MARKEY, Chief Judge, United States Court of Customs and Patent Appeals:

Appeal from a judgment of the district court denying plaintiffs' motion for judgment that Regulation 700.100 of the Food and Drug Administration (FDA) is arbitrary and capricious and not in accordance with law and granting FDA's motion for summary judgment. We reverse.

BACKGROUND

On February 25, 1974, appellee Food and Drug Administration (FDA) in accordance with 21 U.S.C. §§ 321(m), 362(a), and 371(a), initiated informal rulemaking proceedings by publishing a proposed regulation governing hypoallergenic cosmetics, under which:

A cosmetic may be designated in its labeling by words that state or imply that the product or any ingredient thereof is "hypoallergenic" if it has been shown by scientific studies that the relative frequency of adverse reactions in human subjects from the test product is significantly less than the relative frequency of such reactions from each reference product(s). [ 39 F.R. 7291.]

The lynch-pin of the regulation was its requirement for employment of "comparison testing," i. e., for testing the labeled product against "reference product(s)" defined in the regulation as "similar-use competitive products in the same cosmetic product category" and representing a market share of 10%. Adoption of the comparison testing method rested entirely on the Commissioner's adoption of a comparative definition: "the term `hypoallergenic' means to the consumer that the product causes fewer adverse reactions than other, similar-type use products . . .," 39 F.R. 7288, and the feeling that, while use of "hypoallergenic" has expanded over the years, the difference between "hypoallergenic" cosmetics and those not so labeled has become less distinct. 39 F.R. 7288.

Included in the preamble were comments of the Cosmetic, Toiletry and Fragrance Association (CFTA), the Bureau of Consumer Protection of the Federal Trade Commission (FTC), and appellant Almay Corporation (Almay). CFTA alleged that "there is no demonstrated need nor is it practicable for the minimizing of allergic reactions to be an overriding consideration in all aspects of production and marketing of every cosmetic product," and urged adoption of this definition of "hypoallergenic" and similar terms:

A product labeled hypoallergenic or one using terms having similar meaning is one designed, formulated, tested, manufactured, marketed and monitored for the purpose of minimizing:

1. The incidence of allergic response in individuals with a history of allergic reactions.

2. The risk of allergic induction to the normal population. [ 39 F.R. 7290.]

Almay objected to the comparison testing method because the composition of the selected reference products could not be predicted. Urging adoption of the CFTA definition, Almay proposed an objective compliance test, under which a product would have to demonstrate a minimum irritancy.

The FTC letter questioned whether comparison testing might permit labeling now illegal in the eyes of FTC and made a recommendation:

As you know, FTC case law holds that an unqualified claim of superiority, such as "less allergenic," is likely to be understood by consumers as a comparison to "most" other products or to competitive products "generally," Liggett Myers Tobacco Co., 55 F.T.C. 354 (1958). Yet the rule, as presently drafted, would permit an unqualified "less allergenic" claim to be made, based on a showing that the product is less allergenic than 10 percent or less of competing products. Such claims are thus potentially deceptive or misleading under section 5 and section 12 of the FTC Act. To avoid such potential deception, I believe that products making a hypoallergenic claim should be required to define, in labeling, their hypoallergenic claim by specifically stating that the product is "less likely to cause allergic reactions than some competing products." [ 39 F.R. 7291.]

Under 5 U.S.C. § 553(c), the Commissioner solicited the filing of comments with the FDA Hearing Clerk on or before April 26, 1974. The FDA itself filed items characterized as "supportive data and background information":

1. Letter of June 1, 1973 of the Committee on Cutaneous Health and Cosmetics of the American Medical Association,

2. List of complaints on hypoallergenic cosmetics received by the Food and Drug Administration during the period 1969 through 1973,

3. Summaries of complaints on hypoallergenic cosmetics received by the Food and Drug Administration during the year 1973,

4. Two letters on hypoallergenic cosmetics received from consumers by the Food and Drug Administration,

5. Two letters on hypoallergenic cosmetics received from industry by the Food and Drug Administration,

6. Final Report on A Study of Health Practices and Operations, contract no. FDA 66-193 (June 1972).

During the comment period, the FTC filed the results of a consumer survey on hypoallergenic cosmetics, and these comments thereon by the Director of FTC's Bureau of Consumer Protection:

. . . we are filing these additional comments based upon a recent survey, conducted in California for the Federal Trade Commission . . . in October and November 1973 . . .. I wish to emphasize that even the limited conclusions I draw . . . are those of the Bureau of Consumer Protection . ..

The survey consisted of four questions developed by the staff of the Bureau. . . . The survey used a probability sample of California adults 18 or older. . . . [T]he FTC's questions were administered only to the 585 female adults in the sample, and to 150 teenagers.

The first question concerned consumer awareness of . . . "hypoallergenic." The second . . . asked the respondents to state . . . what they believed "hypoallergenic" means. The third . . . asked whether the use of the term was based on official government tests. Finally, . . . asked to select from four definitions the one they believed came closest to describing "hypoallergenic."

[T]he survey is obviously limited . . in . . . population sample and the number of questions and . . . is silent in some important respects, not least . . . the absence of any breakdown . . . between users and non-users of hypoallergenics. Nevertheless, . . . it sheds some light on consumer understanding of . . . "hypoallergenic."

While no tabulation was made, . . . consumer definition of "hypoallergenic" (question 2) indicates that respondents used the words "allergy," "irritation" and "rash" almost equally. This . . . is consistent with the judgment . . . in the preamble that the consumer "lacks sufficient medical knowledge to distinguish between a reaction that results from allergenicity (sensitization), irritation, or some combination of these." 39 F.R. 7288.

More critically . . . the . . . results suggest . . . a good deal of misunderstanding and confusion . . among many consumers about the meaning of hypoallergenic. . . . [W]hen . . . subjects were given a choice of definitions, only 57% of women and 45% of teenagers selected the "correct" definition — "less likely to cause irritation than regular cosmetics." Twenty-three percent of the women and 22% of the teenagers chose "completely free of any ingredients causing skin irritation/allergy" — an interpretation which would render a product misbranded if made by a company marketing cosmetics. . . . [A] still smaller portion . . . was able correctly to define "hypoallergenic" when given no assistance. . . . [O]nly 7.5% of women defined "hypoallergenic" in . . . relative terms as "less likely to cause" or "low incidence of" allergenic reactions. Sixty-five percent used such absolute terms as "doesn't cause allergic reactions," "not harmful to skin," "no harsh chemicals" or "safe for use."[*]

[*] [D]isparity between responses to . . . question (# 2) and . . . question (# 4) is not surprising. Open-end questions . . . work against respondents whose verbal skills are limited . . . Where choices are given, all respondents have to do is choose. Where the subject is . . . to choose . . there is an element of guessing . . .. [T]he probability of . . . guessing the correct meaning . . . is 1 in 4. In this case . . . the probabilities are greater. If respondents who answered question 2 positively . . . behaved consistently, they had a 50% chance . . . since only 2 . . . definitions . . . were positive; one . . . was neutral and the other was negative. (i. e., "more" allergenic) Since most . . . responded positively to the open-end . . . not surprising that the rate of correct responses increased dramatically. Finally, pre-coded answers . . . restrict the range of choice . . . eliciting less divergent responses and fewer non-responses. The funneling of answers . . . to predetermined lines is likely to lead to more correct responses. [Footnote that of the Director.]

The results of our survey should be used with caution. It is plausible . . . that a higher proportion of accurate definitions . . . would be found in . . . the population . . . solely of hypoallergenic users. Nevertheless . . . the survey raises questions as to the likelihood that consumers, without some further assistance, will properly understand what is meant by hypoallergenic. To this extent, at least, . . . our survey provides some additional support for requiring a label statement such as that recommended in my January 11 letter . . . "less likely to cause allergic reactions than some competing products."

Comments by cosmetics and related industries were (1) uniform in their criticism of the comparative definition and comparison test method, and (2) generally supportive of the CFTA definition and an objective test.

Almay, by letter of April 24, 1974, alleged that the proposed regulation was irrational and objectionable, citing five reasons: (1) the standard is constantly changing, (2) comparison testing would lead to unfortunate competitive practices, (3) comparison testing would deprive the consumer of unique services of hypoallergenic manufacturers, (4) ignoring the severity of reactions in accordance with the preamble is unscientific and illogical, and (5) it would be impossible to ascertain reference products. Almay's letter further stated that "the proper focus for testing hypoallergenic products to determine the validity of the claim should be a comparison of such products with an objective standard . . .."

Almay also challenged the FTC survey as an improper basis for FDA's comparative definition:

The FDA is reported to have relied upon a survey conducted in California for its belief that the public ascribes a comparative meaning to the word "hypoallergenic". There are weaknesses in this survey, particularly the parochial nature of the sample used. But this aside — the results of the survey do not support the FDA's proposition that the public does understand the word in a comparative sense. A majority of those answering indicated this only in answer to Question 4 where answers were suggested from which to choose, one of which expressed the comparative idea. However, in answer to Question 2, where no answers were suggested, only about 4% gave an answer involving a comparative concept.

Appellant Clinique Laboratories, Inc. (Clinique), citing essentially the same reasons, alleged that the proposed regulation was illogical and unworkable and urged adoption of the CFTA definition. Regarding FDA's comparative definition, Clinique stated:

. . . to those special consumers who need and want "hypoallergenic" products what is important is not whether a given product has a lower potential for allergic reaction than another product but whether the given product has been especially designed, formulated, tested, marketed and monitored to meet her needs.

By its emphasis on "competitive testing" the FDA proposal has established an artificial and illogical definition of "hypoallergenic" products.

Like Almay, Clinique urged the adoption of the CFTA definition.

Comments were also submitted by a number of dermatologists, consumer groups, and individual consumers. Eight dermatologists favored testing in which a product would have to demonstrate an extremely low potential for allergic reaction to qualify as hypoallergenic. Seven dermatologists opposed the comparison test method, and one was non-committal. Four consumer groups took issue with the comparative definition as likely to cause confusion among users of hypoallergenic cosmetics. One consumer group was in general agreement with the FDA proposal.

After close of the comment period, Almay informed FDA that it planned to convene a panel of leading dermatologists to develop an acceptable series of test procedures for regulation of hypoallergenic cosmetics. The Commissioner responded that he intended to go forward with a final order, based on comments received on the February 25, 1974 proposal. Almay was, however, given until April 25, 1975 to submit further information. An alternate proposal was submitted by Almay on April 22, 1975, supplemented by the comments of dermatologists contacted by Almay.

Also after close of the comment period (apparently in March, 1975) FDA contacted a number of dermatologists for comment on the proposed regulation. Fifteen dermatologists replied (apparently in April). Their replies were placed in the record after the final order was signed.

The final order was signed on May 29, 1975, and published as Regulation 700.100 on June 6, 1975. In its final form the regulation was amended to require that use of "hypoallergenic" on a package be followed by "less likely to cause adverse reactions than some competing products."

On July 15, 1975 Almay and Clinique (the companies) brought an action, in the U.S. District Court for the District of Columbia, for declaratory judgment that Regulation 700.100 of the FDA was arbitrary and capricious in that the administrative record did not factually or legally support the conclusions made therein; and was not in accordance with law. The complaint alleges in part:

(1) The Commissioner's definition and application of "hypoallergenic" is not sustainable on the administrative record;

(2) The Commissioner's finding that the test method prescribed in the regulation is feasible and necessary for compliance with the regulation is not sustainable on the administrative record; and

(3) The extension of the regulation to products labeled in part "allergy tested" is not sustainable on the administrative record.

FDA moved for summary judgment, asserting the absence of any issue of material fact and arguing that its decision to comparatively define "hypoallergenic" as meaning "less likely to produce adverse reactions than some competing products," rendered non-arbitrary its imposition of the comparison testing method of assessing whether a particular cosmetic is in fact less likely to produce adverse reactions than competing products.

FDA justified its decision to define "hypoallergenic" as meaning less allergenic than some competing products on what it considered confusion in the use of the term. The evidence relied on was the preamble to the proposed regulation and the FTC survey. The letters of dermatologists contacted by FDA after close of the comment period were not relied on as support for any aspect of the regulation.

In initially opposing FDA's motion for summary judgment, and believing that significant documents relating to the dermatologists' letters were missing from the administrative record, the companies requested discovery. A supplement to the Certified Record having been submitted by FDA, the companies moved to remand with instructions that FDA review the entire record. The district court denied both discovery and remand.

The companies then moved for summary judgment. The district court found the Commissioner's definition supported by two factors in the administrative record: (1) a significantly greater number of consumers believed that "hypoallergenic" meant "safer than competitors" rather than "very safe," and (2) a comparative definition would be more helpful to consumers because adverse reactions to cosmetic products are relatively rare today overall, according to the AMA Committee on Cutaneous Health and Cosmetic statement. Having found the comparative definition supported, the court held reasonable the Commissioner's decision to impose the comparison test method. The Commissioner's conclusion that "related terms" convey to consumers that a particular cosmetic is comparatively less likely to cause adverse reactions was found to be supported as a rational inference from the FTC survey. The regulation having been found neither arbitrary nor capricious, plaintiffs' motion for summary judgment was denied and FDA's motion for summary judgment was granted.

ISSUES

The dispositive issues on appeal are: (1) the proper scope of review, and (2) whether the administrative record sustains the Commissioner's definition of "hypoallergenic."

OPINION [27] (1) Scope of Review

Involved here is an informal rulemaking proceeding, in which no hearing is required. The scope of review is therefore governed by the "arbitrary or capricious" standards set out in the Administrative Procedure Act, 5 U.S.C. § 706(2)(A). Camp v. Pitts, 411 U.S. 138, at 142, 93 S.Ct. 1241, 36 L.Ed.2d 106 (1973). The companies contend that the standard is met only by an administrative record giving "adequate and substantive support" to the agency's position, arguing in effect for a type of "clearly erroneous" standard. FDA would have us believe that informal rulemaking means that "the record of the proceeding is not required to contain `evidence' to support the ultimate determination made," and that, in promulgating a regulation by informal rulemaking, an agency may rely on its own experience and expertise, so long as those factors appear in the record and the resulting decision is rational, arguing in effect for a type of "super-presumption of correctness" standard.

The scope of review is neither as broad as that for which the companies contend, nor as narrow as that for which FDA contends. The governing case is Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971), in which the Court held that, because § 706(2)(A) applied, the Secretary of Transportation's action did not have to meet the "substantial evidence" test, but that the generally applicable standards of § 706 did require that the reviewing court engage in a substantial inquiry.

Section 706(2)(A) requires a finding that the actual choice made was not "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." . . .. To make this finding the court must consider whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment. . . . Although this inquiry into the facts is to be searching and careful, the ultimate standard of review is a narrow one. The court is not empowered to substitute its judgment for that of the agency. [ 401 U.S. at 416, 91 S.Ct. at 823.]

This court extensively analyzed Overton Park and the scope of review applicable to informal rule-making in Ethyl Corp. v. Environmental Protection Agency, supra, note 4, a case involving technological considerations not present here. It was there said that the Supreme Court "does not intend the `clear error of judgment' phrase to sanction review" as intrusive as the "clearly erroneous" test, and that "in the context of `arbitrary and capricious' review, we shall reverse for a `clear error of judgment' only if the error is so clear as to deprive the agency's decision of a rational basis." 176 U.S.App.D.C. at 407, 541 F.2d at 35.

In Camp v. Pitts, supra, the Court again considered the "arbitrary and capricious" standard of review, in the context of a contemporaneous explanation of an agency decision which indicated the determinative reason for the final action taken. The Court held that "the validity of the . . . [agency's] action must, therefore, stand or fall on the propriety of . . . [the agency's] finding, judged, of course, by the appropriate standard of review," and the finding must be "sustainable on the administrative record made." 411 U.S. at 143, 93 S.Ct. at 1244.

The holding of Camp v. Pitts, supra, was applied to informal rulemaking procedures of FDA by the Second Circuit in National Nutritional Foods Ass'n v. Weinberger, 512 F.2d 688 (2d Cir. 1975). Noting that a different standard of review applies where "notice and comment" procedure is used in rulemaking, rather than an evidentiary hearing, the court said:

The very absence of a detailed record of the type that would be made if an evidentiary hearing were held makes it advisable for the agency, in lieu thereof, to provide a thorough and comprehensible statement of the reasons for its decision. Where the agency's "finding is not sustainable on the administrative record made, then the . . . decision must be vacated and the matter remanded to [the agency] for further consideration." [ 512 F.2d at 701 (citations omitted).]

The fact that an "arbitrary and capricious" standard applies to informal rulemaking, rather than a "substantial evidence" requirement, cannot mean that nothing of an evidentiary nature is needed in the administrative record to support an agency decision. On the contrary, there being no evidentiary hearing, informal rulemaking proceedings are much more susceptible to abuse, and it becomes all the more important that a rational basis for the agency's decision be found in the facts of record.

Review of the Commissioner's decision to use a comparative definition thus requires an inquiry into the facts and a determination of whether his decision was (1) based on "a consideration of the relevant factors," and (2) free of such error as would deprive it of a rational basis.

(2) The Comparative Definition

The companies have not contested the Commissioner's right to set standards assuring that a labeling claim is not false or misleading. Nor do they contest his authority to promulgate testing standards when results of testing are implied by a labeling claim. 40 F.R. 24442. The companies do contest the Commissioner's conclusion that results of comparative testing is implied by "hypoallergenic" on a label, i. e., they contest his definition of "hypoallergenic."

FDA relies first on the preamble to its own proposed regulation in support of the Commissioner's conclusion. Respecting the definition of "hypoallergenic," the preamble is conclusory. The only authorities cited are a dictionary which defines "hypo" to mean "under," "beneath," "down," "less than normal," of "the lowest position in a series of compounds;" and the statement of an AMA Committee on Cutaneous Health and Cosmetics that "the term "hypoallergenic' as applied to cosmetics has outlived its usefulness, is misleading, and should be dropped from the labeling of cosmetic products." 39 F.R. 7288.

The dictionary definition may have supported the Commissioner's statement that "`hypoallergenic" was meant to denote that hypoallergenic products caused fewer reactions than conventional products," 39 F.R. 7288, "conventional" being used in the sense of "normal." The dictionary definition clearly does not support the Commissioner's decision to define "hypoallergenic" as causing "fewer reactions than some [10% of the market] products."

The AMA statement is cited in support of the Commissioner's decision to choose a new definition for "hypoallergenic," not in support of the new definition he actually chose. In light of the AMA report's conclusion that "hypoallergenic" should be dropped entirely, and of its further statement that "little distinction can be made between established cosmetic products as to their sensitization potential," it was inappropriate to cite the AMA report in support of any use whatever of "hypoallergenic," no matter how defined.

FDA relied also on the FTC survey, which is discussed in the preamble to the final regulation:

The Commissioner believes that consumers do not have a single understanding of the meaning of the term "hypoallergenic," but, as indicated by the results of the FTC survey, a significant number believe that it means that the product so labeled will result in fewer adverse reactions.

In drawing inference from the FTC survey, the Commissioner failed to consider a relevant factor — the comments of the FTC's Director of the Bureau of Consumer Protection. The survey questions were developed by the Bureau staff and the survey was conducted for the Bureau, yet its Director seriously questioned the statistical integrity of the survey results. Use of the FTC survey was thus a "clear error of judgment" under the criterion established in Ethyl Corp., supra.

The Commissioner and the district court relied primarily on the FTC survey. Though the district court referred to the AMA report's indication that adverse reactions are rare, it did so in support of the view that a comparative definition would be "helpful" to consumers whose belief as to the meaning of "hypoallergenic" was thought to have been established in the FTC survey. The Director of the survey-sponsoring Bureau, however, stated on the record that the survey: (1) was limited in population sample and number of questions; (2) was silent in important respects; (3) lacked a breakdown between users and nonusers; (4) lacked a tabulation; (5) established that consumers lacked medical knowledge sufficient to distinguish skin reactions; (6) produced results which should be used with caution; and (7) probably produced fewer "correct" definitions because it was not limited to consumers interested in the subject, i. e., hypoallergenic cosmetic users. Finally, the survey defined the "correct" definition as "less likely to cause irritation than regular cosmetics," yet the Commissioner chose a different definition: "less likely to cause adverse reactions than some [10% of market] similar-use competitive products in the same category."

We are fully aware of the caveat that we must not substitute our judgment for that of the regulator, nor shall we. We are equally aware, however, of the need for rationality, in the interest not only of justice, our major concern, but in the interest of the continued viability and public acceptance of the federal regulatory scheme itself. Though doubtless well-intentioned, imposition of the present regulation, over the not-unexpected, but here well stated, strenuous, and factually unrefuted, objections of those who must comply, in the face of indications that compliance would be unworkable, in effective disregard of alternate paths to consumer protection proposed by those who must comply — all on the basis of the flawed survey here involved — cannot contribute to public acceptance of the federal regulatory scheme.

An aura of unreality surrounds the creation of a definition in the present case. In the apparent belief that most products are today non-allergenic, it may have been thought that producers could not find 10% of marketed products producing more reactions and that use of "hypoallergenic" would thereupon cease. If so, the cumbersome method here chosen to achieve that result is an irrational substitute for a direct prohibition of all use of "hypoallergenic." The only bases in the administrative record for the Commissioner's definition are conclusions drawn from a dictionary that does not employ the chosen definition, an AMA report recommending against any use of "hypoallergenic," and an FTC survey, the unreliability of which is apparent on its face and indicated by its sponsor.

Under such circumstances we are compelled to find that the action of the Commissioner was arbitrary and capricious. Accordingly, the judgment of the district court is vacated, and there being no genuine issue of material fact, the case is remanded to the district court with instructions to grant plaintiffs' motion for declaratory judgment.

Order accordingly.