CHEMICAL MFRS ASSN v USEPA 859 F2d 977 (DC Cir 1988)

CHEMICAL MANUFACTURERS ASSOCIATION ET AL PETITIONERS v US

ENVIRONMENTAL PROTECTION AGENCY RESPONDENT

No 86-1718

United States Court of Appeals District of Columbia Circuit

Argued September 12 1988

Decided October 21 1988

Robert M Sussman with whom William K Rawson David F ZollWashington DC and Mark N Duvall were on the brief forpetitioners Gabrielle H Williamson Washington DC enteredan appearance for petitioners

Ashley Doherty Atty US Dept of Justice with whom RogerJ Marzulla Asst Atty Gen Beth Ginsberg Atty US Deptof Justice and Andrew G Gordon Atty USEPA were on thebrief for respondent Michael W Steinberg Atty US Dept ofJustice Washington DC also entered an appearance forrespondent

Petition for Review of a Final Rule of the EnvironmentalProtection Agency

Before WALD Chief Judge and STARR and WILLIAMS CircuitJudges

Opinion for the Court filed by Chief Judge WALD

WALD Chief Judge

[1] Petitioners Chemical Manufacturers Association and fourcompanies that manufacture chemicals[fn1] (collectively CMA)seek to set aside a rule promulgated by the EnvironmentalProtection Agency (EPA or the Agency)[fn2] This Final TestRule was promulgated under section 4 of the Toxic Substances Control Act(TSCA or the Act) 15 USC sectsect 2601-2629 TheFinal Test Rule required toxicological testing to determine thehealth effects of the chemical 2-ethylhexanoic acid (EHA) andit continues to impose on exporters of EHA a duty to file certainnotices with EPA

[2] We uphold EPAs interpretation of TSCA as empowering the Agencyto issue a test rule on health grounds where it finds amore-than-theoretical basis for suspecting that the chemicalsubstance in question presents an unreasonable risk of injury tohealth This in turn requires the Agency to find amore-than-theoretical basis for concluding that the substance issufficiently toxic and human exposure to it is sufficient inamount to generate an unreasonable risk of injury to healthWe hold further that EPA can establish the existence and amount

of human exposure on the basis of inferences drawn from thecircumstances under which the substance is manufactured and usedEPA must rebut industry-supplied evidence attacking thoseinferences only if the industry evidence succeeds in renderingthe probability of exposure in the amount found by EPA no morethan theoretical or speculative The probability of infrequent oreven one-time exposure to individuals can warrant a test rule solong as there is a more-than-theoretical basis for determiningthat exposure in such doses presents an unreasonable risk ofinjury to health Finally we hold that the Agency correctlyapplied these standards in this case and that its findings aresupported by substantial evidence Consequently we affirm theFinal Test Rule

[3] I BACKGROUND

[4] A Statutory Structure

[5] TSCA provides for a two-tier system for evaluating andregulating chemical substances to protect against unreasonablerisks to human health and to the environment Section 6 of theAct permits EPA to regulate a substance that the Agency has foundpresents or will present an unreasonable risk of injury tohealth or the environment 15 USC sect 2605(a) Section 4 of theAct empowers EPA to require testing of a suspect substance inorder to obtain the toxicological data necessary to make adecision whether or not to regulate the substance under section6 The Act provides not surprisingly that the level ofcertainty of risk warranting a section 4 test rule is lower thanthat warranting a section 6 regulatory rule EPA is empowered torequire testing where it finds that the manufacturedistribution processing use or disposal of a particularchemical substance may present an unreasonable risk of injury tohuman health or the environment Id sect 2603(a)(1)(A)(i) TheAgencys interpretation of this statutory standard for testing isthe central issue in this case

[6] One of the chief policies underlying the Act is that mdash

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

[7] Id sect 2601(b)(1) The statute establishes an InteragencyTesting Committee comprised of scientists from various federalagencies to recommend that EPA give certain chemicals priorityconsideration for testing Id sect 2603(e) Under section 4 theAgency shall by rule require that testing [of a particularchemical] be conducted if three factors are present (i)activities involving the chemical may present an unreasonablerisk of injury to health or the environment (ii) insufficientdata and experience exist upon which to determine the effects ofthe chemical on health or environment and (iii) testing isnecessary to develop such data Id sect 2603(a)(1)(A) The

companies that manufacture and process the substance are toconduct the tests and submit the data to the Agency Id sect2603(b)(3)(B) Costs of testing are to be shared among thecompanies either by agreement or by EPA order in the absence ofagreement Id sect 2603(c)(3)(A)

[8] A test rule promulgated under section 4 is subject to judicialreview in a court of appeals pursuant to section 19(a) of TSCA15 USC sect 2618(a) A test rule may be set aside if it is notsupported by substantial evidence in the rulemaking record taken as a whole Id sect 2618(c)(1)(B)(i)

[9] B Facts and Prior Proceedings

[10] EHA is a colorless liquid with a mild odor It is usedexclusively as a chemical intermediate or reactant in theproduction of metal soaps peroxy esters and other products usedin industrial settings EHA itself is totally consumed during themanufacture of these products as a result no products offeredfor sale to industry or to consumers contain EHA See FinalTest Rule 51 FedReg at 40319

[11] The Interagency Testing Committee first designated EHA forpriority consideration for health effects tests on May 29 1984Fourteenth Report of the Interagency Testing Committee to theAdministrator 49 FedReg 22389 22391 (1984) The Committeebased its recommendation in part on the structural similarity ofEHA to chemicals known to cause cancer in test animals and on itsfinding that insufficient information existed concerning thechronic health effects of EHA Id at 22399-400 SubsequentlyEPA held two public meetings on EHA Supplemental Joint Appendix(SJA) 499-504 505-15 During these meetings in whichpersons representing the petitioners made appearances EPA soughtinformation on a variety of issues relating to EHA usesproduction and human exposure SJA 502-04 508-11 513-14

[12] EPA issued a proposed test rule on May 17 19852-Ethylhexanoic Acid Proposed Test Rule 50 FedReg 20678(1985) (the Proposed Test Rule) The rule proposed a series oftests to ascertain the health risks of EHA and it set outproposed standards for the conduct of those tests Id at20682-83 EPA based the Proposed Test Rule on a finding that EHAmay present an unreasonable risk of subchronic toxicity (harmto bodily organs from repeated exposure over a limited period oftime) oncogenicity (tumor formation) and developmental toxicity(harm to the fetus) Id at 20682 As to subchronic toxicityEPA cited studies suggesting that both EHA and chemicalsstructurally similar to it cause harm to the livers of testanimals Id As to oncogenicity EPA cited studies suggestingthat chemicals structurally analogous to EHA cause cancer inlaboratory animals Id at 20681-82 As to developmentaltoxicity EPA cited studies indicating that both EHA and itschemical analogues have produced fetal malformations in testanimals Id at 20682

[13] The Proposed Test Rule also addressed the question of whetherhumans are exposed to EHA a question of critical importance to

this case The Agency acknowledged that since no finishedproducts contain EHA consumer exposure is not a concern Itlikewise discounted the dangers of worker exposure to EHA vaporsId at 20681 The Agency based its Proposed Test Rule solelyon the potential danger that EHA will come in contact with theskin of workers As evidence of potential dermal exposure theAgency noted that approximately 400 workers are engaged in themanufacture transfer storage and processing of 20 to 25 millionpounds of EHA per year Id at 20679 20682 Furtherrebutting claims by industry representatives that gloves areroutinely worn during these activities EPA noted that workerhygiene procedures can vary widely throughout the industrythat workers are not required by existing federal regulations towear gloves and that the industry had not monitored work sitesfor worker exposure to EHA Id at 20680-81

[14] A public comment period commenced with the publication of theProposed Test Rule and ended on July 16 1985 EPA held a publicmeeting on October 8 1985 to discuss issues related to theProposed Test Rule JA 69-71 Industry representativessubmitted extensive comments on July 15 1985 and January 171986 JA 96-97 161-63 Before publication of the Final TestRule EPA received notice of a new study purporting to presentfurther evidence of the potential developmental toxicity of EHAFinal Test Rule 51 FedReg at 40319-21

[15] CMA criticized the toxicology studies cited by EPA and soughtto show that the use of gloves by employees of companies workingwith EHA prevented human exposure to the chemical thus renderingany test rule invalid Before publication of the final rule CMAretained an independent consultant to conduct a survey of gloveuse by the employees of companies working with EHA JA 32-47(the Glove Use Survey) The results of the survey weresubmitted to EPA CMA also submitted the results of an EastmanKodak Co study of the permeability of nitrile and neoprenegloves JA 68 (the Glove Permeability Study)

[16] EPA published the Final Test Rule for EHA on November 6 1986The rule required a 90-day subchronic toxicity test adevelopmental toxicity test and a pharmacokinetics test FinalTest Rule 51 FedReg at 40318 In the preamble to the FinalTest Rule EPA asserted its finding of potential exposure to EHAThe Agency rebutted industry claims of non-exposure bycriticizing the methodology of the Glove Use Survey and the GlovePermeability Study Id at 40320[fn3] The Final Test Rulealso addressed the industry critique of the prior toxicologystudies on EHA Acknowledging some weaknesses in the studies EPAnonetheless asserted that they add to the weight of evidencesupporting findings of potential toxicity and thus helped tojustify further testing Id at 40321

[17] The pharmacokinetics study required by the rule entailed theoral and dermal administration of EHA to experimental animals atlow and high doses Id at 40327-28 The subchronic toxicitystudy involved administering EHA to animals in graduated dailydoses over a period of 90 days Id at 40323 40328-30 Thedevelopmental toxicity tests entailed administering EHA orally in

various doses during the pregnancy of experimental animals Idat 40323 40330 All studies were to be conducted in accordancewith EPA standards Results were to be submitted by certaindeadlines the last of which was 18 months after the effectivedate of the Final Test Rule hence the last deadline was June20 1988[fn4] Id at 40330

[18] The petitioners filed this petition for review on December 231986 moving at the same time for a stay pending review Thiscourt denied the stay motion on January 20 1987

[19] The last of the required test data were submitted to EPA onJune 20 1988 EPA moved this court on June 9 1988 for leave tofile a motion to dismiss the case as moot By order on June 301988 the court denied the motion and directed that the partiesaddress the mootness issue before this panel

[20] II MOOTNESS

[21] EPA argues that this petition for review is moot because all ofthe test results have been submitted CMA does not seek damagesor reimbursement for having conducted the tests Thus the Agencyargues there are no residual effects of the testing requirementto keep the controversy alive We nonetheless hold the case to bejusticiable because the rule continues to impose concreteobligations on companies handling EHA

[22] The doctrine of mootness is one aspect of the requirement thatfederal courts decide only actual cases or controversiesUS Const art III sect 2 In order for a case not to be moot aparty must show a direct injury that gives it a continuing`personal stake in the outcome of the litigation Sample vJohnson 771 F2d 1335 1339 (9th Cir 1985) (quoting Warth vSeldin 422 US 490 498 95 SCt 2197 2205 45 LEd2d 343(1975)) cert denied 475 US 1019 106 SCt 1206 89LEd2d 319 (1986) In this sense the mootness inquiry issimilar to the inquiry of whether a party has constitutionalstanding Sample 771 F2d at 1339 13A C Wright A Miller ampE Cooper Federal Practice amp Procedure sect 35331 at 219 (1984)(assimilating mootness principles with standing principles)[hereinafter Wright amp Miller] Federal courts areconstitutionally forbidden to render advisory opinions or todecide questions that cannot affect the rights of litigants inthe case before them Better Govt Assn v Department ofState 780 F2d 86 90-91 (DCCir 1986) (quoting NorthCarolina v Rice 404 US 244 246 92 SCt 402 404 30LEd2d 413 (1971))

[23] On the other hand a claim for prospective relief remainsjusticiable so long as the conduct complained of has continuingpresent adverse effects See OShea v Littleton 414 US 488496 94 SCt 669 676 38 LEd2d 674 (1974) The fact thatthe principal form of relief that a party seeks is no longeravailable does not in itself render the case nonjusticiable SeeBetter Govt Assn 780 F2d at 91-92 Maine Central R Co vBrotherhood of Railway Maintenance Employees 813 F2d 484 487(1st Cir) cert denied ___ US ___ 108 SCt 91 98

LEd2d 52 (1987) Where one of the several issues presentedbecomes moot the remaining live issues supply the constitutionalrequirement of a case or controversy Powell v McCormack395 US 486 497 89 SCt 1944 1951 23 LEd2d 491 (1969) Thisis so even where the mooted claim is much greater in magnitudethan the remaining ones Id In light of the similarity betweenmootness and standing it should be noted that the distinct andpalpable injury a party must suffer in order to be deemed tohave constitutional standing Warth v Seldin 442 US at 50195 SCt at 2206 need not be tangible or great an`identifiable trifle will do National Wildlife Federation vHodel 839 F2d 694 704 (DCCir 1988) (quoting United Statesv SCRAP 412 US 669 689 n 14 93 SCt 2405 2417 n 14 37LEd2d 254 (1973)) Similarly the continuing personal stakeneeded to rebut an assertion of mootness need not be large inmagnitude to suffice Thus as the Supreme Court has stated acase is not moot if some stake in the outcome remainsnotwithstanding the size of the dispute Firefighters Local1784 v Stotts 467 US 561 571 104 SCt 2576 2584 81LEd2d 483 (1984) All that is necessary is that the partyseeking relief have a concrete interest in the outcome Id

[24] In light of these principles this case should not be dismissedon the ground of mootness Section 12(b) of TSCA provides thatany company exporting a chemical substance that is subject to atest rule under section 4 must notify the Administrator of suchexportation or intent to export and the Administrator shallfurnish to the government of such country notice of theavailability of the data submitted to the Administrator undersuch section for such substance or mixture15 USC sect 2611(b)(1) EPA regulations require an exporter to provide aseparate notice for the first export or intended export to aparticular country in a calendar year 40 CFR sect 70765(a)(2)(1987) See also Final Test Rule 51 FedReg at 40324(summarizing these obligations) There is also indication in therecord that compliance with the notification requirement requiresmore than a minimal amount of work Supplemental Affidavit ofHenry J Sauer para 5

[25] The acknowledgedly modest nature of these obligations does notrequire this court constitutionally to refuse to adjudicate thecase[fn5] If something still remains to be done under a rulethat is usually enough to spare it from the mootness ax SeeCasey v FTC 578 F2d 793 796 (9th Cir 1978) (challenge toagency subpoenas not moot despite submission of all documentswhere documents were not verified in accordance with subpoenas)FTC v Gibson Prods 569 F2d 900 903 (5th Cir 1978) 13AWright amp Miller supra sect 35332 at 258-59 Although thepetitioners stake in the outcome of this case is manifestlysmaller than it was when the case was filed they still have asufficiently concrete interest in the disposition of theirpetition to warrant the exercise of federal courtjurisdiction[fn6]

[26] III STATUTORY INTERPRETATION

[27] The Toxic Substances Control Act requires EPA to promulgate a

test rule under section 4 if a chemical substance inter aliamay present an unreasonable risk of injury to health or theenvironment 15 USC sect 2603(a)(1)(A)(i) The parties bothaccept the proposition that the degree to which a particularsubstance presents a risk to health is a function of two factors(a) human exposure to the substance and (b) the toxicity of thesubstance See Ausimont USA Inc v EPA 838 F2d 93 96(3d Cir 1988) They also agree that EPA must make some sort ofthreshold finding as to the existence of an unreasonable risk ofinjury to health The parties differ however as to the mannerin which this finding must be made Specifically three issuesare presented

[28] The first issue is whether under section 4 of TSCA EPA mustfind that the existence of an unreasonable risk of injury tohealth is more probable than not in order to issue a test ruleCMA argues that the statute requires a more-probable-than-notfinding CMA Br 22 EPA disagrees contending that the statuteis satisfied where the existence of an unreasonable risk ofinjury to health is a substantial probability mdash that is aprobability that is more than merely theoretical speculative orconjectural EPA Br 21[fn7]

[29] The second issue is whether once industry has presentedevidence tending to show an absence of human exposure EPA mustrebut it by producing direct evidence of exposure CMA claimsthat when industry evidence casts doubt on the existence ofexposure the burden of production shifts back to EPA which mustproduce direct evidence documenting actual instances in whichexposure has taken place CMA Br 27 30 EPA on the other handargues that it can make the requisite finding of exposure basedsolely on inferences drawn from the circumstances under which achemical substance is manufactured and used EPA Br 32-33

[30] The third issue is whether the Agency has authority to issue atest rule where any individuals exposure to a substance is anisolated non-recurrent event CMA argues that even if EPApresents direct evidence of exposure the Act precludes issuanceof a test rule where exposure consists only of rare instancesinvolving brief exposure CMA Reply Br 11 n 10 EPA contendson the other hand that the Act does not require in allcircumstances a risk of recurrent exposure EPA Br 28

[31] We will explore each of these three issues in turn In eachinstance we must examine the statute itself and its legislativehistory to determine whether Congress addressed the particularpoint in question If this court ascertains that Congress hasdirectly spoken to the precise question at issue then both thecourt and EPA must give effect to the unambiguously expressedintent of Congress Chevron USA Inc v Natural ResourcesDefense Council Inc 467 US 837 842-43 104 SCt 27782781 81 LEd2d 694 (1984) accord NLRB v United Food amp ComlWorkers Union ___ US ___ 108 SCt 413 421 98 LEd 2d 429(1987) If however the statute is silent or ambiguous withrespect to the specific issue the court is left to decide onlywhether the Agencys construction of the statute is a reasonableone Chevron 467 US at 843 104 SCt at 2782 Congressional

silence is deemed an implicit delegation of power to an agency tomake policy choices consistent with the statutory purpose Idat 844-45 104 SCt at 2782-83 [The] view of the agencycharged with administering the statute is entitled toconsiderable deference and to sustain it we need not find thatit is the only permissible construction that EPA might haveadopted but only that EPAs understanding of this very `complexstatute is a sufficiently rational one to preclude a court fromsubstituting its judgment for that of EPA ChemicalManufacturers Assn v Natural Resources Defense Council Inc470 US 116 125 105 SCt 1102 1107 84 LEd2d 90 (1985)(interpreting the Clean Water Act)

[32] A Required Finding of Unreasonable Risk

[33] As to the first issue in this case the standard of probabilityof an unreasonable risk to health we find that Congress did notaddress the precise question in issue Examining the EPAinterpretation under the second prong of Chevron we find it tobe reasonable and consistent with the statutory scheme andlegislative history Consequently we uphold the Agencysconstruction of TSCA as authorizing a test rule where EPAs basisfor suspecting the existence of an unreasonable risk of injuryto health is substantial mdash ie when there is amore-than-theoretical basis for suspecting that some amount ofexposure takes place and that the substance is sufficiently toxicat that level of exposure to present an unreasonable risk ofinjury to health

[34] 1 Text and Structure of the Statute

[35] Both the wording and structure of TSCA reveal that Congress didnot expect that EPA would have to document to a certainty theexistence of an unreasonable risk before it could requiretesting This is evident from the two-tier structure of the ActIn order for EPA to be empowered to regulate a chemicalsubstance the Agency must find that the substance presents orwill present an unreasonable risk of injury to health or theenvironment TSCA sect 6 15 USC sect 2605(a) The testingprovision at issue here by contrast empowers EPA to act at alower threshold of certainty than that required for regulationSpecifically testing is warranted if the substance maypresent an unreasonable risk of injury to health or theenvironment TSCA sect 4 15 USC sect 2603(a)(1)(A)(i) (emphasisadded) Thus the language of section 4 signals that EPA is tomake a probabilistic determination of the presence ofunreasonable risk

[36] 2 Legislative History

[37] The legislative history of TSCA compels a further conclusionIt not only shows that unreasonable risk need not be a matterof absolute certainty it shows the reasonableness of EPAsconclusion that unreasonable risk need not be established to amore-probable-than-not degree

[38] A House Report on the version of the bill that eventually

became TSCA underscores the distinction between the section 6standard and the section 4 standard To issue a test rule EPAneed not find that a substance actually does cause or present anunreasonable risk

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

[39] HRRep No 1341 94th Cong 2d Sess 17-18 (1976)[fn8]

[40] The House Report also contains signals indicating that Congressexpected EPA to act even when evidence of unreasonable risk wasless than conclusive According to that report the word may insection 4 was intended to focus the Agencys attention onchemical substances about which there is a basis for concernbut about which there is inadequate information to reasonablypredict or determine the effects of the substance or mixture onhealth or the environment Id at 17 (emphasis added) TheConference Committee Report reemphasized that the statutorylanguage focused the Agencys attention on substances aboutwhich there is a basis for concern HRConfRep No 1679 94thCong 2d Sess 61 (1976) US Code Cong amp Admin News 1976 p4546

[41] These indications of congressional intent illustrate that EPAsreading of TSCA is a permissible one Congress intended toauthorize testing where the existence of an unreasonable riskcould not yet be reasonably predicted The Agencysdetermination that it is empowered to act where the existence ofan unreasonable risk cannot yet be said to be more probablethan not is entirely consistent with that expression of intentThe EPA interpretation is likewise consistent with the level ofcertainty suggested by the phrase basis for concern To acceptCMAs position would require the Agency to gather adequateinformation to make a reasonable prediction or determination ofrisk before issuing a test rule To say the least this is notmandated by the statutory history which indicates Congressdesire that EPA act on the basis of rational concern even in theabsence of adequate information that an unreasonable riskexisted[fn9] Section 4 may permissibly be read to authorizeissuance of a test rule on the basis of less thanmore-probable-than-not evidence about a potentially unreasonablerisk to health

[42] This conclusion is further bolstered by the legislative historyunderlying section 6 If CMA were correct that EPA must make amore-probable-than-not finding of risk under section 4s maypresent language then it would logically follow that section 6mdash which contains the term presents or will present anunreasonable risk mdash must require an even stronger thanmore-probable-than-not demonstration of unreasonablerisk[fn10] Yet neither section 6 nor its legislative historyindicate any such super-requirement of certainty Indeed section6 states expressly that the Agency need only find a reasonable

basis to conclude that an unreasonable risk exists 15 USC sect 2605(a)A reasonable basis requirement is certainly no moredemanding than a more-probable-than-not requirement indeed thephrase suggests a less demanding standard This interpretation isconfirmed by the House Report which states that an EPA findingof unreasonable risk under section 6 is not expected to besupported by the same quantum of evidence as is customary inadministrative proceedings

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

[43] Id at 32 In sum the standard Congress set for section 6regulation that the chemical will present an unreasonablerisk is no more rigorous (and arguably is less rigorous) than amore-probable-than-not finding It follows as a matter of coursethat section 4s may present language demands even less

[44] Of course it is also evident from the legislative history thatCongress did not intend to authorize EPA to issue test rules onthe 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 ie it may or may not cause a risk

[45] HRRep No 1341 supra at 18[fn11] Congress obviouslyintended section 4 to empower EPA to issue a test rule only afterit had found a solid basis for concern by accumulating enoughinformation to demonstrate a more-than-theoretical basis forsuspecting that an unreasonable risk was involved in the use ofthe chemical

[46] CMA curiously insists that this very passage in thelegislative history indicates that Congress embraced amore-probable-than-not standard for section 4 CMA Br 21-22 CMAReply Br 5-6 CMA arrives at this conclusion by underlining thewords probability of a risk in the quoted sentence and readingthe phrase as being synonymous with a risk that is more probablethan not This reading however wrenches the targeted phraseout of its immediate context and puts the legislative history asa whole on the rack Standing alone the term probability of arisk might conceivably be read to describe a risk that is atleast 51 percent certain Within its host sentence however thephrase probability of a risk clearly refers to any degree ofprobability except for the one set out in the phrase immediatelyfollowing it Thus the sentence says only that the Agency mustmake a finding respecting probability of a risk and thatfinding must be based on more than conjecture and speculationThe distance between the conjectural-speculative degree ofprobability foreclosed by Congress and the more-probable-than-not

standard that CMA advances was clearly meant to be bridged byEPAs discretion and expertise The Agencys reading of TSCA is areasonable accommodation of the conflicting policies Congresscommitted to its care mdash specifically the need to gatherinformation about suspect chemicals without mandating expensivetests based on little more than a hunch In sum the legislativehistory fails to undermine EPAs choice of a standard for section4

[47] 3 Interpretations in Other Circuits

[48] We note that EPAs interpretation of section 4 is consistentwith the views of the only other two courts of appeals that haveexamined its may present language See Ausimont USA Incv EPA 838 F2d 93 (3d Cir 1988) Shell Chemical Co v EPA826 F2d 295 (5th Cir 1987)

[49] In Ausimont the Third Circuit held that EPA could notrequire testing based on little more than scientific curiosity838 F2d at 97 A test rule could not be sustained the courtstated if the existence of exposure were merely a remotepossibility founded on theoretical factual situations Id TheAgency it said can validly require testing only when anexisting possibility of harm raises reasonable and legitimatecause for concern Id The court however recognized that theAgencys burden in the review of a section 4 test rule differssubstantially from the normal standards of review applied toagency factfinding See id at 96 (Agencys burden is todemonstrate not fact but doubt and uncertainty as to theexistence of an unreasonable risk) Thus we find nothing inthe Third Circuits interpretation of section 4 that isinconsistent with EPAs reading of the statute

[50] The only other circuit which has considered section 4 remandedthe case without reaching any decision as to its correctinterpretation In Shell Chemical Co v EPA 826 F2d 295 298(5th Cir 1987) the record showed that production of the chemicalin question took place in an enclosed controlled environmentwhere potential exposure [was] very limited Id at297-98 During oral argument an EPA lawyer asserted that someadditional evidence of exposure had come to light afterpromulgation of the rule Id at 298 The Fifth Circuitremanded the case to the Agency to explore the additionalevidence observing that without it the case was a close oneId The Shell court did not decide how strong the evidence ofunreasonable risk must be to merit a test rule Despiteindustry efforts to cite it to advantage see CMA Reply Br 10amp n 9 Shell offers little guidance in our case SeeAusimont 838 F2d at 97 (The holding in Shell Chemical is inconclusive and of no real assistance in the dispute beforeus)

[51] CMA cites other cases interpreting language identical orsimilar to the may present term of section 4 as requiring amore-probable-than-not finding of risk These opinions are ofscant assistance however because in the main they involveprovisions authorizing the regulation or prohibition of

substances rather than mere testing[fn12] Such interpretationscan hardly be transplanted into the statutory scheme of TSCAwhere Congress has self-consciously created a lower threshold fortesting rules than for rules regulating chemical substances SeeAusimont 838 F2d at 96 (distinguishing between evidencerequired for test rule and evidence required for regulation underTSCA) HRRep No 1341 supra at 14 (same)

[52] 4 Conclusion

[53] We have determined that Congress did not address the precisenature of the finding of unreasonable risk of injury to healththat must underlie a section 4 test rule EPAs interpretation ofthe statutory standard is consistent with the statutory schemeand its legislative history and it is a reasonable accommodationof the conflicting policies that Congress entrusted to theAgencys care We therefore uphold the Agencys interpretationunder which EPA is empowered to issue a test rule where in lightof the evidence before it the existence of an unreasonable riskof injury to health is a substantial (ie more thantheoretical) probability Since unreasonable risk of injury tohealth is a function of toxicity and exposure this standard canbe restated as follows A test rule is warranted when there is amore-than-theoretical basis for suspecting that some amount ofexposure occurs and that the substance is sufficiently toxic atthat exposure level to present an unreasonable risk of injury tohealth Under the second prong of Chevron we uphold theAgencys interpretation of section 4s standard for issuing atesting rule

[54] B Use of Inferences Versus Direct Evidence of Exposure

[55] The second issue in the case is whether EPA must produce directevidence documenting human exposure in order to rebutindustry-submitted evidence casting doubt on the existence ofexposure[fn13] EPA contends that it need not provide directevidence of exposure even in response to industry evidencerebutting its initial circumstantial case on exposure so long asthe evidence on exposure as a whole provides amore-than-theoretical basis for discerning the presence of anunreasonable risk of injury to health EPA concedes thatexposure is a necessary component of unreasonable risk of injuryto health EPA Br 18-19 The Agency argues however that itcan issue a test rule where the existence of exposure is inferredfrom the circumstances under which the substance is manufacturedand used So long as industry evidence attacking those inferencesfails to negate the Agencys more-than-theoretical basis forinferring the existence of exposure EPA claims a test rule iswarranted After a careful search of the legislative materialswe conclude that Congress did not address this particularissue[fn14] Applying the second prong of Chevron however weconclude that the Agencys construction of section 4 is areasonable one and therefore uphold it

[56] Reasonableness of Agency Interpretation

[57] CMA does not contest the proposition that the use of inferences

to establish exposure is reasonable as a general matter CMAReply Br 10-11 CMA challenges only the Agencys reliance oninferences in the face of industry evidence attacking its initialexposure finding Id In light of our preceding decision on thequantum of proof necessary for a test rule however we see noreason to require EPA to come up with additional evidence ofexposure when the industry challenges its initial finding unlessthe industry evidence effectively reduces the basis for anexposure finding to the realm of theory speculation andconjecture We conclude that it is reasonable for EPA to rely oninferences in issuing a section 4 test rule so long as all theevidence mdash including the industry evidence mdash indicates amore-than-theoretical probability of exposure Whether theAgencys ultimate finding of exposure is supported by sufficientevidence depends on a weighing of all the evidence not onwhether the type of evidence produced by EPA is direct asopposed to circumstantial

[58] Even if EPA were required under CMAs scenario to establishexposure as more probable than not there would be no reason tolimit its proof to direct evidence Yet CMA bases its argumentfor a direct evidence requirement on its claim that exposure mustbe shown to be more-probable-than-not We fail to see thepurported link between these two issues A party may show a factto be more probable than not solely by presenting circumstantial(ie inference-based) evidence Conversely a party may fail tomeet a more-probable-than-not burden even with the help of directevidence See 4 S Gard supra note 13 sect 296(circumstantial evidence is often more reliable than directevidence because the latter is more easily manufactured and itsfalsity difficult to detect) Thus even had we accepted CMAscontention on the more-probable-than-not issue it would notnecessarily follow that industry evidence of lack of exposuremust be rebutted by EPA direct evidence that exposure exists Atany rate we have concluded that a standard less rigorous than amore-probable-than-not standard is reasonable and consistent withthe intent of Congress Accordingly we find EPAs interpretationof section 4 as not requiring direct evidence of exposure to bereasonable and consistent with the statutory scheme[fn15]

[59] C Recurrent Versus Rare Exposure

[60] The third statutory issue is whether section 4 of TSCAauthorizes EPA to issue a test rule where any individualsexposure to a chemical is likely to be a rare brief event CMAcontends that only recurrent exposure warrants a test rule EPAmaintains that it can issue a test rule in the absence ofrecurrent exposure where there is a more-than-theoretical basisfor suspecting that infrequent or single-dose exposure presentsan unreasonable risk of injury to health We find no indicationin the statute or its history that Congress addressed thisparticular issue but once again turning to the second prong ofChevron we deem reasonable the Agencys construction ofsection 4 as permitting a test rule even where exposure is notrecurrent

[61] 1 Reasonableness of Agency Interpretation

[62] Requiring testing in the absence of recurrent or long-termexposure would be unreasonable only if there were not amore-than-theoretical probability that non-recurrent briefexposure could create an unreasonable risk of injury to healthHere the agency has already concluded that even one-time exposureto suspected developmental toxicants (of which EHA is one) may besufficient to cause harm to a developing fetus Guidelines forthe Health Assessment of Suspect Developmental Toxicants 51FedReg 34028 34037 (1986) Where the pool of potentiallyexposed persons consists of hundreds of individuals and wherethe substance in question is toxic in small doses the mere factthat exposure events might occur on an accidental rare oroccasional basis would not prevent the substance from potentiallycausing a significant degree of injury to human health Tointerpret TSCA as not authorizing testing where potentialexposure is brief or non-recurrent would render it ineffective insuch circumstances Certainly this court should not substituteits judgment for that of the Agency on the issue of whetheroccasional exposure to a developmental toxicant can ever merit atest rule We therefore hold the EPA construction of section 4 onthis point to be reasonable and consistent with the purpose ofTSCA[fn16]

[63] 2 Interpretations in Other Circuits

[64] We are not dissuaded from this conclusion by the cases in twoother circuits CMA cites to bolster its claim that recurrentexposure is necessary to justify testing under section 4 of TSCASee Ausimont USA Inc v EPA 838 F2d 93 (3d Cir 1988)Shell Chemical Co v EPA 826 F2d 295 (5th Cir 1987) It istrue that both Ausimont and Shell mention amount of exposureas a relevant factor in a determination of unreasonable riskAusimont 838 F2d at 96 Shell 826 F2d at 298 Neither ofthe options however asserts that a particular quantity ofexposure is needed in all instances to warrant a test rule Nordo they assert that individuals must be exposed repeatedly Thecases lay down no statutory interpretation inconsistent with theone EPA has adopted Assuredly amount or quantity of exposure isrelevant in determining whether there is a more-than-theoreticalbasis for suspecting that the combination of toxicity andexposure is sufficient to present an unreasonable risk of injuryto health Clearly the necessary amount of exposure to meetthat norm will vary depending on the level of exposure at whichthe substances toxicity is likely to present an unreasonablerisk of injury to health None of this is to say however thatEPA can never issue a test rule unless recurrent exposure takesplace

[65] D Summary of Statutory Interpretation

[66] We reject all three of CMAs challenges to EPAs statutoryconstruction of section 4 Summarizing the foregoing discussionwe uphold EPAs conclusion that it is empowered to issue a testrule where the evidence pointing to the presence of anunreasonable risk of injury to health is substantial enough toindicate that the decision to issue a test rule is based on more

than theory speculation and conjecture The Agency must findthat there is a morethan-theoretical basis for concluding thatsome amount of exposure takes place and that toxicity at thatlevel of exposure suffices to present an unreasonable risk ofinjury to health Inferences drawn from the circumstances underwhich a substance is manufactured and used can suffice toestablish the existence and amount of exposure Industry-suppliedevidence attacking those inferences must be rebutted by EPA onlyif the industry evidence renders the probability of exposure at alevel sufficient to present an unreasonable risk no more thantheoretical and speculative So long as there is amore-than-theoretical probability that the toxic substance inrare or single doses presents an unreasonable risk of injury tohealth the statutory standard is met whatever the infrequencyof exposure

[67] We turn now to consider the sufficiency of the evidence in thisrecord to meet these statutory standards

[68] IV STANDARD OF REVIEW FOR AGENCY FINDINGS

[69] The Final Test Rule was promulgated under section 4 of TSCA15 USC sect 2603(a) Judicial review of the rule must proceedaccording to 15 USC sect 2618(c)(1)(B)(i) which specificallystates that the Administrative Procedure Act (APA) standardcustomarily applied in agency cases 5 USC sect 706 isinapposite

[I]n the case of review of a rule under section 2603(a) of this title the standard for review prescribed by paragraph (2)(E) of such section 706 [of title 5] shall not apply and the court shall hold unlawful and set aside such rule if the court finds that the rule is not supported by substantial evidence in the rulemaking record taken as a whole

[70] 15 USC sect 2618(c)(1)(B)(i)

[71] This case therefore does not turn on an interpretation of theterm substantial evidence as it appears in the APA5 USC sect 706(2)(E) but on an interpretation of the term substantialevidence in the rulemaking record taken as a whole as itappears in TSCA 15 USC sect 2618(c)(1)(B)(i) Despite thesimilarity in wording Congress apparently contemplated that theTSCA standard should be viewed as a distinct standard otherwisethere would have been no need to specifically rule out the APAreview standard

[72] The legislative history of TSCA further indicates that Congressperceived some difference between the standard it chose for TSCAand the APAs arbitrary-and-capricious standard[fn17] The HouseReport explained

[I]t is the intent that the traditional presumption of validity of an agency rule would remain in effect [but] [t]he Committee has chosen to adopt the

substantial evidence test for the Committee intends that the reviewing court engage in a searching review of the Administrators reasons and explanations for the Administrators conclusions

[73] HRRep No 1341 supra at 55-56 (emphasis added)[fn18]

[74] The Conference Committee Report further illuminated itsrejection of the arbitrary-and-capricious standard

[T]he conferees have adopted the substantial evidence test because they intend that the reviewing court focus on the rulemaking record to see if the Administrators action is supported by that record Of course the conferees do not intend that the court substitute its judgment for that of the Administrator

[75] HRConfRep No 1679 supra at 96 US Code Cong ampAdmin News 1976 p 4581 Both the House and the ConferenceReport thus indicate that it was Congress intent that review oftest rules under section 4 of TSCA be more searching than thejudicial review undertaken in most agency cases Interpretingcourts have similarly concluded that the standard of review fortest rules under TSCA is a particularly demanding oneAusimont USA Inc v EPA 838 F2d 93 96 (3d Cir 1988)Shell Chemical Co v EPA 826 F2d 295 297 (5th Cir 1987)accord Environmental Defense Fund v EPA 636 F2d 1267 1277(DCCir 1980)

[76] This fairly rigorous standard of record review should nothowever be confused with the substantive statutory standardpreviously discussed at length EPAs permissible interpretationof may present an unreasonable risk works in tandem with thesubstantial evidence standard of record review to effectuate astatutory scheme that empowers the Agency to require testingwhere the existence of an unreasonable risk of injury to healthis not yet more-probable-than-not but at the same time theAgency is required to identify the facts that underlie itsdetermination that there is a more-than-theoretical basis tosuspect the presence of such a risk See Ausimont 838 F2d at96 ([H]ere we look to see if the Administrator producedsubstantial evidence to demonstrate not fact but doubt anduncertainty)

[77] V SUPPORT IN THE RULEMAKING RECORD

[78] This court must affirm the Final Test Rule unless the recordtaken as a whole fails to provide substantial evidencedemonstrating a more-than-theoretical probability of exposure anda more-than-theoretical probability of toxicity at a suspectedlevel of exposure that presents an unreasonable risk of harmWe conclude that the Agencys findings are supported bysubstantial evidence

[79] A Evidence of Exposure to EHA

[80] EPA ultimately concluded that there is a more-than-theoreticalprobability that workers are exposed to EHA After revising itsestimate of the likely amount of exposure downward in response toindustry comments the Agency estimated that workers are likelyto be exposed on infrequent occasions as a result of accidentalspillage and that such exposure could amount to 60 milligrams ofEHA per kilogram of body weight for each incident Final TestRule 51 FedReg 40320 (exposure believed to be incidentalrather than routine)

[81] Before issuing the Proposed Test Rule EPA adduced a number offacts from which it could reasonably have inferred potentialexposure Although industry representatives presented rebuttalevidence purporting to show widespread use of gloves andpurporting to show the effectiveness of gloves in preventingdermal exposure EPA rationally concluded in the Final Test Rulethat this evidence did not overcome its finding of potentialexposure First weaknesses in testing methods suggested that thepurported facts averred by industry had not been establishedreliably Second and more importantly the very facts averred byindustry (even assuming their truth) tended to support ratherthan refute a finding of potential exposure

[82] A number of facts supported EPAs initial inference ofpotential exposure First the number of employees involved inhandling EHA is not small approximately 400 are potentiallyexposed Proposed Test Rule 50 FedReg at 20680 Final TestRule 51 FedReg at 40319-20 Second a large amount of EHA isproduced and used some 20 to 25 million pounds annuallyProposed Test Rule 50 FedReg at 20679 Final Test Rule 51FedReg at 40319 Increases in production are expected to raisethis amount substantially Id at 40322 SJA 516 CfShell Chemical Co v EPA 826 F2d 295 298 (5th Cir 1987) (inreview of a TSCA section 4 test rule remand to EPA isappropriate where EPA asserts post-promulgation increases in usenot anticipated at time of rulemaking) Third EHA is handled notat one or a small number of plants with uniform practicesrather it is handled at two manufacturing plants andapproximately 100 processing plants nationwide giving rise tolegitimate doubts by EPA about the uniformity of safetypractices Proposed Test Rule 50 FedReg at 20680 Final TestRule 51 FedReg at 40319 Fourth certain actions (egdrumming and undrumming cleaning sampling) were deemed likelyby EPA to subject workers to dermal exposure Proposed Test Rule50 FedReg at 20680 Final Test Rule 51 FedReg at 40319Fifth EPA inferred potential exposure from the fact that gloveuse is not required by existing federal regulations ProposedTest Rule 50 FedReg at 20681 Sixth EPA noted that thephysicochemical properties of EHA do not force workers to avoidcontact with EHA mdash it is a colorless liquid with a mild odorFinal Test Rule 51 FedReg at 40319 From these facts EPAlogically could have reached the initial conclusion that therewas a more-than-theoretical probability of human exposure to EHA

[83] As the record must be viewed as a whole contrary evidence mdashsuch as that purportedly contained in the industry Glove UseSurvey and Glove Permeability Study mdash should be given careful

attention The industry studies however did not weaken the EPAfinding of potential exposure Indeed the Glove Use Survey lentsignificant support to that finding

[84] First of all EPA convincingly discredited the Glove UseSurvey to the extent that it purported to show widespread use ofgloves As EPA pointed out in the rulemaking record the surveyentailed no on-site visits or inspections to facilities producingor using EHA Final Test Rule 51 FedReg at 40320 Theindependent researchers conducting the survey did not in any wayverify the responses of company managers Id Nor did theresearchers audit glove use themselves mdash a noteworthy weakness ofthe survey in view of the 20 percent of respondent employers whoindicated that they did not monitor glove use either JA 39The researchers obtained no exposure information from workerswhose incentives to detail non-use would presumably be greaterthan those of management Final Test Rule 51 FedReg at40320 The candor of responses was also diluted by tellingsurvey participants that their responses would be used as thebasis for comments on a proposed EPA test rule JA 434[fn19]

[85] In addition to its skepticism of the Glove Use Survey EPAdiscredited the Glove Permeability Study to the extent that itpurported to show the effectiveness of gloves in preventingexposure to EHA In the study gloves were immersed for sevenhours continuously with no resultant breakthrough of EHA JA68 EPA contends it is possible that an employee would wear apair of EHA-doused gloves for longer than seven hours Final TestRule 51 FedReg at 40320 Indeed the testing standards do notindicate a seven-hour time limit but indicate rather thatimmersion should continue until breakthrough occurs JA 448Additionally continuous immersion testing does not address thepossibility of intermittent exposure over a period of timeespecially with reused gloves JA 320-21 Thirdly theimmersion test evaluated only two types of gloves while workersin fact use five different types of gloves Compare JA 68(nitrile and neoprene gloves tested) with JA 36 (neoprenerubber polyvinyl chloride latex and leatherpalmed cotton glovesused) Final Test Rule 51 FedReg at 40320 Finally the studytested only the permeation of pure EHA while EHA is oftenpresented in mixtures with mineral spirits during processingoperations Id Permeation of some substances in mixtures hasbeen found greater than in pure form JA 452 455 Thus theGlove Permeability Study did not succeed in effectively rebuttingEPAs inference of potential exposure so as to dissipate anyserious concerns about the likelihood of exposure

[86] More importantly however the Glove Use Survey lent positiveevidentiary support to the Agencys exposure finding Surveyrespondents stated that dermal contact with EHA is possible inseveral situations including the emptying and filling of drumssampling and maintenance JA 36 They indicated the use offive different types of gloves JA 36 only one of which wasincluded in the Glove Permeability Study JA 68(Leather-palmed cotton gloves JA 36 could be expected todiffer substantially from the neoprene and nitrile gloves used inthe test) Survey respondents indicated the use of half a dozen

different procedures for ensuring the use of gloves JA 37Because no measurements were made of actual glove use there isno basis for determining the extent to which any of these methodssucceeded Since only 15 percent of respondents used all thelisted methods JA 37 there exists a real possibility thatsome companies do not use the most effective method orcombination of methods for ensuring glove use Five percent ofthe respondents reported that gloves are not routinely wornduring equipment maintenance JA 38 and 20 percent indicatedthey do not monitor glove use JA 39 Eight percent of therespondents do not even instruct their employees to wear glovesJA 36 Most importantly fully 25 percent of the respondentsreported that there are `occasional situations in which dermalcontact occurs JA 39 (emphasis added) Company managersspecifically mentioned small spills accidental leaks andminor splashes that occur when employees do not wearprotective gloves provided JA 39[fn20] Cumulatively thesedata are sufficient to support the conclusion that there is amore-than-theoretical probability of human exposure to EHA Whereone fourth of the companies making or using EHA report occasionaldermal contact the likelihood of exposure cannot becharacterized as speculative or theoretical Cf AusimontUSA Inc v EPA 838 F2d 93 97 (3d Cir 1988) (evidence ofexposure sufficient where not all workers exposed to fugitivegases wear respirators routinely) In sum EPA has pointed tosufficient evidence of dermal exposure to warrant thepromulgation of a test rule

[87] B Evidence of the Toxicity of EHA

[88] The Agency bears the burden of demonstrating that there is amore-than-theoretical basis for determining that EHA is toxicenough at the level of exposure suspected to present anunreasonable risk of injury to health Two purported healthhazards underlay the rule challenged in this case developmentaltoxicity and subchronic toxicity[fn21] EPA predicts theoccurrence of exposure episodes involving up to 60 milligrams ofEHA per kilogram of body weight for each incident After athorough review of the record we uphold the Agencysdetermination that a more-than-theoretical basis exists forsuspecting toxicity at that level of exposure

[89] 1 Evidence of Developmental Toxicity

[90] EPA based its finding as to developmental toxicity on severalfactors none of which has been rebutted successfully by CMAFirst one study (the Ritter Study) reported that EHA caused anincreased incidence of fetal malformations JA 233-36 Another(the Hazelton Laboratories Study) illustrated adverse effectson fetuses of laboratory animals Proposed Test Rule 50 FedRegat 20682 It is true that EPA described the Ritter Study as notbeing state-of-the-art and thus inappropriate for assessinghuman risk Final Test Rule 51 FedReg at 40320 The Agencynoted flaws with the Hazelton Laboratories Study Id at40321 EPA noted at the same time however that the studiesflawed as they concededly were raise[d] sufficient concernabout developmental toxicity to warrant testing Id at 40320

In a statutory structure that calls for a testing rule preciselywhere there are insufficient data and experience on which tobase a conclusion as to toxicity 15 USC sect 2603(a)(1)(A)(ii)the existence of such studies surely provides the basis fordemanding testing in order to assess risk conclusively CfAusimont 838 F2d at 97 ([i]f no doubt existed testing wouldnot be required)

[91] Secondly the agency found that EHA is structurally similar tovalproic acid which has been shown to cause harm to thedeveloping embryo Proposed Test Rule 50 FedReg at 20681-82CMA describes the EPA inference of toxicity drawn from thesefacts as speculative and debatable CMA Br 48 But Congressexplicitly contemplated that EPA would base test rules oncomparisons among structurally similar chemicals See HRRepNo 1341 supra at 17 It is true that CMA criticized thestudies done on valproic acid JA 105-08 But EPA respondedwith its own analysis of those points JA 307-11 Again whileit recognized that existing studies were inconclusive EPAmaintained that they were sufficient to raise enough concern tomerit conducting new studies that would be more adequate JA308 Cf Ausimont 838 F2d at 96 (characterizing similararguments by industry as highlight[ing] the need for testing)

[92] The statutory standard requires EPA to correlate the suspectedtoxicity of a substance with the suspected level of exposure Wefind that the Agency had a more-than-theoretical basis forsuspecting that the toxicity of EHA even in rare or single dosesof 60 milligrams per kilogram of body weight sufficed to presentan unreasonable risk of injury to health

[93] Sufficient evidence existed as to single-dose toxicity TheRitter Study found evidence of developmental toxicity in testanimals after single doses of EHA Final Test Rule 51 FedRegat 40320 Another study (the Nau Study) showed that chemicalsstructurally similar to EHA cause fetal effects in test animalsin single doses Id at 40321 JA 291 292 Single-dosetoxicity is considered by EPA to be a possibility withdevelopmental toxicants generally Guidelines for the HealthAssessment of Suspect Developmental Toxicants 51 FedReg34028 34037 (1986)

[94] The evidence moreover provided sufficient grounds for concernas to single doses in the suspected amount of real-worldexposure 60 milligrams of EHA per kilogram of body weight TheRitter and Nau Studies involved exposing test animals tosubstance amounts 10 to 30 times those expected to occur to EHAworkers[fn22] These large-dose exposures caused high incidenceof fetal toxicity in one study a non-exposed control groupexperienced 44 percent incidence of fetal toxicity while exposedanimals experienced 711 percent incidence of fetal toxicityFinal Test Rule 51 FedReg at 40320 While conceding thatlarge-dose testing does not necessarily provide preciseinformation for predicting toxicity at small dosage levels EPApointed out that a no-observed-effect level for EHA has neverbeen found Id at 40321 EPA reasonably maintainedtherefore that the high-dosage studies provided a rational basis

for requiring tests to study the effects of EHA at lower dosagesIndeed as our examination of the legislative history made clearsection 4 of TSCA authorizes issuance of test rules in theabsence of adequate information to reasonably predict healtheffects Were we to deem the high-dosage studies an insufficientbasis for a test rule EPA would be unable to react to thepotential dangers of a chemical substance until either the Agencyconducted government-funded studies at the expected real-worlddosage level or private researchers fortuitously happened topublish such studies The aim of TSCA was to make producers andusers of chemical substances assume the costs of testing so longas a more-than-theoretical basis existed for finding the presenceof an unreasonable risk of injury to health We are persuadedthat EPAs finding of potential developmental toxicity issupported by substantial evidence

[95] 2 Evidence of Subchronic Toxicity

[96] EPA bases its finding on subchronic toxicity on two studies(Moody I and Moody II) Moody I reported that EHA and similarsubstances adversely affect the livers of test animals JA401-08 Moody II reported that EHA and similar chemicals causeincreases in liver size increases in liver enzyme alterationsand decreases in blood serum fat levels Final Test Rule 51FedReg at 40321 EPA credibly rejected CMAs contention thatthese reactions constituted simple adaptation in the organismId The Agency based its conclusion on the magnitude of thechanges and the short (three-week) duration of the study IdFurther support for a finding of potential subchronic toxicitylies in the fact that valproic acid (as noted above a substancestructurally analogous to EHA) is toxic to the liver of bothanimals and humans Id Thus EPA marshaled substantialevidence of subchronic toxicity which was not appreciablyweakened by CMAs attempts at rebuttal

[97] CONCLUSION

[98] We find this petition for review presents a justiciablenon-moot case We uphold EPAs interpretation of section 4 ofTSCA as authorizing issuance of a test rule where there is amore-than-theoretical basis to suspect the presence ofunreasonable risk of injury to health Exposure may beestablished by means of inferences and exposure to a particularsubject need not be recurrent to warrant testing Finally wefind that EPA has presented substantial relevant evidence ofexposure and toxicity so as to justify the promulgation of a testrule and has carefully and credibly answered the criticisms ofindustry EPAs findings as to exposure subchronic toxicity anddevelopmental toxicity are supported by substantial evidence onthe record viewed as a whole The petition for review istherefore

[99] DENIEDFootnotes omitted