with whom Justice Marshall joins, concurring in the result.
We granted certiorari in these cases in order to review the legal standard employed by the Second Circuit in finding that a generic drug manufacturer is vicariously liable for trademark infringement committed by pharmacists who dispense the generic drug. The Court implicitly endorses the legal standard purportedly employed by the Court of Appeals, ante, at 853-854, but finds that the court erred in setting aside factual findings that were not clearly erroneous. The question whether the Court of Appeals had misapplied the clearly-erroneous rule, however, was not presented in the petitions for certiorari. This was conceded at oral argument.1 Tr. of Oral Arg. 69. Our Rule 21.1(a) states that “[ojnly the questions set forth in the petition or fairly included therein will be considered by the Court.” The majority suggests no reason for ignoring our own Rule. Furthermore, if the issue presented in the petitions for certiorari had been whether the clearly-erroneous standard, although properly invoked, was erroneously applied, it is doubtful in my mind that this fact-bound issue would have warranted certiorari. I nevertheless concur in reversal because I believe that the Court of Appeals has watered down to an impermissible extent the standard for finding a violation of § 32 of the Lanham Act, 15 U. S. C. § 1114.
In its first opinion in this litigation, the Court of Appeals indicated that a “manufacturer or wholesaler would be liable *860under § 32 if he suggested, even if only by implication, that a retailer fill a bottle with the generic capsules and apply Ives’ mark to the label, or continued to sell capsules containing the generic drug which facilitated this to a druggist whom he knew or had reason to know was engaging in the practices just described.” 601 F. 2d 631, 636 (1979) (Ives II). The District Court applied this test but concluded that no violation of § 32 had been shown. On appeal after trial, a majority of the Second Circuit found defendants liable for contributory infringement by revising and expanding the doctrine of contributory trademark infringement. 638 F. 2d 538 (1981) (Ives IV):
“By using capsules of identical color, size, and shape, together with a catalog describing their appearance and listing comparable prices of CYCLOSPASMOL and generic cyclandelate, appellees could reasonably anticipate that their generic drug product would by a substantial number of druggists be substituted illegally .... This amounted to a suggestion, at least by implication, that the druggists take advantage of the opportunity to engage in such misconduct.” Id., at 543 (emphasis added).
Ives II required a showing that petitioners intended illegal substitution or knowingly continued to supply pharmacists palming off generic cyclandelate as CYCLOSPASMOL; Ives IV was satisfied merely by the failure to “reasonably anticipate” that illegal substitution by some pharmacists was likely. In my view, this is an erroneous construction of the statutory law governing trademark protection.
William R. Warner & Co. v. Eli Lilly & Co., 265 U. S. 526 (1924), made clear that a finding of contributory infringement requires proof of either an intent to induce illegal substitution or continued sales to particular customers whom the manufacturer knows or should know are engaged in improper palming off. In that case, it was shown that the manufacturer’s salesmen actively induced, either in direct terms or by insinuation, the filling of requests for Coco-Quinine with *861Quin-Coco. “The wrong was in designedly enabling the dealers to palm off the preparation as that of the respondent.”2 Id., at 530. Coca-Cola Co. v. Snow Crest Beverages, Inc., 64 F. Supp. 980, 989 (Mass. 1946), aff’d, 162 F. 2d 280 (CA1), cert. denied, 332 U. S. 809 (1947), the case upon which the Court of Appeals relied in Ives II, stands for this very proposition. There was no contributory infringement in Snow Crest’s manufacture of a product identical in appearance to that of Coca-Cola. Judge Wyzanski observed that
“any man of common sense knows that in any line of business . . . there are some unscrupulous persons, who, when it is to their financial advantage to do so, will palm off on customers a different product from that ordered by the customer.” 64 F. Supp., at 988-989.
These cases reflect the general consensus. 2 J. McCarthy, Trademarks and Unfair Competition § 25:2 (1973) (“[T]he supplier’s duty does not go so far as to require him to refuse to sell to dealers who merely might pass off its goods”). The mere fact that a generic drug company can anticipate that some illegal substitution will occur to some unspecified extent, and by some unknown pharmacists, should not by itself be a predicate for contributory liability. I thus am inclined to believe that the Court silently acquiesces in a significant change in the test for contributory infringement.
Diluting the requirement for establishing a prima facie case of contributory trademark infringement is particularly unjustified in the generic drugs field. Preventing the use of generic drugs of the same color to which customers had become accustomed in their prior use of the brand name product interferes with the important state policy, expressed in New York and 47 other States, of promoting the substitution of *862generic formulations. See Warner, Consumer Protection and Prescription Drugs: The Generic Drug Substitution Laws, 67 Ky. L. J. 384 (1978-1979).
The Court of Appeals concluded that there was no “persuasive evidence of a legitimate reason” for petitioners to use imitative colors. The District Court, however, had expressly found that for purposes of § 43(a), the capsule colors were functional. With respect to functionality, I fully agree with the Court that the Court of Appeals erred in setting aside factual findings without finding that they were clearly erroneous. The District Court found that capsule color was functional in several respects: patient anxiety and confusion were likely if accustomed medicine were dispensed in a different color; capsule colors assist patients in identifying the correct pill to take; standard colors help physicians identify the drug involved in case of overdose.3 Clearly, the Court of Appeals could not reject these findings merely because it viewed the evidence as less persuasive than did the District Court. Rule 52(a) imposes a stricter standard.
Finally, although the Court states that a “finding of functionality may also be relevant to an action involving § 32,” it does not explicate the relationship of functionality in a § 32 case. It is my view that a finding of functionality offers a complete affirmative defense to a contributory in*863fringement claim predicated solely on the reproduction of a functional attribute of the product. A functional characteristic is “an important ingredient in the commercial success of the product,” 601 F. 2d, at 643, and, after expiration of a patent, it is no more the property of the originator than the product itself. It makes no more sense to base contributory infringement upon the copying of functional colors than on the petitioners’ decision to use the same formulation of the drug, or even to market the generic substitute in the first place. To be sure, the very existence of generic drugs “facilitates” illegal substitution. But Ives no longer has a patent for cyclandelate, “and the defendants have a right to reproduce it as nearly as they can.” Saxlehner v. Wagner, 216 U. S. 375, 380 (1910) (Holmes, J.). Reproduction of a functional attribute is legitimate competitive activity.
I am also mindful that functionality is a defense to a suit under § 43(a) of the Lanham Act alleging damages from a competitor’s “false designation of origin” on a good.4 The use of a product or package design that is so similar to that of another producer that it is likely to confuse purchasers as to the product’s source may constitute “false designation of origin” within the meaning of the Act.5 As the Court of Appeals noted in Ives II, § 43(a) “goes beyond § 32 in making certain types of unfair competition federal statutory torts,” 601 F. 2d, at 641. Section 43(a) offers the direct protection of Ives’ interest in this case, and it is not surprising that the alleged *864§ 43(a) violation was the primary claim in this litigation, as it has been in other cases of this genre. It would be anomalous for the imitation of a functional feature to constitute contributory infringement for purposes of § 32, while the same activity is not a “false designation of origin” under § 43(a).6
I would reverse the decision of the Court of Appeals and remand for review of the District Court’s findings consistent with the principles stated above.
The third question in petitioner Darby Drug Co.’s petition embraced the claim that the Court of Appeals had failed to observe Rule 52(a) in overturning the District Judge’s finding of functionality. As discussed below, I agree with the Court’s invocation of Rule 52 with respect to this aspect of the decision below.
Although Warner and other eases were decided before §32 was enacted, the purpose of the Lanham Act was to codify and unify the common law of unfair competition and trademark protection. S. Rep. No. 1333, 79th Cong., 2d Sess. (1946). There is no suggestion that Congress intended to depart from Warner and other contemporary precedents.
“The reality is that for every link in the distributive chain (from producer to ultimate consumer) the color and shape of drugs dispensed by prescription do perform a function. For each of them, color or shape may be a convenient shorthand code by which to identify the drug and its milligram dosage so that mistakes can be avoided in the interests of pharmaceutical precaution and patient safety. For the patient-user, of course, the constancy of color and shape may be as psychologically reassuring and therefore as medically beneficial as the drug itself; in addition, they also serve to identify the drug for his ingestion . . . .”
“[I]f the generic producer is constrained by § 43(a), trademark law, or the law of unfair competition to adopt a substantially different color . . . the therapeutic value of his generic drug might be seriously impaired and confusion at the pharmacist level could be compounded beyond redemption.” 3 R. Callmann, Unfair Competition, Trademarks and Monopolies § 82.1(m), pp. 217, 213 (Supp. 1981).
See, e. g ., International Order of Job’s Daughters v. Lindeburg & Co., 633 F. 2d 912, 917 (CA9 1980), cert. denied, 452 U. S. 941 (1981); Keebler Co. v. Rovira Biscuit Corp., 624 F. 2d 366, 378 (CA1 1980). See generally Note, The Problem of Functional Features: Trade Dress Infringement Under Section 43(a) of the Lanham Act, 82 Colum. L. Rev. 77, 81 (1982) (“Over the past three years the rule that functionality of a copied feature bars relief in section 43(a) claims for trade dress infringement or product imitation has become the plurality view”).
See, e. g., Truck Equip. Serv. Co. v. Fruehauf Corp., 536 F. 2d 1210 (CA8), cert. denied, 429 U. S. 861 (1976); Warner Bros., Inc. v. Gay Toys, Inc., 658 F. 2d 76 (CA2 1981). See also Note, 82 Colum. L. Rev., supra, at 78-80.
This is not to suggest that the copying of a functional feature protects a defendant from § 32 liability predicated on active inducement of trademark infringement or protects a defendant who has also reproduced nonfunctional features.