The labeling of economic poisons is regulated by the Federal Insecticide, Fungicide and Rodenticide Act (FIFRA).1 All such poisons distributed in interstate commerce must be registered with the Administrator of the Environmental Protection Agency.2 The Administrator is authorized by FIFRA to refuse, or to cancel, the registration of any poison that is misbranded.3

*****

* * *

Since 1878 petitioner has been selling phosphorous paste for home use as a rat and roach poison. Its English and Spanish language labels, as modified from time to time, have been registered since shortly after the registration requirement became effective in 1947. On January 4, 1971, the Administrator cancelled these registrations on the ground that phosphorous paste is too poisonous for use in the home except by commercial pest control operators.

The cancellations were precipitated by a review of petitioner's labels, but it is fair to state that the contents of the labels were irrelevant to the determination that the product was too dangerous to be permitted in the home. The evidence plainly showed that phosphorous paste is extremely toxic, that it possesses "great potential for harm," as the Hearing Examiner found, and that both adults and children have been killed or hospitalized by misuse of the product. On the other hand, there is no finding, and little or no evidence, of mortality or morbidity resulting from the use of the product in compliance with the directions on the label.4 The theory of the cancellation order was succinctly explained by the Government's principal witness who testified "that the general public is incapable of handling these things and following directions."5

Petitioner's challenge to the cancellations raises both procedural and substantive questions of first impression.6 At the heart of the controversy is the question whether FIFRA includes a "substantive standard of product safety7," and if so, what that standard is. Because of the novelty and importance of the issues, we shall state the facts in some detail, then review the statute, and finally test the findings against it.

Stearns Electric Brand Paste is an inexpensive and effective killer of rodents, roaches, water bugs and similar pests. The active ingredient in Stearns Paste is white phosphorus, a highly toxic substance, for which-as is true of most poisons-there is no known antidote.8 Phosphorous paste is the only kind of poison which is sold as both a rodenticide and roach killer. It is also the only product sold by petitioner. If an adult swallows over half a tube, "the odds are" that the ingestion will be fatal. Ingestion of much smaller quantities may prove fatal to children. The record plainly supports the finding that the product possesses "great potential for harm." (A. 269, Judicial Officer's decision.)

Petitioner markets its product through distributors and by direct mail. About 300,000 tubes are sold annually at a retail price of about 69 cents per tube. Since each tube contains enough paste for about 50 baits, it is estimated that 72 million baits of Stearns Paste have been used in the past five years. The Company receives about 1500 letters a year from customers who have moved and cannot find a local source of the product; many of these letters state that Stearns Paste is the best rat or roach killer the writer has used. The effectiveness of the product is not disputed; indeed, the question is whether it is too poisonous to be permitted in homes.

The tube is plainly labeled in black and red print as "POISON," with the skull and crossbones symbol and instructions for use printed in black and red letters. The tube is sold in a paper carton, which also contains an explanatory insert. The insert and carton, like the tube, prominently display the poison warnings.9

The record indicates that other products which will kill rats, and other products which will kill roaches, are available. Petitioner's evidence tended to show that the alternatives were less effective and more expensive. The evidence also indicated that rats and roaches pose a significant health problem, particularly in low income areas. Apparently the services of a commercial pest control operator cost at least $35 per visit. The findings do not specifically consider the magnitude of the danger from rats or roaches. Although the findings recognize the availability of various other rodenticides and pesticides, the record contains no square finding that any other product is as effective as phosphorous paste,10 or, assuming that cost is relevant, that any other product performing the same function is equally available.11

As a result of an interagency agreement negotiated in 1964,12 the Department of Agriculture began to refer economic poison labels to the Food and Drug Administration for toxicological review. In making such review, the FDA applied a "general policy that a product that can produce serious injury or death in a small child from an average swallow of about 4 1/2 to 5 cubic centimeters should not be used around the home particularly if there are safer equally effective alternate materials."13

On March 8, 1968, in response to a request from Agriculture for a comment on a proposed label revision, FDA objected to the reregistration of petitioner's product pursuant to its general policy against the home use of products that can produce serious injury or death in a small child from an average swallow. Petitioner was then advised by respondent14 that it would reevaluate its "registration policy with respect to products containing phosphorus paste for use in the home."

On October 7, 1968, respondent issued a notice of proposed interpretation "with respect to labeling of phosphorus paste products," and invited comments thereon.15 The interpretation stated that home use would be unacceptable, but use by government agencies and professional pest control operators would be permitted. Petitioner filed written comments and suggestions which were duly considered and rejected, and on March 19, 1969, Interpretation No. 26 was issued effective 60 days after publication in the Federal Register.16

On May 23, 1969, respondent issued notices of cancellation of both of petitioner's registrations.17 Petitioner filed timely objections and on October 6 and 7, 1970, an evidentiary hearing was held. The Hearing Examiner ruled that the burden of proof was on petitioner, who thereafter presented one witness and 21 exhibits. Respondent's case included the testimony of 8 witnesses and 9 exhibits. After considering the record and arguments of counsel, the Hearing Examiner filed a recommended decision including proposed findings of fact and conclusions. His recommendations were adopted by a judicial officer of the Department of Agriculture, who entered the order of cancellation on January 4, 1971. We stayed the operation of the order pending review in this court.

Respondent's evidence consisted primarily of expert opinion as to the toxicity of phosphorous paste and such data as was available concerning the harm which it has actually caused humans. The statistical evidence included data from the National Clearinghouse for Poison Control Centers18 and reports from state agencies.

The witness from the National Clearinghouse estimated that there are approximately one million ingestions of harmful substances each year, of which only about 115,000 are reported to a poison control center. The reported ingestions include a wide variety of substances, such as aspirin, kerosene, detergents, and other household products, as well as economic poisons. When measured strictly by the number of reported fatalities, aspirin is the most lethal substance used in the home environment.19 The National Clearinghouse records placed in evidence in Government Exhibit IV cover the period from January 7, 1962, through August, 1968. Those records describe 207 ingestions of phosphorous paste products, of which 147 were accidental, 51 were suicidal, and 9 were not classified as either accidental or suicidal. A total of 15 deaths resulted, 9 in the suicidal category, 5 in the accidental, and 1 that was unclassified. Forty of the accidental ingestions, 38 of which involved children, required hospitalization. The 5 accidental deaths all involved children.20

Stearns Paste accounted for a substantial portion of the total during the 6 1/2 year period. Of the 5 accidental deaths, 3 were attributed to Stearns Paste; of the 51 suicidal ingestions, 20 were Stearns; of the total 207 ingestions, 86 were Stearns.21

Since only about 10% of all ingestions of harmful substances are reported to a poison control center, it is reasonable to infer that there may have been a significantly larger number of ingestions of phosphorous paste than is revealed by the record. Since there is presumably a greater likelihood that a poisoning which resulted in death or hospitalization would be reported (at least if correctly diagnosed), it is not clear whether the same inference may be drawn with respect to unreported fatalities. The National Clearinghouse records do establish, however, that at least three children were killed by petitioner's product in the 1962-68 period. Data collected from state agencies for the years 1952 through 1968 indicate that phosphorous paste was responsible for 40 deaths.22

The evidence describing how most of the ingestions actually occurred is sparse. However, the Medical Examiner of Dade County, Florida, and the State Medical Examiner for Maryland have kept better records than most agencies; they provided the examples which respondent stressed in its presentation to the Hearing Examiner.

Dr. Davis, from Dade County, expressed the firm opinion that phosphorous paste is too toxic to be permitted in the home. He was a sponsor of an ordinance in the City of Miami which outlawed the use of the product in that jurisdiction.23 His data reflects 13 deaths in Dade County in the years 1956 through 1968, of which 10 were probably suicides. The three accidents all involved children. In 1956 a 10-year old boy apparently ingested some paste applied in an apartment building by a professional exterminator; in 1958 a young child found a poison bait in a neighbor's garbage container and swallowed it; in 1960 a 3-year old apparently obtained access to a tube of paste in a home and swallowed part of its contents.

The Maryland records describe 22 deaths in the years 1950 through 1966, of which 15 were suicides, 5 were accidental ingestions by adults (apparently intoxicated), and 2 by children. In 1954 a 4-year old girl found an open can of paste which had apparently been left in an apartment by a former tenant; in 1961, 3 children shared some baits found in or near a garbage can in a neighbor's yard separated from the children's own yard by a four-foot wire fence. One died and two were hospitalized. The Maryland doctor was not asked to produce records indicating the relative importance of phosphorous paste as a cause of death in the home environment, but commented that aspirin is the most common cause of poisoning in children. He estimated that aspirin or other medicinal poisonings killed 2 or 3 children each year in Maryland, which he characterized as a "low incidence."

The record discloses that at least some of the deaths caused by phosphorous paste products followed application by commercial pest control operators; it does not disclose how many children were killed by rats or roaches, or how many were saved from harm by the use of petitioner's product.

After finding that petitioner's product had actually caused significant mortality and morbidity, and that it presented a great potential for harm, the Judicial Officer concluded, in conformity with the Hearing Examiner's recommendations: (1) that petitioner had the burden of proving that the registrations should not be cancelled; (2) that the use of phosphorous paste insecticides and rodenticides in and around the home could not be rendered safe by any label;24 (3) that petitioner's warning statement, even if complied with, is inadequate to prevent injury to living man;25 and (4) that the product is "misbranded" because when used "in accordance with commonly recognized practice" it is "injurious to living man";26 and (5) that he had taken in consideration the fact that effective and less toxic insecticides and rodenticides are available on the market.

Petitioner contends (1) that there were procedural defects in the administrative proceedings; and (2) that Stearns Paste was not "misbranded" within the meaning of the statute. Before discussing these contentions, we shall review the history of FIFRA.II.

The Insecticide Act of 1910 prohibited the interstate sale of any insecticide or fungicide which was adulterated or misbranded within the meaning of the statute. 36 Stat. 331. The text of the Act makes it plain that Congress was primarily concerned with the effectiveness of such products and protecting purchasers from deceptive labeling. The Act contained criminal sanctions and provisions for seizure of misbranded or adulterated items, but neither a registration requirement nor a safety oriented labeling requirement.

FIFRA, which repealed the 1910 statute, was enacted in 1947. Like its predecessor, its text indicates a primary interest in protecting consumers from the purchase of ineffective products. However, the coverage of the statute was substantially broadened,27 a purpose to protect the public from the hazards associated with the use of economic poisons was implemented,28 and, for the first time, all economic poisons were required to be registered with the Secretary of Agriculture. The registration requirement was included as an aid to enforcement.29 If the Secretary disapproved of the applicant's proposed labeling, the applicant nevertheless had an absolute right to have his product registered under protest. Thereafter, unless the Secretary could prove in a judicial proceeding that the product was either misbranded or adulterated, he had no authority to exclude it from commerce.30

In 1964 the statute was amended to give the Secretary the power to refuse to register a new product, or to cancel an existing registration, if he found that the product was either adulterated or misbranded. The House Report on the 1964 changes and what little floor discussion there was indicate quite clearly that the only major change31 contemplated was elimination of the registration under protest procedure. The change was made for two basic reasons: (1) to settle the question of compliance with the act before the economic poison could be marketed and (2) to place the burden of proof of safety and effectiveness on the applicant for registration.32 The concern remained with efficacy and safety "when used as directed."33 No changes were made in the language of 7 U.S.C. Sec. 135(z) (2) (c), (d) and (g). the provisions with which we are primarily concerned.

*****

* * *

As a matter of procedure, petitioner contends that Interpretation 26 is invalid because it was promulgated without a prior public hearing, and that it was error to require a registrant to assume the burden of proving that the proposed cancellations were improper.34

We agree with respondent's characterization of Interpretation 26 as a mere announcement of the agency's position which did not have the legal effect of a regulation. It is true, as petitioner argues, that the policy expressed in the Interpretation led to the issuance of the cancellation notices, but the Interpretation was not self-executing. Although not required by the Administrative Procedure Act,35 we think the agency acted properly in soliciting comments on its proposed policy statement before issuing Interpretation 26, and that it was appropriate to give an industrywide notice of its proposed position even though the Interpretation had no immediate legal effect.36 We are satisfied that the availability of an evidentiary hearing before a cancellation order is effective, together with the safeguard of appellate review, adequately protects a registrant's procedural rights.

We also reject petitioner's contention that respondent, as the proponent of the cancellation order, should have assumed the burden of proof. The 1964 amendment to FIFRA was clearly and specifically intended to shift the burden of proof from the Secretary (now the Administrator) to the registrant.37 It is true that most of the legislative comment concerned new registration rather than cancellation of existing registrations, but we do not believe the statute was intended to differentiate between the two situations.38 In view of the agency's continuing obligation to review the propriety of existing registrations, we are also satisfied that the purpose of the 1964 amendment is applicable to cancellation proceedings. The Examiner properly held that Stearns should assume the burden of proving that its registrations complied with the statute.39

*****

* * *

We must therefore decide whether petitioner proved a prima facie case and, if so, whether the right to continue the registration was overcome by respondent's evidence. These issues require identification of the statutory standards for registration.

To be eligible for the registration under FIFRA, the product must be "an economic poison."40 The statute has no application to products which are completely safe, or to products like aspirin and detergents which are safe enough in normal use but endanger children when used negligently. Thus, the first element of a registrant's prima facie case is proof that his product is dangerous; by hypothesis a poison is not completely safe.

The poison may not be registered if it is either "adulterated" or "misbranded." These terms embody more than one standard.

First, the product must be effective; more precisely, it must be at least as effective as the registrant claims it to be. Since all economic poisons are intended to kill some form of plant or animal life, in a sense the statute includes a minimum standard of deadliness. Petitioner's uncontradicted evidence of the effectiveness of its product met that standard.

Second, the product must satisfy certain safety standards. Although the definition of the term "adulterated" in other legislation embodies safety considerations,41 in FIFRA it is the definition of the term "misbranded" that identifies the statutory standards of product safety. There are slight variations in the language used in different subsections of the Act, but two principal standards are identified: (1) the label accompanying the product must contain directions for use and a warning or caution statement which "if complied with [is] adequate to prevent injury to living man and other vertebrate animals, vegetation, and useful invertebrate animals";42 and (2) an insecticide, nematocide, fungicide, or herbicide (but not a rodenticide), is misbranded if "when used as directed or in accordance with commonly recognized practice it shall be injurious to living man or other vertebrate animals, or vegetation, except weeds to which it is applied, or to the person applying such economic poison."43 The italicized phrases are those most relevant to the issues in this case.

The first of these standards focuses on the safety of the product when used in compliance with directions. We think petitioner's evidence of a long history of use of Stearns Paste, involving broad distribution and numerous repeat orders, coupled with the absence of claims or evidence that injury had actually resulted from use of the product in compliance with directions, was sufficient to make a prima facie showing of satisfaction of this statutory standard.

The second standard relates to the use of Stearns Paste as a roach and water bug killer. For this use, the directions specify pieces of paper, rather than scraps of food, as bait. It was therefore less hazardous than when used as a rodenticide. Moreover, from petitioner's evidence it would be reasonable to infer that the "commonly recognized practice" in applying Stearns Paste was consistent with the directions on the label. Accordingly, we believe petitioner also made a prima facie showing of compliance with this standard.

Petitioner's prima facie case was, of course, subject to being overcome by respondent's evidence of misbranding. Whether it has been overcome in this case depends largely on a proper formulation of the standard for finding a violation of FIFRA. Respondent, in effect, relies on a substantive standard of product safety which has little, if any, relevance to the contents of the label.44 Respondent states the test thusly:

"Thus, the final decision with respect to initial or continued registration of a product depends on the intricate balance struck between the benefits and dangers to the public health and welfare resulting from its use. More specifically, the Administrator must determine and weigh (1) the nature and magnitude of the foreseeable hazards associated with use of a particular product against (2) the nature of the benefit conferred by the use of the product, or, put another way, against the magnitude of the social cost of foregoing the use of the product."45

Respondent explains that this test finds its source in the opinion of the Court of Appeals for the District of Columbia in Environmental Defense Fund, Inc. v. Ruckelshaus, 142 U.S.App. D.C. 74, 439 F.2d 584, 594 (1971), rather than in language employed by Congress.46 An important distinction between Ruckelshaus and this case should be noted. Ruckelshaus involved possible cancellation of the registration of DDT, an insecticide which has an impact on the environment even when used in complete conformity with the manufacturer's directions.47 That product, when properly used, has known benefits and a potential for harm that is not precisely known. To determine whether DDT is "injurious to man" requires, as the District of Columbia Circuit has fairly stated, a delicate balancing of its benefits against its adverse effects.48 Does the net result constitute injury to man within the meaning of FIFRA? If so, it is misbranded. In short, we think the standard as stated by respondent gives proper effect to the statutory language if used to measure the net injury resulting from use of an economic poison in compliance with directions.

A different situation is presented when the harm is entirely, or at least primarily, attributable to misuse of the product. To apply respondent's balancing test to such a situation is to ignore completely the concept of misbranding. Although it is consistent with the statutory language and purpose to apply a substantive standard of product safety to the use of a product in compliance with its manufacturer's directions, there is no statutory support for the application of that standard to misuse of a product. Without such support, the formulation of substantive standards of product safety by an administrative agency expands the scope of administrative discretion beyond permissible limits.

There are other objections to respondent's application of the "intricate balance" test to the problem presented by this case. In the DDT situation, the benefits of the poison are ascertainable with a reasonable degree of certainty; it is the other side of the balance that is difficult to weigh accurately. Moreover, the injury from DDT is to "man" in a collective sense-that is, to the total environment in which he lives. The obverse situation is present here. The cost to the community at large of depriving the homeowner and apartment dweller of an inexpensive rat poison cannot be measured on this record.49 The injury which respondent seeks to avoid in this proceeding is to specific individuals rather than to the total environment. That harm is largely attributable to willful misuse (in the case of suicide ingestions), wanton recklessness, or at least negligent behavior. Thus, on one side of the balance is a relatively small number of incidents of individual harm resulting from misuse by the comparative few; on the other side is the cost of depriving the prudent majority of a known but only vaguely defined benefit. In short, the conflicting interests are not identified sufficiently in the findings to determine whether the counterbalancing factors have been assigned proper weights.

The "intricate balance" test is inappropriately applied in this case for yet another reason. The Hearing Examiner placed important reliance on the absence of adequate information about the incidence of phosphorous paste poisonings and the actual circumstances in which it occurs without making any attempt to classify the data, either by estimate, by extrapolation, or by specific example, as between ingestions of registered and unregistered products, as between products purchased at retail and paste applied by commercial exterminators, or even as between misuse and use in accordance with instructions. He did endeavor, in discussing the National Clearinghouse data, to identify the portion of total ingestions which involved petitioner's product, and he also identified those examples of deliberate misuse that fell in the suicide category. But his affirmative reliance on the lack of adequate information as possibly supporting an inference that the danger may be ten times as great as the available data actually disclosed, cannot satisfy the test characterized as an "intricate balance." Furthermore, the Judicial Officer's findings contain no analysis of the actual or potential injury to man resulting from rats and roaches.

In any balancing test used to measure the acceptability of public sale of poisonous substances, it is imperative that the emotional impact of dramatic but unfortunate tragedies not be permitted to weigh too heavily on the scales. The spectacle of a young child suffering a violent death by poisoning offers a compelling justification for avoiding the danger of recurrence by banning future use of the poison forthwith.50 Unfortunately, however, such tragedies are a common occurrence in today's complex society and must be appraised as dispassionately as possible. Whether they justify a particular prohibition involves a policy choice which, under our scheme of government, must be made by a legislature or by an agency to which the legislature has delegated the responsibility for making principled decisions in accordance with its basic statement of policy. The fact that a legislature may react slowly to obvious dangers, such as the holocaust on our highways,51 the creeping infection of our environment, and the consumption of deleterious substances in the home, cannot justify an agency's policy determinations that are not authorized by statute.

The danger of misuse is, of course, a proper subject of regulatory concern. But unless the statutory concept of misbranding has itself been misbranded, under FIFRA that danger must be related to the form of the label.52 Neither the language of the statute nor its legislative history focuses directly on the problem of misuse, but there can be no doubt that the agency was intended to supervise the form and content of labels.53 An obvious purpose of such supervision is to minimize the risk of misuse.

The Hearing Examiner did note the relevance of the label in certain of his conclusions. Thus, he stated that the label warnings had "not been adequate to prevent injury to living man" since injuries and fatalities had actually been caused by phosphorous paste products.54 He thus implied that a product might be misbranded whenever its label failed to prevent injury to man. Such a standard of total prevention is manifestly too strict; it would require the agency to prohibit the use of phosphorous paste by commercial pest control operators, and would be broad enough to authorize cancellation of any poison registration whenever an incident involving fatal misuse occurred.55

We do not believe an isolated incident of misuse causing harm, or even death, to a particular individual is contemplated by the "injury to man" language in the statute. The word "man" is used in a collective sense, or perhaps with a typical connotation, as in the "reasonable and prudent man" concept familiar to negligence lawyers. In that sense the adequacy of the label may be judged by its tendency to protect against misuse. The judgment appropriately takes into account the toxicity of the product, the clarity of the warnings and the directions, and the ability of the user or purchaser to comprehend and thereby to avoid misuse.

To a limited extent these factors were considered by the Hearing Examiner; but again, we believe he implicitly adopted a test which was more strict than Congress intended. He appears to have accepted the expert's view that "the general public" is incapable of following instructions. That view would justify-indeed, might require-exclusion of all economic poisons from home use. Of greater importance, it is contrary to the premise which Congress must have accepted in the enactment of FIFRA. A statute which is primarily a regulation of labels necessarily assumes that the general public does heed warnings. We believe a fair respect for the statute requires rejection of a test of misbranding predicated on total illiteracy or universal disregard of instructions.

The adequacy of a label will, of course, be affected by the nature of the message to be conveyed and the ability of the reader to comprehend its meaning. Thus, if a product is not safe unless intricate or esoteric instructions printed in small type are followed with precision, use by laymen, even if reasonably careful, would create an obvious risk of injury to man.56 On the other hand, a conspicuous "POISON-KEEP AWAY FROM CHILDREN" warning in large red letters, prominently accompanied by skull and crossbones symbols, conveys a message which even the illiterate can understand. Disregard of such a simple warning would certainly constitute gross negligence.

It is not our function, however, to articulate in the first instance the standards which may support a finding of misbranding based primarily on evidence of misuse.57 The agency must direct its attention to that problem in its administration of a statute which is principally a regulation of labels rather than substances. We are persuaded that it has not yet done so but instead has indiscriminately applied a balancing test which is appropriate as a measure of proper use to this case which primarily involves misuse.

We express no opinion on whether the evidence in this record discloses such a probability of misuse of petitioner's product that a finding of "misbranded" would be supportable under standards consistent with FIFRA. Nor, of course, do we express any opinion on the policy issue of whether phosphorous paste should be banned from the home environment regardless of whether or not the products are misbranded within the meaning of FIFRA. We are persuaded, however, that the tests of misbranding, to the extent that they have been articulated, employed in this proceeding go beyond the authority which Congress has delegated to the agency in FIFRA.

In reaching this conclusion we have considered respondent's reliance on the "commonly recognized practice" phrase found in Sec. 135(z) (2) (g), as well as the standard which assumes use in compliance with directions. Perhaps the phrase indicates that whenever misuse occurs with sufficient frequency to be considered a common practice, a finding of misbranding is required. Such an interpretation, however, would attach no significance to the word "recognized." We believe a fair reading of the phrase relates to common practices which are "recognized" in the sense that they are approved by widespread custom or practice.58 In this case there has been no finding that misuse of Stearns Paste is either a common practice or a commonly recognized practice; furthermore, this provision would be inapplicable to the use of the paste as a rat poison, since subsection (g) does not apply to rodenticides.

We therefore hold that the cancellation orders must be set aside. After developing and articulating standards consistent with the authority delegated by FIFRA for determining when a label inadequately avoids the danger of harmful misuse, respondent may again propose cancellation of petitioner's registrations. Since the propriety of adducing additional evidence cannot be determined until the standards have been articulated, we express no view on whether or not the record should be reopened. We merely hold that petitioner's prima facie case has not yet been overcome.