In this consolidated appeal, petitioners Standard Oil Company of California (Standard), and Batten, Barton, Durstine & Osborn, Inc. (BBD&O), challenge a decision and a cease and desist order issued by the Federal Trade Commission (FTC or Commission). The Commission found that Standard and BBD&O had broadcast advertising that was false, misleading, and deceptive, in violation of section 5 of the Federal Trade Commission Act, 15 U.S.C. § 45. Standard and BBD&O contend that the advertising was accurate and in all respects lawful. They further state that even if the advertising were improper, the cease and desist order issued by the FTC is overly broad. We affirm the Commission’s finding that the advertising violated section 5 of the Act, but we agree that the cease and desist order is unwarranted and therefore direct that the scope of the order be modified.
The case concerns three television commercials that promoted a gasoline additive known as F-310. The commercials were broadcast on television from January 9 to June 9, 1970. Two commercials were based on demonstrations which were designed to afford viewers a visual comparison of automobile exhaust before and after using F— 310. See Appendix I. In one, a large clear balloon was attached to the exhaust pipe of an idling automobile. The balloon was shown inflating with black, opaque vapor, while the announcer described it as “filling with dirty exhaust emissions that go into the air and waste mileage.” The announcer then stated that “Standard Oil of California has accomplished the development of a remarkable gasoline additive, Formula F-310, *656that reduces exhaust emissions from dirty engines.” He informed the viewer that the same car was run on six tankfuls of Chevron F-310 and the result was “no dirty smoke, cleaner air.” To prove the point, a clear balloon was again shown being attached to the car. This time the balloon inflated with transparent vapor. In conclusion the television viewer was told: “Chevron with F-310 turns dirty smoke into good, clean mileage. There isn’t a car on the road that shouldn’t be using it.”
The second commercial, known as the bag demonstration, was similar to the commercial just described except that the automobile was completely enclosed in a transparent bag. In the before segment of the demonstration, the automobile was hidden from view by black exhaust smoke. In the after segment the automobile was plainly visible, thus illustrating the effects of using the F-310 additive.
The third commercial considered by the Commission focused on a meter dial labeled “exhaust emissions.” The dial showed a scale from 0 to 100. The word “clean” labeled the 0 extremity, and the word “dirty” was printed under 100. Once again a before-and-after format was used, with the meter pointing to 100 before use of F-310 and to 20 afterwards. The announcer’s accompanying message was: “[AJfter just six tankfuls of Chevron F-310 — exhaust emissions reduced.” The commercial again closed with the message that: “There isn’t a car on the road that shouldn’t be using it.”
On May 19, 1970 the FTC advised Standard that it objected to the advertisements because it was not clear that the car depicted in the before segment of each commercial had been driven previously with a gasoline that was deliberately formulated to accelerate carbon deposits, resulting in an especially dirty engine. The FTC expressed concern that the commercials did not indicate either that the degree of improvement in gasoline mileage and pollution reduction would vary according to the condition of the engine, or that the gasoline used to prepare the test cars for the before part of the sequence caused much dirtier engines than would have been caused by normal gasoline.
Standard submitted an assurance of voluntary compliance to the Commission on May 25, 1970. Beginning June 10 the commercials were altered by superimposing on the film phrases such as, “Very Dirty Engine Purposely Used to Provide Severe Test,” and “Degree of Improvement in Your Car Depends on Condition of Engine.”
On December 29, 1970 the Commission filed a complaint against Standard and BBD&O based on the commercials as they were presented prior to June 10,1970, specifying eleven charges of false advertising. The charges were sweeping. The FTC alleged not only that the ads were misleading in the respects indicated above, but also that F-310 did not reduce pollution to any significant degree.1 After extended hear*657ings, the administrative law judge found that F-310 was effective as claimed, and that it did result in a significant reduction in exhaust emissions. He further found that the commercials accurately depicted the product’s performance. Accordingly he recommended that the administrative complaint be dismissed in its entirety.
The Commission adopted substantially all of the findings of the administrative law judge pertaining to the efficacy of the product. The FTC’s charges that the product did not reduce pollution were found to be wholly without merit. The Commission agreed they should be dismissed and stated that the development of F-310 was “laudable.”
The Commission did not, however, adopt the findings of the administrative law judge as to certain of the alleged disparities between the claims which the commercials were interpreted to make and the data offered to substantiate those claims. Specifically, the Commission found that the commercials falsely represented (1) that use of F-310 would result in a complete reduction of air pollutants; (2) that all cars would show the same degree of improvement as was illustrated by the reduction of pollution in the exceptionally dirty engine; (3) that the use of F-310 would affect all types of exhaust emissions; and (4) that the meter used in the third commercial portrayed an eighty percent reduction of pollution with F-310. Based on these findings the Commission held that Standard and BBD&O had violated section 5 and issued broad cease and desist orders against them.
Before turning to an analysis of the cease and desist orders and a determination of their propriety, we examine the petitioners’ contention that there is no substantial evidence in the administrative record to support the Commission’s finding that the commercials were misleading and deceptive, and BBD&O’s additional contention that there is no support for the finding that it knew or should have known that the three advertisements were deceptive. The court has jurisdiction over this appeal under section 5(c) of the FTC Act, 15 U.S.C. § 45(c).
The Commission’s Findings
Although the law is intended to protect “that vast multitude which includes the ignorant, the unthinking and the credulous,” Aronberg v. FTC, 132 F.2d 165, 167 (7th Cir. 1942), neither the courts nor the Commission should freely speculate that the viewing public will place a patently absurd interpretation on an advertisement. On the facts before it, the Commission could not properly have found that these ads claimed that F-310 removed all harmful chemical emissions from automobile exhaust. We do not think that any television viewer would have a level of credulity so primitive that he could expect to breathe fresh air if he stuck his head into a bag inflated by exhaust, no matter how clean it looked. If that were the whole case, we would agree with the petitioners and overturn the Commission’s order in its entirety. But the Commission’s opinion made a different point. The Commission was concerned with the ads’ implicit statements about the degree of reduction in auto emissions and the effect of that reduction on the pollution problem.
Exhaust emissions are of two types: visible and transparent. The visible elements (black smoke and particulate matter) contribute less to pollution than the transparent elements. The Commission determined that the public is not aware of this fact. In support of this - finding, it cited a study made by BBD&O with respect to the very commercials in question here. BBD&O, in reporting on this study to Standard, concluded from a survey it had made that only fourteen percent of motorists are aware *658that most polluting elements in automobile exhaust are invisible.2
Hydrocarbons and carbon monoxide, two common exhaust pollutants which generally are transparent, were reduced by fifty percent and thirty-three percent, respectively, after use of F-310 in tests run on cars with very dirty engines. The additive had no effect at all on lead compound auto emissions which are also prime contributors to air pollution. The ads gave the public no indication whatsoever of the degree of reduction of invisible pollutants with use of F-310. Moreover, since the public, at least before it analyzes the question fully, is unaware that the visible pollutants eliminated by F-310 are not the most troublesome part of exhaust pollution, the complete disappearance of black smoke may indicate a greater pollution reduction than is actually achieved by the product. References in the ads to “no dirty smoke, cleaner air” and “turns dirty smoke into good, clean mileage” reinforced that erroneous impression.
We would be less confident in upholding the Commission’s finding of a section 5 violation were there not other, more palpable, misrepresentations besides the overall misleading impression as to F-310’s impact on the problem of air pollution. But there were other misrepresentations, and these lend support to the Commission’s conclusion that the advertising had a capacity and tendency to mislead.
The Commission found that a substantial portion of the viewing audience would believe that the automobiles in all three advertisements were representative of most automobiles, and petitioners do not cite evidence in the record to contradict that conclusion. That representation, clear and implicit in the ads, simply was not true. Heavy engine deposits were built up in the test vehicles by use of a special, extra-dirty fuel to show a more dramatic difference between the before-and-after sequences. It is conceded that only a small percentage of cars on the road would have similar engine conditions before using F-310. Thus most cars, although they would be cleaned by the detergent effects of F-310, would not show as large an improvement as was demonstrated by the ad. The statement made in each advertisement that “[t]here isn’t a car on the road that shouldn’t be using it” seemed to promise similar results for all cars, thus reinforcing the misleading impression that the condition of the demonstration vehicle was representative of the condition of most automobiles. There is adequate evidence, therefore, to support the Commission’s finding that the balloon and bag ads misrepresented the effect of F — 310 on most automobiles. That misrepresentation provides an independent basis for the Commission’s order, and it further suggests the capacity of the ads to misrepresent the efficacy of F-310 in reducing pollution generally.
Finally, the Commission concluded that the meter advertisement was misleading. The ad showed a change of 80 units on a dial of 100 units after use of F-310. It is undisputed that the meter was a Beekman Infra-red Analyzer used by federal and state authorities to measure exhaust emissions. That instrument measures the level of hydrocarbon emissions. By reference to this measurement, one also can draw an inference as to the degree of reduction of carbon monoxide emissions, since the two pollutants generally come from the same source. It is undisputed that the meter was *659not rigged. The Commission’s finding that it conveyed a misleading impression as used in the ad was based on the total absence of any explanation of the calibration shown on the meter dial. The drop from 100 units (“dirty”) to 20 units (“clean”) did not correspond to a percentage reduction or, indeed, to any other scale of measurement with which the public is familiar. As we have previously stated, the hydrocarbon reduction was fifty percent, not eighty percent. Equally important, the ad gave no indication at all that some emission pollutants not measured by the meter where wholly unaffected by F-310. These considerations support the Commission’s determination that the meter ad violated section 5.
The issue before us as to each of the ads is whether the record supports the Commission’s interpretation of the meaning the commercial would have to the average viewer, and the Commission’s determination of the accuracy of that message.
The Commission found that the public is acutely aware of the air pollution problem and that it is a matter of public importance for advertisements which play upon this concern to be accurate and precise. Implicit in the Commission’s opinion is a finding that commercial messages might lead the average viewer, in his anxiety to help solve the pollution problem, to overreact even though upon careful reflection he might see for himself the limitations inherent in the advertiser’s claim. Since it is now clear that the area of commercial speech is not exempt from first amendment protection, Bates v. State Bar of Arizona, 433 U.S. 350, 97 S.Ct. 2691, 53 L.Ed.2d 810 (1977); Virginia State Board of Pharmacy v. Virginia Citizens Consumer Council, Inc., 425 U.S. 748, 96 S.Ct. 1817, 48 L.Ed.2d 346 (1976), comment on a subject of public interest must not be stifled even when it is presented in an advertising format. However, when discussion of a matter of public concern becomes a vehicle for sale of a product, the representations which bear on the characteristics of the product may take on increased importance in the mind of the public, and it is appropriate for the Commission to consider this factor in determining whether the advertising is misleading or deceptive.
Here the Commission found that the predominant visual message was misleading, and that it was not corrected or contradicted by the accompanying verbal message in the advertisements. “[T]he Commission may determine for itself [the deceptive nature of the representations] through visual inspection and analysis.” United States Retail Credit Ass’n v. FTC, 300 F.2d 212, 217 (4th Cir. 1962); Carter Products, Inc. v. FTC, 268 F.2d 461, 495 (9th Cir.), cert. denied, 361 U.S. 884, 80 S.Ct. 155, 4 L.Ed.2d 120 (1959). In making this determination the Commission is entitled to “draw upon its own experience in order to determine (even) in the absence of consumer testimony, the natural and probable result of the use of advertising expressions.” Carter Products, Inc., 268 F.2d at 495 (parenthetical in original). Based on those principles and our review of the record, we find substantial evidence to support the Commission’s finding that the three advertisements were misleading.
BBD&O contends the Commission acted improperly in holding it liable under section 5. The standard of care to be exercised by an advertising agency in determining what express and implied representations are contained in an ad and in assessing the truth or falsity of those representations increases in direct relation to the advertising agency’s participation in the commercial project. See Doherty, Clifford, Steers & Shenfield, Inc. v. FTC, 392 F.2d 921, 928 (6th Cir. 1968). The degree of its participation is measured by a number of factors including the agency’s role in writing and editing the text of the ad, its work in creating and designing the graphic or audio-visual material, its research and analysis of public opinions and attitudes, and its selection of the appropriate audience for the advertising message. Precisely these factors were weighed in reaching the conclusion that BBD&O knew or should have known of the deceptive nature of the F-310 advertising. The Commission determined that:
*660Representatives of BBD&O were involved in the development of the F-310 advertising from the very earliest stages. They carefully reviewed all the test results and were active participants in numerous meetings in which alternative advertising approaches were evaluated and ultimately accepted or rejected. The final determination to use the demonstration format of the January 1970 advertisements was a joint decision of representatives of BBD&O and Standard, and after the final joint decision was made, BBD&O actively participated in the filming of the pictorial portions of the advertisements, the drafting of the verbal texts, the preparation of layouts and the promotion and distribution of the advertisements.
Record, vol. V, at 3480 (footnotes omitted).
The advertising agency argues vigorously that it was entitled to rely on the elaborate safeguards, including independent laboratory tests and procedures for high level review, that preceded this advertising campaign. The network advertising here in question was approved by the following persons and departments: Chevron Research; Standard’s Product Engineering, Advertising, and Law Departments; the highest level of Standard’s corporate management; executives of BBD&O who handle Standard’s account; BBD&O’s corporate management and legal counsel; Scott Research Laboratories, a highly reputable independent laboratory which reviewed each of the auto emissions tests; and Scott Carpenter, a former NASA astronaut and qualified engineer, who made an independent review of the technical reports on F-310 before agreeing to act as the announcer for the advertisements.
These safeguards and procedures are highly relevant to the petitioners’ common defenses, including the truth of the message which the ads conveyed, and to BBD&O’s claim that it exercised diligence in investigating the accuracy of the advertising. Nevertheless, review of the record indicates that while the technical accuracy of the tests was fully investigated, the review procedures did not have a specific focus on the accuracy of the implicit representations the ads conveyed to the viewing public. No specialized engineer was needed to put BBD&O on notice that a gauge which drops from a reading of 100 (“dirty”) to 20 (“clean”) implies a sweeping representation with reference to the change in level of pollution discharge. In light of the advertising agency’s active participation in developing this advertising, it was BBD&O’s responsibility to assure itself not only that the gauge was not rigged, but also that use of the gauge did not convey a distorted impression. The evidence is fully adequate to support the Commission’s findings that, given the degree of participation by this advertising agency, it knew or should have known that the F-310 ads were misleading in the respects previously indicated.
Breadth of the Order
Although the violation with which Standard and BBD&O were charged was based on three commercials advertising a single product, the cease and desist order issued by the Commission related to any product and enjoined both petitioners from all advertising which creates a misleading impression by use of tests or demonstrations or by visual means generally.3 An all-prod*661ucts order against one who has violated section 5 is not unprecedented, see, e. g., FTC v. Colgate-Palmolive Co., 380 U.S. 374, 85 S.Ct. 1035, 13 L.Ed.2d 904 (1965); however, the order issued in this case was exceptionally burdensome by reason of its breadth and generality.
The over-breadth of the order results from its coverage of “any . . . product in commerce” which is advertised by Standard or BBD&O. The products manufactured and sold by Standard and its subsidiaries number in the thousands and include, in addition to fuel and lubricant products, waxes, fertilizers, pesticides, garden equipment, and cook books. The FTC order, by its terms, requires petitioners to submit a written report “setting forth in detail the manner and form of . compliance” with the order. By letter, the FTC directed that such report include: “All advertising (TV, radio, newspapers, etc.), brochures, mailers, promotional material, etc., relating to all products covered by the Order, together with documentation in support of each claim.” A duty to compile such materials on the large number of Standard’s diverse products is nothing short of oppressive. The same applies to BBD&O, a national firm representing customers in almost every commercial endeavor which relies upon advertising. The agency misused its discretion in issuing the order.
The order’s generality adds to its burdensome character. Various paragraphs of the Commission’s order are framed in language which amounts to a restatement of general principles of fair advertising. In effect the petitioners are simply enjoined from violating the law.4 The Supreme Court has expressly disapproved of such nonspecific FTC orders:
The severity of possible penalties prescribed by the [1959 amendments to section 11 of the Clayton Act] for violations of orders which have become final underlies the necessity for fashioning orders which are, at the outset, sufficiently clear and precise to avoid raising serious questions as to their meaning and application.
FTC v. Henry Broch & Co., 368 U.S. 360, 367-68, 82 S.Ct. 431, 436, 7 L.Ed.2d 353 (1962). See Country Tweeds, Inc. v. FTC, 326 F.2d 144, 149 (2d Cir. 1964). Orders worded in such general language are disfavored because they alter the scheme of penalties and enforcement procedures defined by the Act without specific identification of the proscribed conduct. A cease and desist order transfers responsibility for enforcement of the Act from the Commission to the district courts, American Home Products Corp. v. FTC, 402 F.2d 232, 237 (6th Cir. 1968), thereby causing a short-circuit in the normal procedure which gives an alleged section 5 violator a full hearing before the Commission.
The Commission argues that any vagueness in the order may be cured if petitioners seek the Commission’s advice as to the application or proper interpretation of the order. But this suggestion does not reach the core of the problem. The vice inherent in an overly broad order lies not only in the *662imposition of an additional penalty for any violation, but also in the requirement that one who is subject to its terms must “either expose his major business decisions to a Commission veto, or remain in the dark regarding their legal consequences . . . Joseph A. Kaplan & Sons v. FTC, 121 U.S. App.D.C. 1, 7, 347 F.2d 785, 791 (1965).
We recognize that the Commission is allowed wide latitude in structuring its remedies based on its expertise in determining what is necessary to eliminate unfair practices and its administrative responsibility to devise remedies according to a consistent statutory policy. FTC v. National Lead Co., 352 U.S. 419, 428-29, 77 S.Ct. 502, 1 L.Ed.2d 438 (1957); Carter Products, Inc. v. FTC, 268 F.2d 461, 498 (9th Cir.), cert. denied, 361 U.S. 884, 80 S.Ct. 155, 4 L.Ed.2d 120 (1959); see Resort Car Rental System, Inc. v. FTC, 518 F.2d 962, 964 (9th Cir.), cert. denied, 423 U.S. 827, 96 S.Ct. 41, 46 L.Ed.2d 42 (1975). However, the FTC’s discretion to formulate cease and desist orders is not unlimited. The remedy selected by the Commission must have a “reasonable relation to the unlawful practices found to exist.” FTC v. Colgate-Palmolive Co., 380 U.S. 374, 394-95, 85 S.Ct. 1035, 1048, 13 L.Ed.2d 904 (1965); National Lead, 352 U.S. at 429, 77 S.Ct. 502; Carter Products, 268 F.2d at 498. “The propriety of a broad order depends upon the specific circumstances of the case . . . .” Colgate, 380 U.S. at 394, 85 S.Ct. at 1048; see Feitler v. FTC, 201 F.2d 790, 794 (9th Cir.), cert. denied, 346 U.S. 814, 74 S.Ct. 24, 98 L.Ed. 341 (1953). Among the circumstances which should' be considered in evaluating the relation between the order and the unlawful practice are whether the respondents acted in blatant and utter disregard of the law, and whether they had a history of engaging in unfair trade practices. National Lead, 352 U.S. at 429-30, 77 S.Ct. 502; Floersheim v. FTC, 411 F.2d 874, 875 n. 1, 878 (9th Cir. 1969), cert. denied, 396 U.S. 1002, 90 S.Ct. 551, 24 L.Ed.2d 494 (1970); see Joseph A. Kaplan & Sons v. FTC, 121 U.S.App.D.C. 1, 5 n. 8, 347 F.2d 785, 789 n. 8 (1965).
Moreover, first amendment considerations dictate that the Commission exercise restraint in formulating remedial orders which may amount to a prior restraint on protected commercial speech. Beneficial Corp. v. FTC, 542 F.2d 611, 619-20 (3d Cir. 1976), cert. denied, 430 U.S. 983, 97 S.Ct. 1679, 52 L.Ed.2d 377 (1977); see Bates v. State Bar of Arizona, 433 U.S. 350, 97 S.Ct. 2691, 53 L.Ed.2d 810 (1977); Virginia State Board of Pharmacy v. Virginia Citizens Consumer Council, Inc., 425 U.S. 748, 96 S.Ct. 1817, 48 L.Ed.2d 346 (1976). The cease and desist orders issued by the Commission in this case unquestionably burden Standard’s and BBD&O’s advertising capacity. Although it is generally accepted that false commercial advertising may be prohibited, see Virginia State Board of Pharmacy, 425 U.S. at 771 & n. 24, 96 S.Ct. 1817; FTC v. Simeon Management Corp., 532 F.2d 708, 713 (9th Cir. 1976), and cases cited therein, these orders go far beyond elimination of the specific misrepresentations which were made and also beyond what in fairness could be deemed necessary to deter future unlawful conduct.5 In the emerging law of first amendment protection for commercial speech, one of the first principles is that “such speech serves individual and societal interests in assuring informed and reliable decisionmaking.” Bates, 433 U.S. at 364, 97 S.Ct. at 2699. At a minimum, administrative agencies may not pursue rigorous enforcement to the extent of discouraging advertising with no concomitant gain in assuring accuracy and truthfulness.
We hold that the blanket order issued by the Commission is wholly unwarranted by the circumstances of this case. *663Publication of the three advertisements in question was not a blatant disregard of the law. Petitioners’ error was to miscalculate the effect which the televised commercials would have on the public; the major claims made for the product were valid. The Commission itself found that the additive was a “laudable” product, effective in eliminating at least some buildup of dirt in all car engines.6 The FTC’s charges that F-310 does not significantly reduce auto emission pollution were eventually dismissed, and the Commission expressly found that
the reduction in automotive air pollution, which would result from the general use of F-310, would exceed in most instances and approximate in others the reductions achieved from various emission control programs adopted, enforced or approved by government agencies.
Record, vol. V, at 3451 (adopting finding 31 of the administrative law judge). The misrepresentations in the ads pertained to the extent of pollution reduction; they did not amount to a wholly false claim about an inferior product.
This case is thus distinguishable from FTC v. Colgate-Palmolive Co., 380 U.S. 374, 85 S.Ct. 1035, 13 L.Ed.2d 904 (1965), where the Supreme Court upheld an all-products order since “the Commission [had] a sufficient basis for believing that the respondents would be inclined to use similar commercials with respect to the other products they advertise.” Id. at 395, 85 S.Ct. at 1048. The violation in Colgate justified a comprehensive cease and desist order because the advertising indicated the company’s utter disregard of the law. In contrast, the vice of petitioners’ advertising was a failure to provide sufficient information for the public fully to evaluate the efficacy of the product. Moreover, in Colgate the violation involved use of a deceptive mock-up, a technique that was directly applicable to the advertising of many other types of products. In the instant case the petitioners’ violations involved use of a visual image which was misleading by implication because of the specific subject matter of the advertising. The violations were not a technique of deception that easily could be transferred to an advertising campaign for some other product.
Standard has never before been accused of false advertising, and BBD&O, the nation’s third largest domestic advertising agency, has had one FTC consent order entered against it. The breadth of these orders cannot be justified by characterizing petitioners as habitual violators of the Act for whom special deterrence is required. See Grove Laboratories v. FTC, 418 F.2d 489 (5th Cir. 1969); American Home Products Corp. v. FTC, 402 F.2d 232 (6th Cir. 1968).
The circumstances surrounding the violation, including the nature of the offense, the procedures used to evaluate the ads before broadcasting, the good faith attempt to eliminate rapidly the implied misstatements, and the previous good record of the parties, indicate that Standard and BBD&O did not act in blatant disregard of their responsibilities under the Act. We therefore hold that the exceptionally broad orders issued by the FTC in this case bear no “reasonable relation to the unlawful practices found to exist.” Both orders shall be modified so that they refer only to future advertising of F-310. As so modified, the order of the Commission will be affirmed.
*664APPENDIX I
Television advertisement as it appeared after June 10 with explanatory phrases superimposed on screen.
. As summarized by the Commission, the complaint alleged that the advertisements contained the following misleading and deceptive representations:
(1) F-310 in Chevron gasolines is a revolutionary development in the reduction of air pollution; (2) Chevron with F-310 will produce motor vehicle exhaust which is generally pollution free; (3) Chevron with F-310 will significantly reduce the total amount of air pollution; (4) Chevron with F-310 will significantly reduce air pollution caused by motor vehicles; (5) Chevron with F-310 will significantly reduce emissions of carbon monoxide (CO) and unburned hydrocarbons (HC) from every motor vehicle in which it is used; (6) the Balloon and Bag demonstrations constitute proof, or accurately or visually demonstrate that Chevron with F-310 reduces motor vehicle emissions of HC and CO and significantly reduces air pollution caused by motor vehicles; (7) every motor vehicle will emit black exhaust as pictured in the Balloon, Bag and other demonstrations if operated on motor fuels other than Chevron with F-310; (8) the building identified as Standard Oil Company of California Research Center in some of the advertisements is owned, occupied, or used for research by Standard; (9) the machine (Meter) pictured in some of the advertisements is used by the Federal Government to measure the total amount of pollution emitted by a motor vehicle; (10) tests or demonstrations had been performed before publication or dissemination of the advertisements which proved representations (2), (3), (4) and (5) above, and also that every *657purchaser of Chevron with F-310 will obtain significantly better mileage than with any other commercially available gasoline; (11) F-310 or Chevron with F-310 will clean or keep clean all engines and engine components.
Record, vol. V, at 3465-66.
. BBD&O reached this conclusion on the basis of answers received from motorists who were asked the following question:
Sometimes a car, idling at a stoplight, gives off black smoke that you can see. Do you believe this is contributing more to air pollution than other cars?
Eighty-three percent of the motorists interviewed answered yes, and fourteen percent said no. BBD&O now interprets these responses as indicating that most people correctly believe that there is a relationship between black smoke and the degree of hydrocarbon and carbon monoxide emissions present in exhaust. Brief for Appellant BBD&O at 29. While this interpretation of the survey is certainly a reasonable one, we think that the Commission was entitled to take into consideration the inferences drawn by BBD&O at the time the survey was taken.
. The final order relating to Standard required, among other things, that the company refrain from making certain representations with respect to the characteristics of F-310 “or any other product in commerce” unless every such representation is true and is completely substantiated by competent scientific tests; from advertising any product by a visual means which directly or by implication creates a misleading impression as to the true state of material facts; and from advertising any product through use of a test or experiment which purports to confirm a product characteristic or the truth of a representation but which does not accurately demonstrate that characteristic or representation. The cease and desist order directed to BBD&O included a prohibition against advertising F-310 or “any other product in commerce” by using a test or experiment to prove a fact or product characteristic when the test does not accurately demonstrate the fact or characteristic, unless BBD&O can estab*661lish that it neither knew nor on reasonable inquiry could have known of the inaccuracy; and from advertising any product by a visual means which directly or by implication creates a misleading impression as to the true state of facts, unless BBD&O can establish that it could not reasonably have known the true facts.
. For example, as to all products sold or advertised by Standard, the order directs the company to cease and desist from:
Using any pictorial or other visual means of communication with or without an accompanying verbal text which directly or by implication creates a misleading impression in the minds of viewers as to the true state of material facts which are the subject of said pictures or other visual means of communication.
Record, vol. V, at 3453. Similarly, BBD&O was ordered to cease and desist from:
Using any pictorial or other visual means of communication with or without an accompanying verbal text which directly or by implication creates a misleading impression in the minds of viewers as to the true state of material facts which are the subject of said pictures or other visual means of communication unless the respondent can establish it neither knew nor had reason to know nor upon reasonable inquiry could have known the true facts.
Id. at 3456.
. Compare Warner-Lambert Co. v. FTC, 183 U.S.App.D.C. 230, 249-52, 562 F.2d 749, 768-71 (1977) (supplemental opinion on petition for rehearing), appeal dismissed,-U.S.-, 98 S.Ct. 1575, 55 L.Ed.2d 800 (1978). That case held that an FTC order requiring a § 5 violator to do corrective advertising did not violate the first amendment where the facts made it “eminently clear” that the important government objective of protecting citizens against deception could not have been accomplished by any less restrictive means.
. [It] removes and reduces the buildup of deposits in carburetors, intake manifolds, intake ports and on intake valves and PCV valves, and minimizes the buildup of sludge and varnish on pistons, piston rings, valve lifters, oil screens, oil pump relief valves and throughout the crankcase area of dirty engines. The degree of improvement necessarily depends upon the condition of the engine. In the case of new cars or cars with clean engines, F-310 will prevent or minimize the accumulation of such deposits.
Record, vol. V, at 3451 (adopting finding of the administrative law judge).