioaccumulative hazardous constituents," and the Bill therefore says that if a specified waste contains significant concentrations of one or more hazardous constituents that is highly toxic, mobile, or bioaccumulative, then that waste cannot be disposed on the land unless there will be no migration of the highly toxic, mobile, or bioaccumulative hazardous constituents, at any concentration.

The Senate Report explicitly sets forth this interpretation of the Senate Bill. The Report instructs the EPA to carry out a two-step assessment with respect to hazardous wastes. First, the Administrator must consider a waste's inherent characteristics, and determine whether it contains significant concentrations of one or more hazardous constituents that are highly toxic, highly mobile, or have a strong propensity to bioaccumulate. If it does, it is presumptively prohibited from land disposal. S.Rep. No. 284, 98th Cong., 1st Sess. 14 (1983). The presumption for prohibition may be overcome with respect to a particular method of land disposal if the Administrator determines that the method will be safe. Id. The Report then states, "[t]his determination may be made if the Administrat[or] finds, to a reasonable degree of certainty, that no migration of the highly mobile, highly toxic, or highly bioaccumulative constituents will occur from the disposal unit or injection zone, for as long as the waste remains hazardous." Id. (emphasis added). The Senate Report uses the phrase "highly mobile, highly toxic, or highly bioaccumulative constituents" where the Senate Bill uses the term "such constituents."

The Senate Report definitively refutes the majority's interpretation of the Senate Bill and verifies that the vital element of the final no migration standard--that it means no migration of hazardous constituents, at any concentration--was already present in the Senate Bill. The Senate Report comments on this standard extensively; the Conference Report thus had no need to make further comment.7

C. The Policy Behind the No Migration Standard

Because the majority accepts the EPA's interpretation of the "no migration" standard, it does not need to respond to the EPA's argument that the NRDC's interpretation would produce an anomalous result: although no hazardous constituents could migrate from the injection zone for as long as the wastes within remained hazardous, once the wastes within the injection zone were no longer hazardous, hazardous constituents could migrate--even at hazardous concentrations. In fact, this result would not be permitted under the statute, and understanding why helps one to understand the purpose behind the "no migration" standard.

Section 6924 contains two safety standards that a method of land disposal of hazardous wastes must meet. First, the method must be "protective of human health and the environment for as long as the waste remains hazardous."8 This general standard gives EPA the power and the duty to ensure the safety of all methods of land disposal. The "no migration" standard, which requires that there be "no migration of hazardous constituents from the disposal unit or injection zone for as long as the wastes remain hazardous," is a second, more specific requirement that a disposal method must also meet. If a waste disposal method satisfies the "no migration" standard, but for some other reason is unsafe, the EPA must prohibit it under Sec. 6924's general safety standard.

The following example proves that the general safety standard and the "no migration" standard are separate requirements. Suppose a manufacturer wishes to dispose of a hazardous waste that contains no hazardous constituents (it is a hazardous waste because it displays a hazardous characteristic). Suppose the manufacturer plans to inject the waste into an injection zone from which the waste will migrate while still hazardous. This method of disposal certainly satisfies the "no migration" standard: no matter how one reads the standard, it does not bar the migration of a waste containing no hazardous constituents.

Must the EPA therefore permit the disposal method? Of course not. The EPA has properly concluded, see Brief for Respondent at 58-59, that such a method must be prohibited. Although the method satisfies the no-migration-of-hazardous-constituents standard, it does not satisfy the general safety standard, that disposal methods must be protective of human health and the environment for as long as the waste remains hazardous.

As this example shows, Congress did not need the "no migration" standard to ensure that hazardous wastes not migrate from the injection zone. The general safety standard already does that. The "no migration" standard is an overprotective standard, inserted by Congress to ensure that not even hazardous constituents migrate from a body of waste known to be hazardous. It does not lead to any anomalous results, because the general standard prevents such results from occurring. It reflects Congress' evident concern that once hazardous constituents are known to be migrating from a body of waste known to be hazardous, the uncertainties involved are too great to allow the EPA to determine that the migration is at an "acceptable" level.

While there is no explicit expression of this specific concern in the legislative history proper, the statute does reflect a concern with "the long-term uncertainties associated with land disposal." Sec. 6924(d)(1)(A). I note also that eleven members of Congress, including several who were highly influential in the passage of HSWA, wrote a letter to the EPA Administrator Lee M. Thomas during an EPA rulemaking proceeding expressing this concern. They stated:

The requirement for proof of "no migration" is to be interpreted literally....

* * *

We specifically rejected the concept of an acceptable level of migration because of the scientific uncertainties associated with determining what is an "acceptable" level. For example, predicting the character and rate of migration, the fate and transport of the contaminants, and points of present and future human and environmental exposure are subject to significant error. The scientific uncertainty makes reliance on such predictions inconsistent with the statutory presumption against land disposal and in favor of treatment.

* * *

The phrase "as long as the waste remains hazardous" describes the time frame for which EPA must ascertain whether migration from the disposal unit will occur; it does not describe a substantive standard.

Letter from Rep. Dingell, Sen. Chafee, et al. to Lee M. Thomas 4-5 (Mar. 4, 1986), reprinted in Joint Appendix 374, 377-78. Five Senators expressed a similar concern in a letter during a later rulemaking proceeding, specifically objecting to the EPA's proposal to allow hazardous constituents to migrate from an injection zone in concentrations that the agency considers safe. See Letter from Sens. Durenburger, Baucus, Stafford, Mitchell, and Chafee to Lee M. Thomas 3 (Mar. 10, 1988), reprinted in Joint Appendix 1350, 1352.

Of course, these letters, being merely the post-enactment statements of individual members of Congress, deserve little weight in our task of statutory interpretation. I do not cite them as evidence of the statute's meaning; such evidence is abundant in the text of the statute and its legislative history proper. I cite them to show that the strict "no migration" standard that is clearly expressed in the text of the statute is not some wild, counterintuitive result that we should strain to avoid, see Brief for Respondents at 57-59, but rather is the natural result of a belief that migration is an inherently uncertain process.D. Conclusion

Congress did not use the term "hazardous constituents" idly. Congress understood that the term referred to substances on a certain list regardless of the concentration at which they were present; Congress did not want that meaning varied. Congress clearly understood that "hazardous constituents" was a technical term with a meaning distinct from the term "hazardous wastes." The EPA's conclusion notwithstanding, Congress, in enacting the no migration standard banning migration of "hazardous constituents," did not mean to ban only migration of "hazardous wastes."

Under Chevron, it is our duty to correct administrative agencies when they disregard the clear mandate of Congress. In this case, Congress has used well-established terms of art to express its meaning with great precision. I respectfully dissent from the majority's holding that the EPA permissibly interpreted RCRA's no migration standard.

Before WALD, Chief Judge; MIKVA, EDWARDS, RUTH B. GINSBURG, SILBERMAN, BUCKLEY, WILLIAMS, D.H. GINSBURG, SENTELLE, THOMAS, HENDERSON and RANDOLPH, Circuit Judges.

Oct. 5, 1990.

Petitioners' Suggestion for Rehearing En Banc has been circulated to the full court. The taking of a vote was requested. Thereafter, a majority of the judges of the court in regular active service did not vote in favor of the suggestion. Upon consideration of the foregoing it is

ORDERED, by the court en banc, that the suggestion is denied.

A statement of Chief Judge WALD, joined in by Circuit Judges MIKVA and EDWARDS, is attached.

Circuit Judge BUCKLEY would grant the suggestion for rehearing en banc.

WALD, Chief Judge, dissenting from the denial of rehearing en banc, with whom MIKVA and EDWARDS, Circuit Judges, join:

I believe that rehearing en banc should be granted in this case because of its far-reaching impact upon the entire land disposal schemata for hazardous wastes contained in Sec. 3004 of the Resource Conservation and Recovery Act. The keystone of Congress' mandate in this section is that in any land disposal method, i.e. deep wells, landfills, waste piles and salt dome formations, there be "no migration of hazardous constituents from the disposal unit ... for as long as the wastes remain hazardous" (emphasis added). The majority and EPA have changed that directive to permit leakage and spill of hazardous constituents so long as they do not leach in sufficient quantity to themselves qualify as hazardous wastes at the point they are leaving the unit. (Ironically, EPA and the majority do not even use the same definition of hazardous wastes.) This reading appears an unnatural one, since it is wastes that are injected in the first place, and Congress clearly knew how to distinguish between constituents and wastes. It is also, obviously, a much less protective standard than the literal language of the statute requires. And it does not accord with the known behavior of many kinds of wastes.

I believe this to be one of a rare breed of "Chevron I " cases where "Congress has directly spoken" but the agency has not listened, with critical consequences for the environment. It was because Congress could not begin to predict the behavior of hazardous wastes 10,000 years from now that it adopted the stringent "constituent," no migration bar. Whether the reformulation of its environmentally protective standard through the tortured interpretation of EPA and the majority is warranted is surely and issue of extraordinary importance, justifying en banc treatment.