International Fabricare Institute v. U.S. Environmental Protection Agency

OPINION PER CURIAM. PER CURIAM: In these consolidated cases, petitioners challenge regulations promulgated by the Environmental Protection Agency (“EPA” or “Agency”) pursuant to the Safe Drinking Water Act. The regulations establish permissible concentration levels for contaminants occurring in drinking water. Petitioners claim that the EPA committed both substantive and procedural errors in formulating the regulations. Before turning to the merits of these challenges, we first consider the EPA’s contention that one of the consolidated petitioners, International Fabricare Institute, lacks standing to challenge the regulations because its claims of injury are too speculative. We reject this contention and conclude that all petitioners satisfy both the constitutional and prudential requirements for standing. Petitioners raise a general challenge to the EPA policy rejecting the existence of safe threshold levels for carcinogens in the absence of contrary evidence. Petitioners then present procedural and substantive challenges to the permissible concentration levels established for two contaminants— dibromochloropropane and ethylene dibromide. Petitioners also allege procedural defects in the regulations pertaining to perchloroethylene. Finally, petitioners assert that the EPA ignored the Administrative Procedure Act’s notice and comment requirements when it chose a method for measuring polychlorinated biphenyls in water samples, and that we should therefore remand to the EPA both the choice of the method and the permissible contaminant level predicated on that method. We conclude that the EPA complied with notice and comment requirements and adequately explained its rulemaking. Regarding the substantive challenges, we cannot find that the EPA’s actions were arbitrary and capricious. We therefore deny the consolidated petitions in all respects. I. Background The EPA promulgated the disputed regulations to implement the Safe Drinking Water Act (“SDWA”1 or “Act”), 42 U.S.C. §§ 300f-300j-25 (1988), which requires the Agency to establish national drinking water standards and to issue regulations to prevent the harmful contamination of public water systems. The regulations are to identify contaminants occurring in drinking water that may have an adverse effect on health, and to regulate them to the extent cost and technology permit. The regulations are to establish a “maximum contaminant level goal” (“MCLG”) for each identified contaminant. Id. § 300g-l(a)(3). An MCLG is a non-enforceable goal that is to be set “at the level at which no known or anticipated adverse effects on the health of persons occur and which allows an adequate margin of safety.” Id. § 300g-1(b)(4). The regulations for most contaminants also specify an enforceable “maximum contaminant level” (“MCL”) that sets the maximum permissible level of the contaminant in water delivered to any user of a public water system. See id. § 300f(3). MCLs must be set “as close to the [MCLG] as is feasible.” Id. § 300g-1(b)(4). “Feasi — 335 — ble” means using “the best technology, treatment techniques and other means ... available (taking cost into consideration).” Id. § 300g-l(b)(5). On November 13, 1985, the EPA published a notice of proposed rulemaking for the regulation of specific water contaminants. See National Primary Drinking Water Regulations; Synthetic Organic Chemicals, Inorganic Chemicals and Microorganisms, 50 Fed.Reg. 46,936 (1985) (“1985 Proposed Rule”). The EPA resubmitted the rulemaking on May 22, 1989, to reflect revisions to the Act made by Congress in amendments enacted in 1986. See National Primary and Secondary Drinking Water Regulations, 54 Fed.Reg. 22,062, 22,-068 (1989) (“Proposed Rule”). This revised proposal contained MCLGs and MCLs for thirty-eight organic and inorganic chemicals. Id. at 22,064. The Proposed Rule also responded to the comments the EPA had received on the 1985 proposal. The EPA received approximately 170 comments on the Proposed Rule. The EPA held a public hearing and issued its response to them. See Comment/Response Document on 1989 Proposed Rule (Dec. 31, 1990), reprinted in Joint Appendix (“J.A.”) at 1131-1483. Then in January 1991, the EPA promulgated the final regulations protested in this petition. See National Primary Drinking Water Regulations—Synthetic Organic Chemicals and Inorganic Chemicals; Monitoring for Unregulated Contaminants; National Primary Drinking Water Regulations Implementation; National Secondary Drinking Water Regulations, 56 Fed.Reg. 3,526 (1991) (to be codified at 40 C.F.R. parts 141-143) (“Final Rule”). On March 28, 1991, several petitioners filed challenges to the final regulations. The petitions involve four contaminants. Petitioners Dow Chemical Company, Shell Oil Company, and Occidental Chemical Corporation seek review of the MCLs and MCLGs for 1,2 dibromo-3-chloropropane (“DBCP”). Dow also challenges the regulation with respect to ethylene dibromide (“EDB”). International Fabricare Institute (“IFI”) and intervenor Halogenated Solvents Industry Alliance (“HSIA”) ask us to set aside the regulations concerning tetrachloroethylene, also known as perchloroethylene (“perc”). Finally, National Electrical Manufacturers Association (“NEMA”) and Chemical Manufacturers Association (“CMA”) seek review of the rule relating to polychlorinated biphenyls (“PCBs”). We have consolidated these various petitions. See Order Granting Motion to Consolidate (D.C.Cir. May 14, 1991). As part of its methodology for establishing MCLGs, the EPA has developed a classification scheme that sorts contaminants into “Groups” (A to E) and “Categories” (I to III). See Final Rule, 56 Fed.Reg. at 3,532. Group A contaminants are considered to be known human carcinogens based on sufficient human epidemiologic evidence. Id. Group B1 substances are considered probable human carcinogens based on limited human epidemiological evidence. Id. Group B2 chemicals are classified as probable human carcinogens “based on a combination of sufficient evidence in animals and inadequate data in humans.” Id. Group C contaminants are defined as “[p]ossible human carcinogen[s] based on limited evidence of carcinogenicity in animals [and] the absence of human data.” Id. Group D is reserved for substances that the EPA finds to be unclassifiable due to lack of data or inadequate evidence of carcinogenicity. Id. Finally, Group E chemicals are those for which the EPA finds no evidence of carcinogenicity based on adequate animal tests or animal and human epidemiological studies. Id. In most cases, the EPA places Group A, Bl, and B2 contaminants into Category I, Group C into Category II, and Groups D and E into Category III. Id. The EPA sets MCLGs for Category I chemicals at zero because it assumes, “in the absence of other data, that there is no known threshold” at which these known or probable carcinogens can be safely tolerated. Id. at 3.533. In contrast, Category II and III substances are set at levels “likely to be without an appreciable risk of deleterious health effects during a lifetime.” Id. at 3,532. — 336 — Consolidated petitioners first raise a general challenge to this method for establishing MCLGs of zero for known or suspected carcinogens. They assert that the EPA erred by failing to address authoritative evidence indicating that there are safe threshold levels for carcinogens. Turning to the specific contaminants, petitioners claim that the EPA acted arbitrarily and capriciously in setting the MCLGs and MCLs for DBCP and EDB, largely because of the EPA’s alleged failure to address the available human epidemiological data on those chemicals. Regarding perc, petitioners challenge the EPA’s action in setting the MCLG at zero, arguing that this was arbitrary and capricious because the EPA ignored its own procedures when it placed perc, a Group C chemical, in Category I. Finally, petitioners assert that the methodology used to measure PCBs and the MCL based upon that technique should be vacated because the EPA did not provide an opportunity for public comment on the crucial data that supported the EPA’s conclusion that the methodology was reliable. We consider petitioners’ standing and each of these challenges in turn. II. Discussion A. Standard of Review Pursuant to the Administrative Procedure Act (“APA”), we will reverse an EPA action only if it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A) (1988). This highly deferential standard of review “presumes agency action to be valid.” Ethyl Corp. v. EPA, 541 F.2d 1, 34 (D.C.Cir.) (en banc), cert. denied, 426 U.S. 941, 96 S.Ct. 2662, 49 L.Ed.2d 394 (1976). The rationale for deference is particularly strong when the EPA is evaluating scientific data within its technical expertise: “[I]n an area characterized by scientific and technological uncertainty^] ... this court must proceed with particular caution, avoiding all temptation to direct the agency in a choice between rational alternatives.” Environmental Defense Fund, Inc. v. Costle, 578 F.2d 337, 339 (D.C.Cir.1978). Despite this deferential standard, we must ensure that the EPA has examined the relevant data and has articulated an adequate explanation for its action. See Motor Vehicle Mfrs. Ass’n of the United States, Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443 (1983) (“State Farm”). The “EPA is required to give reasoned responses to all significant comments in a rulemaking proceeding.” PPG Indus., Inc. v. Costle, 630 F.2d 462, 466 (6th Cir.1980). We will therefore overturn a rulemaking as arbitrary and capricious where the EPA has failed to respond to specific challenges that are sufficiently central to its decision. American Mining Congress v. EPA, 907 F.2d 1179, 1191 (D.C.Cir.1990). B. Standing As a threshold matter, the EPA contests the standing of petitioner IFI, an association of dry cleaning businesses, to challenge the regulations concerning perc. Moreover, prior to oral argument we advised the other petitioners in this action that we would consider their standing to challenge the Final Rule. The judiciary is constitutionally constrained from exercising its authority save when a party can “allege personal injury fairly traceable to ... allegedly unlawful conduct and likely to be redressed by the requested relief.” Hazardous Waste Treatment Council v. EPA, 861 F.2d 277, 282 (D.C.Cir.1988) (citation omitted), cert. denied, 490 U.S. 1106, 109 S.Ct. 3157, 104 L.Ed.2d 1020 (1989). “The harm can be actual or threatened, but it must be ‘distinct and palpable’; mere harm to an ideological interest will not suffice.” Competitive Enterprise Institute v. NHTSA, 901 F.2d 107, 112 (D.C.Cir.1990) (citations omitted). An organization such as IFI may have standing, even when the organization itself has not suffered an injury, if it can show that “(a) its members would otherwise have standing to sue in their own right; (b) the interests it seeks to protect are germane to the organization’s purpose; and (c) neither the claim asserted nor the relief requested requires the participation of individual members in the lawsuit.” — 337 — Hunt v. Washington State Apple Advertising Comm’n, 432 U.S. 333, 343, 97 S.Ct. 2434, 2441, 53 L.Ed.2d 383 (1977). Finally, there must be “some indicia—however slight—that the litigant before the court was intended to be protected, benefited or regulated by the statute under which suit is brought.” Copper & Brass Fabricators Council, Inc. v. Department of the Treasury, 679 F.2d 951, 952 (D.C.Cir.1982). We are satisfied that petitioners have cleared these hurdles. As explained above, the EPA uses MCLGs to establish enforceable MCLs that apply to operators of public water systems. Petitioners Shell, Dow, Occidental, and the National Electrical Manufacturers Association have alleged that they or their members operate public water systems and therefore will be directly regulated by the new drinking water standards established pursuant to the challenged rulemaking. Clearly, then, they have standing to bring their challenges. Additionally, IFI has met the standing requirements by showing that it or its members will likely face greater liability under the Comprehensive Environmental Response, Compensation and Liability Act (“CERCLA”), 42 U.S.C. §§ 9601-9675, if the EPA sets MCLGs for perc at the standard established in the challenged rulemaking. IFI alleges that it and some of its members have been subjected to lawsuits seeking cleanup costs under CERCLA for contamination caused by perchloroethylene, a chemical commonly used and disposed of in the dry cleaning industry. Among other things, CERCLA imposes liability for releases of hazardous substances (such as perc) and provides a general framework for establishing the level of cleanup liable parties must achieve at a contaminated site. Section 107 of CERCLA imposes joint and several liability for cleanup costs on persons who own or operate a site from which hazardous substances are released, or who arrange for the disposal of hazardous substances that are released. 42 U.S.C. § 9607(a); see generally McSlarrow, Jones & Murdock, A Decade of Superfund Litigation, 21 Envt.L.Rep. (Envt.L.Inst.) 10,-367, 10,390-94 (1991) (discussing scope of CERCLA liability) [hereinafter Superfund Litigation]. Section 121 then provides that cleanups paid for by liable parties must “assure[ ] protection of human health and the environment.” 42 U.S.C. § 9621(d)(1). A cleanup has complied with this directive when the levels of contaminants at a site have fallen to the applicable regulatory standards, which are borrowed from other statutory schemes such as the SDWA. Id. § 9621(d)(2)(A); see Superfund Litigation, at 10,378-79 (discussing method by which cleanup standards are chosen under § 121 of CERCLA). For example, at a site contaminated with perc, CERCLA would require the liable parties to pay for and perhaps execute a cleanup designed to remove perc down to a level based on the MCLG—the lower the MCLG, the more extensive, and expensive, the cleanup. Thus, MCLGs play a pivotal role not only in the regulation of public water systems, but also in the development of the cleanup standards that must be met by individuals found liable under CERCLA for releasing hazardous substances. We therefore conclude that since IFI has alleged that it and its members face a genuine threat of CERCLA liability for the cleanup of perc, and since a successful challenge could result in a higher MCLG for perc, thereby reducing the cost of any cleanup, IFI has standing to bring the present action. C. Zero Maximum Contaminant Level Goals The Act requires the EPA to set “[e]ach maximum contaminant level goal ... at the level at which no known or anticipated adverse effects on the health of persons occur and which allows an adequate margin of safety.” 42 U.S.C. § 300g-1(b)(4). Consistent with its past practice, see, e.g., NRDC v. EPA, 824 F.2d 1211, 1214-16 (D.C.Cir.1987); 50 Fed.Reg. 46,895 (1985), the EPA here set a goal of zero for each chemical it considered a known or probable human carcinogen (Category I contaminants). The EPA previously adopted the zero goal approach in another rulemaking under this statute. See National Primary Drinking Water Regulations; Volatile — 338 — Synthetic Organic Chemicals, 50 Fed.Reg. 46,880 (1985). It did so in part based on its reading of the legislative history of the Act, see id. at 46,884 col. 2 (quoting H.R.Rep. No. 1185, 93d Cong., 2d Sess. 20 (1974)), and in part based on a conclusion that, since current science could not ascertain exactly how a carcinogen caused cancer, “it was conservatively believed” that exposure to any dose could have adverse health effects. Id. On petition for review, this court determined that, in adopting its zero goal approach, the “EPA made an expert judgment” based on a “reasoned determination.” NRDC v. EPA, 824 F.2d at 1215. We upheld that judgment. Id. Perhaps in light of the foregoing, petitioners today argue neither that the EPA’s use of a zero goal approach is precluded by Congress's clear intent as expressed in the Act, nor that it is per se an impermissible exercise of the agency’s discretion under the Act. Rather, they assert that new scientific evidence exists that might make the EPA’s approach untenable, and the agency arbitrarily and capriciously failed to consider and address that evidence. Brief for Petitioners at 30-33. The evidence consisted primarily of a three-page letter to the editor, written by Drs. Bruce N. Ames and Lois Swirsky Gold of the University of California at Berkeley, published in Science magazine. Ames & Gold, Pesticides, Risk, and Applesauce, 244 Science 755 (May 19, 1989). The only other evidence submitted was the declaration of Dr. Gori. Comment to the EPA, Declaration of Dr. Gio Batta Gori (Aug. 17, 1989), reprinted in J.A. at 320-36. Both admitted that the causes of human cancer were still unknown. 244 Science at 757; Gori Deck at 4, reprinted in J.A. at 323. The Ames letter advanced the argument that “low doses of carcinogens appear to be ... less hazardous than is generally thought,” 244 Science at 757; Gori pointed out the difficulties inherent in drawing conclusions about humans from studies done on animals. Gori Deck at 6-10, reprinted in J.A. at 325-29. Neither document, however, reflected the results of any new empirical studies or laboratory experiments. Neither offered a new statistical analysis of existing data. Both simply pointed out certain shortcomings in the methods and data generally relied on by the scientific establishment. (Ames’s letter was substantially similar to his 1985 submission to the California Legislature regarding a bill then under consideration. Letter from Dr. Bruce N. Ames to Hon. Art Torres (Nov. 11, 1985), reprinted in J.A. at 352-54.) The “new scientific evidence” petitioners point to, see Brief for Petitioners at 31, boils down to the opinions of a few scientists who, however qualified, are in their own words at odds with what is “generally thought” about the subject. 244 Science at 757; cf. Gori Decl. at 9, reprinted in J.A. at 328. In the context of the EPA’s treatment of carcinogens generally, and its rulemakings under this statute in particular, we find the Agency’s response to the comments adequate. After noting that it had heard before most of what was now being said, Final Rule, 56 Fed.Reg. at 3,533, the EPA stated that it was sticking with its zero goal policy “when current scientific data do not show a safe threshold.” Id. at 3,535. Necessarily implicit in that statement, we think, is an Agency conclusion that the Gori declaration and the Ames letter did not require a contrary result. The EPA stated that it would continue to consider “whether it is possible to define levels that have little or no meaning in terms of cancer risk,” and that if such levels were found it would reconsider its zero goal approach. Id. While in some circumstances an agency must undertake a more detailed re-justification of its prior position, see, e.g., Bechtel v. FCC, 957 F.2d 873 (D.C.Cir.1992), when the agency’s decision is as recent (and as recently upheld) as this one, that duty cannot be triggered by the submission of comments consisting of little more than assertions that in the opinions of the commenters the agency got it wrong. Given that neither submission here presented any “data,” Final Rule, 56 Fed.Reg. at 3,535, we cannot fault the EPA for not paying more attention to them. D. Regulation of DBCP 1. Background The pesticide DBCP was widely used in the continental United States until 1979, — 339 — when the EPA banned it because of health risks. DBCP was marketed by a number of companies, including petitioners Dow, Shell, and Occidental. Because DBCP in runoff water will migrate through soil, it may leach into aquifers used for drinking water. In 1985, the EPA concluded that the available data indicate that DBCP has non-carcinogenic and carcinogenic effects in animals. Based on potential adverse health effects and occurrence and potential occurrence in drinking water, EPA is proposing to regulate DBCP. The [MCLG] will be based upon carcinogenic effects and an [MCLG] of zero is proposed. 1985 Proposed Rule, 50 Fed.Reg. at 46,989. Four years later, the EPA noted that the epidemiology data base is inadequate to refute or demonstrate that DBCP causes tumors in humans. EPA believes there are sufficient data to conclude that DBCP is carcinogenic in animals since the compound has been shown to be carcinogenic in both rats and mice. EPA therefore has classified DBCP in Group B2: probable human carcinogen. Proposed Rule, 54 Fed.Reg. at 22,083. The Final Rule reiterated this analysis, Final Rule, 56 Fed.Reg. at 3,545, placing DBCP in Category I with an MCLG of zero and an MCL of 0.0002 mg/1. Id. at 3,528 (Table 3). 2. Discussion Petitioners Dow, Shell, and Occidental argue that the EPA did not adequately explain its action in setting the MCLG and MCL for DBCP. Petitioners also assert that the EPA failed to address their comments and improperly rejected data presented by them. Petitioners focus on the EPA’s response to human epidemiological studies submitted during the comment period, claiming that the EPA rejected human epidemiological evidence on DBCP “without any explanation.” Brief for Petitioners at 33. Petitioners refer primarily to two epidemiological studies, the Hearn and Wong studies, of workers exposed to DBCP in the work environment. Id. at 13. Shell presented these studies in comments submitted to the EPA in 1986. See Comments of Shell Oil Co. (May 8,1986), reprinted in J.A. at 859. Shell and Dow presented them again in 1989 in response to the Proposed Rule. See Comments of Shell Oil Co. and The Dow Chemical Co. (Aug. 19, 1989), reprinted in J.A. at 172. Shell and Dow stated that the studies “do[ ] not show any statistically significant increase in either overall cancer rates or any specific cancer type” in workers exposed to the chemical. Id. at 204. Petitioners assert that the EPA made only one response to the evidence— that it “believe[d] the epidemiology data base is inadequate to refute or demonstrate that DBCP causes tumors in humans[.]” Brief for Petitioners at 34 (quoting Final Rule, 56 Fed.Reg. at 3,545). A conclusory statement, of course, does not in itself provide the “satisfactory explanation” required in rulemaking. State Farm, 463 U.S. at 43, 103 S.Ct. at 2866. The EPA maintains, however, that petitioners ignore many of its responses to the epidemiological data. See Brief for Respondent at 34-36. The EPA points to a report it prepared in 1987 in response to Shell’s 1986 comments. See EPA Final Report: Review of the Epidemiologic Literature on the Carcinogenic Effects of DBCP (Mar. 5, 1987), reprinted in J.A. at 1852. That report contained the following assessment of the Hearn study: The limited number of years of followup and the limited statistical power are two serious limitations to this study. Continued follow-up is essential to firmly conclude that there is no or little excess risk of cancer in workers with DBCP exposure. The absence of more definitive exposure data limits the utility of this study for risk assessment. Id. at 16-17, reprinted in J.A. at 1868-69. Regarding the Wong study, the report stated: No qualitative or industrial hygiene data are available to indicate the extent and level of exposure to [DBCP]. However, the authors suggest that the exposure was relatively low compared to exposure — 340 — to [other] brominated chemicals. As such, this study offers very limited evidence about the potential risks associated with DBCP, since the exposure levels are unknown and are likely to be extremely low. Id. at 18, reprinted in J.A. at 1870. The EPA, then, found that both the Hearn and Wong studies presented insufficient exposure data. We conclude that the EPA adequately addressed the human epidemiological data on DBCP. The data was collected from workers exposed to DBCP primarily through inhalation in the workplace, not from individuals exposed to DBCP through ingestion of drinking water. In response to studies of the effects of the chemical on animals, the EPA emphasized that “toxicity may depend on dosing route ..., [therefore] inhalation data should not be relied upon in a risk assessment for oral exposure.” EPA, Response Document (revised July 19, 1989) at 4-127, reprinted in J.A. at 1104. This is particularly true where, as here, the EPA has assessed the inhalation studies and has concluded that they fail to provide adequate exposure data. Dow, Shell, and Occidental raise several other challenges to the EPA’s rulemaking on DBCP. First, they contest the EPA’s reliance on animal studies on the ground, among others, that the EPA did not adequately respond to petitioners’ comments. These comments had questioned the EPA’s use of studies that reviewed the effects of DBCP on animals when inhaled into the lungs. See id. at 4-125, reprinted in J.A. at 1102. The EPA responded that “[inhalation toxicity data on DBCP were evaluated but played no critical role in EPA’s risk assessment for DBCP in water.” Id. at 4-127, reprinted in J.A. at 1104 (emphasis added). Shell and Dow also commented that “some experiments” on which the EPA relied involved extremely high doses of DBCP and that in these experiments the tumors tended to develop at the site of contact. Comments of Shell Oil Co. and The Dow Chemical Co. at 10-11, reprinted in J.A. at 184-85. The EPA responded that its analysis of DBCP was an exhaustive survey that was not limited to one particular type of study: Based on positive animal data in two species (rats and mice) by three routes of exposure (oral, inhalation and dermal), including site of contact and distant site tumors, DBCP is considered to have a sufficient level of evidence for carcinogenicity in animals. EPA Drinking Water Criteria Document for DBCP (Mar. 1988) at V-88, reprinted in J.A. at 1613 (emphasis added). Considering the EPA’s responses, we cannot agree with petitioners’ assertion that the “EPA has given no indication of how it evaluated the evidence Petitioners submitted as to the reliability of certain of the animal studies[.]” Brief for Petitioners at 34. We conclude that the EPA adequately responded to the commentary and sufficiently justified its reliance on animal studies. Petitioners also maintain that the EPA failed to respond adequately to data demonstrating that DBCP naturally degrades over time. Petitioners noted that DBCP “has an approximate half-life of six years and is no longer being introduced into the environment.” Comment/Response Document on 1989 Proposed Rule at 9-14, reprinted in J.A. at 1464 (citation omitted). We agree with the EPA that petitioners have failed to establish the relevance of DBCP degradation to the MCL level. If DBCP concentration has already naturally degraded to levels below the MCL in many water sources, then plants using this water will not have to treat for DBCP. As the EPA explained in its response to the “natural degradation” commentary, “[o]ccurrence is a different issue than the level at which the MCL is set.” Id. at 9-15, reprinted in J.A. at 1465. While it is true that the Agency’s rule may require the construction of costly treatment facilities in order to handle a hazard that will soon cease to exist, we cannot quarrel with its conclusion that this is nevertheless required by the Act. Finally, petitioners argue that the EPA did not establish that the MCL for DBCP — 341 — was “feasible” in light of the high costs to drinking water systems that will have to lower the DBCP concentration to no more than 0.0002 mg/1. The SDWA mandates an MCL that “is as close to the [MCLG] as is feasible.” 42 U.S.C. § 300g-1(b)(4). The Act also requires the EPA, in making this feasibility determination, to “tak[e] cost into consideration.” Id. § 300g-1(b)(5). Petitioners argue that the EPA erred in its cost analysis by, inter alia, focusing on the cost estimates for large water systems as opposed to small ones. We begin with the meaning of “feasible” as used in the Act. The statute states: [T]he term “feasible” means feasible with the use of the best technology, treatment techniques and other means which the [EPA] finds ... are available (taking cost into consideration). For the purpose of [§ 300g-l(b)(4) ], granular activated carbon is feasible for the control of synthetic organic chemicals[.] Id. The EPA’s projected cost of treatment for DBCP is based on the use of granular activated carbon technology. See Final Rule, 56 Fed.Reg. at 3,577 (Table 30 and n. 1). Because DBCP is a synthetic organic chemical, the use of granular activated carbon is per se “feasible” under section 300g-l(b)(5). We therefore conclude that the EPA adequately explained why the prescribed treatment was feasible within the meaning of the statute. E. Regulation of EDB 1. Background The compound EDB is used both as a pesticide and as an additive in leaded gasoline. The EPA banned EDB as a pesticide in 1983 and regulates its use in leaded gasoline. The EPA discussed EDB in 1985, proposing an MCLG of zero based on “available data indicating] that EDB has carcinogenic effects in animals.” 1985 Proposed Rule, 50 Fed.Reg. at 46,996. The EPA reiterated this position in the 1989 Proposed Rule: “EPA believes that the available data are adequate to classify EDB in Group B2, probable human carcinogen and therefore an MCLG of zero is appropriate.” Proposed Rule, 54 Fed.Reg. at 22,086. The Final Rule established an MCLG of zero and an MCL of 0.00005 mg/1 for EDB. Final Rule, 56 Fed.Reg. at 3,528 (Table 3). 2. Discussion Dow challenges the MCLG and MCL for EDB on the grounds that the EPA both failed to address the comments submitted on EDB and improperly rejected material contained in those comments. Dow complains that the EPA failed to consider a human epidemiological study conducted by Professor M.G. Ott in 1980 and submitted by the CMA in 1986. See Comments of CMA (Apr. 14, 1986) at 5-6, 9, reprinted in J.A. at 1053-54, 1057. The Ott study investigated the experience of 156 male workers exposed to EDB (primarily through inhalation) in manufacturing facilities. See EPA Drinking Water Criteria Document for EDB (July 1987) at VI-2, reprinted in J.A. at 1702. Contrary to Dow’s assertions, the EPA specifically analyzed and responded to the Ott study, concluding that the results should be regarded as equivocal because interpretation is complicated by the small population size, poorly characterized exposure concentrations, and concomitant exposure to numerous other chemicals. Id. at VI-4, reprinted in J.A. at 1704. Furthermore, the EPA did not stop with Ott’s work. See id. at VI-4 to VI-10, reprinted in J.A. at 1704-10. After reviewing additional epidemiological studies of the effects of EDB on humans, the EPA summarized its findings: Mortality studies conducted on workers exposed to EDB are inconclusive with respect to cancer risk or death by other specific target organ effects (Ott et al., 1980; Turner, 1976, 1977). Because of limitations of scope and design, these epidemiologic studies are not considered to provide definitive results. In particular, small cohort size and lack of, or poorly characterized, exposure estimates are serious drawbacks. Id. at VI-10, reprinted in J.A. at 1710. Thus, it is simply untrue that the EPA failed to consider the available human epi — 342 — demiological data. Not only did it respond to the Ott study that Dow claims it ignored, it also analyzed other epidemiological data that Dow itself now ignores. Moreover, as with DBCP, we note that these studies involved exposure primarily through inhalation, not ingestion, and that the EPA has cautioned that “inhalation data should not be relied upon in a risk assessment for oral exposure.” EPA, Response Document at 4-127, reprinted in J.A. at 1104. We find no basis for concluding that the EPA had ignored the epidemiological data on EDB. Dow also asserts that the EPA failed to address and improperly rejected alternative risk assessments of EDB. The comments including the alternative risk assessments were directed against the EPA’s proposal to establish the MCLG at zero. See Comments of CMA at 5-9, reprinted in J.A. at 1053-57. In essence, then, these studies were used to support petitioners' general argument that there are safe threshold levels for carcinogens. The EPA responded to the alternative risk assessments by reiterating that there is no safe threshold for carcinogenic substances. See EPA, Response Document at 4-154 to 4-155, reprinted in J.A. at 1112-13. As discussed above, the EPA has adequately explained its no-threshold position on carcinogenic substances. Finally, Dow alleges that it appears that the EPA ignored all relevant scientific information on EDB’s carcinogenicity and instead simply set the MCL as a multiple of the lowest detection limit for EDB. The record, however, provides evidence that the EPA relied not only on EDB’s detectibility but also on its assessment of (1) the best available technology (“BAT”) for removing EDB, Final Rule, 56 Fed.Reg. at 3,556; (2) the feasibility (including costs) of the BAT for EDB, id.-, and (3) the carcinogenic risks of EDB, Comment/Response Document on 1989 Proposed Rule at 9-16, reprinted in J.A. at 1466. We therefore cannot accept Dow’s final argument that the rule bears little if any relationship to the record. F. EPA’s Actions Regarding Perchloroethylene 1. Background Perc is a man-made volatile organic chemical used principally as a solvent in the dry cleaning business and also as a degreasing agent in industry. Because of its past disposal in landfills, perc has leached into groundwater and poses a potential contamination threat to drinking water supplies. EPA Office of Drinking Water, Tetrachloroethylene (perchloroethylene) Occurrence in Drinking Water, Food, and Air at i-iii (1983), reprinted in J.A. at 1484. Perc has long been a subject of the EPA regulatory consideration, with the Agency proposing that it regulate perc as a Category I contaminant in 1985. At that time, the EPA recognized that perc had been classified in Group C, which usually results in the EPA assigning the chemical to Category II, but remarked that a 1985 study by the National Toxicology Program (“NTP”) suggested that perc should be placed in Category I. National Primary Drinking Water Regulations; Volatile Synthetic Organic Chemicals, 50 Fed.Reg. 46,880, 46,887-88 (1985). In light of the NTP study, the EPA deferred final action on perc until it could consider additional comments. Id. at 46,888. Nothing further occurred concerning perc until 1989 when the EPA proposed the rule at issue here. In the proposed rule, the EPA reviewed perc’s classification history and explained that no scientific consensus had yet developed as to whether perc should fall within Group B2 or C. Proposed Rule, 54 Fed.Reg. at 22,091. Accordingly, the EPA proposed two alternatives for setting an MCLG for perc: The EPA could take a “limited” view of the evidence and treat perc as a Group C contaminant with an MCLG of 0.01 mg/1; or, as it preferred, the EPA could treat perc as a Group B2 chemical with an MCLG of zero. Id. In the Final Rule, the EPA admitted that it was still deliberating about whether to classify perc in Group B2 or C. Final Rule, 56 Fed.Reg. at 3,541. However, the Agen — 343 — cy felt compelled by a consent order, see id. at 3,531, to promulgate a final perc standard by December 31, 1990; therefore, the EPA concluded that based on its “careful review of the comments received in response to the May, 1989 notice and the Agency’s evaluation of scientific evidence available since the proposal, it remains the EPA’s view that there is strong evidence of carcinogenicity through ingestion and that [perc] is a Category I chemical for purposes of establishing an MCLG under the SDWA.” Id. at 3,541. After promising that its action would not prejudice the ongoing review of perc’s proper classification as a Group B2 or C contaminant, the EPA then went on to answer comments attacking the scientific studies it had relied upon in placing perc in Category I. Id. at 3,541-42. IFI filed a timely petition for judicial review of the perc standard, and this court granted the motion by the HSIA to intervene in support of IFI. 2. Discussion Attacking the EPA’s regulation of perc for alleged procedural defects, IFI argues that the EPA has since 1985 followed a clear, consistent policy of placing Group A and B contaminants in Category I and Group C contaminants in Category II for purposes of establishing SDWA MCLGs.2 According to IFI, perc is now, and has been throughout this rulemaking, classified in Group C, despite an aborted EPA proposal to reclassify perc as a Group B2 contaminant. The EPA has a legal obligation to follow its own procedures. Vitarelli v. Seaton, 359 U.S. 535, 539-40, 79 S.Ct. 968, 972-73, 3 L.Ed.2d 1012 (1959). If an agency action “is based upon a defined procedure, even though generous beyond the requirements that bind such agency, that procedure must be scrupulously observed.” Id. at 546-47, 79 S.Ct. at 976 (Frankfurter, J., concurring & dissenting). IFI asserts that if the EPA had followed its past practice, it would have placed perc in Category II and assigned it a non-zero MCLG. Therefore, IFI contends, the EPA acted arbitrarily and capriciously when it placed perc, a Group C contaminant, in Category I with a resulting MCLG of zero. We are not convinced that the EPA failed to follow its procedures when it placed perc in Category I and assigned it an MCLG of zero. Perc’s classification group has not been definitively determined. While a 1985 Health Assessment Document assigned perc to Group C, we find that the EPA reopened this designation based on a subsequent NTP study. See Hazardous Waste and Superfund Programs; Amendment to Preambles, 56 Fed.Reg. 643, 643-44 (1991) (discussing history of the EPA’s effort to place perc in a group and stating that the “EPA is continuing its deliberations concerning the weight-of-evidence classification for [perc].”). The EPA was careful to note in the proposed rule that a scientific consensus had not developed on whether perc should be classified in Group B2 or C, and that consequently it would “fully consider both approaches before promulgation.” Proposed Rule, 54 Fed.Reg. at 22,091. The EPA asserts that the categorization of perc in the Final Rule will have no effect on the separate effort to classify perc. Moreover, when the EPA completes its classification, the Agency represents that it may reconsider perc’s categorization. Brief for Respondent at 61-62. As such, IFI overstates its case by contending that the EPA has forced a Group C contaminant into Category I or that the EPA preordained a zero MCLG in the proposed rule. More accurately, the EPA has placed a Group CB2 contaminant in Category I and assigned it an MCLG of zero, without waiting for a final classification. The determinative question, then, is whether the EPA violated its own regulations by assigning perc to a category before it had completed perc’s classification. National Conservative Political Action Committee v. FEC, 626 F.2d 953, 959 (D.C.Cir.1980) (“Agencies are under an obligation to follow their own regulations, procedures, and precedents, or provide a rational explanation for their departures.”). — 344 — The EPA contends that its action did not violate agency procedures since the SDWA-specific contaminant categories do not follow in lockstep the agency-wide, weight-of-the-evidence cancer classification groups. Brief for Respondent at 63 (“The cancer classification thus does not dictate the drinking water category, or vice versa.”). As examples, the EPA lists three contaminants that are classed in Groups A or B, but were placed in SDWA Category II for the Final Rule because there were studies showing these contaminants did not pose a high risk from the ingestion of drinking water. Final Rule, 56 Fed.Reg. at 3,532, 3,535-37 (discussing categorization of asbestos, cadmium, and chromium). Additionally, the EPA relates that styrene, like perc, did not have a cancer classification and was therefore categorized, like perc, based on available scientific evidence. Id. at 3,532, 3,540-41. We note that the examples of deviations the EPA provides us occurred in this rule-making, thereby diminishing their value as rebuttal evidence to IFI’s charge that the EPA departed from its past practice in this rulemaking. Moreover, the record indicates that the EPA devised the three-category approach in 1985 to complement the agency-wide classification system that had been proposed in 1984 and would become final in 1986 — the categories and classifications are thus not the total strangers the EPA would suggest. See 1985 Proposed Rule, 50 Fed.Reg. at 46,948; see also Cancer Risk Assessment Guidelines, 51 Fed. Reg. 33,992, 33,999-34,000 (1986) (“Cancer Guidelines”) (promulgating as a final rule the cancer risk classification system first proposed in 1984). In fact, when the EPA discussed the possible MCLGs for perc in the May 1989 Proposed Rule, it spoke in terms of whether perc fell within Group C or B2. Proposed Rule, 54 Fed.Reg. at 22,-091. However, IFI has not directed us to anything within the EPA’s procedures prohibiting the Agency from assigning an SDWA Category when the cancer classification for a contaminant is unresolved. Accordingly, this petition is distinguishable from those cases requiring the invalidation of agency actions that conflict with or depart from the plain meaning of an agency’s regulations. See, e.g., Exportal Ltda. v. United States, 902 F.2d 45, 49-50 (D.C.Cir.1990) (rejecting agency interpretation of its own regulation that conflicted with “plain meaning” of the regulation). Instead, we believe this petition turns on whether the EPA reasonably interpreted its categorization procedures as allowing it to place an unclassified contaminant in a category based on a review of the scientific evidence. As explained by the Agency when it established the SDWA Categories, “Category I includes those chemicals which have sufficient human or animal evidence of carcinogenicity to warrant their regulation as probable human carcinogens.” 1985 Proposed Rule, 50 Fed.Reg. at 46,948. The EPA interpreted this as making the critical inquiry not whether the Agency has completed its classification process, but rather whether a review of the scientific evidence leads to a determination that the contaminant is a probable human carcinogen. We find this interpretation of the categorization procedures permissible and fully consonant with the dictates of notice-and-comment rulemaking. Chemical Manufacturers Ass’n v. EPA, 919 F.2d 158, 170 (D.C.Cir.1990) (“ ‘[A]n agency’s interpretation of its own regulations will be accepted unless it is plainly wrong.’ ” (quoting General Carbon Co. v. Occupational Safety and Health Review Comm’n, 860 F.2d 479, 483 (D.C.Cir.1988))). When the EPA proposed an MCLG for perc, it furnished a full opportunity for comment on perc’s carcinogenicity based on publicly available scientific studies. Proposed Rule, 54 Fed.Reg. at 22,091. It also advised the public that “[bjecause no scientific consensus yet exists, it is necessary for the Agency to make a judgment based on a reasonable weighing of evidence from the data at hand.” Id. So despite the proposed rule’s discussion of classifying perc in Group C or B2, IFI cannot complain that it was denied an opportunity to comment on the studies relied on by the EPA, or that it was unaware that a finding of probable human carcino — 345 — 'genicity based on a “reasonable weighing of evidence” would result in perc being placed in Category I. Therefore, we are satisfied that the EPA acted in an entirely permissible manner without depriving the petitioner of its right to notice and comment in this rulemaking. Finally, we reject IFI's contention that the studies relied on by the EPA do not support a finding that perc is a probable human carcinogen. Faced with a similar dispute regarding the 1985 categorization of trichloroethylene, we upheld the EPA’s action, noting that “[hjappily, it is not for the judicial branch to undertake comparative evaluations of conflicting scientific evidence. Our review aims only to discern whether the agency’s evaluation was rational.” NRDC v. EPA, 824 F.2d at 1216. We apply that standard here and find that IFI has not shown that the EPA acted irrationally. The Agency responded to comments such as those now iterated by IFI and provided a detailed discussion of its reasons for placing perc in Category I. Final Rule, 56 Fed.Reg. at 3,541-42. For example, the Agency noted that “[indications of cancer activity were seen in mice and rats, in both sexes, by inhalation and oral exposure.” Id. at 3,542. The EPA acknowledged the controversy surrounding the occurrence of mouse liver tumors, mononuclear cell leukemia in rats, and male rat kidney tumors, but concluded that these events were “relevant for consideration in the overall weight of evidence for potential [perc] human health hazards” and that “the totality of the animal evidence is judged by EPA to be sufficient to view [perc] as a Category I contaminant.” Id. Critical to this conclusion was the EPA’s belief that although there may be “other plausible mechanistic explanations” for the reported occurrences of cancer, “the experimental animal evidence identifies the potential for a carcinogenic response in humans unless there is evidence to the contrary”; that is, “lack of key information does not support the use of the uncertainties to discount the sufficient level of animal evidence.” Id. In sum, the EPA considered a variety of animal evidence suggesting the carcinogenicity of perc. Though other factors may also have played a role in causing carcinogenic responses, the EPA’s methodology has consistently been to accept findings of carcinogenicity absent contrary findings. NRDC v. EPA, 824 F.2d at 1217. Therefore, we will not disturb the EPA’s regulation of perc as a Category I contaminant.3 G. Method 508A Before the EPA can set the enforceable limit (maximum contaminant level) for a chemical, it first must ascertain how low a concentration of that chemical reliably can be measured when testing water to determine compliance with the limit. This figure — the practical quantitation level — depends upon the methodology adopted to test for the chemical. Because some methods, though technically capable of measuring extremely low concentrations, prove too unreliable for regular enforcement use, the agency normally tests each method at different laboratories to ensure that it yields reasonably consistent results. Petitioners contend that the method the EPA chose to measure polychlorinated biphenyls found in water samples — Method 508A — was adopted without proper notice and comment, and so both the choice of that method, and the maximum contaminant level predicated on it, must be vacated and remanded to the agency. In the Notice of Proposed Rulemaking, the EPA proposed a two-step approach for detecting and measuring PCBs. It planned to use Methods 505 and 508 to “screen” water samples, i.e., to determine whether PCBs were present. If they were, Method 508A would be employed to determine the concentration of PCBs in a particular sample. Proposed Rule, 54 Fed.Reg. at 22,099346 100. The EPA planned to use Method 508A largely because that method could measure much lower concentrations of the chemical than could other methods. However, the EPA acknowledged in the Notice that Method 508A was but recently developed, id. at 22,099, and that its accuracy had been verified in only one laboratory, id. at 22,104. Accordingly, the Agency requested comments on “the proposed approach to measure PCBs ... and on the proposed analytical methods.” Id. at 22,-100. Petitioners NEMA and CMA (among others) obliged the Agency, submitting detailed comments on PCBs issues. They argued that because the EPA had tested Method 508A in only one laboratory, not in interlaboratory tests, it could not rely on the method to establish the enforceable standard. The EPA then did what petitioners requested. In developing its responses to their comments, the Agency reviewed the results of Water Studies 23-25, conducted by private laboratories. In the EPA’s view, the results of those studies adequately confirmed the reliability of Method 508A, and so the Agency adopted both the method and the maximum contaminant level as proposed. Final Rule, 56 Fed.Reg. at 3,551 col. 1. Petitioners now complain that the EPA’s use of the Water Studies, without prior notice that they would be relied on, violated its notice-and-comment obligations. See 5 U.S.C. § 553(c) (1988). They urge us to remand the matter to the Agency so that they may submit criticisms of Water Studies 23-25, and arguments about why Method 508A, even in light of the new studies, remains insufficiently reliable. Because petitioners had fair notice of, and full opportunity to comment on, the issue actually decided by the EPA, we reject their request.4 The core of this issue was whether Method 508A was sufficiently reliable, and whether it or some other method would be adopted by the EPA. In relying on Water Studies 23-25, the EPA did no more than provide support for the same decision it had proposed to take. The conclusion the EPA reached was one petitioners both had and took the opportunity to criticize. Most of their criticism centered around the lack of evidence of reliability — the very evidence the Agency then amassed to support its proposed conclusion. The EPA did not perform an unheralded about-face in light of a recalculation of data unknown to petitioners, cf. Solite Corp. v. EPA, 952 F.2d 473, 499 (D.C.Cir.1992), nor did it fail to “identify the ... methodology used” in actually developing the standard at the outset, cf. Portland Cement Ass'n v. Ruckelshaus, 486 F.2d 375, 392 (D.C.Cir.1973), cert. denied, 417 U.S. 921, 94 S.Ct. 2628, 41 L.Ed.2d 226 (1974), or reveal information “critical” to its determination only after promulgation of the final rule, id. at 393. We have in similar circumstances refused to make the Agency subject each and every study to notice and comment, see Community Nutrition Inst. v. Block, 749 F.2d 50, 57-58 (D.C.Cir.1984), and have even upheld an agency’s change of course as a “logical outgrowth” of the proposed rule although the change was supported in part by a study not subject to public comment. See City of Stoughton, Wis. v. EPA, 858 F.2d 747, 753 (D.C.Cir.1988). Those decisions reflect the commonsense recognition that Congress, in providing for notice and comment under the APA, could not have intended to subject the agencies— and the public on whose behalf they regulate — to the sort of interminable back-and-forth petitioners would have us require. — 347 — See id.; cf. Solite, 952 F.2d at 496, citing NRDC v. Thomas, 838 F.2d 1224, 1242 (D.C.Cir.1988), in turn citing International Harvester Co. v. Ruckelshaus, 478 F.2d 615, 632 n. 51 (D.C.Cir.1973). We will not require that today. III. Conclusion In ruling on these petitions, our responsibility is limited to determining whether the EPA’s interpretations of the Safe Drinking Water Act are permissible, and whether in applying the Act, the Agency has abided by the requirements of the Administrative Procedure Act. As we are not scientists and must defer to the Agency’s judgments on matters within its technical competence, our task is to assure that they be reasoned, not that they be right. Because we find that the EPA’s interpretations of the Act are permissible, that it has complied with the procedures mandated by the APA, and that its conclusions have been neither arbitrary nor capricious, we deny the petitions for review. So ordered. APPENDIX Glossary of Abbreviations and Acronyms APA Administrative Procedure Act BAT Best available technology CERCLA Comprehensive Environmental Response, Compensation & Liability Act CMA Chemical Manufacturers Association DBCP Dibromochloropropane EDB Ethylene dibromide EPA Environmental Protection Agency HSIA Halogenated Solvents Industry Alliance IFI International Fabricare Institute MCL Maximum contaminant level MCLG Maximum contaminant level goal NEMA National Electrical Manufacturers Association NTP National Toxicology Program PCBs Polychlorinated biphenyls perc Perchloroethylene (a/k/a tetrachloroethylene) SDWA Safe Drinking Water Act Footnotes [1] A glossary of abbreviations and acronyms appears at the end of this opinion. [2] See Background discussion of "group" and "category” classification, supra at 335. [3] We also note that the EPA’s representation that it may reconsider perc’s categorization once the classification has been completed answers IFI’s request that we remand for additional evidence based on a Science Advisory Board study that had not been completed when EPA promulgated the Final Rule. Brief for Petitioners at 54-55. Since the study will be a part of the agency-wide classification of perc, the EPA’s reconsideration of perc based upon the new classification will by necessity include consideration of the study. [4] We decline to address petitioners’ additional argument that the EPA failed to prove that Method 508A was "feasible” as defined by the statute. See 42 U.S.C. § 300g—1(b)(5). So far as we can tell, this objection was never made to the EPA in petitioners' comments; petitioners' brief provides no citation to the argument before the EPA. This was not an objection that petitioners needed the Water Supply studies to make, for if — as they assert — Method 508A has not been shown to be "feasible” even in light of the studies, it could have been attacked with even more force on those same grounds before the studies were introduced. Moreover, petitioners have adduced nothing — beyond their general assertions of unreliability — to show that Method 508A is not feasible.