These are petitions to review an order of the Secretary of Health, Education and Welfare delisting certain coal-tar colors (FD & C Orange No. 1, Orange No. 2 and Red No. 32) manufactured by petitioners. Prior to this order the dyes in question had been certified since 1939 as “harmless” and placed on the approved list for unrestricted use in foods, drugs and cosmetics. Prompted by hearings before a Select Committee of Congress to Investigate the Use of Chemicals in Food concerning the toxicity and possible carcinogenicity of coal-tar colors, respondent conducted extensive experimental research with laboratory animals which conclusively showed that the colors in question produced substantial deleterious effects and sometimes death. This information and experimental data was given to petitioners sufficiently in advance of the hearing to permit them to offer contradictory evidence. After the hearing on notice, at which petitioners offered no opposing scientific data, the Secretary found the dyes to be “not harmless and suitable for use in coloring food or for use in coloring drugs or cosmetics intended for other than external application,” and ordered that they be deleted from the listing.
This order was made November 10, 1955, effective February 10, 1956. Petitioners filed the instant Petition to Review on February 7, 1956, three days before their time would have expired. On January 27, 1956, only twelve days before, they made a motion before the Secretary to reopen the hearings to receive evidence of the maximum extent to which these colors are used in foods (but not of drugs or cosmetics) under normal conditions of use, which were represented to be far less than the levels of administration causing damage to the test animals. This motion was denied on February 20, 1956, on a number of grounds, viz., (1) as to the color Red 32 (a) because no safe levels were found for animals and (b) this color recently caused 196 persons (adults and children) to become acutely ill after eating popcorn with this color added; (2) man appears to be much more susceptible to color dyes than test animals; (3) the proposed proof was inaccurate and incomplete and did not take into account all uses; (4) the statute did not permit tolerances for toxic colors; (5) although the tests proved the colors toxic they did not establish the extent of toxicity to a certainty so as to permit the establishment of safe tolerances, and (6) the Secretary had no authority to establish tolerances with regard to colors.
Petitioners suggest that the basic question for review is whether the statutory language “harmless and suitable for use”, Federal Food, Drug, and Cos*868metic Act, § 406(b), 21 U.S.C.A. § 346 (b), is to be interpreted in an absolute sense or in a relative sense meaning “incapable of producing harm under normal conditions of use.” Subsidiary to this, they say, is the question whether the Secretary is authorized to impose limitations or tolerances under which colors may continue to be certified. It is their position that the legislative intent can be found in the prior construction by respondent’s predecessors which Congress was aware of since it knew that coal-tar colors were toxic and must have realized that their authorized use would have to have some relation to the quantities used. All of this follows, it is argued, since many substances, e. g., salt and vinegar, are concededly not harmless in an absolute sense. The second string to their bow is the argument that the Secretary should authorize limitations or tolerances for these colors for normal use the same as the Secretary does with regard to poisonous or deleterious substances added to food but which are necessary or unavoidablé in good manufacturing practice. Federal Food, Drug, and Cosmetic Act, § 406(a), 21 U.S.C.A. § 346(a).
We turn to the legislative history of the act in an effort to resolve the question whether “harmless” should be construed in a relative or absolute sense. The 1906 Act provided that food should be deemed adulterated “if it contain [s] any added poisonous or other added deleterious ingredient which may render such article injurious to health.” Food and Drugs Act of 1906, § 7, 34 Stat. 769. Although no specific provisions were then enacted dealing with coal-tar colors, the Food and Drug Administration early recognized the necessity of special treatment for these dyes. Without express statutory authority it issued regulations to the effect that only harmless colors might be used and proceeded to certify for use only those coal-tar dyes which satisfied its specifications. This practice continued until the act was amended in 1938. Under the amended statutory scheme a food is deemed adulterated, “If it bears or contains any poisonous or deleterious substance which may render it injurious to health; but in case the substance is not an added substance such food shall not be considered adulterated under this clause if the quantity of such substance in such food does not ordinarily render it injurious to health * * Federal Food, Drug, and Cosmetic Act, § 402(a), 21 U.S.C.A. § 342(a). Section 402(c), 21 U.S.C.A. § 342(c), was added to deal specifically with the problem of coal-tar colors. It provides flatly that a food shall be deemed adulterated if it contains a coal-tar color other than one from a batch certified under § 406(b), 21 U.S.C.A. § 346(b). This latter section authorizes the Secretary to “promulgate regulations providing for the listing of coal-tar colors which are harmless and suitable for use in food and for the cerification of batches of such colors * * Section 406(a), 21 U.S.C.A. § 346(a), gives the Secretary permission to set up tolerances for added poisonous or deleterious substances which are required or cannot be avoided by good manufacturing practice. At least to the extent that the amendment authorized the certification of coal-tar colors, it codified the administrative practice of the past 30 years.1
*869Petitioners argue that the 1938 amendment, insofar as these sections are concerned, amounts to nothing more than legislative recognition of existing practice and procedure; that the word “harmless” must be equated to “added poisonous or other added deleterious ingredient which may render such article injurious to health”, and that these words must be given the meaning assigned to them in Wood Mfg. Co. v. United States, 7 Cir., 1923, 286 F. 84. The Wood case adopted a standard of relativity and held that an infinitesimal quantity of arsenic in a coal-tar color was not a poisonous or deleterious substance, injurious to health, as ordinarily used, and that the government was required to prove that the substance was in fact injurious as used. Respondent, on the other hand, urges that the use of a different word — “harmless”—instead of the more familiar statutory language is evidence of Congressional intent to provide an absolute standard. We are unable to agree with either of these interpretations.
However, we do agree that the word “harmless” as used in § 406(b) must have some relation back to § 402 (a). But the provisions of that section, as amended, are substantially different from those of its predecessor — § 7 of the 1906 Act. The new provision treats deleterious substances separately depending upon whether they are added to or are an integral part of the food. Thus, if the deleterious substance is inherent in the food there nevertheless is no adulteration if the quantity of the substance used does not ordinarily render it injurious to health. But if the complained of ingredient is an added one, then the food is adulterated if the substance may render it injurious to health. This, so far as added substances are concerned, would seem to indicate legislative discontent with the Wood case, and a return to the principles enunciated by the Supreme Court in United States v. Lexington Mill & Elevator Co., 1914, 232 U.S. 399, 34 S.Ct. 337, 58 L.Ed. 658, wherein it was said:
“It is not required that the article of food containing added poisonous or other added deleterious ingredients must affect the public health, and it is not incumbent upon the government in order to make out a case to establish that fact. The act has placed upon the government the burden of establishing * * * that the added poisonous or deleterious substances must be such as may render such article injurious to health. The word ‘may’ is here used in its ordinary and usual signification, there being nothing to show the intention of Congress to affix to it any other meaning. It is, says Webster, ‘an auxiliary verb, qualifying the meaning of another verb, by expressing ability * * * contingency or liability, or possibility or probability.’ In thus describing the offense, Congress doubtless took into consideration that flour may be used in many ways, in bread, cake, gravy, broth, etc. It may be consumed, when prepared as a food, by the strong and the weak, the old and the young, the well and the sick; and it is intended that if any flour, because of any added poisonous or other deleterious ingredient, may possibly injure the health of any of these, it shall come within the ban of the statute. 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. at page 411, 34 S.Ct. at page 340.
By restricting the ordinary usage test to non-added substances, the inference is inescapable that Congress meant to similarly restrict the Wood case and to correspondingly restore the doctrine of the Lexington Mill case. Definition of “harmless” along these lines can scarcely be considered too strict in view of the stringent provisions relating to coal-tar colors. Section 402(c) deems food adul*870terated if it contains such a. dye, regardless of whether it could possibly cause any harm, if it does not come from a certified batch.
Thus, the government need not prove that the use of Orange 1 and 2, and Red 32, is in fact injurious to health. It is enough if it shows that such addition “might render the article of food injurious.” United States v. Lexington Mill & Elevator Co., supra, 232 U.S. at page 410, 34 S.Ct. at page 340. This it has done by the overwhelming weight of the evidence. Without analyzing the scientific data in the record, suffice it to say that the uncontradicted testimony shows that the least toxic of the three colors in question had serious immediate effect on human beings and that the overall effect on the laboratory animals by proof on autopsy was alarming. Consequently, the delisted colors are not “harmless” within the meaning of § 406(b).
The subsidiary question raised by the denial of a rehearing, relating to the Secretary’s authority to impose tolerances under which the colors may continue to be used need not be answered. All we need' decide is whether or not petitioners can require him to establish such tolerances. Under the circumstances disclosed in the record we hold that they have no right to do so. There are two reasons for this. In the first place there is nothing in the record to indicate how much of these coal-tar colors a human being could ingest over a period of years, without any harmful effects, nor did petitioners propose to introduce any. The respondent’s conclusion in this connection is as follows:
“While a safe level of administration to test animals is disclosed by the record in the case of FD & C Orange No. 1, the record also discloses that the color has adverse effects upon man at a level well below the safe level of administration to test animals. The safe level of administration of FD & C Orange No. 2 to test animals is well below the level at which FD & C Orange No. 1 was found safe to test animals. It is therefore even more toxic than FD & C Orange No. 1. No safe level of administration was found even in test animals for FD & C Red No. 32.”
Furthermore, it may be impossible to demonstrate the precise cumulative effect over 20 or 30 years of consumption of specified amounts of color and therefore impossible to prescribe safe tolerances. We think that the Secretary’s findings that such tolerances cannot be determined should be accepted, and that they are dispositive of the case. The colors here involved have been shown to be alarmingly toxic. To be sure, it is not certain that continued consumption at present rates will be harmful to a normal consumer. But it is sufficiently likely to make the Secretary’s decision a completely reasonable one. Failure on his part to provide an adequate margin of safety might well constitute an abuse of discretion. Cf. Atlas Powder Co. v. Ewing, 3 Cir., 1952, 201 F.2d 347, 355, certiorari denied Glyco Products Co. v. Federal Sec. Adm’r, 1953, 345 U.S. 923, 73 S.Ct. 781, 97 L.Ed. 1355. The tests were conducted for the sole purpose of determining whether or not the colors were harmless. 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. It must be remembered that while petitioners did nothing to obstruct the accumulation of data on this point they certainly did nothing to advance the cause of science. The industry made no pharmacological investigation of its own after 1938, submitted no evidence and readily accepted the results of the studies made by the Food and Drug Administration.
But quite aside from this lack of any evidence in the record indicating a safe level of consumption, there is, as a practical matter, a far more serious objection to requiring a schedule of tolerances. Relying on Exhibit A to the petition to *871reopen the hearings it appears that the colors in question are used in cakes, cookies, pies, bread, cheese, ice cream, frankfurters, bologna, spreads, oranges, canned and frozen vegetables, candy, desserts and puddings, soft drinks, condiments, soups, pickles, olives and prepared dishes, as well as in unenumerated drugs and cosmetics. Assuming for the moment that the Secretary could determine that X miligrams of coal-tar color per day could be taken for 25 years with absolute safety — How, considering the vast number of items which contain the dyes and the extremely diverse consumer habits of the American public, could he possibly guard against the ingestion of more than X milligrams per day? The short answer, of course, is that he could not. He has made the following specific finding:
“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, FD & C Red No. 32 for use in color-added oranges.”
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. Hearings Before the Subcommittee on Health and Science of the House Committee on Interstate and Foreign Commerce, 84th Cong., 2d Sess. at 13 (1956).
In view of the impossibility of controlling the intake of coal-tar colors, even assuming maximum tolerances per item could be established, the Secretary acted within his discretion in refusing to reopen the hearings. Cf. Federal Security Administrator v. Quaker Oats Co., 1943, 318 U.S. 218, 63 S.Ct. 589, 87 L. Ed. 724.
Order affirmed.
. Parenthetically it may be noted that on July 9, 1956, the President signed into law a bill further amending § 402 by the addition of a provision permitting the use of Red 32 on the skins of oranges not intended for processing, such permission to last only until March 1, 1959 or until another suitable color is listed under § 406, whichever is sooner. Pub.L. No. 672, 84th Cong., 2d Sess., Ohapt. 530 (July 9, 1956). What is significant, however, is the vast difference between the bill as proposed and the bill as passed. The former provided for the promulgation of regulations respecting the listing of coal-tar colors for use on orange skins “which are safe in the manner in which used and suitable for such use.” The Secretary objected to this but did not oppose the bill in its final form. Hearings Before the Subcommittee on Health and Science of the House Committee on Interstate and Foreign Commerce, 84th Cong., 2d Sess. at 1-3 (1956).