Nine years ago Congress amended the Food, Drug, and Cosmetic Act, 52 Stat. 1040 (1938), by enacting the Color Additive Amendments of 1960, 74 Stat. 397. Nearly three years later, after appropriate rule-making proceedings, the Food and Drug Administration (FDA) published its Regulations thereunder, 28 F.R. 6439. This litigation about the validity of their provisions concerning diluents, finished cosmetics and hair-dyes has continued ever since. An initial effort by the Government to prevent the judicial review of the Regulations sought by the cosmetics industry, on the ground of lack of "ripeness," was rebuffed by Judge Tyler in the District Court for the Southern District of New York, 235 F. Supp. 648 (1964). A year later, when the case was nearly ready for trial, the Government renewed its effort to obtain dismissal for lack of ripeness on the basis of an allegedly conflicting decision by another circuit, with no success beyond obtaining a certificate from the district judge for an interlocutory appeal under 28 U.S.C. § 1292(b) which we allowed. We sustained the district court save in one respect no longer material, 360 F.2d 677 (1966). Both sides successfully sought certiorari but the Supreme Court affirmed, 387 U.S. 167, 87 S. Ct. 1526, 18 L. Ed. 2d 704 (1967). The case then returned to the district court where Judge Tyler rendered a decision favorable to the plaintiffs, 278 F. Supp. 786 (1968). This time the defendants were some seventeen months in bringing their appeal to us.

Finished Cosmetic Products as Color Additives

The key provision of the Color Additive Amendments is 21 U.S.C. § 376, providing for what the parties term "premarketing clearance." This begins by saying in subsection (a):

A color additive shall, with respect to any particular use (for which it is being used or intended to be used or is represented as suitable) in or on food or drugs or cosmetics, be deemed unsafe for the purposes of the application of sections 342(c), 351(a) (4), or 361(e) of this title, as the case may be, unless —

(1) (A) there is in effect, and such additive and such use are in conformity with, a regulation issued under subsection (b) of this section listing such additive for such use, including any provision of such regulation prescribing the conditions under which such additive may be safely used, and (B) such additive either (i) is from a batch certified, in accordance with regulations issued pursuant to subsection (c) of this section, for such use, or (ii) has, with respect to such use, been exempted by the Secretary from the requirement of certification; or

(2) such additive and such use thereof conform to the terms of an exemption which is in effect pursuant to subsection (f) of this section.

Subsection (b) (1) directs the Secretary to provide by regulation

for separately listing color additives for use in or on food, color additives for use in or on drugs, and color additives for use in or on cosmetics, if and to the extent that such additives are suitable and safe for any such use when employed in accordance with such regulations.

Section 8.4(c) of the Regulations sets forth the requirements for a listing application in elaborate detail. These include a complete description of the chemical and physical properties of the color additive, the chemical and physical tests relied on to identify it, and the methods and ingredients used for its production; a description of practical methods to determine the amount of the color additive, and of any substance formed by its use, contained in the final product; full reports of investigation made with respect to its safety; complete data to allow the Commissioner to consider the probable consumption of or other exposure to the additive and any substance formed therefrom; and proposed tolerances. If an exemption from batch certification is requested, supporting data are to be furnished. Each listing application must be accompanied by a fee of $2600 to cover the FDA's costs in processing the application and performing tests. Other provisions of the Regulations, § 8.22-§ 8.30, contain equally detailed requirements with respect to certification, including a provision, § 8.26, that "the person to whom a certificate is issued shall keep complete records showing the disposal of all the color additive from the batch covered by such certificate."

The parties are in basic disagreement concerning what constitutes a color additive requiring the premarketing clearance just described. On this the statute informs us, § 321(t) (1):

(1) The term "color additive" means a material which —

(A) is a dye, pigment, or other substance made by a process of synthesis or similar artifice, or extracted, isolated, or otherwise derived, with or without intermediate or final change of identity, from a vegetable, animal, mineral, or other source, and

(B) when added or applied to a food drug or cosmetic, or to the human body or any part thereof, is capable (alone or through reaction with any other substance) of imparting color thereto; except that such term does not include any material which the Secretary, by regulation, determines is used (or intended to be used) solely for a purpose or purposes other than coloring.

The disputed Regulation, 21 C.F.R. § 8.1 (f), after repeating this definition, adds two provisions of which the industry complained. One is "This includes all diluents"; the other is:

A substance that, when applied to the human body results in coloring, is a "color additive," unless the function of coloring is purely incidental to its intended use, such as in the case of deodorants. Lipsticks, rouge, eye makeup colors and related cosmetics intended for coloring the human body are "color additives."

The industry contended, and the district court held, that the Regulation exceeded the statutory authority in requiring premarketing clearance of diluents1 and of finished cosmetic products. The Government does not claim error in the conclusion that § 8.1(f) is invalid in defining diluents as color additives and thereby making them subject to premarketing clearance. But it vigorously attacks the holding that the section is similarly invalid in so defining finished cosmetic products.

In a valiant effort to overcome the instinctive reaction that when Congress spoke of a "color additive" it scarcely meant the entire product, the Government has developed an ingenious parsing of the definition in § 321(t). A finished cosmetic product, it says, while not a dye or pigment within subdivision (A), is an "other substance." And while such a product is not "added or applied to a food, drug or cosmetic" within subdivision (B), it is applied "to the human body or any part thereof" and "is capable (alone or through reaction with any other substance) of imparting color thereto." Hence the finished cosmetic product is itself a color additive. Indeed, says the Government, any other construction would fail to comply with the canon that every word in a statute must be given effect, since, as it claims, the industry has been unable to point to any "other substance" which, when added or applied to the human body, is capable of imparting color thereto.2

The district judge rejected the Government's construction on several grounds, 278 F. Supp. at 790-791. He pointed to the numerous instances where Congress used the word "cosmetics" when that was what it intended. See 21 U.S.C. §§ 331, 374, 381. He countered the defendants' canon with that of ejusdem generis; the natural interpretation of "other substance" is a coloring ingredient having properties similar to a dye or pigment, see fn. 2. Defendants' reading "other substance" to include a cosmetic would cause the definition to refer to "a dye, pigment, or other substance (including a cosmetic)" which, when added to "a food, drug or cosmetic" is capable of imparting color, a result he deemed absurd. When Congress wished to employ a comprehensive term for finished products, it used the word "articles," 21 U.S.C. § 321, whereas "substance" was used to connote a component or ingredient, 21 U.S.C. §§ 321(s), 346.

The Government answers the last argument rather effectively by pointing to other instances where the word "articles" was used to refer to components, §§ 321(f) and (i) and 355(b). It also contends that if the industry's construction is right, the exemption for pesticides in § 321(t) (3) was unnecessary, to which the industry responds that this was inserted on the request of pesticide manufacturers out of abundant caution, and that in any case it would have been more logical to call pesticides and fertilizers which might alter the color of plants "color additives" than to apply the term to finished cosmetics. The Government argues from the rejection of many similar requests to limit the definition that it was intended to have broad scope, while the industry asserts that many of these requests stemmed from such absurd fears — for example, that the peas in a can of peas and carrots might be "color additives" because they are not the same color as the carrots — that they reinforce Judge Tyler's restricted reading of the phrase "dye, pigment, or other substance."

Coming closer to the critical issue, the Government points to § 361(e), hereafter discussed, as showing that the statutory scheme places at least some kinds of finished cosmetic products under the rubric of color additives. It also directs our attention to § 362, which provides that "A cosmetic shall be deemed to be misbranded * * *, (e) if it is a color additive" unless its packaging and labelling accord with prescribed regulations. Although § 321(i) defines "cosmetics" to include the components of finished cosmetics as well as the finished cosmetics themselves, and therefore may make it necessary for "color additives" to be called "cosmetics" even when they are not meant to be used directly by consumers, these provisions scarcely show that all finished cosmetics were thought to be color additives requiring premarketing clearance.

While we consider the industry to be ahead in this exchange of textual volleys, we find more enlightenment in the history and declared purpose of the Color Additive Amendments. The definition of color additives in the 1960 amendment did not spring full-blown from the congressional brain. The Food, Drug, and Cosmetic Act of 1938, 52 Stat. 1040, had prohibited the sale, § 301(a) (c), of food, § 402(c), drugs, § 501(a) (4), and cosmetics (other than hair dyes), § 601(e), bearing or containing a coal-tar color other than one from a batch certified, in accordance with regulations as authorized for the three categories respectively in §§ 406(b), 504, and 604, to be "harmless and suitable for use."3 The Regulations, 21 C.F.R. § 135.1 (1949 ed.), defined "coal-tar color" as meaning "articles which (1) are composed of or contain any substance derived from coal-tar * * *, and (2) when added or applied to a food, drug, cosmetic, or the human body or any part thereof, are capable (alone or through reaction with other substance) of imparting color thereto." There is no indication that the FDA ever considered this definition to require the listing and certification of finished cosmetics as distinguished from the coal-tar colors themselves. The burden thus falls on the Government to show that when, in the Color Additive Amendments, Congress employed words similar to those previously in the FDA's regulations, it meant them to have a different effect. The history seems rather to reinforce the view that it did not.

The occasion for the legislation of 1960 was the Supreme Court's ruling of December 15, 1958, in Flemming v. Florida Citrus Exchange, 358 U.S. 153, 79 S. Ct. 160, 3 L. Ed. 2d 188, which this court had anticipated in Certified Color Industry Committee v. Secretary of Health, Ed. & Welfare, Folsom, 2 Cir., 236 F.2d 866 (1956). These decisions sustained the Secretary's new position, dating from 1950, that he could certify only coal-tar colors which were harmless without regard to quantity or concentration, regardless of the fact that they were "safe-for-use" within specified quantity limits.4

Shortly after this court's decision the color industry began steps to meet the threat with which it was confronted. The common goal of the industry and the FDA was, on the one hand, to restore the "safe-for-use" principle that had been in effect from 1906 to 1950 and, on the other, to deal with all colors rather than only with those derived from coal-tar. A revised industry draft, introduced by Representative Curtis in 1957, 85th Cong. 1st Sess., H.R. 8945, defined "color additive" as "a material which (1) is composed of or contains any dye, pigment or other substance * * *, and (2) is added or applied to a food, drug, cosmetic, or the human body or any part thereof, for the purpose of imparting color thereto alone or through reaction with any other substance." Obviously premarketing clearance of all cosmetic products containing color was the last thing in the industry's mind, as the FDA must have been well aware. In an extended letter, dated June 27, 1958, to the Chairman of the House Committee on Interstate and Foreign Commerce, the Assistant Secretary of Health, Education and Welfare endorsed the general principles of the bill. Nothing in the letter suggested that the Department conceived of the bill as requiring premarketing clearance of finished cosmetic products. Instead the letter used such language as "the proposal that we be permitted to admit colors for use under appropriate tolerance and other prescribed conditions of use in specified foods, drugs, and cosmetics," and "the bill, as we interpret it, authorizes the Secretary to decide which foods, drugs, or cosmetics may bear a color and in what amounts."

The Supreme Court's decision of December 15, 1958, created a new urgency. On January 13, 1959, Secretary Flemming issued a statement on the need for legislation. This, he said, "should place upon the users of colors the responsibility of proving to the satisfaction of the Food and Drug Administration that coloring materials in foods, drugs, and cosmetics are safe in the quantities they propose to use" (emphasis supplied). A draft bill circulated by the FDA to the industry on April 10 defined color additive in substantially the same manner as the Curtis bill except that in the second clause it used the phrase "is capable of imparting color" instead of "for the purpose of imparting color." Ten days later the Secretary characterized the FDA bill, among other things, as requiring "industry to prove to our satisfaction that its use of any color would be without harm" and as allowing the FDA "to place a limit on the amount of a color that may be used, or to specify the products in which it may be used, or both." Nothing in his statement indicated an intention that finished cosmetic products would have to be listed and certified. However, in response to fears by some members of industry, the words "is composed of or contains" in the first clause of the definition were deleted by the Department in the bill submitted to Congress. In his letter of transmittal the Secretary characterized the bill as providing "a scientifically sound and uniform system for the listing of color additives of any kind which may safely be used in foods, drugs, or cosmetics, subject, when necessary, to appropriate tolerance limitations and other conditions of use and to official certification of batches of color so as to assure the safety of such use to the consumer." 2 U.S.C. Cong. & Adm.News, 86th Cong. 2d Sess., p. 2914 (1960). A detailed analysis of the bill spoke of the need for "permitting the listing of colors under safe tolerances," for extending "the pretesting requirement * * * to those non-coal-tar colors, especially those used in cosmetics and drugs, to which it does not now apply," and also for applying to all colors "the requirement of certification of batches of color, where necessary for the protection of the public health." Id. at p. 2919. Nothing in the Department's analysis gave either the industry or the legislators any reason to think that by adopting the bill Congress would be subjecting finished cosmetics products to a premarketing clearance never previously imposed, and the Government has not directed our attention to any point in the hearings, reports or debates where any such intention was indicated.

We do not mean to imply that all or even many of the Congressmen who voted for the Color Additive Amendments in 1960 had such arcana vividly in mind. But it is a good principle that before a court will hold Congress to have made a basic change in regulatory procedures, legislators must either use plain language or give other clear manifestation of intent. Cf. Ozawa v. United States, 260 U.S. 178, 194, 43 S. Ct. 65, 67 L. Ed. 199 (1922). In the absence of this it is reasonable to assume that both those legislators who were knowledgeable and those who voted on faith wished to leave things as they were,

There is still further evidence in favor of the district court's construction. The basic structure of the Color Additive Amendments is to apply the same requirements to foods, drugs and cosmetics which contain coloring materials. Yet the Government's exegesis would result in finished cosmetics with coloring ingredients being subject to listing and certification whereas foods and drugs bearing such ingredients would not be. Yet, if there is a difference in hazard, foods and drugs, which are ingested, would seem to outrank cosmetics; indeed cosmetics were not brought into the regulatory scheme until 32 years after foods and drugs.

We cannot agree with the Government that the district court's invalidation of the portion of § 8(f) which we have quoted frustrates the "safe-for-use" principle which concededly underlay the Color Additive Amendments. Other provisions of 21 U.S.C. § 376 confer all authority needed to that end. Although the Secretary "may list any color additive for use generally in or on food. or in or on drugs, or in or on cosmetics, if the Secretary finds that such additive is suitable and may safely be employed for such general use," 21 U.S.C. § 376 (b) (2) (A), he also may list an additive "only for any more limited use or uses for which it is suitable and may safely be employed." § 376(b) (2) (B). He is authorized to prescribe broadly "the conditions under which such additive may be safely employed for such use or uses," including tolerance limitations. § 376 (b) (3). This power and the further directives in § 376(b) (4) and (5) are ample authority for the Secretary to prescribe that a particular "dye, pigment or other substance" may be used with certain diluents but not with others, or at higher concentrations with some than with others. It may well be that, in some cases, a cosmetics manufacturer might be obliged to supply an amount of information concerning a finished cosmetic equivalent to what would be required for listing. But it does not follow from this that Congress meant to subject the many thousands of cosmetic formulations to premarketing clearance — both listing and certification — at enormous and apparently unnecessary cost both to the industry and to the Government. "In our anxiety to effectuate the congressional purpose of protecting the public, we must take care not to extend the scope of the statute beyond the point where Congress indicated it would stop." 62 Cases, More or Less, Each Containing Six Jars of Jam v. United States, 340 U.S. 593, 600, 71 S. Ct. 515, 521, 95 L. Ed. 566 (1951).5

The teeth of the control over cosmetics are in 21 U.S.C. § 361, prescribing that "A cosmetic shall be deemed to be adulterated" under any one of five conditions there prescribed. The introduction of an adulterated cosmetic into interstate commerce is a prohibited act, § 331, which is subject to injunction, § 332, and constitutes a crime, § 333, and the adulterated goods are subject to seizure, § 334.

The two provisions of § 361 relevant to the FDA's hair-dye regulation are subdivisions (a) and (e). They say that a cosmetic shall be deemed adulterated

(a) If it bears or contains any poisonous or deleterious substance which may render it injurious to users under the conditions of use prescribed in the labeling thereof, or under such conditions of use as are customary or usual: Provided, That this provision shall not apply to coal-tar hair dye, the label of which bears the following legend conspicuously displayed thereon: "Caution — This product contains ingredients which may cause skin irritation on certain individuals and a preliminary test according to accompanying directions should first be made. This product must not be used for dyeing the eyelashes or eyebrows; to do so may cause blindness," and the labeling of which bears adequate directions for such preliminary testing. For the purposes of this subsection and subsection (e) of this section the term "hair dye" shall not include eyelash dyes or eyebrow dyes.

(c) If it is not a hair dye and it is, or it bears or contains, a color additive which is unsafe within the meaning of section 376(a) of this title.

Subdivision (a) was part of the original legislation concerning cosmetics, 52 Stat. 1054 (1938), and was not altered in any manner in 1960. Subdivision (e) had previously read

(e) If it is not a hair dye and it bears or contains a coal-tar color other than one from a batch that has been certified in accordance with regulations as provided by section 604.

The Regulation held invalid by the district court, 21 C.F.R. § 8.1(u), provides:

(u) The "hair-dye" exemption in section 601(a) of the act applies to those articles intended for use in altering the color of the hair and which are, or which bear or contain, color additives with the sensitization potential of causing skin irritation in certain individuals and possible blindness when used for dyeing the eyelashes or eyebrows. The exemption is permitted with the condition that the label of any such article bear conspicuously the statutory caution and adequate directions for preliminary patch-testing. If the poisonous or deleterious substance in the "hair dye" is one to which the caution is inapplicable and for which patch-testing provides no safeguard, the exemption does not apply; nor does the exemption extend to poisonous or deleterious diluents that may be introduced as wetting agents, hair conditioners, emulsifiers, or other components in a color shampoo, rinse, tint, or similar dual-purpose cosmetics that alter the color of the hair.

Although the Regulation appears to be limited to § 361(a), the court's decree was broader. The decree not only invalidated the Regulation "in its entirety" but also

any other provisions of the Regulations which have or may have the purpose or effect of defining or treating hair dyes or hair coloring products as color additives, or of requiring the listing or certification of such products, or of limiting or changing in any way the statutory exemption for hair dye products, or which may cause or result in having the provisions of Sections 601(a), 601(e), 602(e), 706, 301, 303, 304 of the Act, or of regulations issued or purported to be issued thereunder, applied to hair dyes or to hair coloring products, or to the ingredients thereof, which are exempt from the provisions of said Sections of the Act.

Taking first things first, we agree with the invalidation of so much of the Regulation as sought to deprive coal-tar hair dyes of the exemption conferred by § 361(a) in cases where, in the view of FDA, the coal-tar color ingredient carries a danger for which patch-testing provides no safeguard. The Government's argument should indeed be appealing to a legislator — what good is the warning to make a patch test if the test will not disclose the danger? But a court must take the statute as it is, and Congress wrote with great specificity. Whether it relied solely on the patch test warning because it was unaware in 1938 that coal-tar dyes might have damaging effects not detectable by such a test, as the Government asserts but the industry denies, or because it thought such instances so rare as not to warrant indentation of the exemption, the language is too clear for us to read it as meaning something different from what it so plainly says, at least in the absence of persuasive legislative history. The case fits perfectly Mr. Justice Brandeis' famous stricture, "What the Government asks is not a construction of the statute, but, in effect, an enlargement of it by the court, so that what was omitted, presumably by inadvertence, may be included within its scope." Iselin v. United States, 270 U.S. 245, 251, 46 S. Ct. 248, 250, 70 L. Ed. 566 (1926). Cf. Commissioner of Internal Revenue v. Acker, 361 U.S. 87, 80 S. Ct. 144, 4 L. Ed. 2d 127 (1959).

It is equally plain that the exemption of § 361(a) does not apply to coloring ingredients in hair dyes not derived from coal-tar; indeed we do not read the decree as prohibiting a regulation which would make that clear. We likewise see no basis for invalidating the portion of the Regulation which says that the exemption does not apply to poisonous or deleterious diluents. It is inconceivable that Congress meant to deprive the FDA of its ordinary powers with respect to other ingredients simply because they were combined with a coal-tar dye.

We think the court also erred in excluding from § 361(e) color additives in hair dyes other than those made from coal tar. The 1938 Act applied only to coal-tar colors, and the exemption of hair dyes in subdivision (e) was logical since coal-tar colors were dealt with by subdivision (a) in a supposedly adequate fashion. The modification of subdivision (e) in 1960 was part of a program to regulate all colors, and not merely coal-tar colors. But, if the statute be read with entire literalness, the unaltered introductory provision in subsection (e) now would have the effect of excluding any sanction for the use in a hair dye of an unlisted or uncertified coloring ingredient although not within the proviso to § 361 (a). In the absence of any legislative history indicating an intention to broaden the exemption in § 361(e), the most sensible construction is that, despite the retention of the introductory words "if it is not a hair dye," Congress did not mean to exempt non-coal-tar color additives used in hair dyes from the requirement of listing and certification. There is no inconsistency between this ruling and that with respect to coal-tar color additives in hair dyes. As to the latter we hold that, in the absence of any indication of a changed Congressional intent, § 361(a) continued to mean what it said and had always been thought to mean. As to non-coal-tar hair dyes we refuse to rule that the retention of an introductory phrase, appropriate when enacted in 1938, should produce a result wholly inconsistent with the basic purpose of the 1960 amendments.

We therefore affirm the holding of the district court with respect to § 8.1(f) of the Regulations, affirm in part and reverse in part with respect to § 8(u), and remand for proper modification of the decree. No costs.