Chemical Manufacturers Association v. U.S. Environmental Protection Agency

Opinion for the Court filed by Chief Judge WALD.

Petitioners, Chemical Manufacturers Association and four companies that manufacture chemicals 1 (collectively, “CMA”), seek to set aside a rule promulgated by the Environmental Protection Agency (“EPA” or “the Agency”). 2 This Final Test Rule was promulgated under section 4 of the Toxic Substances Control Act (“TSCA” or “the Act”), 15 U.S.C. §§ 2601-2629. The Final Test Rule required toxicological testing to determine the health effects of the chemical 2-ethylhexanoic acid (“EHA”), and it continues to impose on exporters of EHA a duty to file certain notices with EPA.

We uphold EPA’s interpretation of TSCA as empowering the Agency to issue a test rule on health grounds where it finds a more-than-theoretical basis for suspecting that the chemical substance in question presents an “unreasonable risk of injury to health.” This, in turn, requires the Agency to find a more-than-theoretical basis for concluding that the substance is sufficiently toxic, and human exposure to it is sufficient in amount, to generate an “unreasonable risk of injury to health.” We hold, further, that EPA can establish the existence and amount of human exposure on the basis of inferences drawn from the circumstances under which the substance is manufactured and used. EPA must rebut industry-supplied evidence attacking those inferences only if the industry evidence succeeds in rendering the probability of exposure in the amount found by EPA no more than theoretical or speculative. The probability of infrequent or even one-time exposure to individuals can warrant a test rule, so long as there is a more-than-theoretical basis for determining that exposure in such doses presents an “unreasonable risk of injury to health.” Finally, we hold that the Agency correctly applied these standards in this case and that its findings are supported by substantial evidence. Consequently, we affirm the Final Test Rule.

I. BACKGROUND

A. Statutory Structure

TSCA provides for a two-tier system for evaluating and regulating chemical substances to protect against unreasonable risks to human health and to the environment. Section 6 of the Act permits EPA to regulate a substance that the Agency has found “presents or will present an unreasonable risk of injury to health or the environment.” 15 U.S.C. § 2605(a). Section 4 of the Act empowers EPA to require testing of a suspect substance in order to obtain the toxicological data necessary to make a decision whether or not to regulate the substance under section 6. The Act provides, not surprisingly, that the level of certainty of risk warranting a section 4 test rule is lower than that warranting a section 6 regulatory rule. EPA is empowered to require testing where it finds that the manufacture, distribution, processing, use or *980 disposal of a particular chemical substance “may present an unreasonable risk of injury to human health or the environment.” Id. § 2603(a)(l)(A)(i). The Agency’s interpretation of this statutory standard for testing is the central issue in this case.

One of the chief policies underlying the Act is that—

adequate data should be developed with respect to the effect of chemical substances and mixtures on health and the environment and that the development of such data should be the responsibility of those who manufacture and those who process such chemical substances and mixtures.

Id. § 2601(b)(1). The statute establishes an Interagency Testing Committee, comprised of scientists from various federal agencies, to recommend that EPA give certain chemicals “priority consideration” for testing. Id. § 2603(e). Under section 4, the Agency “shall by rule require that testing [of a particular chemical] be conducted” if three factors are present: (i) activities involving the chemical “may present an unreasonable risk of injury to health or the environment”; (ii) “insufficient data and experience” exist upon which to determine the effects of the chemical on health or environment; and (iii) testing is necessary to develop such data. Id. § 2603(a)(1)(A). The companies that manufacture and process the substance are to conduct the tests and submit the data to the Agency. Id. § 2603(b)(3)(B). Costs of testing are to be shared among the companies, either by agreement or by EPA order in the absence of agreement. Id. § 2603(c)(3)(A).

A test rule promulgated under section 4 is subject to judicial review in a court of appeals, pursuant to section 19(a) of TSCA, 15 U.S.C. § 2618(a). A test rule may be set aside if it is not “supported by substantial evidence in the rulemaking record ... taken as a whole.” Id. § 2618(c)(l)(B)(i).

B. Facts and Prior Proceedings

EHA is a colorless liquid with a mild odor. It is used exclusively as a chemical intermediate or reactant in the production of metal soaps, peroxy esters and other products used in industrial settings. EHA itself is totally consumed during the manufacture of these products; as a result, no products offered for sale to industry or to consumers contain EHA. See Final Test Rule, 51 Fed.Reg. at 40,319.

The Interagency Testing Committee first designated EHA for priority consideration for health effects tests on May 29, 1984. Fourteenth Report of the Interagency Testing Committee to the Administrator, 49 Fed.Reg. 22,389, 22,391 (1984). The Committee based its recommendation in part on the structural similarity of EHA to chemicals known to cause cancer in test animals and on its finding that insufficient information existed concerning the chronic health effects of EHA. Id. at 22,399-400. Subsequently, EPA held two public meetings on EHA. Supplemental Joint Appendix (“S.J. A.”) 499-504, 505-15. During these meetings, in which persons representing the petitioners made appearances, EPA sought information on a variety of issues relating to EHA uses, production and human exposure. S.J.A. 502-04, 508-11, 513-14.

EPA issued a proposed test rule on May 17, 1985. 2-Ethylhexanoic Acid Proposed Test Rule, 50 Fed.Reg. 20,678 (1985) (the “Proposed Test Rule”). The rule proposed a series of tests to ascertain the health risks of EHA, and it set out proposed standards for the conduct of those tests. Id. at 20,682-83. EPA based the Proposed Test Rule on a finding that EHA “may present an unreasonable risk” of subchronic toxicity (harm to bodily organs from repeated exposure over a limited period of time), oncogenicity (tumor formation) and developmental toxicity (harm to the fetus). Id. at 20,682. As to subchronic toxicity, EPA cited studies suggesting that both EHA and chemicals structurally similar to it cause harm to the livers of test animals. Id. As to oncogenicity, EPA cited studies suggesting that chemicals structurally analogous to EHA cause cancer in laboratory animals. Id. at 20,681-82. As to developmental toxicity, EPA cited studies indicating that both EHA and its chemical analogues have produced fetal malformations in test animals. Id. at 20,682.

*981 The Proposed Test Rule also addressed the question of whether humans are exposed to EHA, a question of critical importance to this case. The Agency acknowledged that, since no finished products contain EHA, consumer exposure is not a concern. It likewise discounted the dangers of worker exposure to EHA vapors. Id. at 20,681. The Agency based its Proposed Test Rule solely on the potential danger that EHA will come in contact with the skin of workers. As evidence of potential dermal exposure, the Agency noted that approximately 400 workers are engaged in the manufacture, transfer, storage and processing of 20 to 25 million pounds of EHA per year. Id. at 20,679, 20,682. Further, rebutting claims by industry representatives that gloves are routinely worn during these activities, EPA noted that worker hygiene procedures “can vary widely throughout the industry,” that workers are not required by existing federal regulations to wear gloves, and that the industry had not monitored work sites for worker exposure to EHA. Id. at 20,680-81.

A public comment period commenced with the publication of the Proposed Test Rule and ended on July 16, 1985. EPA held a public meeting on October 8,1985, to discuss issues related to the Proposed Test Rule. J.A. 69-71. Industry representatives submitted extensive comments on July 15, 1985, and January 17, 1986. J.A. 96-97, 161-63. Before publication of the Final Test Rule, EPA received notice of a new study purporting to present further evidence of the potential developmental toxicity of EHA. Final Test Rule, 51 Fed. Reg. at 40,319-21.

CMA criticized the toxicology studies cited by EPA and sought to show that the use of gloves by employees of companies working with EHA prevented human exposure to the chemical, thus rendering any test rule invalid. Before publication of the final rule, CMA retained an independent consultant to conduct a survey of glove use by the employees of companies working with EHA. J.A. 32-47 (the “Glove Use Survey”). The results of the survey were submitted to EPA. CMA also submitted the results of an Eastman Kodak Co. study of the permeability of nitrile and neoprene gloves. J.A. 68 (the “Glove Permeability Study”).

EPA published the Final Test Rule for EHA on November 6, 1986. The rule required a 90-day subchronic toxicity test, a developmental toxicity test, and a pharma-cokinetics test. Final Test Rule, 51 Fed. Reg. at 40,318. In the preamble to the Final Test Rule, EPA asserted its finding of potential exposure to EHA. The Agency rebutted industry claims of non-exposure by criticizing the methodology of the Glove Use Survey and the Glove Permeability Study. Id. at 40,320. 3 The Final Test Rule also addressed the industry critique of the prior toxicology studies on EHA. Acknowledging some weaknesses in the studies, EPA nonetheless asserted that they “add to the weight of evidence” supporting findings of potential toxicity and thus helped to justify further testing. Id. at 40,321.

The pharmacokinetics study required by the rule entailed the oral and dermal administration of EHA to experimental animals, at low and high doses. Id. at 40,327-28. The subchronic toxicity study involved administering EHA to animals in graduated daily doses over a period of 90 days. Id. at 40,323, 40,328-30. The developmental toxicity tests entailed administering EHA orally in various doses during the pregnancy of experimental animals. Id. at 40,323, 40,330. All studies were to be conducted in accordance with EPA standards. Results were to be submitted by certain deadlines, the last of which was 18 months after the effective date of the Final Test Rule; hence, the last deadline was June 20, 1988. 4 Id. at 40,330.

*982 The petitioners filed this petition for review on December 23, 1986, moving at the same time for a stay pending review. This court denied the stay motion on January 20, 1987.

The last of the required test data were submitted to EPA on June 20, 1988. EPA moved this court on June 9, 1988, for leave to file a motion to dismiss the case as moot. By order on June 30, 1988, the court denied the motion and directed that the parties address the mootness issue before this panel.

II. Mootness

EPA argues that this petition for review is moot because all of the test results have been submitted. CMA does not seek damages or reimbursement for having conducted the tests. Thus, the Agency argues, there are no residual effects of the testing requirement to keep the controversy alive. We nonetheless hold the case to be justiciable, because the rule continues to impose concrete obligations on companies handling EHA.

The doctrine of mootness is one aspect of the requirement that federal courts decide only actual “cases” or “controversies.” U.S. Const, art. Ill, § 2. In order for a case not to be moot, a party must show a direct injury that gives it a continuing “ ‘personal stake’ ” in the outcome of the litigation. Sample v. Johnson, 771 F.2d 1335, 1339 (9th Cir.1985) (quoting Warth v. Seldin, 422 U.S. 490, 498, 95 S.Ct. 2197, 2205, 45 L.Ed.2d 343 (1975)), cert. denied, 475 U.S. 1019, 106 S.Ct. 1206, 89 L.Ed.2d 319 (1986). In this sense, the mootness inquiry is similar to the inquiry of whether a party has constitutional standing. Sample, 771 F.2d at 1339; 13A C. Wright, A. Miller & E. Cooper, Federal Practice & Procedure § 3533.1 at 219 (1984) (assimilating mootness principles with standing principles) [hereinafter “Wright & Miller”]. Federal courts are constitutionally forbidden to render advisory opinions or “to decide questions that cannot affect the rights of litigants in the case before them.” Better Gov’t Ass’n v. Department of State, 780 F.2d 86, 90-91 (D.C.Cir.1986) (quoting North Carolina v. Rice, 404 U.S. 244, 246, 92 S.Ct. 402, 404, 30 L.Ed.2d 413 (1971)).

On the other hand, a claim for prospective relief remains justiciable so long as the conduct complained of has “continuing, present adverse effects.” See O’Shea v. Littleton, 414 U.S. 488, 496, 94 S.Ct. 669, 676, 38 L.Ed.2d 674 (1974). The fact that the principal form of relief that a party seeks is no longer available does not in itself render the case nonjusticiable. See Better Gov’t Ass’n, 780 F.2d at 91-92; Maine Central R. Co. v. Brotherhood of Railway Maintenance Employees, 813 F.2d 484, 487 (1st Cir.), cert. denied, — U.S. -, 108 S.Ct. 91, 98 L.Ed.2d 52 (1987). “Where one of the several issues presented becomes moot, the remaining live issues supply the constitutional requirement of a case or controversy.” Powell v. McCormack, 395 U.S. 486, 497, 89 S.Ct. 1944, 1951, 23 L.Ed.2d 491 (1969). This is so even where the mooted claim is much greater in magnitude than the remaining ones. Id. In light of the similarity between mootness and standing, it should be noted that the “distinct and palpable injury” a party must suffer in order to be deemed to have constitutional standing, Warth v. Seldin, 442 U.S. at 501, 95 S.Ct. at 2206, “need not be tangible or great: an ‘identifiable trifle’ will do.” National Wildlife Federation v. Hodel, 839 F.2d 694, 704 (D.C.Cir.1988) (quoting United States v. SCRAP, 412 U.S. 669, 689 n. 14, 93 S.Ct. 2405, 2417 n. 14, 37 L.Ed.2d 254 (1973)). Similarly, the continuing “personal stake” needed to rebut an assertion of mootness need not be large in magnitude to suffice. Thus, as the Supreme Court has stated, a case is not moot if some stake in the outcome remains, “notwithstanding the size of the dispute.” Firefighters Local 1784 v. Stotts, 467 U.S. 561, 571, 104 S.Ct. 2576, 2584, 81 L.Ed.2d 483 (1984). All that is necessary is that the party seeking relief “have a concrete interest in the outcome.” Id.

In light of these principles, this case should not be dismissed on the ground of mootness. Section 12(b) of TSCA provides that any company exporting a chemical *983 substance that is subject to a test rule under section 4 must “notify the Administrator of such exportation or intent to export and the Administrator shall furnish to the government of such country notice of the availability of the data submitted to the Administrator under such section for such substance or mixture.” 15 U.S.C. § 2611(b)(1). EPA regulations require an exporter to provide a separate notice “for the first export or intended export to a particular country in a calendar year.” 40 C.F.R. § 707.65(a)(2) (1987). See also Final Test Rule, 51 Fed.Reg. at 40,824 (summarizing these obligations). There is also indication in the record that compliance with the notification requirement requires “more than a minimal amount of work.” Supplemental Affidavit of Henry J. Sauer 115.

The acknowledgedly modest nature of these obligations does not require this court constitutionally to refuse to adjudicate the case. 5 If something still remains to be done under a rule, that is usually enough to spare it from the mootness ax. See Casey v. FTC, 578 F.2d 793, 796 (9th Cir.1978) (challenge to agency subpoenas not moot despite submission of all documents, where documents were not verified in accordance with subpoenas); FTC v. Gibson Prods., 569 F.2d 900, 903 (5th Cir.1978); 13A Wright & Miller, supra, § 3533.2 at 258-59. Although the petitioners’ stake in the outcome of this case is manifestly smaller than it was when the case was filed, they still have a sufficiently concrete interest in the disposition of their petition to warrant the exercise of federal court jurisdiction. 6

III. Statutory Interpretation

The Toxic Substances Control Act requires EPA to promulgate a test rule under section 4 if a chemical substance, inter alia, “may present an unreasonable risk of injury to health or the environment.” 15 U.S.C. § 2603(a)(l)(A)(i). The parties both accept the proposition that the degree to which a particular substance presents a risk to health is a function of two factors: (a) human exposure to the substance, and (b) the toxicity of the substance. See Ausimont U.S.A., Inc. v. EPA, 838 F.2d 93, 96 (3d Cir.1988). They also agree that EPA must make some sort of threshold finding as to the existence of an “unreasonable risk of injury to health.” The parties differ, however, as to the manner in which this finding must be made. Specifically, three issues are presented.

The first issue is whether, under section 4 of TSCA, EPA must find that the existence of an “unreasonable risk of injury to health” is more probable than not in order to issue a test rule. CMA argues that the statute requires a more-probable-than-not finding. CMA Br. 22. EPA disagrees, contending that the statute is satisfied where the existence of an “unreasonable risk of injury to health” is a substantial probability — that is, a probability that is more than merely theoretical, speculative or conjectural. EPA Br. 21. 7

*984 The second issue is whether, once industry has presented evidence tending to show an absence of human exposure, EPA must rebut it by producing direct evidence of exposure. CMA claims that, when industry evidence casts doubt on the existence of exposure, the burden of production shifts back to EPA, which must produce direct evidence documenting actual instances in which exposure has taken place. CMA Br. 27, 30. EPA, on the other hand, argues that it can make the requisite finding of exposure based solely on inferences drawn from the circumstances under which a chemical substance is manufactured and used. EPA Br. 32-33.

The third issue is whether the Agency has authority to issue a test rule where any individual’s exposure to a substance is an isolated, non-recurrent event. CMA argues that, even if EPA presents direct evidence of exposure, the Act precludes issuance of a test rule where exposure consists only of rare instances involving brief exposure. CMA Reply Br. 11 n. 10. EPA contends, on the other hand, that the Act does not require in all circumstances a risk of recurrent exposure. EPA Br. 28.

We will explore each of these three issues in turn. In each instance, we must examine the statute itself and its legislative history to determine whether Congress addressed the particular point in question. If this court ascertains that Congress has directly spoken to the precise question at issue, then both the court and EPA “must give effect to the unambiguously expressed intent of Congress.” Chevron U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 842-43, 104 S.Ct. 2778, 2781, 81 L.Ed.2d 694 (1984); accord NLRB v. United Food & Com’l Workers Union, — U.S. -, 108 S.Ct. 413, 421, 98 L.Ed.2d 429 (1987). If, however, the statute is silent or ambiguous with respect to the specific issue, the court is left to decide only whether the Agency’s construction of the statute is a reasonable one. Chevron, 467 U.S. at 843, 104 S.Ct. at 2782. Congressional silence is deemed an implicit delegation of power to an agency to make policy choices consistent with the statutory purpose. Id. at 844-45, 104 S.Ct. at 2782-83. “[The] view of the agency charged with administering the statute is entitled to considerable deference; and to sustain it, we need not find that it is the only permissible construction that EPA might have adopted, but only that EPA’s understanding of this very ‘complex statute’ is a sufficiently rational one to preclude a court from substituting its judgment for that of EPA.” Chemical Manufacturers Ass’n v. Natural Resources Defense Council, Inc., 470 U.S. 116, 125, 105 S.Ct. 1102, 1107, 84 L.Ed.2d 90 (1985) (interpreting the Clean Water Act).

A. Required Finding of “Unreasonable Risk”

As to the first issue in this case, the standard of probability of an unreasonable risk to health, we find that Congress did not address the precise question in issue. Examining the EPA interpretation under the second prong of Chevron, we find it to be reasonable and consistent with the statutory scheme and legislative history. Consequently, we uphold the Agency’s construction of TSCA as authorizing a test rule where EPA’s basis for suspecting the existence of an “unreasonable risk of injury to health” is substantial — i.e., when there is a more-than-theoretical basis for suspecting that some amount of exposure takes place and that the substance is sufficiently toxic at that level of exposure to present an “unreasonable risk of injury to health.”

1. Text and Structure of the Statute Both the wording and structure of TSCA reveal that Congress did not expect that *985 EPA would have to document to a certainty the existence of an “unreasonable risk” before it could require testing. This is evident from the two-tier structure of the Act. In order for EPA to be empowered to regulate a chemical substance, the Agency must find that the substance “presents or will present an unreasonable risk of injury to health or the environment.” TSCA § 6, 15 U.S.C. § 2605(a). The testing provision at issue here, by contrast, empowers EPA to act at a lower threshold of certainty than that required for regulation. Specifically, testing is warranted if the substance “may present an unreasonable risk of injury to health or the environment.” TSCA § 4, 15 U.S.C. § 2603(a)(l)(A)(i) (emphasis added). Thus, the language of section 4 signals that EPA is to make a probabilistic determination of the presence of “unreasonable risk.”

2. Legislative History The legislative history of TSCA compels a further conclusion. It not only shows that “unreasonable risk” need not be a matter of absolute certainty; it shows the reasonableness of EPA’s conclusion that “unreasonable risk” need not be established to a more-probable-than-not degree.

A House Report on the version of the bill that eventually became TSCA underscores the distinction between the section 6 standard and the section 4 standard. To issue a test rule, EPA need not find that a substance actually does cause or present an “unreasonable risk.” H.R.Rep. No. 1341, 94th Cong., 2d Sess. 17-18 (1976). 8

Such a finding requirement would defeat the purpose of the section, for if the Administrator is able to make such a determination, regulatory action to protect against the risk, not additional testing, is called for.

The House Report also contains signals indicating that Congress expected EPA to act even when evidence of “unreasonable risk” was less than conclusive. According to that report, the word “may” in section 4 was intended to focus the Agency’s attention on chemical substances “about which there is a basis for concern, but about which there is inadequate information to reasonably predict or determine the effects of the substance or mixture on health or the environment.” Id. at 17 (emphasis added). The Conference Committee Report reemphasized that the statutory language focused the Agency’s attention on substances “about which there is a basis for concern.” H.R.Conf.Rep. No. 1679, 94th Cong., 2d Sess. 61 (1976), U.S. Code Cong. & Admin.News 1976, p. 4546.

These indications of congressional intent illustrate that EPA’s reading of TSCA is a permissible one. Congress intended to authorize testing where the existence of an “unreasonable risk” could not yet be “reasonably predicted.” The Agency’s determination that it is empowered to act where the existence of an “unreasonable risk” cannot yet be said to be more probable than not is entirely consistent with that expression of intent. The EPA interpretation is likewise consistent with the level of certainty suggested by the phrase “basis for concern.” To accept CMA’s position would require the Agency to gather “adequate” information to make a reasonable prediction or determination of risk before issuing a test rule. To say the least, this is not mandated by the statutory history, which indicates Congress’ desire that EPA act on the basis of rational concern even in the absence of “adequate” information that an unreasonable risk existed. 9 Section 4 *986 may permissibly be read to authorize issuance of a test rule on the basis of less than more-probable-than-not evidence about a potentially unreasonable risk to health.

This conclusion is further bolstered by the legislative history underlying section 6. If CMA were correct that EPA must make a more-probable-than-not finding of risk under section 4’s “may present” language, then it would logically follow that section 6 — which contains the term “presents or will present an unreasonable risk” — must require an even stronger than more-probable-than-not demonstration of “unreasonable risk.” 10 Yet neither section 6 nor its legislative history indicate any such super-requirement of certainty. Indeed, section 6 states expressly that the Agency need only find a “reasonable basis” to conclude that an “unreasonable risk” exists. 15 U.S.C. § 2605(a). A “reasonable basis” requirement is certainly no more demanding than a more-probable-than-not requirement; indeed the phrase suggests a less demanding standard. This interpretation is confirmed by the House Report, which states that an EPA finding of “unreasonable risk” under section 6 is not expected to be supported by the same quantum of evidence as is customary in administrative proceedings.

A finding by the Administrator that there is such a reasonable basis must include adequate reasons and explanations for the Administrator’s conclusion. It does not, however, require the factual certainty of a “finding of fact” of the sort associated with adjudication.

Id. at 32. In sum, the standard Congress set for section 6 regulation, that the chemical “will present an unreasonable risk,” is no more rigorous (and arguably is less rigorous) than a more-probable-than-not finding. It follows as a matter of course that section 4’s “may present” language demands even less.

Of course, it is also evident from the legislative history that Congress did not intend to authorize EPA to issue test rules on the basis of mere hunches. The House Report states:

[T]he term “may” ... does not permit the Administrator to make a finding respecting probability of a risk on the basis of mere conjecture or speculation, i.e., it may or may not cause a risk.

H.R.Rep. No. 1341, supra, at 18. 11 Congress obviously intended section 4 to empower EPA to issue a test rule only after it had found a solid “basis for concern” by accumulating enough information to demonstrate a more-than-theoretical basis for suspecting that an “unreasonable risk” was involved in the use of the chemical.

CMA, curiously, insists that this very passage in the legislative history indicates that Congress embraced a more-probable-than-not standard for section 4. CMA Br. 21-22; CMA Reply Br. 5-6. CMA arrives at this conclusion by underlining the words “probability of a risk” in the quoted sentence and reading the phrase as being synonymous with “a risk that is more probable than not.” This reading, however, wrenches the targeted phrase out of its immediate context and puts the legislative history as a whole on the rack. Standing alone, the term “probability of a risk” might conceivably be read to describe a risk that is at least 51 percent certain. Within its host sentence, however, the phrase “probability of a risk” clearly refers to any degree of probability except for the one set out in the phrase immediately following it. Thus, the sentence says only that the Agency must *987 make “a finding respecting probability of a risk,” and that finding must be based on more than conjecture and speculation. The distance between the conjectural-speculative degree of probability foreclosed by Congress and the more-probable-than-not standard that CMA advances was clearly meant to be bridged by EPA’s discretion and expertise. The Agency’s reading of TSCA is a reasonable accommodation of the conflicting policies Congress committed to its care — specifically, the need to gather information about suspect chemicals without mandating expensive tests based on little more than a hunch. In sum, the legislative history fails to undermine EPA’s choice of a standard for section 4.

3. Interpretations in Other Circuits

We note that EPA’s interpretation of section 4 is consistent with the views of the only other two courts of appeals that have examined its “may present” language. See Ausimont U.S.A., Inc. v. EPA, 838 F.2d 93 (3d Cir.1988); Shell Chemical Co. v. EPA, 826 F.2d 295 (5th Cir.1987).

In Ausimont, the Third Circuit held that EPA could not require testing “based on little more than scientific curiosity.” 838 F.2d at 97. A test rule could not be sustained, the court stated, if the existence of exposure were “merely a remote possibility founded on theoretical factual situations.” Id. The Agency, it said, can validly require testing only “when an existing possibility of harm raises reasonable and legitimate cause for concern.” Id. The court, however, recognized that the Agency’s burden in the review of a section 4 test rule differs substantially from the normal standards of review applied to agency factfinding. See id. at 96 (Agency’s burden is “to demonstrate not fact, but doubt and uncertainty” as to the existence of an “unreasonable risk”). Thus, we find nothing in the Third Circuit’s interpretation of section 4 that is inconsistent with EPA’s reading of the statute.

The only other circuit which has considered section 4 remanded the case without reaching any decision as to its correct interpretation. In Shell Chemical Co. v. EPA, 826 F.2d 295, 298 (5th Cir.1987), the record showed that production of the chemical in question took place “in an enclosed, controlled environment where potential exposure ... [was] very limited.” Id. at 297-98. During oral argument, an EPA lawyer asserted that some additional evidence of exposure had come to light after promulgation of the rule. Id. at 298. The Fifth Circuit remanded the case to the Agency to explore the additional evidence, observing that, without it, the case was a “close one.” Id. The Shell court did not decide how strong the evidence of “unreasonable risk” must be to merit a test rule. Despite industry efforts to cite it to advantage, see CMA Reply Br. 10 & n. 9, Shell offers little guidance in our case. See Ausimont, 838 F.2d at 97 (“The holding in Shell Chemical ... is inconclusive and of no real assistance in the dispute before us.”).

CMA cites other cases interpreting language identical or similar to the “may present” term of section 4 as requiring a more-probable-than-not finding of risk. These opinions are of scant assistance, however, because in the main they involve provisions authorizing the regulation or prohibition of substances, rather than mere testing. 12 Such interpretations can hardly *988 be transplanted into the statutory scheme of TSCA, where Congress has self-consciously created a lower threshold for testing rules than for rules regulating chemical substances. See Ausimont, 838 F.2d at 96 (distinguishing between evidence required for test rule and evidence required for regulation under TSCA); H.R.Rep. No. 1341, supra, at 14 (same).

4. Conclusion

We have determined that Congress did not address the precise nature of the finding of “unreasonable risk of injury to health” that must underlie a section 4 test rule. EPA’s interpretation of the statutory standard is consistent with the statutory scheme and its legislative history, and it is a reasonable accommodation of the conflicting policies that Congress entrusted to the Agency’s care. We therefore uphold the Agency’s interpretation, under which EPA is empowered to issue a test rule where, in light of the evidence before it, the existence of an “unreasonable risk of injury to health” is a substantial (i.e., more than theoretical) probability. Since “unreasonable risk of injury to health” is a function of toxicity and exposure, this standard can be restated as follows: A test rule is warranted when there is a more-than-theoretical basis for suspecting that some amount of exposure occurs and that the substance is sufficiently toxic at that exposure level to present an “unreasonable risk of injury to health.” Under the second prong of Chevron, we uphold the Agency’s interpretation of section 4’s standard for issuing a testing rule.

B. Use of Inferences Versus Direct Evidence of Exposure

The second issue in the case is whether EPA must produce direct evidence documenting human exposure in order to rebut industry-submitted evidence casting doubt on the existence of exposure. 13 EPA contends that it need not provide direct evidence of exposure, even in response to industry evidence rebutting its initial circumstantial case on exposure, so long as the evidence on exposure as a whole provides a more-than-theoretical basis for discerning the presence of an “unreasonable risk of injury to health.” EPA concedes that exposure is a necessary component of “unreasonable risk of injury to health.” EPA Br. 18-19. The Agency argues, however, that it can issue a test rule where the existence of exposure is inferred from the circumstances under which the substance is manufactured and used. So long as industry evidence attacking those inferences fails to negate the Agency’s more-than-theoretical basis for inferring the existence of exposure, EPA claims, a test rule is warranted. After a careful search of the legislative materials, we conclude that Congress *989 did not address this particular issue, plying the second prong of Chevron, however, we conclude that the Agency’s construction of section 4 is a reasonable one and therefore uphold it. 14 Ap-

Reasonableness of Agency Interpretation

CMA does not contest the proposition that the use of inferences to establish exposure is reasonable as