Environmental Defense Fund, Inc. v. Environmental Protection Agency and Russell E. Train, Administrator, Velsicol Chemical Corporation and Pineapple Growers Association of Hawaii, Intervenors. Earl L. Butz, Secretary of Agriculture of the United States v. Russell E. Train, Administrator of the Environmental Protection Agency, Velsicol Chemical Corporation, Intervenor. Velsicol Chemical Corporation v. Environmental Protection Agency, Russell E. Train, Administrator, Velsicol Chemical Corporation v. Russell E. Train, Administrator, Environmental Protection Agency

This case involves the pesticides heptachlor and chlordane. Consolidated petitions seek review of an order of the Environmental Protection Agency (EPA) suspending the registration of those pesticides under the Federal Insecticide, Fungicide and Rodenticide Act (FIFRA)1 for certain uses. The Administrator of EPA issued an order on December 24, 1975. The order prohibited further production of these pesticides for the suspended uses, but permitted the pesticides' continued production and sale for limited minor uses. Even as to the suspended uses, the Order tempered its impact in certain respects: It delayed until August 1, 1976, the effective date of the prohibition of production for use on corn pests; and it permitted the continued sale and use of existing stocks of registered products formulated prior to July 29, 1975.

One petition to review was filed by Earl L. Butz, Secretary of Agriculture of the United States (U.S.D.A.). Secretary Butz and intervenor Velsicol Chemical Corporation, the sole manufacturer of heptachlor and chlordane, urge that the EPA order as to chlordane be set aside on both substantive and procedural grounds.2 They contend that substantial evidence does not support the Administrator's conclusion that continued use of chlordane poses an "imminent hazard"3 to human health, and that the Administrator made critical errors in assessing the burden of proof and in weighing the benefits against the risks of continued use of chlordane.

The other petition, filed by Environmental Defense Fund, urges that the Order did not go far enough to protect against the hazards of heptachlor and chlordane use. EDF sought an injunction against the provisions permitting continued production and use of the pesticides on corn pests until August 1, 1976. EDF also challenges the Administrator's decision to allow continued use of the stocks of the two pesticides existing as of July 29, 1975,4 contending that EPA should have provided for retrieval and controlled disposal of such stocks. EDF also contends that the Administrator erred in failing to suspend certain "minor uses" of chlordane and heptachlor.5

On the issue of retrieval of existing stocks, we remand for further consideration. In all other respects, we affirm the Administrator's Order. In view of our conclusion, we denied EDF's request for a stay pending appeal of the provision delaying the effective date for use on corn pests.6 In effect, we approved the delay of effective date (as to corn) until August 1, 1976, and in this opinion we set forth our reasons for that conclusion. (Part II, B, 1)

I. STATUTORY FRAMEWORK AND STANDARD OF REVIEW

The issues posed by administrative action pursuant to FIFRA are not new to this court,7 and we have previously extensively described the statutory framework for such actions. What is involved here is a suspension of registration of two pesticides during the pendency of the more elaborate cancellation of registration proceeding, initiated in this case by a November 18, 1974, notice of intent to cancel. This 1974 notice stated that there existed "substantial questions of safety amounting to an unreasonable risk to man and the environment" from continued use of heptachlor and chlordane. Public cancellation hearings pursuant to that notice were not expected to commence for some time.8 On July 29, 1975, the Administrator issued a Notice of Intent to Suspend the registrations of most uses of the two pesticides. The Administrator then commented on that expected delay in completing the cancellation hearings, and cited "new evidence . . . which confirms and heightens the human cancer hazard posed by these pesticides." On August 4, 1975, registrant Velsicol Chemical Corporation requested an expedited adversary hearing on the suspension question pursuant to § 6 of FIFRA, 7 U.S.C. § 136d(c). Administrative Law Judge Herbert L. Perlman presided over the cancellation hearings beginning August 12. Evidence was limited to human health issues and the benefits of continued use of heptachlor and chlordane. The record was closed December 4, 1975, and on December 12, the ALJ recommended against suspension, stating that he was unable to find that "heptachlor and chlordane are conclusively carcinogens in laboratory animals."9 The Administrator reversed that decision on December 24, 1975, and suspended most uses of chlordane and heptachlor.

The Administrator is authorized to suspend the registration of a pesticide where he determines that an "imminent hazard" is posed by continued use during the time required for cancellation. Section 6(c) of FIFRA, 7 U.S.C. § 136d(c)(1). An "imminent hazard" exists where continued use during the time required for the cancellation proceeding would be likely to result in "unreasonable adverse effects on the environment." Section 2(l ) of FIFRA, 7 U.S.C. § 136(l ). The term "unreasonable adverse effects on the environment" is, in turn, defined as "any unreasonable risk to man or the environment, taking into account the economic, social, and environmental costs and benefits of the use of any pesticide." Section 2(bb) of FIFRA, 7 U.S.C. § 136(bb).

As in our previous suspension case involving aldrin/dieldrin,10 the primary challenge raised by Velsicol and USDA goes to the adequacy of the evidentiary basis of EPA's finding that the suspended pesticides present an imminent hazard during the time required for cancellation. The standard against which we test that challenge is defined in Section 16(b) of FIFRA:

The court shall consider all evidence of record. The order of the Administrator shall be sustained if it is supported by substantial evidence when considered on the record as a whole.

The standard of substantial evidence has been defined as:11

something less than the weight of the evidence . . . (T)he possibility of drawing two inconsistent conclusions from the evidence does not prevent an administrative agency's finding from being supported by substantial evidence.

In applying this principle of review12 in the specific context of a suspension of pesticides, this court has reiterated that "the function of the suspension decision is to make a preliminary assessment of evidence, and probabilities, not an ultimate resolution of difficult issues. We cannot accept the proposition . . . that the Administrator's findings . . . (are) insufficient because controverted by respectable scientific authority. It (is) enough at this stage that the administrative record contain(s) respectable scientific authority supporting the Administrator." Environmental Defense Fund v. EPA (Shell Chemical Co., et al.), 167 U.S.App.D.C. 71, 77, 510 F.2d 1292, 1298 (1975) quoting Environmental Defense Fund v. EPA, 150 U.S.App.D.C. 348, 357, 465 F.2d 528, 537 (1972).

These decisions of our court also point out that the Administrator is not required to establish that the product is unsafe in order to suspend registration, since FIFRA places "(t)he burden of establishing the safety of a product requisite for compliance with the labeling requirements . . . at all times on the applicant and registrant."13 Velsicol and USDA urge that this allocation of burden of proof relied on by the Administrator is inconsistent with the explicit terms of FIFRA. They rely on FIFRA's specific incorporation of subchapter II of the Administrative Procedure Act, which provides in relevant part that "Except as otherwise provided by statute, the proponent of a rule or order shall have the burden of proof." 5 U.S.C. § 556(d).

The EPA regulation governing the burden of proof in suspension proceedings provides:

At the hearing, the proponent of suspension shall have the burden of going forward to present an affirmative case for the suspension. However, the ultimate burden of persuasion shall rest with the proponent of the registration. 40 C.F.R. § 164.121(g) (1975).

Assuming that the Administrator is the "proponent" of a suspension order and is governed by § 556(d), the legislative history of that provision indicates that it allocates the burden of going forward rather than the burden of ultimate persuasion and is consistent with the EPA's apportionment of burden:

That the proponent of a rule or order has the burden of proof means not only that the party initiating the proceeding has the general burden of coming forward with a prima facie case but that other parties, who are proponents of some different result, also for that purpose have a burden to maintain. S.Doc.No.248, 79th Cong., 2d Sess. 208, 270 (1946).

This allocation of the burden of going forward14 structures evaluation of the factual evidence adduced for both the Administrator and the reviewing court, and is consistent with the traditional approach that this burden normally falls on the party having knowledge of the facts involved.

In urging that the ultimate burden of proof in a suspension proceeding rests on the Administrator, Velsicol and USDA assert that the suspension decision is a drastic step differing fundamentally from both the registration and cancellation decisions made under FIFRA. But we have already cautioned that the "imminent hazard" requisite for suspension is not limited to a concept of crisis: "It is enough if there is substantial likelihood that serious harm will be experienced during the year or two required in any realistic projection of the administrative process."15 "FIFRA confers broad discretion" on the Administrator to find facts and "to set policy in the public interest." Wellford v. Ruckelshaus, 142 U.S.App.D.C. 88, 91, 439 F.2d 598, 601 (1971). This broad discretion was conferred on the implicit assumption that interim action may be necessary to protect against the risk of harm to the environment and public health while a fuller factual record is developed in the cancellation proceeding. This avenue of protective relief would be effectively foreclosed if we accepted Velsicol's argument that the Administrator must prove imminent hazard16 apparently in some sense of weight of the evidence, going beyond substantial likelihood. But as we have already pointed out, the basic statutory directive requires affirmance of the Administrator's decision if supported by substantial evidence, and this requires "something less than the weight of the evidence."17 We reject that renewed invitation to exercise increased substantive control over the agency decision process, and turn to a consideration of whether the Administrator's decision to suspend most uses of heptachlor and chlordane is supported by substantial evidence.

II. SUBSTANTIAL EVIDENCE SUPPORT FOR THE ADMINISTRATOR'S DECISION

To evaluate whether use of a pesticide poses an "unreasonable risk to man or the environment," the Administrator engages in a cost-benefit analysis that takes "into account the economic, social, and environmental costs and benefits of the use of any pesticide." 7 U.S.C. § 136(bb). We have previously recognized that in the "preliminary assessment of probabilities" involved in a suspension proceeding, "it is not necessary to have evidence on . . . a specific use or area in order to be able to conclude on the basis of substantial evidence that the use of (a pesticide) in general is hazardous." EDF v. EPA, 160 U.S.App.D.C. at 130, 489 F.2d at 1254, quoted in EDF v. EPA (Shell Chemical Co.), 167 U.S.App.D.C. at 80, 510 F.2d at 1301. "Reliance on general data, consideration of laboratory experiments on animals, etc." has been held a sufficient basis for an order cancelling or suspending the registration of a pesticide. Id. Once risk is shown, the responsibility to demonstrate that the benefits outweigh the risks is upon the proponents of continued registration. Conversely, the statute places a "heavy burden"18 of explanation on an Administrator who decides to permit the continued use of a chemical known to produce cancer in experimental animals. Applying these principles to the evidence adduced in this case, we conclude that the Administrator's decision to suspend most uses of heptachlor and chlordane and not to suspend others is supported by substantial evidence and is a rational exercise of his authority under FIFRA.A. Risk Analysis Carcinogenicity of Heptachlor and Chlordane

Velsicol and USDA contend that the laboratory tests on mice and rats do not "conclusively" demonstrate that chlordane is carcinogenic to those animals; that mice are too prone to tumors to be used in carcinogenicity testing in any case; and that human exposure to chlordane is insufficient to create a cancer risk. They place strong reliance on the Administrative Law Judge's refusal to recommend suspension because he was "hesitantly unwilling at this time to find that heptachlor and chlordane are conclusively carcinogens in laboratory animals." Rec.Dec. at 86 (emphasis in original). The ALJ recognized however, that on the basis of the record made the Administrator "could determine that the pesticides involved pose potential or possible carcinogenic risk to man" and that he could "find that heptachlor and chlordane are conclusively carcinogenic in laboratory animals."19 While adopting the ALJ's factual findings, the Administrator concluded that the ALJ had applied an erroneous legal standard in requiring a conclusive rather than probable showing that the pesticides were animal carcinogens, and concluded in any case that the evidence showed heptachlor and chlordane to be animal carcinogens. We affirm.

An ultimate finding in a suspension proceeding that continued use of challenged pesticides poses a "substantial likelihood of serious harm" must be supported by substantial, but not conclusive, evidence. In evaluating laboratory animal studies on heptachlor and chlordane there was sufficient "respectable scientific authority" upon which the Administrator could rely in determining that heptachlor and chlordane were carcinogenic in laboratory animals.

We start by rejecting Velsicol's argument that the "cancer principles" EPA relied on in structuring its analysis of the mice and rat studies improperly biased the agency's open-minded consideration of the evidence.20 In brief form, the principles accept the use of animal test data to evaluate human cancer risks; consider a positive oncogenic effect in test animals as sufficient to characterize a pesticide as posing a cancer risk to man; recognize that negative results may be explained by the limited number and sensitivity of the test animals as compared to the general human population; note that there is no scientific basis for establishing a no-effect level for carcinogens; and view the finding of benign and malignant tumors as equally significant in determining cancer hazard to man given the increasing evidence that many "benign" tumors can develop into cancers. The Agency's reliance on these principles did not come as a surprise to Velsicol; they were included in the Administrator's Notice of Intent to Suspend; and as recognized in EDF v. EPA, 167 U.S.App.D.C. at 77-78, 510 F.2d at 1298-99, form part of the Agency's "scientific expertise." Velsicol was properly given an opportunity to put in evidence contesting those principles, but failed to demonstrate anything more than some scientific disagreement with respect to them. Velsicol's principal complaint that mice are inappropriate test animals was specifically rejected by the Administrator, citing statements by the National Academy of Sciences' Food Protection Committee, the World Health Organization, HEW's Commission on Pesticides and their Relationship Environmental Health, FDA Advisory Panel on Carcinogenesis, International Agency for Research on Cancer, and Director of the National Cancer Institute's Carcinogenesis Program.21 Unlike the failure to adduce critical methodology that we criticized in International Harvester,22 EPA's specific enunciation of its underlying analytic principles, derived from its experience in the area, yields meaningful notice and dialogue, enhances the administrative process and furthers reasoned agency decisionmaking.

The animal experiments the Administrator relied on tested heptachlor, its chief metabolite heptachlor epoxide and chlordane. Technical or commercial heptachlor and chlordane both contain the substances tested.23 Five studies involved mice and one involved rats. Velsicol urges that only the chlordane studies can be relied on to support a finding of carcinogenicity for chlordane, and more broadly, that there was insufficient objective evidence of carcinogenicity in laboratory animals to find a cancer risk to man. The objective evidence Velsicol charges is lacking is evidence that the tumors induced during the feeding studies had properties of invasion (spreading of a tumor into adjacent tissues) and metastasis (spreading of malignant cells into non-adjacent tissues) characteristics two of their witnesses considered essential definitional elements of "cancer." By contrast, the other pathologists consulted during the suspension proceeding generally believe that cancer can be reliably diagnosed by observation of tissue cells under a microscope (histopathological evidence) without evidence, or very much evidence, of invasion or metastasis. The Administrator concluded, on what we view as respectable scientific authority, that evidence of invasion and metastasis is not essential to the diagnosis of cancer in mice, and also found that metastasis and invasion had been reported in some of the studies where protocols for the study did not preclude such analysis. That a low percentage of metastasis was found in those studies was also explained by the failure to use the more sensitive examination techniques developed since the studies relied on here were carried out.

In reviewing and evaluating the studies relied on by the Administrator, five EPA witnesses and five Velsicol consultants agreed that animals fed chlordane and animals fed heptachlor/heptachlor epoxide in fact underwent cellular changes indicating malignancy. Dr. Reuber reevaluated the FDA C3H Mouse Study of 1965 and found "a statistically significant increase in carcinomas" in the mice treated with heptachlor/heptachlor epoxide. The ALJ concluded that no substantial refutation of his results was made.24 The 1959 Kettering CFN Rat Study tested heptachlor epoxide; five pathologists found that the original evaluation had underdiagnosed liver cancer, the biological significance of which was found by Judge Perlman to be "especially significant because of the rarity of this finding."25 Histological review of the IRDC CD-1 Mouse Study, using a hardy non-inbred strain of mice with low incidence of spontaneous tumors, resulted in the finding of "a highly statistically significant incidence of hepatocellular carcinomas" for those mice fed heptachlor/heptachlor epoxide,26 and basically the same finding for mice fed chlordane.27 After considering design problems of control group and dosages, the Administrator also relied on the preliminary results of National Cancer Institute B6C3F1 Mouse Studies of Heptachlor and Chlordane, showing a highly statistically significant increase in carcinomas in a strain of mice with a low spontaneous incidence of tumors.28 The Administrator has adequately explained his reliance on these test results which show significant carcinoma development in treated animals. None of the tests yielded negative results; chlordane was shown to be independently carcinogenic, as well as to contain a carcinogenic component (heptachlor/heptachlor epoxide). We think it plain that the foregoing establishes substantial evidence supporting the Administrator's result, and that Velsicol cannot be said to have met its burden of overcoming EPA's prima facie case by showing that chlordane and heptachlor are not carcinogenic in laboratory animals.

Human epidemiology studies so far attempted on chlordane and heptachlor gave no basis for concluding that the two pesticides are safe with respect to the issue of cancer.29 To conclude that they pose a carcinogenic risk to humans on the basis of such a finding of risk to laboratory animals, the Administrator must show a causal connection between the uses of the pesticides challenged and resultant exposure of humans to those pesticides. He made that link by showing that widespread residues of heptachlor and chlordane are present in the human diet and in human tissues. Their widespread occurrence in the envir- onment30 and accumulation in the food chain31 is explained by their chemical properties of persistence, mobility and high solubility in lipids (the fats contained in all organic substances). Residues of chlordane and heptachlor remain in soils and in air and acquatic ecosystems for long periods of time.32 They are readily transported by means of vaporization, aerial drift, and runoff of eroding soil particles.33 The residues have been consistently found in meat, fish, poultry and dairy products monitored in the FDA Market Basket Survey and are also frequent in components of animal feeds. This evidence supports a finding that a major route of human exposure is ingestion of contaminated foodstuffs.34 EPA's National Human Monitoring Survey data shows that heptachlor epoxide and oxychlordane, the principal metabolites of heptachlor and chlordane respectively, are present in the adipose tissue of over 90% of the U.S. population.35

The population's exposure to these pesticides, in large part involuntary, can be divided into agricultural and nonagricultural related routes. Seven million pounds of heptachlor and chlordane were used as corn soil insecticide in 1975, producing residues which persist in the soil for several years after application. These residues are taken up by such food, feed, and forage crops as soybeans, barley, oats, and hays typically rotated with corn.36 By volatilization the pesticides contaminate corn and other plant leaves.37 And root crops like potatoes, carrots and beets directly absorb the pesticides from the soil.38 Other sources of agricultural-related residues include exposure to contaminated dust particles and agricultural runoff containing eroded soil particles.

Velsicol urges that the dietary exposure resulting from agricultural uses of the pesticides is insignificant, and that current exposure is well below "safe" dose levels as calculated by the Mantel-Bryan formula, or by the World Health Organization's Acceptable Daily Intake figures. Mantel himself criticized the use of the formula for a persistent pesticide, and the Administrator rejected the concept of a "safe" dose level defined by mathematical modeling because of "the incomplete assumptions made by the registrant's witnesses about the sources of human exposure in the environment, the natural variation in human susceptibility to cancer, the lack of any evidence relating the level of human susceptibility to cancer from heptachlor and chlordane as opposed to that of the mouse, and the absence of precise knowledge as to the minimum exposure to a carcinogen necessary to cause cancer."39 That explanation is within the reasonable bounds of the agency's expertise in evaluating evidence. And it is confirmed by the common sense recognition that reliance on average "safe" dietary levels fail to protect people with dietary patterns based on high proportional consumption of residue-contaminated foods (e.g., children who ingest greater quantities of milk than the general population).40

There are several non-agricultural uses which involve a large volume of heptachlor and chlordane as well as significant human exposures. For example, the record shows that approximately six million pounds of chlordane are used annually on home lawns and gardens. The Administrator found that these uses involve high risks of human intake "due to the many avenues which exist for direct exposure, through improper handling and misuse, inhalation, and absorption through the skin from direct contact."41 Velsicol asserts that the mice studies showing carcinogenic effects after ingestion of chlordane do not warrant an inference about the carcinogenic effects of inhaling it or absorbing it through the skin, and that consequently nonagricultural routes of exposure cannot be considered to present a cancer risk. They rely on Reserve Mining Co v. EPA, 514 F.2d 492 (8th Cir. 1975) (en banc). That reliance is misplaced. In that case, the court was concerned with the propriety of the district court's granting the immediate relief of shutting down a plant discharging asbestos fibers into the City of Duluth's drinking water source. It instead ordered cessation of dumping within a "reasonable time" pursuant to the unstructured equitable discretion given the court under the Federal Water Pollution Control Act, even though it had concluded that continued discharge posed a hazard to health. By contrast, the FIFRA statutory scheme mandates explicit relief the suspension of registration when an unreasonable risk to health is made out. We have previously held that it is not necessary to have evidence on a specific use to be able to conclude that the use of a pesticide in general is hazardous. Once the initial showing of hazard is made for one mode of exposure in a suspension proceeding, and the pesticide is shown to be present in human tissues, the burden shifts to the registrant to rebut the inference that other modes of exposure may also pose a carcinogenic hazard for humans.42 Velsicol has totally failed to meet that burden here. Although it was put on notice in the Notice to Suspend of EPA's intent to rely on direct inhalation and dermal exposure as reasons to suspend household lawn and turf uses of chlordane, it failed to offer even a medical theory as to why the significant inhalation or dermal exposure associated with such uses would not pose a carcinogenic threat. In view of the general failure to understand the mechanics of carcinogenicity, the lack of hypothetical explanation may be based on Velsicol's own data that exposure to vapors of chlordane and heptachlor in the work place, leads (as dietary exposure leads) to storage of oxychlordane, heptachlor epoxide, and other components in the fat tissue, and to circulation of these compounds in the blood, with consequent exposure to other organs in the body.43 Nor did Velsicol focus on the individual user's intense inhalation exposure associated with lawn and turf uses in its response to the point made in the EPA Staff's exceptions to the ALJ recommended decision, that the evidence showed that an individual using these chemicals for lawn and turf applications is subjected to a marked intensity of inhalation. Instead Velsicol attacked as inconsistent with the minimal amounts of chlordane and heptachlor normally found in ambient air, the EPA Staff's proposed reliance on inhalation as a major route of human exposure for the general population. However, the Administrator did not proceed on this basis. And if Velsicol hypothesized that chlordane residues are safe so long as they reach the tissue only through inhalation (even intense inhalation) it should have presented witnesses expressing that hypothesis. Instead they argue, in general and procedural terms, that the evidence presented by the Administrator was not sufficient to meet his full burden, and this in our view seeks to impose a broader burden on the Administrator than is appropriate in a suspension proceeding.

Velsicol and USDA challenge the Administrator's finding that the benefits derived from the suspended uses of chlordane do not outweigh the harms done. EDF urges that the Administrator's decision to continue some uses was not justified by evidence that the risk of harm was outweighed by benefits from the continued uses.

Heptachlor and chlordane were used on an estimated 3.5% of the total corn acreage in the United States in 1975, largely in an effort to control black cutworm. Cutworms sporadically infest 2 to 8% of total U.S. corn farms, and occur most often in lowland, river bottom areas. Chlordane and heptachlor are used as preplant treatments to insure against possible infestations. The Administrator found, with record support, that no macro-economic impact will occur as a result of suspending those pesticides. He also found that crop surveillance or "scouting" for infestations during the early weeks of plant growth, together with application of post-emergence baits or sprays where necessary, provide an effective alternative to the more indiscriminate prophylactic use of chlordane and heptachlor. Velsicol urges that this approach is not as effective as the persistent protection provided by chlordane. Especially in the absence of proof of a serious threat to the nation's corn, there is no requirement that a pesticide can be suspended only if alternatives to its use are absolutely equivalent in effectiveness. The Administrator reasonably took into account that a transition period would be necessary to implement post-emergent techniques of control and concluded that the challenged pesticides could continue in use for corn protection until August 1, 1976.44 This evaluation of alternatives and the time required to implement them is supported by substantial evidence, and we find no basis to disturb the Administrator's balancing of costs and benefits.

The Administrator suspended a number of agricultural uses where the record was insufficient to support any finding that benefits outweigh costs of continued use of heptachlor or chlordane on these crops.45 Possibly the lack of benefits evidence reflected readily available alternatives, possibly a relative lack of interest in lesser-volume uses. In any event, the registrant's failure to carry its burden of adducing sufficient evidence on benefits in effect leaves the Administrator nothing to weigh in his cost-benefit analysis except the evidence that the use of the challenged pesticides in general is hazardous. That evidence of general hazard is sufficient to support a suspension of uses.

3. Non-Agricultural Uses Suspended by the Administrator

Chlordane is a common household, lawn, garden, and ornamental turf insecticide, with over 7.5 million pounds (36% of total use) so employed in 1974. The ALJ and Administrator found on the basis of substantial evidence that the "efficaciousness of the substitutes for control of household and lawn insects is not really at issue"46 and that when lack of evidence of substantial benefits from continued use is weighed against the special hazards of exposure presented by the possibilities of inhalation, dermal absorption, and the increased dangers associated with improper handling, suspension of those uses was justified. Similarly, on the basis of evidence in the record, the Administrator could reasonably find that the residual capacity of chlordane was not necessary to control either structural pests or ticks and chiggers, given the existence of effective alternatives to each of those uses.474. The Administrator's Refusal to Suspend Certain Uses

EDF challenges the Administrator's refusal to suspend use of chlordane or heptachlor on strawberries, for seed treatment, pineapples, the white fringed beetle, Florida citrus, white grubs in Michigan, narcissi bulbs, harvester ants, imported fire ant, Japanese beetle quarantine, and black vine weevil quarantine in Michigan. Following the recommendations of the ALJ, the Administrator found that for each use the benefits outweighed the risks for the limited time under consideration, effective alternatives were generally not available, and that the exposure risk arising from the use was minimal. EDF counters that the total exposure resulting from these "minor" uses is in fact significant, and that the Administrator continued these uses whenever a "colorable" case of benefits had been made out.

Once the Administrator has found that a risk inheres in the use of a pesticide, he has an obligation to explain how the benefits of continued use outweigh that risk. We are satisfied that he has met that obligation here, and that substantial evidence supports his decision. We note, however, that we come to this conclusion in the context of a suspension proceeding where perforce the Administrator is engaged in making a "preliminary assessment" of the evidence; a more careful exploration of economic impact and available alternatives would be required to support continued registration in a cancellation proceeding.

C. Continued Sale and Use of Existing Stocks of Chlordane and Heptachlor for Suspended Uses

Although we have no doubt that the Administrator has the power under FIFRA to exempt from a suspension order the use of existing stocks (in this case stocks existing as of July 29, 1975), the Administrator acted arbitrarily when he failed to even inquire into the amount of stocks left, and the problem of returning and disposing of them. Some evidence must be adduced before an exemption decision is made, and it is the responsibility of the registrant to provide it.48 It may be that the lapse of time has lessened the current significance of this issue but we are in no position to do other than remand for further consideration.

We affirm the Agency's suspension order of December 24, 1975, as clarified by the order of January 19, 1976, except for the exemption of the sale and use of existing stocks. The record is remanded for further consideration of that issue.

So ordered.

Supplemental Opinion on Petition for Rehearing

On consideration of the petition of Velsicol Chemical Corporation for rehearing it is

ORDERED by the Court that the aforesaid petition for rehearing is denied.

Supplemental Opinion for the Court filed by Circuit Judge LEVENTHAL.

Velsicol argues in its petition for rehearing that in upholding the Administrator's allocation of the burden of proof to the registrant this Court misinterpreted § 7(c) of the Administrative Procedure Act. Velsicol points out that the interpretation adopted by this Court was not advanced by any of the parties and that, in fact, the problem of interpreting the APA as opposed to section 6(c)(2) of FIFRA was not briefed or argued in this proceeding.

In light of these contentions, we have reexamined our analysis of the APA and FIFRA. After further study, we are still of the view that the Administrative Procedure Act does not preclude the assignment of the burden of persuasion to the opponent of an order. So far as this case is concerned, however, Velsicol's position is undermined by still another doctrine, that in any event the controlling statute is FIFRA. We hold that in light of the legislative history of FIFRA, and the numerous cases holding that its 1964 amendment was specifically intended to shift the burden of proof from the Secretary to the registrant, this case is one where the location of the burden of proof is, in the language of the APA, "otherwise provided by statute." 5 U.S.C. § 556(d).

* As to the APA, our opinion holds that the "burden of proof" it casts upon the "proponent" is the burden of coming forward with proof, and not the ultimate burden of persuasion.

Velsicol claims that the precedents of this Circuit1 and case law in other circuits2 establish that under the APA the proponent of an order bears the burden of persuasion.

We have examined the cited cases and find no support for that view. Time does not permit discussion of each case. We shall refer to a few as illustrations.

Velsicol calls attention to National Airlines, Inc. v. Civil Aeronautics Board, 112 U.S.App.D.C. 119, 300 F.2d 711 (1962). There we affirmed an order that held that once a complaining City had demonstrated the inadequacy of existing airline service, the airline bore the burden of proving the economic infeasibility of expanding service. The court explained:

Petitioner argues that the Administrative Procedure Act, § 7(c), 60 Stat. 241 (1946), 5 U.S.C.A. § 1006(c), which places the burden of proof upon the proponent of an order, prohibits the Board from placing the burden of proving economic infeasibility upon the carriers. We think that once the Board has established that existing service is inadequate to meet the needs of a community, it has discharged its burden of proof. Economic infeasibility is more like an affirmative defense than an element of the Board's case. Accordingly we think that the Board properly placed the burden of producing evidence of economic infeasibility upon petitioner.