JONES, Circuit Judge.
In the nineteen-thirties growers and packers of oranges in Florida and Texas began the practice of adding color to the rind of oranges. Early oranges, maturing during warm weather, will ripen to full maturity with the skin still green in color. Oranges maturing in the late winter or spring may fully ripen and undergo a “regreening” of the rind. The orange-purchasing public requires an orange-colored orange and will not accept those with green rinds. Representatives of a substantial number of those engaged in the growing and marketing of citrus fruits in the two states assert that severe economic reverses would result from a prohibition of adding color to the fruit; indeed it has been asserted that the future of the orange industry of these states is dependent upon the continued coloring of its product.
In 1938 the Congress enacted the present Federal Food, Drug, and Cosmetic Act. 52 Stat. 1040, 21 U.S.C.A. §§ 301-392. The adulteration of food in interstate commerce and the introduction of adulterated food into interstate commerce were prohibited. 21 U.S.C.A. § 331. Congressional standards for determining when food would be deemed adulterated were set up.1 Provisions [853]*853for tolerances of poisonous and deleterious substances were included.2
The Secretary of Agriculture, to whom the administration of the Act was originally committed, issued regulations on May 4, 1939, effective May 9, 1939, governing the listing of coal-tar colors.3 Among those eligible for certification were listed F D & C Orange No. 1, F D & C Orange No. 2, and F D & C Red No. 32.4 In 1954 the Secretary of Health, Education and Welfare, succeeded to the administration of the Act in the place originally held by the Secretary of Agriculture.5 The Secretary reached the conclusion that the three colors are not “harmless” as that word is used in Section 406(b) of the Act. 21 U.S.C.A. § 346(b). An order promulgating a further regulation6 was issued by which the three colors were deleted from the list approved for certification for use in foods and in drugs to be taken internally. By Section 701(f) of the Act,7 any person who will be adversely affected by an order of the Secretary may file a petition for review with the court of appeals for the circuit wherein such person resides. Petitions for review were filed in three circuits. In the Seventh Circuit the proceeding for review was dismissed. In the Second Circuit, by a decision which we shall hereafter discuss, the Secretary’s order was affirmed. Certified Color Industry Committee v. Secretary of Health, Education and Welfare, 2 Cir., 1956, 236 F.2d 866. Two petitions for review have been filed in this Circuit, one by Florida Citrus Exchange and others, orange growers and packers of Florida and Texas, and the other by Frank R. Schell, the holder of patents on the process for the coloring of citrus fruits with Orange 2 and Red 32. The two proceedings for review in this Court have been consolidated. Red 32 is the only one of the three colors with which we are concerned in this proceeding.
The three colors were on the original list, prepared in 1939, of coal-tar colors available for certification. Some of the coal-tar colors on the list became suspect as toxic and possibly carcinogenic. This prompted further investigation which culminated in the hearings which formed the bases of the order which we here review. At the hearing evidence was intro[854]*854duced that rats did not survive a diet of 1000 parts per million of Red 32. A dog fed 100 parts per million of Red 32 sickened and ultimately died. Tais was the lowest percentage tested of Red 32. Evidence was introduced at the hearing which showed that Red 32, when used for coloring oranges penetrated the skin for a depth of about Wooo of an inch; that peel of the colored orange had from 17.63 to 34.26 parts per million of color; candied orange peel had 7.4 parts per million of color; orange marmalade analyzed 1.8 parts per million of color; and the squeezed juice of colored oranges had a color content range of from fioo of one part per million to %oo of one part per million. A spokesman for the entire coal-tar color industry stated that the sale of Red 32 had been discontinued for any purpose except the coloring of oranges. The principal witness for the Secretary was Dr. Vos, Assistant Chief of the Division of Pharmacology of the Food and Drug Administration. He reviewed the tests made upon animals. He testified that on the basis of the experiments and his education and experience he was of the opinion that Red 32 was not a “harmless coal-tar color”. On cross-examination he stated that he used the word “harmless” in the absolute sense that the color was capable of producing harm. This was developed by his saying that bad results might follow the taking of two or three ounces of salt but, salt being essential, it would not be regarded as harmful. He had no evidence as to whether the color would produce harmful effects at ordinary levels of use, although he gave an affirmative answer to the question “Aren’t the amount and method of proposed use of a material necessary data for deciding whether the material is harmless for the use intended?”
The order of the Secretary which we here review made findings of fact based upon the evidence of the tests of which Dr. Vos testified. At the conclusion of nine paragraphs of specific findings as to the effects of the various quantities of the colors given to animals with mention of one episode of illness resulting to man from eating popcorn colored with Orange 1 where the percentage level did not appear, the Secretary stated:
“There was no evidence on which findings could be made concerning how much of the three colors is likely to be ingested by man from his food, drugs, and cosmetics. Some interested persons, taking their own products, attempted to show that the amounts ingested would be small to the point of insignificance. But those contentions leave aside the occurrence of the colors in the products of others, as well as the fact that upon certification of a color the Department has no means of controlling the amounts of colors used in a variety of food, drugs, and cosmetics. Nor is there authority to limit a color, once certified, to a single food—for example, F D & C Red No. 32 for use in color-added oranges.”
By the order it was determined that Red 32 and the other two suspect coal-tar colors were not harmless and suitable for use in food, drugs and cosmetics and these colors were deleted from the approved list. The legal theory upon which the order.was based is found in paragraph 2 of the Secretary’s conclusions where it was held:
“Sections 406(b), 504, and 604 of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 346(b), 354, and 364) provide for the listing of coal-tar colors that are harmless and suitable for use in food, drugs, and cosmetics. The act does not provide any method for listing toxic colors for specific food, drug, or cosmetic uses so as to limit their total use to small enough amounts that the toxicity might be disregarded. Under the statute a toxic color cannot be classified as a harmless color.”
The primary attack of the petitioners is upon the correctness of the principle as stated in the foregoing quotation.
[855]*855When the order of the Secretary was entered an effort was made to procure an amendment to the Act to require the Secretary to permit Red 32 to be used in coloring oranges until a more acceptable color should be made available. The amendment as originally offered was opposed by the Department of Health, Education and Welfare. The measure, as finally agreed upon by its proponents and representatives of the Department, amended Section 402(c) of the Act by adding thereto the following:
“Provided further, That this paragraph shall not apply to oranges meeting minimum maturity standards established by or under the laws of the States in which the oranges were grown and not intended for processing (other than oranges designated by the trade as ‘packing house elimination’), the skins of which have been colored at any time prior to March 1, 1959, with the coal-tar color certified prior to the enactment of this proviso as F.D. & C. Red 32, or certified after such enactment as External D. & C. Red 14 in accordance with section 21, Code of Federal Regulations, part 9: And provided further, That the preceding proviso shall have no further effect if prior to March 1, 1959, another coal-tar color suitable for coloring oranges is listed under section 406”. 70 Stat. 512, 21 U.S. C.A. § 342(c).
While this amendment was pending before the House of Representatives where it had its origin,8 a committee hearing was held. Before the committee was a letter from the Secretary of Health, Education and Welfare where it was stated that the evidence so far available does not establish what the lowest safe dosage of Red 32 would be to test animals, neither does it establish a likelihood of injury to man from the use of the color on the exterior of oranges at the levels of use involved. The studies, said the Secretary, were not designed to determine whether a safe tolerance could be established for a single food, such as oranges, so that the amounts that might be consumed as color on the exterior surface of oranges would pose no undue risk.9 The Commissioner of Food and Drugs appeared and gave a statement at the hearing similar to and in some respects amplifying those contained in the Secretary’s letter. He reiterated that the tests made were not designed to learn the effect upon human beings of the minute quantities of the color that might be ingested by its use as an orange coloring. He stated that although the color had been used on oranges since in the early 1930’s, no evidence has been found that injury has resulted from the consumption of oranges so colored. This use, said the Commissioner, presented a strong equity for conducting tests. He stated, as did the Secretary, the view of the department that if the color was “harmful” its use must be prohibited and that no tolerances could be fixed.10 Among the purposes of the amendment, as explained in the report of the Committee of the House of Representatives, was the allowing of time for further exploration of the toxicity of Red 32. In the report it was stated that the practice of coloring the skins of oranges has become an economic necessity for a major segment of the orange industry.11 In the report of the Senate Committee on Labor and Public Welfare, to whom the bill was referred, the same conclusions were reached.12 On this factual situation and background we review the order of the Secretary.
The petitioners list twelve specifications of error upon which they ground their attack upon the Secretary’s order. [856]*856The basic and underlying contention of the petitioners is that the Secretary has erred in construing “harmless”, as used in Section 406(b) of the Act, in an absolute rather than in a relative sense. If the position of the Secretary is correct, he has no choice but to refuse to permit a coal-tar color to be used in food in minute, safe and harmless quantities if, when used in much larger quantities it would or might be injurious or dangerous to health. If the position of the Secretary is correct he has no power to permit a particular coal-tar color to be used for a specific and restricted purpose even though, when so used, it would not be injurious or dangerous to health.
The Secretary has urged that the questions- presented for our review are the identical questions which were decided by the Second Circuit in Certified Color Industry Committee v. Secretary of Health, Education and Welfare, supra, and that we should follow that decision unless we are convinced it was clearly wrong. Where an identical issue is presented and no different principle is involved, the decision of another court of appeals is entitled to great weight and affords persuasive argument.13 Although the appeal in the Certified Color Industry case was from the same order and was upon the same record as the case before us, the identical issue is not presented and a different principle is involved. In the Second Circuit case three coal-tar colors were involved and the uses were all inclusive. The difference was recognized by the Court of Appeals for the Second Circuit. In its opinion it is said:
“Thus the problem is far different from the one presented recently to Congress when the act was amended to permit the use of Red 32 on orange skins not intended for processing. With only one product contaminated it was not too difficult to argue that such small amounts were
involved that a man would have to drink 5000 gallons of orange juice a day before experiencing any adverse effects.” 236 F.2d 871.
The problem we consider is not that which the Second Circuit had and the decision there is, as we hope to show, not in conflict with what we decide here.
In a brief filed with us by amicus curiae, he quotes from a letter to him from the Food and Drug Administration where it was said:
“Let me say at the outset that this Administration is in agreement with your conclusion that the usé of added color is inherently deceptive and is contrary to the interests of both consumers and producers.” This letter is not in the record and would not be the subject of comment by us but for a suggestion that it evidences a hostile and prejudiced attitude of the Food and Drug Administration. We think the record shows the contrary to be the case. It is shown that all oranges to which color is added are so marked and that immature fruit is not colored. The record reflects a departmental attitude of fairness and objectivity, and discloses that the ill-advised statement of the writer of the letter does not express the view of the Secretary or of the department.
It has been contended that the officers charged with the administration of the Act had, from the time the Act was passed until the time of the investigation which culminated in the hearings and in the order we here review, construed the Act as permitting, if not requiring, the fixing of tolerance for harmless uses of Red 32. Such administrative construction should be persuasive, so we are told, if not binding as an administrative construction of an ambiguous provision. The Secretary, on the other hand, asserts that the administrative interpretation has not altered, and that the change of position resulted from the [857]*857different findings of fact required by the recent investigations and experiments. We need not pursue this inquiry. An administrative construction which is clearly erroneous is not binding on the administrative agency nor is it persuasive in judicial proceedings,14 and this should be particularly true where the purpose of the Act is to protect the health of the public.15 The doctrine of giving weight to an administrative construction of a statute does not apply to an administrative determination of the administrator’s statutory powers.16 This, in essence, is one of the questions before us.
The decision we are to make is whether the Secretary of Health, Education and Welfare is required or permitted to determine if there is a minimum quantity of the coal-tar color designated as F D & C Red No. 32 which can be used in adding color to the skins of mature oranges without danger of impairing the health of those who consume such oranges; and if so required or permitted, and if it be determined that there is such minimum, should the Secretary list and certify such color for such use. In the construction of statutes courts will look to the entire act and to its objects and policy.17 Results which are incongruous18 or absurd19 are to be -avoided. Where a particular construction of a statute will occasion great inconvenience or produce inequality and injustice, that view is to be avoided if another and more reasonable interpretation is present in the statute.20 In the construction of a legislative measure, statutes in pari materia, amendatory acts, and their legislative history may be examined.21 These rules are the primary guides to our solution of the problem before us.
In the leading and often cited Lexington Mill case 22 the Supreme Court declares the purpose of the Act, saying:
“The statute upon its face shows, that the primary purpose of Congress was to prevent injury to the public health by the sale and transportation in interstate commerce of misbranded and adulterated foods. * * * As against adulteration, the statute was intended to protect the public health from possible injury by adding to articles of food consumption poisonous and deleterious substances which might, render such articles injurious to the health of consumers.” 232 U.S. 409, 34 S. Ct. 340.
The purpose to protect the public health is an important aim of the statute,23 and touches phases of the lives and health of people which are largely beyond self-protection.24
[858]*858We are here reviewing the order of the Secretary only with respect to the coal-tar color F D & C Red No. 32 as used in the coloring of the skins of oranges. If there be other like situations, then of course that which we hold will be applicable to them. We are not confronted with the broad questions which were before the Second Circuit in the Certified Color Industry case, supra.
In some of the cases25 the de minimis doctrine has been recognized as applicable or has been applied in adulterated food cases. The contentions of the petitioners suggest that thejr regard the doctrine applicable here. Neither the Secretary’s findings nor the evidence on which they were based calls for any consideration of the doctrine upon the record before us.
The words “harm”, “harmful”, and “harmless” are terms of relation. In this sense they resemble “wrong” and “wrongful”.26 Red 32 is harmless, i. e. not harmful, to all persons under the protection of the Act while it remains in the vat or the vial. An injury must occur before harm results. The color Red 32 becomes harmful, i. e. not harmless, when it is consumed, and then only if consumed in such quantity that injury or harm might result. If there: can be a use of the color in such small quantities that it can be consumed without risk of injury or harm then, in such quantities, it is harmless. A person Í3 as unharmed after consuming a minute harmless quantity of Red 32 (if such there be) as he would be had he consumed none of it.
In United States v. Lexington Mill & Elevator Co., supra, the Supreme Court reviewed the condemnation of a lot of flour to which had been added a poisonous ingredient but in a quantity so small that the health of a consumer could not be thereby injured. The statute then provided that an article of food should be deemed to be adulterated if it “contain any added poisonous or other added deleterious ingredient which may render such article injurious to health”. It was contended there, as it is contended here, that it is the character-—-not the quantity—of the added substance which is to determine the case. This contention was rejected. The Court said;
“If it cannot by any possibility, when the facts are reasonably considered, injure the health of any consumer, such flour, though having a small addition of poisonous or deleterious ingredients, may not be condemned under the act.” 232 U.S. 411, 34 S.Ct. 341.
As in the Lexington Mill case, it was held in the Seventh Circuit that “injurious” is a relative rather than an absolute term. W. B. Wood Mfg. Co. v. United States, 7 Cir., 1923, 286 F. 84. We think “harmless” is also to be so construed.
The Secretary would concede, apparently, the possibility that a person might consume Red 32 in such small quantities that no hazard to health would result. His guarded statement is that “the evidence so far available does not establish a likelihood of injury to man from the minute amount likely to find its way into the human diet from the consumption of such colored oranges at the level of use involved.” 27 The Commissioner of Food and Drugs has admitted that there was no evidence of injury to consumers of colored oranges.28 The Secretary has made no effort to ascertain whether or not there is, in fact, any [859]*859likelihood of injury to the health of persons who consume oranges colored with Red 32. Taking the position that no quantitative test is authorized, the Secretary disclaims any power to fix a tolerance for the use of Red 32 in the coloring of oranges or for any other particular use. Under this theory it would make no difference whether or not there was a minimum quantity of Red 32 which would not be injurious. Malum in uno, malum in omnibus, is the departmental canon of construction.
Unless the quantities of Red 32 used in the coloring of oranges are dangerous to public health, the objects and purposes of the Act will not be promoted by a construction which prohibits that use in such quantities. The statute which we construe provides for the fixing of tolerances for added poisonous and deleterious substances required in the production of food 29 and, by a 1954 amendment,30 for the fixing of tolerances for poisonous pesticides on raw agricultural commodities. We cannot see how the objects and purposes of the statute would foe furthered by a construction which permits safe, i. e. harmless, quantities of poisons other than coal-tar colors to be added to food where required in its production, and which permits safe tolerances of poisons for pesticides, but prohibits the use of a coal-tar color in the most minute and harmless quantity because it is harmful and injurious in large quantities. Such a construction would be an unwarranted discrimination, not so much against the coal-tar color and the manufacturers of it, but against that important segment of the orange producers who are economically dependent upon it. The construction of the Secretary results, or may result, in inequality and injustice. There is, we think, a more reasonable interpretation in the Act.
The Secretary found that Red 32 is toxic; in other words that it is poisonous. The evidence justified the finding. If the Secretary is correct in his construction of “harmless”, Section 402(a) of the Act is inapplicable to coal-tar colors. If his construction is erroneous, Section 402(a) would apply. If Section 402(a) applies, the colored oranges would be “adulterated” only if the added poisonous substance was unsafe within Section 406(a). Under this section poisonous substances are permitted to be added to food if it “is required in the production thereof” within such limits as to quantity or extent as the Secretary finds necessary for the protection of public health. The Committees of the Senate and the House af Representatives each found that the practice of coloring oranges has become an economic necessity for a major segment of the orange industry sinee a high percentage of the oranges grown in Florida and Texas would meet with strong consumer resistance if not colored.31 It is as important for a food to be marketable as for it to be palatable and, if the findings of the Committees be correct, then the addition of color is “required in the production”, as that phrase is used in the Act. Section 402(a) and Section 402(c) are not mutually exclusive. A coal-tar color may not be added unless from a certified batch. By Section 406(b) only those coal-tar colors may be certified which the Secretary has listed as harmless. The Secretary and his colleagues in the Department did not oppose the enactment of the 1956 amendment to Section 402(c). A reason given was that during the period of the moratorium a “nontoxic color” might be developed. It is the contention of the petitioners that no colors other than Red 32 and the others banned by the Secretary’s order are effective. If another color be found which is not a coal-tar color but is as much or more toxic, then we see no reason to doubt that the Secretary would be required to deal with it under Section 406(a) and to provide tolerances for its use in the event it could be used in any limited quantity without being injurious [860]*860to the public health. The Secretary, in his order, has stated that “the Department has no means of controlling the amounts of colors used in a variety of foods, drugs and cosmetics”. It is not apparent that this is a problem with respect to other toxic substances which are added to foods under tolerances fixed by the Secretary and nothing is shown to indicate that it would be a problem in dealing with a coal-tar color. We see no reason for supposing that Congress intended to permit-the use of toxic substances, including toxic colors other than coal-tar products, in safe and harmless quantities and to prohibit the use of toxic coal-tar colors in safe and harmless quantities. We do not ascribe such an intention to Congress. Another reason given by the Secretary for not opposing the 1956 amendment was that the moratorium for Red 32 would permit further exploration of its toxicity. It is the view of the Secretary that if it be found that the coloring of oranges with Red 32 is not injurious to humans and can be continued with safety, further legislation will be required. With this conclusion we do not agree.
We construe the Act as requiring the Secretary, under Section 406(a), to determine whether the use of the col- or Red 32 is required in the production for market of oranges grown by a substantial segment of the orange producing industry; and if he finds that such usé is so required, then to determine the quantity, if any, that can be tolerated as safe, and by regulation to limit the quantity to such extent as he finds necessary for the protection of public health. With “harmless”, as used in Section 406(b), construed as a relative rather than an absolute term, regulations promulgated by the Secretary under that section as well as under Section 406(a) could be issued if the prerequisite findings as to the necessity for using color had been made and the tolerance limits had been fixed.
In the order of the Secretary it is stated that there is no “authority to limit a color, once certified, to a single food—for example, F D & C Red No. 32 for use in color-added oranges”. Although there may be no express statutory provision authorizing the limiting of the use of a color to a single food, it is not prohibited. If the adding of a toxic substance to one food may be harmful, but harmless when added to another food in limited quantities fixed by a predetermined tolerance, we are of the opinion that the Secretary is empowered to permit its use in the latter instance. The Secretary at one time seemed of the same opinion. In the Food and Drug Administration Regulations promulgated in 1955, a number of coal-tar colors were listed for certification for use in food, drugs and cosmetics. Among these are colors designated as “lakes”. The regulations provide that “No lake listed in paragraph (a) of this section shall be certified for any use in food except external application to shell eggs.” 32 We shall not indulge in speculation upon the comparative porosity of egg shell and orange peel. We merely observe that the Secretary restricted the use of a group of coal-tar colors to a single food, and we think he was fully empowered to do so.
We are not unmindful of the admonition of the Second Circuit in the Certified Color Industry case, supra, that:
“Now that their toxicity has been amply demonstrated it would be unconscionable for any court to require the Secretary to permit their use without the clearest and most uncompromising evidence that usage at certain levels was absolutely safe.” 236 F.2d 870.
With this statement we are in accord. The findings of fact are to be made by the Secretary. Such findings, under Section 701(f) (3) of the Act,33 are conclusive if supported by substantial evi[861]*861denee.34 We again advert to the Secretary’s letter to the Chairman of the Committee of the House of Representatives at the time it was considering the 1956 amendment. The Secretary said:
“While the scientific evidence so far available does not establish what the lowest safe dosage would be to test animals, neither does it establish a likelihood of injury to man from the use of this color [Red 32] on the exterior of oranges at the levels of use involved. Before final conclusions can be drawn on this point, however, it is necessary to conduct adequate scientific studies of chronic toxicity on laboratory animals over their life span, which will involve feeding tests at levels related to the quantities of such externally applied color that might enter man’s diet. Such tests and studies will require approximately 3 years. We assume that the industry will make such studies.” 35
Since the Secretary is authorized by Section 702(a) of the Act36 to conduct examinations and has a technical staff already familiar with the problem, we think it not amiss, in view of our decision here, that we suggest consideration be given by the Department to continuing or resuming the tests and studies from which the evidence was obtained which was adduced before the departmental hearing.
In the Certified Color Industry case, supra, the court rejected the Secretary’s construction of the word “harmless”. It holds that the government need not prove that a coal-tar color is in fact injurious to health but it is enough if it shows that it “might render the article of food injurious”, adopting the quoted words from the Lexington Mill case. Because, in the quantities tested, the possibility of injury to health was indicated, the court sustained the order. It held that “under the circumstances disclosed in the record” the petitioning manufacturers of coal-tar colors could not require the Secretary to establish tolerances fixing safe levels of use. For this position two reasons were assigned, first that safe levels have not been established, and second, the department could not control uses within tolerance limits if such limits were prescribed. In connection with the first of these grounds the court comments that the industry made no investigation as to the effect of the colors at low levels of usage. The record bears out the court’s observation, but under the Secretary’s construction of the Act the effect of the colors at minimal levels would be immaterial. The inability to control the quantity used is no different, as we see it, in dealing with a coal-tar color than with respect to other toxic or poisonous substances where the Secretary is required to establish tolerances. In any event, as the court in the Certified Color Industry case points out, the situation before it was different from the case of a single color such as Red 32 applied to a single use such as the coloring of oranges. Such a case is the one before us.
After a hearing in 1939 Red 32 was found to be harmless and suitable for use in foods, drugs and cosmetics. The situation as presented on this review was briefly and accurately characterized in a letter from the Commissioner of Food and Drugs to Florida Senators and Congressmen written February 8, 1955. In speaking of Red 32 the Commissioner said:
“Recent investigations show that these colors when fed in substantial amounts, show evidence of toxicity. There is, however, no evidence that, in the amounts used, and in the manner of use, in the coloring of citrus fruit, the product so colored is not safe for human consumption.”
Unless there is evidence that, in the amounts used, and in the manner of use, oranges colored with Red 32 are unsafe for human consumption, the 1939 find[862]*862ing should not be supplanted by a contrary ruling by which such use is prohibited. It may be parenthetically observed that “unsafe” is the yardstick for determining tolerances under Section 406(a). A determination may be made by the Secretary on his own initiative or upon the petition of any interested person37 as to whether coloring is required and, if so, as to whether Red 32 is a harmless coal-tar color for use in coloring oranges and to fix the tolerances, if any, as may be proper for such use. .
At the outset of this proceeding, upon agreement, a stay order was entered to permit the use of Red 32 in coloring oranges while the proceeding was pending. Except as to duration, the substance of the stay order’s provisions is followed in our disposition of the matter.
The order promulgated by Marion B. Folsom, Secretary of Health, Education and Welfare, on November 10, 1955, published in the Federal Register on November 16, 1955, (20 F.R. 8492) will be set aside in so far as said order removes the coal-tar color F D & C Red No. 32 from the list of colors which may be certified for use in coloring the skin of oranges meeting minimum maturity standards prescribed in the State of Florida and Texas; provided, that nothing herein or in the judgment of this Court entered pursuant hereto shall restore said coal-tar color to the list of colors which may be certified for unrestricted use in food, drugs and cosmetics but shall operate to authorize the certification of batches of said color conforming to the specifications for the color appearing at 21 C.F.R. 135.3 (1949 ed.) for the purpose of coloring the skin of mature oranges only; provided further, that the Secretary shall be required to certify only sufficient batches of F D & C Red No. 32 as may be necessary to col- or the skin of mature oranges from time to time; provided further, that the certificates issued for batches of F D & C Red No. 32 may be limited by their certificate for use in coloring mature oranges only; and provided further, that nothing herein or in the judgment of this Court entered pursuant hereto shall be deemed to restrict the Secretary from making further investigations and conducting hearings for a determination of whether the use of Red 32 is required in the production of oranges and to determine the tolerances, if any, that are safe and harmless, as harmless is herein construed and defined.
Judgment accordingly will be entered setting aside in part, as herein set forth, the aforesaid order of the Secretary.