This case involves the validity of an order issued by the Administrator of the Environmental Protection Agency (EPA) on October 1, 1974, suspending the registration and prohibiting the manufacture and sale1 of the pesticides aldrin and dieldrin.2 EPA permitted the sale and use of existing stocks manufactured prior to August 2, 1974, the date of the issuance of the Notice of Intention to Suspend.

The validity of the suspension of registration is attacked by Shell Chemical Company, the sole United States manufacturer of the pesticides, which raises general questions as to the basis of the order and stresses the importance of these pesticides for the 1975 corn crop. It is also attacked by Florida Citrus Mutual (FCM), an association of citrus growers, and by the Secretary of Agriculture;3 in addition to adopting the general attack made by Shell, the Secretary stresses the need for the continued registration of aldrin/dieldrin for certain minor uses, including the protection of citrus fruits, onions, strawberries, pineapples, sugar cane, bananas, cranberries, and nursery stock, and use as a seed treatment.

The Environmental Defense Fund (EDF) and the National Audubon Society attack the EPA's decision to permit continued sale and use of existing stocks.

The court has taken into account the need for an expeditious determination, and has, to the extent permitted by its other pressing obligations, expedited the appeal and oral argument, and the issuance of its opinion. While the court has set forth its reasons it has not provided a full elaboration. The court has considered, though it has not spelled out in detail, all the contentions of the various petitioners. It rejects those contentions except that, in the case of the point raised by EDF, the court remands the record for further consideration.

On December 3, 1970, EDF first petitioned EPA for the immediate suspension of aldrin/dieldrin and the initiation of cancellation proceedings for all existing registrations. On March 18, 1971, the Administrator issued notices of intent to cancel, under § 4 of the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), for all products containing the pesticides, on the basis of his finding that 'a substantial question as to the safety' of the chemicals existed.4 Section 4 permits the suspension of registration while a cancellation hearing is pending when 'the Administrator . . . finds that action is necessary to prevent an imminent hazard to the public,' but the Administrator declined to take this further step.5 Registrants objected to the notices of intent to cancel, and requested the appointment of a scientific advisory committee and the commencement of a public hearing.6 When EDF sought review in this court of the refusal to suspend, we remanded for further consideration in light of the Report of the Advisory Committee, which was issued on March 28, 1972. Environmental Defense Fund, Inc. v. Environmental Protection Agency, 150 U.S.App.D.C. 348, 465 F.2d 528 (1972). After considering the Report and further public comments, the EPA issued an order on December 7, 1972, which affirmed its previous decisions to issue a notice of intent to cancel, without interim suspension.

Cancellation hearings began before Chief Administrative Law Judge (ALJ) Perlman on August 7, 1973. Twelve months into the hearings, on August 2, 1974, the Administrator issued a notice of intent to suspend on the ground that evidence developed since December 1972 indicated that the continued use of aldrin/dieldrin presented an 'imminent hazard' to the public. Shell and USDA requested a public hearing on the suspension question. The hearing began before ALJ Perlman on August 14, 1974, and was concluded on September 12, 1974. ALJ Perlman recommended suspension, and, on October 1, 1974, the Administrator suspended the registrations.

We will first develop the general purpose and validity of the order, with a broad overview of its reasoning and the supporting evidence. Then we shall turn to certain particular objections presented by the parties.

Turning first to the broad question of validity raised by cases like this, the court concludes: The EPA's order is a rational exercise of discretion, rather than arbitrary agency action. It is supported by the reasoning of the agency, and by substantial evidence in the record.

The primary challenge raised by Shell, FCM, and the USDA goes to the adequacy of the evidentiary basis of the EPA's finding that aldrin/dieldrin presents 'an imminent hazard (to man) during the time required for cancellation.'

We have cautioned that the term 'imminent hazard' 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.' Environmental Defense Fund, Inc. v. EPA, supra, 150 U.S.App.D.C. at 360, 465 F.2d at 540 (emphasis added.) '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). See also Environmental Defense Fund, Inc. v. EPA, supra, 150 U.S.App.D.C. at 354, 465 F.2d at 534 (1972). It does not require the Administrator to establish that the product is unsafe, but 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.' Environmental Defense Fund, Inc. v. EPA, supra, 150 U.S.App.D.C. at 352, 465 F.2d at 532.

Section 16(b) of FIFRA defines the scope of judicial review of EPA orders made after public hearing:7

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.8

The standard of 'substantial evidence' means

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.9

In applying this principle to review of a suspension decision, this court has said, '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, Inc. v. EPA, supra, 150 U.S.App.D.C. at 357, 465 F.2d at 537.

Although the cancellation hearing encompasses a broad range of issues concerning the effect of aldrin/dieldrin on the environment as well as on human beings,10 the supension hearing was confined to whether the pesticides present a cancer hazard to man.11 The Administrator concluded that aldrin/dieldrin presented an 'imminent hazard' to man on the basis of data indicating that it is carcinogenic in five strains of mice and, as corroboration, indications that 'there is a strong probability that Aldrin-Dieldrin is a carcinogen in rats as well as mice.'12

Shell attacks the Administrator's reliance on mice data on the ground that the inadequacy of present knowledge regarding cancer and the difficulty of extrapolating from mice to men render his decision speculative.

The Administrator's failure to determine a threshold level of exposure to aldrin/dieldrin does not render his determination improper, for he has concluded that the concept of a threshold exposure level has no practical significance where carcinogens are concerned. This is due in part to the irreversibility and long latency period of carcinogens. '(W)here the matter involved is as sensitive and fright-laden as cancer,'13 and the statute places the burden on the registrant to establish the safety of his product, we shall not, assuming a substantial showing of danger, require the Administrator to make impossible proofs. In reviewing administrative actions, courts 'cannot fairly demand the perfect at the expense of the achievable.'14 The Administrator's conclusion is within the scientific expertise of the agency, and is not infected by error of law. Compare Environmental Defense Fund, Inc. v. Ruckelshaus, supra, 142 U.S.App.D.C. at 86, 439 F.2d at 596.

The validity of extrapolation to humans from data derived from tests on animals is also a matter within the agency's expertise. There was testimony before the Administrator to support such extrapolation,15 and this court has acknowledged the significance of test animal data when cancer is involved.16 Use of animal data is particularly appropriate where, as here, accurate epidemiological studies cannot be conducted because the virtually universal contamination of humans by residues of aldrin/dieldrin17 make it impossible to establish an uncontaminated human control group. The long latency period of carcinogens further hinders epidemiological research, and the ethical problems of conducting cancer experiments on human beings are too obvious to require discussion. Although extrapolation of data from mice to men may be quantitatively imprecise, it is sufficient to establish a 'substantial likelihood' that harm will result. Cf. Society of Plastics Industry, Inc. v. OSHA, 509 F.2d 1301, at 1308 (2d Cir., Jan. 31, 1975).

Shell claims that tests based on mouse data are not substantial evidence, because mouse livers are unusually susceptible to cancer. Still, Shell's data--of statistically significant incidence of malignant liver tumors--were in strains of mice that were, as was noted by the Administrator, unusually resistant to such tumors.18 In any event, Shell's objections are outweighed by the substantial evidence supporting EPA's determination that mice are not uniquely susceptible to carcinogens, but are, in fact, good predictors of carcinogenic hazard to man. The Administrator found that rodents are particularly useful experimental animals, in part because of the similarity of their response to carcinogens to the response of man, their short lifespan, and our relatively well-developed understanding of the pathological cevelopment of tumors in mice and rats. Respected research institutions such as the National Cancer Institute have used mice extensively19 because they have found mice to be an accurate predictor of cancer in other species.

Shell stresses EPA's two earlier refusals to suspend the registration of aldrin/dieldrin despite evidence of its carcinogenicity in mice, and attacks the order under review as an unexplained departure fron prior agency policy. To begin, an agency is not required to adhere to a prior policy with iron rigidity; all that the law requires is that it explain the reasons for its modification.20 The doctrine permitting reconsideration has full vitality as to suspension decisions, for here 'the administrative process is a continuing one . . . (that) calls for continuing reexamination at significant junctures.' Environmental Defense Fund, Inc. v. EPA, supra, 150 U.S.App.D.C. at 361, 465 F.2d at 541. The agency's previous determinations are not fixed or permanent policy decisions, but merely earlier stages in an ongoing review and re-evaluation of the evidence.

The EPA's decision makes it clear that what changed here was not EPA's policy but the nature of the evidence. The EPA decision was supported in part by a re-analysis, by several pathologists, of slides prepared in the course of three studies conducted by the Food and Drug Administration (FDA) in the early 1960's.21 Re-evaluation of existing data would be adequate in itself to support the modification of a prior decision,22 but EPA's decision was also based on data that had only recently come to the agency's attention. For instance, improved analytic techniques revealed higher contamination of major food categories by dieldrin than had previously been contemplated.23 Current results fron ongoing projects concerning air sampling and human tissue data were also available. His previous refusals to suspend, on a different record, did not strip the Administrator of the discretion to make a policy judgment that, because of this new information, the unexpected length of the cancellation hearing, and a threat by Shell to commence production of aldrin/dieldrin for the 1975 season, the risk posed by the registration of the pesticides had increased significantly and suspension had become necessary.

The Administrator cited data that he interpreted as indicating a 'strong probability' that aldrin/dieldrin is a carcinogen in rats. The rat data was derived from three tests, two by the FDA and one confirmatory test from Shell's Tunstall laboratories. At least six witnesses reviewing these studies found a carcinogenic effect or a strong probability of one.24

Shell characterizes the Administrator's conclusion as a departure from the ALJ that is insufficiently explained. The difference between the ALJ and the Administrator does not concern evidentiary facts, but is rather a difference in policy concerning the program which the facts warrant--peculiarly a matter for the Administrator to determine. The ALJ was 'hesitantly unwilling at this time to find that dieldrin is conclusively a carcinogen in the rat although there are indications that this is so.'25 The Administrator thought this caution warranted, but, after 'an intensive re-examination of the statistics and testimony,' concluded 'that there is a strong probability that Aldrin-Dieldrin is a carcinogen in rats.'26 Thus, both decision-makers appeared to find substantial, if not conclusive, evidence that aldrin/dieldrin is carcinogenic in the rat. The Administrator did not say that the rat data alone resulted in his suspension order. He relied on the rat data as corroborating the finding of substantial likelihood of serious harm based on the various mice experiments. The Administrator considered what the ALJ had said about rats, and his decision to take the rat data into account, as corroborative of the need for the order recommended by the ALJ, was well within his authority.

C. Causal Connection To Contamination of Man

Shell further challenges the Administrator's finding that an 'imminent hazard' exists on the ground that the Administrator failed to establish a causal connection between the uses of aldrin/dieldrin that Shell defends (primarily implantation in the ground soil at the base of plants) and the ingestion of pesticide residues by humans. Shell claims that human exposure has resulted primarily from spraying, a use that has been discontinued, and points to the lack of aldrin/dieldrin residues in corn grown on treated soil.

The record supports the EPA's finding of 'substantial likelihood' that serious harm will result from the uses defended by Shell.

Treatment of corn soil insects has consistently been the most prominent use of the pesticides since 1955.27 Corn soil usage accounted for as much as 80 percent of aldrin usage in 1971.28 Other suspended uses may have accounted for as much as 6 percent of aldrin sales and approximately half of the dieldrin sales.29 Moreover, the FDA Market Basket Survey has revealed a consistently high, and increasing, incidence of dieldrin residues in meat, fish, poultry, and dairy products.30 The EPA National Human Monitoring Survey shows dieldrin residues in 96.5 percent, 99.5 percent, and 98.2 percent of the human fat samples tested during the years 1970--1972.31

EPA's conclusion that the prohibition of the predominant use would reduce the likelihood of increased exposure is not unreasonable. It is supported by the evidence of record as follows: Aldrin/dieldrin are highly mobile and persistent chemicals that are not lost by dilution in the inorganic components of the environment. The pesticides persist in the soil for several years, where they are absorbed by the roots and transported to the aerial parts of crops, such as soybeans, which are rotated with corn. Many of these products are important feed components for animanls. The pesticide residues are thus incorporated, directly and indirectly, into the milk, meat, poultry, and soy products consumed by humans.

Shell sees inconsistency in EPA's exemption from suspension of the use of soil-implanted aldrin/dieldrin as a termiticide. EPA explains that, when used for this purpose, the pesticide is buried deep beneath the surface of the land, where it remains undisturbed for years. When used for crop protection, however, aldrin/dieldrin is applied to the top few inches of the soil, in lands typically subject to frequent disturbance through plowing and disking.

There is substantial evidence, plainly sufficient to support the suspension order, at least where, as here, the registrant has failed to come forward with proof showing that no causal connection exists. Shell did not even protest the evidence on causal connection in its argument to the ALJ. Nor did it contest the ALJ's finding of a causal relationship in its objections filed with the Administrator. The ALJ's causation findings are the implied assumption of the Administrator's order. His failure to be explicit on the point yields no basis for legal attack, especially in view of the lack of objection on this ground.

Shell, FCM, and USDA claim that the Administrator has failed to show evidence of the existence of an 'imminent hazard' and a causal connection for each suspended use of aldrin/dieldrin. They would place the burden on the agency to bring forth material on each crop and each geographical area touched by the suspension order. But '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.' Environmental Defense Fund, Inc. v. EPA, supra, 160 U.S.App.D.C. at 130, 489 F.2d at 1254. 'Reliance on general data, consideration of laboratory experiments on animals, etc.,' has been found to provide a sufficient basis for an order cancelling the registration of a pesticide. Id. The same principle applies to a suspension proceeding, where the agency makes only a 'preliminary assessment of . . . probabilities.'

Most of the minor uses either share the critical factor of implantation in the soil close to the ground surface,32 which identifies causal connection with human ingestion, or take the form of basal and foliar sprays, which may present an even greater risk to man.33

Shell, FCM and the USDA further challenge the Administrator's finding that the benefits derived from the suspended uses of aldrin/dieldrin do not outweigh the harms done.

The responsibility to demonstrate that the benefits outweigh the risks is upon the proponents of continued registration.34 The statute places

a heavy burden on any administrative officer to explain the basis for his decision to permit the continued use of a chemical known to produce cencer in experimental animals.35

In our 1972 opinion, Environmental Defense Fund, Inc. v. EPA, supra, we said that 'a mere recitation of a pesticide's uses does not suffice as an analysis of benefits' where the EPA has refused to initiate suspension proceedings despite evidence of carcinogenicity and a submission that alternative pest control mechanisms exist. We sought a further 'elucidation of basis' from the agency to ensure that the evidence of harm was indeed outweighed by benefits flowing from the continued use of the pesticide. Where, as in that case, the agency declines to act in the face of evidence of carcinogenicity it bears the burden of justifying its lack of action:

By definition, a substantial question of safety exists when notices of cancellation issue. If there is no offsetting claim of any benefit to the public, then the EPA has the burden of showing that the substantial safety question does not pose an 'imminent hazard' to the public.

150 U.S.App.D.C. at 359, 465 F.2d at 539. In the present case, in contrast, the agency has decided to act, and the burden is on the registrant to establish that continued registration poses no safety threat.

The Administrator's conclusion that alternative methods are sufficiently efficacious in controlling corn pests is supported by data from studies comparing aldrin/dieldrin treatment of black cutworms and wireworms with other techniques,36 as well as by the registration of alternatives for these purposes.

The finding that adequate alternatives will be available for the 1975 planting season is supported by evidence that other chemical pesticides are being produced and nonchemical techniques are available. Alternatives are not available in equal volume, say petitioners. However, the Administrator has determined that no pound-for-pound substitution is necessary because aldrin/dieldrin has been overused in the past as a prophylactic measure and because the threat of corn soil insects is greatly reduced at this time--a conclusion supported by the evidence and one that will not be disturbed by this court. Shell protests that the certain post-emergent treatments impose a much greater work burden on the farmer. When the subject is risk of cancer, convenience may be relevant but it does not weigh heavy in the scales. We can hardly overturn the Administrator's conclusion that the alternatives were adequate on the ground that he did not give compelling weight to convenience.

Testimony that carbamate and organophospate alternatives do not share the persistence, lipidsolubility, and bioconcentration in animal or human tissues characteristic of aldrin/dieldrin--and thus do not pose the same cancer risk--supports the finding that the alternatives are environmentally suitable.

USDA and FCM challenge the Administrator's failure to provide extensive risk-benefit analysis for each crop and each geographical area for which aldrin/dieldrin has been suspended. They fault the order on the grounds that it makes conclusory findings for minor uses without discussing the contrary evidence and that it lacks substantial record evidence to support its rationale. However, the expedited nature of the suspension proceeding imposes limitations on the degree of detail that can be expected from the Administrator's findings at this stage of the administrative process. A more careful exploration of the availability of alternatives for minor uses would be contemplated for the final determination on cancellation vel non, but we cannot demand an extended discussion of the evidence for every use at this emergency, provisional stage.

The record evidence as to the relative risks and benefits of each use is a mixed bag, but it provides substantial support for the Administrator's conclusion. The order cites California experience with alternative pesticides that have proved effective against Fuller's Rose Beetle, the Florida citrus pest controlled by aldrin/dieldrin, and there is also evidence that effective foliar sprays will be available for use should an emergency arise during the suspension period. Moreover, aldrin/dieldrin is used on less than 5 percent of the total citrus acreage, and the ALJ noted that much of that use was a kind of 'just in case' insurance, applied even in the absence of knowledge that the pest exists in the pertinent grove. As to other crops, the record indicates, for example, that effective registered alternatives are available; that, in the case of pineapples and cranberries, aldrin/dieldrin offer multi-year protection, so the benefits of past applications will continue; and that the present oversupply in the cranberry market diminishes the prospect of hardship from the suspension in regard to that crop.

Shell protests that heptachlor/chlordane, pesticides that demonstrate carcinogenicity in mice and are stored in human tissue in the same levels as aldrin/dieldrin, will, in practice, be used in place of the suspended pesticides. Because heptachlor presents an identical cancer risk, Shell argues, the Administrator's suspension of aldrin/dieldrin does not 'prevent' an imminent hazard as required by the statute. Heptachlor is also the subject of cancellation proceedings. There is no law that says that all evils must be attacked at the same time and at the same rate. So far as the public interest is concerned, it sufficies to note that there is evidence that heptachlor is not available in large amounts comparable to aldrin/dieldrin stocks of past years, so that, in any event, the EPA suspension will achieve a total reduction in the use of harmful pesticides.37

Shell and the other petitioners contend that the Administrator's findings are too incomplete to be adequate for a suspension order. This is not a substantial evidence case, they put it, so much as a challenge to the insufficiency of the findings. Time and again petitioners revert to the Administrator's failure to identify explicitly certain contentions, and his disposition thereof, or to his failure to discuss petitioners' evidence.

Under § 6(c)(2) of FIFRA, the Administrator has a maximum of seven days38 after the ALJ's decision in which to issue his opinion. This time constraint--part of the statutory plan for expedition in reaching suspension decisions--is material in appraising how much Congress contemplated would be required of the Administrator's findings. Compare International Harvester Co. v. Ruckelshaus, 155 U.S.App.D.C. 411, 426--27, 478 F.2d 615, 630--31 (1973). The hearing transcript--not including statements by witnesses, exhibits, and over 1,000 pages of briefs--was almost 4,000 pages long,39 and even longer portions of the cancellation hearing transcript were also incorporated. The USDA's complaint that the Administrator failed to give 'attentive consideration' in his discussion of the minor uses to some sixty exhibits and 2,000 pages of testimony by USDA witnesses leads us to comment that we cannot accept the view that the law requires explicit evaluation of all relevant testimony when that is not necessary to deal with the salient grounds of objection and would interpose a barrier that would preclude practical use of the suspension provisions.

Under these circumstances, we think it sufficient that there is substantial evidence in the record and that the court is able to discern the fair import of the Administrator's reasoning.40 Even in full-dress proceedings without time constraints a court will accept findings that are not wholly articulate, if they can 'discern the path' of the agency.41 What impresses us about the suspension order under review is the Administrator's sensible effort to write an opinion that emphasized his consideration of the issue that most concerned the parties--carcinogenicity. And his discussion and findings on the other contested matters is adequate of satisfy a reviewing court that his decision was not made arbitrarily and capriciously.

Perhaps a paradigm of petitioners' extremism and contentiousness is the objection that the Administrative Procedure Act requires findings of the agency 'on all the material issues' and that the Administrator's opinion is not 'salvageable' by reference to the discussion of the ALJ.42

In our view, it is a matter not for condemnation, as was suggested in argument, but for commendation, that the Administrator's opinion consumes only 45 pages, whereas the ALJ wrote 109 pages. It would have been desirable for the Administrator to have said explicitly what is clearly implicit in and indeed suffuses his entire opinion, that he accepts the ALJ's findings and reasoning except where a difference in commentary is made explicit. But, in the case of a suspension order issued under such time pressure, we cannot stand on ceremony to the extent of vacating the order, or remanding for further findings because this was not recited in so many words.

We turn, finally, to the point most ardently pressed by Shell at argument, and adopted by the other petitioners, that the EPA Order is tainted by ex parte communications in that members of the agency's enforcement staff, who were involved in the continuing cancellation hearing at the time, presented arguments to the Administrator favoring the issuance of a Notice of Intent to Suspend. Shell concedes that communications between the prosecutorial and adjudicative staff of an agency are appropriate prior to the initial filing of an administrative complaint, but argues that this rule does not apply where, as here, the communications take place while another phase of the 'same case' is underway. We reject Shell's contention.

Suspension and cancellation hearings are separate proceedings in the respect, critical here, that the decisions in the two proceedings are made under different legal standards.43 To the extent that they are related, we do not find this to be a bar to the kind of communications engaged in here. It may happen that during the course of an agency proceeding against two individuals the 'prosecuting' staff discerns from the evidence that proceedings should also be instituted against, or the initial proceeding broadened to include, a third individual. The prosecutorial staff would not be debarred from consulting with the agency head about these steps by the mere fact that a related proceeding was already under way. The same conclusion is applicable where there is no new party but the emerging evidence indicates that a new charge or a broadened charge is appropriate.

Congress has not accepted the view that the possibilities of unfairness require prohibition of an administrative structure that permits the same agency to issue the notice that begins a proceeding and to make the ultimate determination.44 It has accepted a pragmatic view that the need for effective control by the agency head over the commencement of proceedings requires an ability to conduct consultations in candor with an investigative section on the question whether a notice should be issued and a proceeding begun, and this notwithstanding any residual possibilities of unfairness.

In this case the agency respected the internal separation of functions provided by Congress in its combination of fairness and pragmatism; there is no claim of consultation between the agency's 'prosecutors' and the agency head except on the issue whether to issue a new notice--whether to start a suspension proceeding. There is no allegation of communication between 'prosecutor' and agency head regarding the final decision in either the cancellation proceeding or the suspension proceeding.45

The Administrator's indications in the Notice that he is 'persuaded that there exists an 'imminent hazard" and that he finds that 'a situation exists in which the manufacture of Aldrin and Dieldrin during the coming months will be 'likely to result in unreasonable adverse effects' on man and the environment'46 do not represent prejudgment of the merits of the decision to suspend. The Administrator was merely making a determination to begin a suspension proceeding under § 6(c)(1) of FIFRA,47 accompanied by the prefatory findings required by the law, without prejudice in any way to the consideration that would be given to the suspension record and to the result that would be reached in the light of that record. The proponents of continued registration were given a full hearing on their objections.

Because the suspension hearing was expedited,48 the ALJ incorporated nearly 11,000 pages of transcript and more than 350 written exhibits from the cancellation hearing. Shell protests that these procedures made the hearing inadequate to protect its interests in that the EPA prosecutorial staff and EDF were allowed to incorporate the presentation of their medical and benefits case from the more leisurely cancellation proceeding, where they had discovery and subpoena powers, while the proponents of registration, who had not yet reached that phase of their presentation in the cancellation hearing, were compelled to respond in much less time without similar discovery powers. At the time of the suspension notice, Shell had already called 125 witnesses in the cancellation proceeding.49 Shell did not make a showing that the shortness of the duration of the suspension hearing precluded a fair disposition, or that more time would have been needed by someone seeking, on an expedited basis, to make a presentation in favor of continued registration. Indeed, although Shell was allotted 12 of the 15 days scheduled for the expedited suspension hearing, it used only 8 1/2 of the twelve days, and did not take advantage of ALJ Perlman's offer to permit the presentation without cross-examination of the written statements of witnesses that could not be heard orally within the allotted time.

EDF charges that the EPA's decision to exempt the sale and use of existing stocks of aldrin/dieldrin from the general suspension is arbitrary and capricious. EPA has responded that this decision was based on an assumption that no appreciable and realistically retrievable stocks existed at the time of the order, and that any denial of procedural rights was harmless error. EPA counsel have informed us that EPA was presented in January 1975 with estimates that approximately 5 percent of the total 1974 amount of aldrin granules will be available for use in 1975, and that EPA intends to investigate the matter further, an ongoing re-evaluation that is entirely appropriate.

We affirm the agency's suspension order of October 1, 1974, except for the exemption of the sale and use of existing stocks. The record is remanded for further consideration of that issue.

So ordered.