This is an appeal from a final order of the district court granting appellee CIBA-GEIGY Corporation ("CIBA") a permanent injunction against appellant Bolar Pharmaceutical Company ("Bolar"). CIBA had previously obtained a preliminary injunction against Bolar after an evidentiary hearing before the district court. This court affirmed the preliminary injunction on appeal. After the affirmance, and based on the same record developed in the evidentiary hearing for the preliminary injunction, the district court granted a permanent injunction. The question now presented is whether the district court erred in granting a permanent injunction. Because we believe the district court properly granted CIBA a permanent injunction, we will affirm.

Since our decision affirming the preliminary injunction, Bolar has been eager to take this case to the Supreme Court where Bolar believes this matter would be resolved in its favor. Unfortunately, because Bolar has repeatedly sought a quick determination by the lower courts, this case might be a classic example of the old axiom that "haste makes waste."

Bolar petitioned the Supreme Court for a writ of certiorari after our affirmance of the preliminary injunction. CIBA-GEIGY Corp. v. Bolar Pharmaceutical Co., 719 F.2d 56 (3d Cir.1983). That petition was denied. --- U.S. ----, 104 S. Ct. 1444, 79 L. Ed. 2d 763 (1984) Bolar returned to the district court and asked that the district court convert its hearing concerning the preliminary injunction into a final trial on the merits.

On December 23, 1983, Judge Sarokin, after reviewing the prior litigation history and this court's affirmance of his preliminary injunction, entered an order which provided, inter alia:

the parties having consented to converting the hearings held on plaintiff's motion for preliminary injunction to the final trial on the merits; and the Court having concluded that on the present record and for the reasons set forth in its opinion of August 11, 1982, Ciba-Geigy is entitled to final relief;

It is on this 22nd day of December, 1983, ORDERED, ADJUDGED and DECREED

A. That the preliminary injunction issued on October 4, 1982 is made permanent.1

CIBA-GEIGY Corp. v. Bolar Pharmaceutical Co., No. 82-788 (D.N.J. Dec. 23, 1983). Bolar, the loser below, petitioned this court for summary affirmance in order to get this case to the Supreme Court. We denied that request.

My reference to the axiom "haste makes waste" concerns the confusion created when a district court converts an opinion granting a preliminary injunction into an opinion granting a permanent injunction without expressly recasting its findings accordingly. The standard for this court reversing the issuance of a preliminary injunction is different from the standard for reversing a permanent injunction. Because the district court did not expressly recast its findings in terms of the legal standard applicable to a permanent injunction, that mechanical task is left to this court. Although we are able to so convert the district court's opinion, we would much prefer that the district court recast its own opinion in the language of the standard it is applying. This would eliminate the possibility of confusion as to what the district court intended and would, in the long run, promote judicial economy.

As tedious as the process might be for us, we believe it is necessary to ensure that the district court's conversion of its preliminary injunction does not become the basis of another appeal.

When issuing the findings and opinion on the preliminary injunction, the district court used traditional preliminary injunction language. As an example, it said, "even assuming that SK & F's 'reasonable anticipation' standard does not survive Ives, CIBA-GEIGY has demonstrated a likelihood of success in establishing that Bolar copied its trade dress with the intention, at least in part, to induce or permit illegal substitution." CIBA-GEIGY Corp. v. Bolar Pharmaceutical Co., 547 F. Supp. 1095, 1116 (D.N.J.1982) (emphasis added). At other places in its opinion the district court used similar language couched in a preliminary injunction context.

When the parties agreed to "convert" the hearing on the plaintiff's motion for preliminary injunction to the final trial on the merits, the district court did not expressly change the language of its opinion preliminarily enjoining Bolar. It did so by reference, and the parties agreed to the district court disposing of the case in this manner. This reading is bolstered by the district court's order of December 16, 1983 wherein it stated,

[t]his court having held hearings on April 22, May 12 and June 15, 1982 on the motion of plaintiff Ciba-Geigy Corporation for preliminary injunction, plaintiff being satisfied with the sufficiency of its proof therein for purposes of a trial on the merits, and defendant believing it submitted sufficient evidence and made offers of proof therein to defeat plaintiff's claim at a trial of the merits, it is this 16th day of December, 1983,

ORDERED upon consent and stipulation of the parties that

....

(2) The hearing held on April 22, May 12 and June 15, 1982 on plaintiff's motion for preliminary injunction shall constitute the trial on the merits in this action.

CIBA-GEIGY Corp. v. Bolar Pharmaceutical Co., No. 82-788 (D.N.J. Dec. 16, 1983) (emphasis added).

It is almost absurd to believe that the parties did not realize that the district court had implicitly converted its findings of a "likelihood of success" on the merits to findings of "success" on the merits. Moreover, none of the parties made any suggestion in their briefs or at oral argument on the appeal from the permanent injunction that the findings of the district court should be construed in any fashion different from the manner in which we have construed them. The parties in this case were concerned primarily with the sufficiency of the evidence and with the significance of Inwood Laboratories v. Ives Laboratories ("Ives"), 456 U.S. 844, 102 S. Ct. 2182, 72 L. Ed. 2d 606 (1982) to their case.

CIBA believed it had met its burden of proof under any reading of Ives while Bolar believed that it had presented sufficient evidence to defeat CIBA's claim for the preliminary and final injunction. Indeed both parties were so convinced that they felt little need to alter in any significant respect the briefs they had filed in the appeal of the preliminary injunction. For example, Bolar stated,

[e]xcept for minor changes which reflect the conversion of the preliminary injunction into a permanent injunction, its present brief is substantially identical to [its earlier brief.]

Bolar Motion to Consolidate at 3.

For the above reasons, in analyzing this case we have treated the district court's findings that plaintiff has demonstrated "likelihood of success" on the merits as findings that the plaintiff "succeeded" on the merits.

In 1970 CIBA received a patent on the drug hydrazaline hydrochloride/hydrochlorothiazide, an anti-hypertensive composed of hydralazine hydrochloride, a vasodilator, and hydrochlorothiazide, a diuretic. Until its patent was dedicated to the public in 1981, CIBA retained the exclusive right to produce and market the drug, which it sold under the registered trademark APRESAZIDE. CIBA chose to market its 25 mg./25 mg. dosage in a blue and white capsule, its 50 mg./50 mg. in a pink and white capsule, and its 100 mg./50 mg. dosage in a pink flesh and white capsule.

In August 1976, two months before CIBA first marketed APRESAZIDE, another company, Pfizer, Inc., first marketed another anti-hypertensive, Minipress. Minipress is sold in two dosages, one in pink and white capsules, the other in blue and white capsules. Due to the differences in size, the blue and white Minipress capsule does not resemble the blue and white APRESAZIDE capsule. Although the pink and white Minipress capsule is similar in appearance to the pink and white APRESAZIDE capsule, the drugs are chemically distinct and prescribed to distinct groups of patients.

After CIBA's patents were dedicated to the public, Bolar began to manufacture hydrazaline hydrochloride/hydrochlorothiazide capsules the same size, shape and color as the 25 mg./25 mg. and 50 mg./50 mg. APRESAZIDE capsules.2 Bolar produced the capsules with the intention that they would resemble APRESAZIDE capsules. Bolar capsules are identical to APRESAZIDE capsules except that each is imprinted with the name of its respective manufacturer.

CIBA, in marketing APRESAZIDE, followed normal industry practice. As APRESAZIDE can be obtained only with a prescription, CIBA directed advertising to pharmacists and physicians rather than directly to purchasers of anti-hypertensives. Bolar, in marketing its generic drug, has also followed industry practice by directing its advertising to wholesalers who sell to pharmacists who, in turn, may substitute the generic drug for APRESAZIDE on the advice of the prescribing physician and with the consent of the patient.

A pharmacist, when dispensing APRESAZIDE or its generic equivalent, removes the capsules from their original container and places them in another container bearing the pharmacist's label. As a consequence, the consumer sees no identifying marks other than the small imprints on the capsules themselves.

CIBA brought this action in the United States District Court for the District of New Jersey under section 43(a) of the Trademark Act of 1946 ("Lanham Act" or the "Act"), 15 U.S.C. Sec. 1125(a) (1982), and under New Jersey's unfair competition law, N.J.S.A. 56:4-1 (1964 and Supp. 1984). CIBA asserted two violations of Sec. 43(a), one derived from allegations that Bolar falsely designated the origin of its generic drug by manufacturing it in capsules identical to those CIBA used for APRESAZIDE. The other derived from allegations that Bolar's manufacture and distribution of hydralazine hydrochloride/hydrochlorothiazide capsules identical to CIBA's APRESAZIDE capsules, with reasonable anticipation or the intent that they would be illegally substituted for CIBA's APRESAZIDE capsules, would constitute contributory trademark infringement (passing off) within the meaning of the federal statute.

The district court granted CIBA's initial request for a preliminary injunction and later granted CIBA's request for a permanent injunction. Bolar is thus prevented from advertising, distributing, or marketing its present generic version of APRESAZIDE.3 547 F.Supp. at 1118.

CIBA proved (1) that the trade dress of APRESAZIDE was non-functional; (2) that APRESAZIDE had acquired secondary meaning with an identifiable group of users of anti-hypertensives; (3) that Bolar could reasonably anticipate that some pharmacists provided with Bolar's generic hydralazine hydrochloride/hydrochlorothiazide would attempt to illegally substitute it for CIBA's APRESAZIDE; and (4) that Bolar copied CIBA's APRESAZIDE trade dress with the intent, at least in part, to facilitate the illegal substitution of Bolar's generic drug for CIBA's brand name drug. Bolar initially moved this court for a summary affirmance so it could hurry to the Supreme Court. We denied Bolar's motion because this appeal involves a permanent injunction, which requires us to examine anew evidence that we found sufficient to uphold a preliminary injunction and which requires us to apply a different legal standard.

In deciding whether a permanent injunction should be issued, the court must determine if the plaintiff has actually succeeded on the merits (i.e. met its burden of proof). If so, the court must then consider the appropriate remedy. See Evans v. Buchanan, 555 F.2d 373 (3d Cir.1977).

Because this case involves claimed violations of federal law (Sec. 43(a) of the Trademark Act of 1946 (Lanham Act)) and New Jersey law (N.J.S.A. 56:4-1) as support for the grant of a permanent injunction, we must review both the state and federal claim in order to determine the proper scope of the injunction. If the district court correctly found that CIBA succeeded on the merits in its state claim, the district court's grant of a permanent injunction can only prevent Bolar's activity as it relates to New Jersey. That is, Bolar cannot necessarily be prevented from manufacturing and selling its product in California simply because its activity is unlawful in New Jersey. If, however, the district court correctly found CIBA succeeded on its federal claims, we must uphold the permanent injunction in its entirety so long as the balance of equities favors injunctive relief. We will consider first the claimed violations of New Jersey law, then the claimed violations of federal law.

CIBA claims two violations of New Jersey law by Bolar: unprivileged imitation and passing off.

To establish unprivileged imitation under New Jersey law, a plaintiff must show (1) that an imitated feature of its product is non-functional, (2) that the imitation is likely to cause confusion as to the product's origin, and (3) that the product has acquired secondary meaning. French American Reeds Manufacturing Co. v. Park Plastics Co., 20 N.J.Super. 325, 333-34, 90 A.2d 50, 54-55 (1952), Press Publishing Co. v. Atlantic County Advertiser, 108 N.J.Super. 75, 81, 260 A.2d 6, 9 (1969).

We begin our analysis with consideration of the district court's finding that the features imitated by Bolar are non-functional.

In SK&F, Co. v. Premo Pharmaceutical Laboratories, 625 F.2d 1055 (3d Cir.1980), this court applied the correct standard for making a finding of nonfunctionality under New Jersey law. We stated:

Proof of nonfunctionality generally requires a showing that the element of the product serves no purpose other than identification. In the case of a drug, for example, the allegedly nonfunctional element must not enhance efficacy.

Id. at 1063.

Judge Sarokin's finding of nonfunctionality is a finding of fact. A finding of fact is clearly erroneous when the reviewing court is left with the " 'definite and firm conviction that a mistake has been committed.' " Inwood, 102 S. Ct. at 2189 (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S. Ct. 525, 542, 92 L. Ed. 746 (1948)). We believe in this case appellant Bolar failed to prove that the district court's finding was clearly erroneous.

Regarding the issue of nonfunctionality, the district court found that CIBA had arbitrarily chosen to manufacture APRESAZIDE in blue & white and pink & white capsules and that no medical or business considerations compelled its choice. The district court noted that Reid-Provident Laboratories, Inc., a third corporation, was successfully producing the same drug in a different color. Similarly, the district court found that no medical or business considerations compelled the size and shape of CIBA's APRESAZIDE capsule. As CIBA had asserted, the district court found that APRESAZIDE's capsule color, size and shape serve no purpose other than to identify APRESAZIDE as CIBA's product. The district court rejected Bolar's arguments concerning the alleged functionality of APRESAZIDE's trade dress, stating, "[t]he fact that a manufacturer uses different colors to distinguish between its own products does not make those particular colors functional as to other manufacturers." 547 F. Supp. at 1111. The district court also rejected Bolar's other functionality arguments, which were premised primarily on the assumption that physicians and pharmacists would prescribe a drug based on its appearance only. Based on this evidence we cannot say that the district court's finding of non-functionality was clearly erroneous.

We turn to the second necessary element of a successful claim of unprivileged imitation under New Jersey law--that is, proving that the imitation is likely to cause confusion as to the product's origin in the eyes of the consuming public. Our standard of review on this issue is again the clearly erroneous standard. After carefully reviewing the record, we do not consider the district court's finding on this issue to be clearly erroneous.

The district court found that CIBA's APRESAZIDE capsules could not be distinguished from Bolar's generic imitation by the typical patient utilizing such medication. The district court reasoned as follows:

Realistically, the likelihood of confusion cannot be assessed by a side-by-side comparison of the plaintiff's and defendant's products. It is the overall physical appearance of defendant's trade dress which is critical. The vast majority of patients who take this type of medication do not or cannot identify their medication, or its source, by reference to the matter imprinted on the drug capsule or tablet. Most of such patients are over the age of 60 years and it is unlikely that they will have sufficiently acute eyesight to detect the difference between their usual medication and Bolar's products based on the imprintations on Bolar's capsules.

547 F.Supp. at 1103.

We do not find any error in the district court's finding that Bolar's imitation of CIBA's APRESAZIDE capsules is likely to confuse the relevant consuming public as to the product's origin.

Having decided that CIBA satisfied the first two essential elements of a successful claim of unprivileged imitation under New Jersey law, we turn to the final element--whether the district court's finding that the trade dress of APRESAZIDE capsules had acquired a secondary meaning was clearly erroneous. The district court found that CIBA met its burden. We agree.

In French American Reeds, 20 N.J.Super. 325, 90 A.2d 50 (1952) the court set forth the standard for proving secondary meaning under New Jersey law. It stated that plaintiff must show that the article in question " 'has become associated in the public mind with the first comer as manufacturer or source ....' " Id. at 334, 90 A.2d at 55 (quoting Crescent Tool Co. v. Kilborn & Bishop Co., 247 Fed. 299, 300-301 (2d Cir.1917) (L. Hand, J.)). Similarly, as this court stated in SK&F, if "[t]he only value of the trade dress was in identifying the goods with their source, ... that value suffices in New Jersey courts to establish secondary meaning." 625 F.2d at 1064.

Thus, the district court applied the correct standard when it said, "it is no more necessary here for plaintiff to prove that APRESAZIDE patients identify those products with [CIBA] than it was for plaintiff to prove in SK&F that Dyazide patients identified that product with SK&F." 547 F.Supp. at 1113. Because the district court applied the correct legal standard, we must decide only whether the district court's finding that APRESAZIDE patients identify those products with CIBA is clearly erroneous. If we cannot say with firm conviction that its finding was clearly erroneous, we cannot reverse on this issue.

In concluding that secondary meaning had been acquired by CIBA's APRESAZIDE, the district court considered the duration of CIBA's exclusive manufacturing, the marketing and selling of APRESAZIDE, CIBA's exclusive right to use APRESAZIDE as a trademark, the nature and extent of CIBA's advertising and promotions featuring APRESAZIDE's trade dress, the number of sales under APRESAZIDE's trade dress, and the extent of the copying of APRESAZIDE's trade dress. See generally Ideal Toy Corp. v. Plawner Toy Mfg. Corp., 685 F.2d 78 (3d Cir.1982).

Regarding these factors, the district court made the following observations:

Since October 1976, CIBA-GEIGY has continuously marketed its APRESAZIDE products in the same trade dress, and for over five years, from October 1976 to March 1982, it was the sole pharmaceutical company which marketed these combination products in the particular trade dress described.

From October 1976 to date, CIBA-GEIGY, through its CIBA Pharmaceutical Company division, has spent and continues to spend substantial sums of money advertising and promoting its APRESAZIDE products to physicians, pharmacists and hospitals. During the period October 1976 through 1981, CIBA-GEIGY's expenditures for the distribution of free samples and for promotional and medical literature, films, symposia and seminars promoting its APRESAZIDE products were in excess of $9 million. In addition, during this same period, the cost attributable to the efforts of the CIBA Pharmaceutical Company sales force in promoting APRESAZIDE products was over $16 million.

547 F.Supp. at 1101 (footnote omitted). Moreover, CIBA's sales of APRESAZIDE capsules have increased steadily since their market introduction. In addition to making these findings, the district court reasoned that "[a] product is generally considered to have acquired distinctiveness and secondary meaning when its trade dress has been copied." Id. at 1113. The district court "specifically [found] that Bolar's principal motivation in deliberately copying APRESAZIDE's trade dress was to increase sales of its hydralazine hydrochloride/hydrochlorothiazide combinations at the expense of [CIBA's] APRESAZIDE products." Id. at 1102. The district court also placed great significance in Bolar's "consistent use of hydralazine hydrochloride/hydrochlorothiazide as the generic name of the product in suit," 547 F. Supp. at 1114, in finding that patients associate APRESAZIDE with one source. See Salton Inc. v. Cornwall Corp., 477 F. Supp. 975, 986 (D.N.J.1979).

In this case, association between APRESAZIDE's trade dress and the trade name APRESAZIDE and CIBA is sufficient to prove secondary meaning. The name APRESAZIDE is more than a descriptive term. APRESAZIDE is a federally registered trademark that carries a presumption of validity. See E.I DuPont de Nemours & Co. v. Yoshida International, Inc., 393 F. Supp. 502, 516 (E.D.N.Y.1975); Coca Cola Co. v. Pace, 283 F. Supp. 291, 293 (W.D.Ky.1968). The district court rejected Bolar's claim that APRESAZIDE had become the generic appellation, thus CIBA alone is entitled to use "APRESAZIDE" as the name of its product.

Given (1) CIBA's exclusive sale of APRESAZIDE; (2) the fact that "APRESAZIDE" is a registered trademark; (3) the extent of advertising and sales of APRESAZIDE; (4) Bolar's copying, exactly, of APRESAZIDE's trade dress; and (5) Bolar's purpose in copying, we cannot say with firm conviction that a mistake was made by the district court in finding secondary meaning in APRESAZIDE's trade dress. Because the district court's finding of secondary meaning was not clearly erroneous, it therefore must be upheld.

In addition to establishing its New Jersey claim for unprivileged imitation, the district court concluded that CIBA succeeded on the merits of its New Jersey claim of passing off. The requisite proofs for passing off under New Jersey law have been recently articulated.

[I]t has been held that it is actionable conduct under New Jersey law for a drug manufacturer to put a product in the hands of a pharmacist in a form in which the manufacturer can reasonably anticipate that it may be passed off as another product even if the manufacturer does nothing else to encourage passing off.

SK & F, 625 F.2d at 1062. See also Boehringer Ingelheim v. Pharmadyne Laboratories, 532 F. Supp. 1040, 1061 (D.N.J.1980) (applying New Jersey law); Merrell-National Laboratories v. Zenith Laboratories, 194 U.S.P.Q. 157, 160 (D.N.J.1977) (applying New Jersey law).

As noted previously, the exact copying of APRESAZIDE's trade dress and dosage levels for the blue & white and pink & white capsules has been established. Bolar copied CIBA's APRESAZIDE 25 mg./25 mg. blue & white capsule and 50 mg./50 mg. pink & white capsule. The district court also found that Bolar's purpose in copying was to increase sales of its hydralazine hydrochloride/hydrochlorothiazide at the expense of [CIBA's] APRESAZIDE products. The district court found that Bolar's drug could be mistaken for APRESAZIDE. The district court took judicial notice of many convictions of pharmacists in New York for illegally substituting a generic drug for a brand name drug. 547 F. Supp. at 1108 n. 9.4 Finally, the district court found that Bolar's copying of APRESAZIDE was intended, at least in part, to facilitate pharmacists' illegal substitution of Bolar's look-alike for CIBA's APRESAZIDE.

It was not clearly erroneous for the district court to find, based on this evidence, that Bolar copied APRESAZIDE with the reasonable anticipation that pharmacists would illegally substitute Bolar's drug for CIBA's.

Having established that Bolar violated New Jersey's laws of unprivileged imitation and passing off, we turn to the equitable considerations to determine whether a permanent injunction was properly granted. The district court found that Bolar's look-alike drug deprives the patient of the principal means of differentiating between the drug she/he desires and that which is dispensed. The evidence supports a finding that the practice of illegal substitution occurs and that its full extent cannot be known. We agree with the district court that it may be unfair to the many ethical and honest pharmacists to predicate a rule of law upon an assumption that some pharmacists do and may illegally substitute generic drugs in order to make a greater profit, but we see very little harm to the drug manufacturers and pharmacists in this case where generic equivalents to APRESAZIDE can be manufactured in different sizes, shapes or colors without added cost.

Because of the significance of a permanent injunction, we quote the district court's reasoning regarding the public interest served by enjoining Bolar's imitation of APRESAZIDE's trade dress. The district court explained:

Even in the case of legal substitution, Bolar's conduct endangers the public interest. If Bolar's products look confusingly similar to CIBA-GEIGY's, the patient may never learn or fully understand that they are not made by the same manufacturer as the products he had previously been taking--thereby jeopardizing his right to give his informed consent to the substitution.

A patient who has had a satisfactory experience with a drug may more readily agree to a generic substitute of the same appearance. At first blush, such conclusion may appear consistent with Bolar's arguments regarding functionality. However, such similarity may cause the patient to make the unwarranted assumption that the medication is identical and from the same manufacturer as the prior medication. In this area fraught with risks to health, no procedure should be condoned which affords to the patient anything less than all of the pertinent information and conditions which will prompt an inquiry for such information. There is no harm to the public in differentiating between the pills; there may be great harm in not.

Id. at 1109 (footnote omitted).

We believe that the district court's equitable finding concerning the public interest was not in error.5 We also believe that the district court did not err in finding that Bolar violated the New Jersey laws prohibiting unprivileged imitation and passing off. Consequently, we will affirm the grant of the permanent injunction based on the district court's decision concerning New Jersey law.6

CIBA also asserted claims under section 43(a) of the Lanham Act, 15 U.S.C. Sec. 1125(a), for unprivileged imitation and passing off.

To establish unprivileged imitation under the Act, a plaintiff must prove the nonfunctionality of the imitated feature and must prove secondary meaning of the imitated feature. These proofs are essentially the same as those required for unprivileged imitation under New Jersey law.7 SK & F Co. v. Premo Pharmaceutical Laboratories, 625 F.2d at 1065. Because the district court's findings of fact with respect to the New Jersey claim were not clearly erroneous, its findings of fact necessary to the proof of CIBA's federal claim were also not clearly erroneous. Thus, because CIBA succeeded on the merits of its New Jersey unprivileged imitation claim, it also succeeded on the merits of its claim under federal law. Even if an authoritative interpretation of New Jersey law suggests that the facts established do not prove a claim for unprivileged imitation under N.J.S.A. 56:4-1, an appropriate interpretation of Sec. 43(a) of the Lanham Act indicates that the district court did not err in granting injunctive relief on the basis of CIBA's unprivileged imitation claim under the federal statute. In upholding this claim, we rely on our discussion of the evidence in the New Jersey law section of unprivileged imitation.

Bolar asserts that Ives changed the standard of proof for passing off under Sec. 43(a) of the Lanham Act. It contends that Ives, properly read, requires a showing of intent on the part of the manufacturer to induce retailers to pass off imitative products when a Sec. 43(a) claim is asserted. While it is debatable whether Ives altered the standard of proof for passing off under Sec. 43(a), we need not decide the standard of proof issue. The district court's factual findings make that debate unnecessary. The factual findings and our review of them in this case differ markedly from the findings and their appellate review in Ives.

In Ives, the Court of Appeals for the Second Circuit was reversed because it "erred in setting aside findings of fact that were not clearly erroneous." 102 S. Ct. at 2190. The Supreme Court stressed that Fed.R.Civ.P. 52(a) "recognizes and rests upon the unique opportunity afforded the trial court judge to evaluate the credibility of witnesses and to weigh the evidence." Id. 102 S.Ct. at 2188-89. In evaluating the credibility of witnesses and weighing the evidence, in Ives, the district court's factual findings precluded any holding that there was a violation of section 32 of the Lanham Act. The trial judge in that case held that the materials before him could not "fairly be read" to suggest trademark infringement. Ives Laboratories v. Darby Drug Co., 488 F. Supp. 394, 397 (E.D.N.Y.1980).

In contrast to the findings made by the trial judge exculpating the generic drug manufacturer in Ives, the findings made by the trial judge in this case implicate Bolar and constitute the prerequisite predicate for holding that there was a violation of section 43(a) of the Lanham Act. Here, inter alia, the district court expressly found that Bolar's generic alternative could be mistaken for CIBA's APRESAZIDE, that many pharmacists had been convicted of illegal substitution in New York, that Bolar copied APRESAZIDE's trade dress with intent to increase sales of its generic alternative at the expense of CIBA, and that Bolar copied CIBA's trade dress with the intent, at least in part, to induce illegal substitution.8 Irrespective of the analysis by the trial court as to the significance of footnote 13 in the Ives opinion, on the basis of the record before us, we cannot say with definite and firm conviction that the trial court's finding of intent to facilitate illegal substitution was in error. Just as the trial judge should not have been reversed by the Court of Appeals in Ives, Judge Sarokin in this case, having had the opportunity to evaluate the credibility of witnesses and weigh the evidence, must be sustained on this issue because his findings are not clearly erroneous. We therefore affirm the district court's disposition of CIBA's federal passing off claim.

Finally, the district court was not prevented from granting the injunction by CIBA's unclean hands. A successful unclean hands defense to an injunction proceeding requires a showing by defendant that plaintiff's conduct is inequitable and that it involves the subject matter of the plaintiff's claim. Edelman v. Salomon, 559 F. Supp. 1178, 1187 (D.Del.1983). Bolar's assertions that CIBA marketed APRESAZIDE as a "step-two" drug when it is in fact a "step-three" drug,9 that CIBA sold two adulterated batches of APRESAZIDE in 1976, and that APRESAZIDE's markings do not conform with FDA standards do not implicate transgressions so egregious as to merit the preclusion of an injunction. Moreover, the alleged wrongdoings do not relate to the subject matter of Bolar's claim and, consequently, we cannot say the district court committed reversible error in rejecting Bolar's unclean hands defense.

For the reasons stated above, we will affirm the order of the district court granting CIBA a permanent injunction against Bolar.10

GILES, District Judge, dissenting.

I am unable to agree with the conclusions reached by the majority on the issues of secondary meaning and passing off. Accordingly, I respectfully dissent. Although the court below articulated the appropriate legal standard for secondary meaning under both New Jersey and federal law, it failed to apply properly that standard to the record evidence. With respect to passing off, I conclude that the majority should have decided what effect, if any, Inwood v. Ives, 456 U.S. 844, 102 S. Ct. 2182, 72 L. Ed. 2d 606 (1982), has on the controlling precedent in this Circuit, SK&F Co. v. Premo Pharmaceutical Laboratories, Inc., 625 F.2d 1055 (3d Cir.1980). Resolution of this issue would provide necessary guidance to the lower courts within this Circuit. In addition, I conclude that there was insufficient evidence to support the district court's finding that Bolar copied the trade dress of APRESAZIDE with the intent to induce illegal substitution. Therefore, I would reverse the decision below and dissolve the permanent injunction.

As the majority points out, because the court below converted the injunction from preliminary to permanent without making any additional findings, we must view the previous determinations of likelihood of success on the merits as ultimate findings of success on the merits. That quantum of evidence which suffices for the grant of a preliminary injunction will not necessarily be sufficient for permanent injunctive relief. Therefore, we are obligated to review the record in this case afresh.

To prevail on a claim of unprivileged imitation, CIBA must prove both nonfunctionality and secondary meaning. This is true under both the Lanham Act and New Jersey common law. See e.g., SK & F, 625 F.2d at 1062-63, 1065; French American Reeds Mfr. Co., Inc. v. Park Plastics Co., Inc., 20 N.J.Super. 325, 333-34, 90 A.2d 50, 54-55 (1952). I agree with the majority's analysis and conclusions with respect to functionality. However, I am unable to agree that CIBA has proven that the trade dress of APRESAZIDE has acquired secondary meaning. Although the federal and state law is almost identical on this issue, I will analyze each.

To establish secondary meaning, "a manufacturer must show that, in the minds of the public, the primary significance of a product feature or term is to identify the source of the product, rather than the product itself." Ives, 102 S. Ct. at 2186 n. 11 (citing Kellogg Co. v. National Biscuit Co., 305 U.S. 111, 118, 59 S. Ct. 109, 113, 83 L. Ed. 73 (1938)). See also Freixenet, S.A., et al. v. Admiral Wine & Liquor Co., et al., 731 F.2d 148, at 152 (3d Cir.1984); Ideal Toy Corp. v. Plawner Toy Mfr. Corp., 685 F.2d 78, 82 (3d Cir.1982); Scott Paper Co. v. Scott's Liquid Gold, Inc., 589 F.2d 1225, 1228 (3d Cir.1978). This court stated in Ideal Toy; "[w]hen the primary significance of the trade dress to a consumer is in designating not the product but its producer, the trade dress has acquired secondary meaning." Ideal Toy, 685 F.2d at 82 (emphasis added). At the very least, it must be proven that the consumer links the trade dress with a single, albeit unnamed, source. See Processed Plastic Co. v. Warner Comm., Inc., 675 F.2d 852, 856 (7th Cir.1982); Scott Paper, 589 F.2d at 1228. Various factors have been held to be probative of secondary meaning, including the extent of sales and advertising, the length of use, buyer association and the fact of copying. Ideal Toy, 685 F.2d at 82.

In finding that CIBA had established secondary meaning, the district court relied primarily upon the factors identified in Ideal Toy, including the extensive advertising to promote APRESAZIDE, the continuous use of the trade dress, the drug's great commercial success and the fact of copying. See CIBA-GEIGY v. Bolar Pharmaceutical Co., Inc., 547 F. Supp. 1095, 1101-102, 1105, 1113-14 (D.N.J.1982). Likewise, the majority relies upon these factors in affirming the district court's conclusions on this issue. See Majority Opinion at 852-853. However, these factors are merely indicia of secondary meaning. Focus should not be diverted from the main inquiry--whether the trade dress signifies the producer or source of the product in the minds of the public. The district court found that "patients taking APRESAZIDE products have come to associate the trade dress of those capsules with the brand name APRESAZIDE as representing a single source of supply although not with the particular manufacturer." CIBA, 547 F. Supp. at 1102.1 Thus, the association between the trade dress and the name APRESAZIDE was deemed sufficient to denote a single source. However, there was simply no evidence that patients link the word APRESAZIDE with a single, albeit anonymous source. Nor was there any evidence to suggest that patients recognized that the word APRESAZIDE was a brand name and registered trade-mark.2 Rather than denoting the source or a source, the word APRESAZIDE probably identifies the product and its function, (e.g., "my blood pressure medicine"), which is not sufficient for secondary meaning. See Ives, 102 S. Ct. at 2186 n. 11; Ideal Toy, 685 F.2d at 82.

The promotional techniques necessarily employed by the pharmaceutical industry set it apart from more consumer-oriented fields. All promotional efforts are directed at physicians and pharmacists, leaving the ultimate consumer uneducated as to the source of the prescription drug. On this score, I am persuaded by the observations and reasoning of Judge Nickerson in the district court opinion in Ives. He stated:

No doubt to physicians and pharmacists the word "Cyclospasmol" signifies a source. But there is no reason to suppose this is so as to patients. They have not been the target of the sales campaigns, and they play no role in choosing what drugs to buy. Nothing in this record suggests that they 'understand by the word anything more than a kind of drug to which for one reason or another they [have] become habituated.' Bayer Co. v. United Drug Co., 272 Fed. 505, 510 (S.D.N.Y.1921) (L. Hand, J.). Such patients as read their prescriptions do not attribute to Cyclospasmol "any other meaning than as an ingredient in a general compound, to which faith and science might impart therapeutic virtue." Id.

Thus, one cannot infer from the fact that patients recognize the name Cyclospasmol that they appreciate that the drug they receive comes from a single source. If they do not know that the word designates a single source, it follows that a color they associate with the name does not signify to them a single source and cannot acquire a secondary meaning.

Ives Laboratories, Inc. v. Darby Drug Co., Inc., 488 F. Supp. 394, 400 (E.D.N.Y.1980), rev'd on other grounds, 638 F.2d 538 (2d Cir.1981), rev'd sub nom., Inwood v. Ives, 456 U.S. 844, 102 S. Ct. 2182, 72 L. Ed. 2d 606 aff'd on remand, 697 F.2d 291 (2d Cir.1982).

In the context of the pharmaceutical industry, the Ideal Toy factors also become somewhat less meaningful. The amount of advertising and promotion, which normally helps to familiarize consumers with a product, is not relevant as it is never seen by the public. Similarly, the volume of sales is not telling where the consumer does not choose the particular drug. The continuous use of the trade dress is unimportant unless the public recognizes that only one manufacturer uses that color, shape and size. The act of copying, standing alone, is not sufficient to prove secondary meaning.

In summary, there is no evidence on this record that the consuming public links the trade dress of APRESAZIDE with a single source, named or unnamed. Accordingly, I would hold that CIBA has failed to prove secondary meaning under federal law.3

Under New Jersey law secondary meaning has been defined as the "identification of an object ... with a single source--even though unnamed--and an interest on the part of the consuming public in buying because of that source." Squeezit Corp. v. Plastic Dispensers, Inc., 31 N.J.Super. 217, 223, 106 A.2d 322 (1954). Similarly, in French American Reeds Mfr. Co. v. Park Plastics Co., Inc., 20 N.J.Super. 325, 90 A.2d 50 (1952), the Superior Court of New Jersey stated:

[I]t is an absolute condition to any relief whatever that the plaintiff ... show that the appearance of his wares has in fact come to mean that some particular person--the plaintiff may not be individually known--makes them, and that the public cares who does make them .... The critical question of fact at the outset always is whether the public is moved in any degree to buy the article because of its source and what are the features by which it distinguishes that source.

Kilborn & Bishop Co., 247 Fed. 299, 300-31 (2d Cir.1917). Although the New Jersey test for secondary meaning is virtually the same as the federal standard, New Jersey law adds a second prong to the inquiry--public interest. Therefore, CIBA would have to show not only the link with a single source, but also that the public wants that product because of the source. This would seem to entail some recognition that APRESAZIDE is a brand name, rather than a generic drug, and that its status as such makes it more desirable to the public. Although desirability may be inferred, it brings CIBA no closer to proving that the public associates the trade dress of APRESAZIDE with a particular or single source.

In French American Reeds, the court listed factors similar to the ones relied upon by the district court, including length of use, sales volume, advertising and "efforts to promote a conscious connection in the minds of the public between the product and its source."4 20 N.J.Super. at 337, 90 A.2d at 56. However, the court was quick to point out that these are only factors to be considered, not dispositive criteria establishing secondary meaning. The court explained that "[t]he ultimate test of secondary meaning must always remain whether plaintiff's product in its distinctive trade dress has become broadly known to the public as indicating a certain origin." Id. at 338, 90 A.2d at 57. I have previously concluded that there is no evidence to support a finding that the public associates the appearance of APRESAZIDE with a particular or single source. Therefore, CIBA has also failed to prove secondary meaning under the law of New Jersey.

In SK&F, this court held that liability could be imposed for passing off under section 43(a) of the Lanham Act, 15 U.S.C. Sec. 1125(a) (1982), and New Jersey law where it is proven that a drug is placed "in the hands of a pharmacist in a form in which the manufacturer can reasonably anticipate that it may be passed off as another product even if the manufacturer does nothing else to encourage passing off." SK&F, 625 F.2d at 1062, 1065-66. The SK&F court, like the court below and the majority here, concluded that it was reasonable to anticipate that duplication of a trade dress would facilitate passing off. SK&F, 625 F.2d at 1063; CIBA, 547 F. Supp. at 1116; Majority Opinion at 854. With this conclusion, I have no quarrel. A manufacturer can always reasonably anticipate that the act of copying a trade dress will make illegal substitutions easier. I question the continuing vitality of this "reasonable anticipation" test in light of the United States Supreme Court decision in Ives. Further, I conclude that the district court's alternate finding of intentional inducement under Ives is not supported by the record evidence. Moreover, it is necessary to consider what effect the demise of the reasonable anticipation standard has on this court's prediction of how the Supreme Court of New Jersey would decide state law on the subject.

Stating that "it is debatable whether Ives altered the standard of proof for passing off under Sec. 43(a)...." the majority opinion refuses to reach the issue, relying instead upon the district court's alternate finding of intentional inducement. See Majority Opinion at 854. I share neither the majority's doubt nor hesitancy on this issue.

Ives was also a "look-alike" drug case. However, it was decided on the narrow ground of the appropriate standard of review.5 Justice White, in a concurring opinion, took the majority to task for not deciding the issue presented in the petitions for certiorari--whether the Second Circuit erred in applying the reasonable anticipation standard to a claim of contributory infringement under the Lanham Act. Ives, 102 S. Ct. at 2190-91 (White, J., concurring). Citing Warner & Co. v. Eli Lilly & Co., 265 U.S. 526, 44 S. Ct. 615, 68 L. Ed. 1161 (1924), Justice White chracterized the reasonable anticipation standard as "watered down." Id. 102 S.Ct. at 2191. He explained that "[t]he 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." Id. 102 S.Ct. at 2192.

In response to Justice White's criticism, the majority stated that if the Second Circuit had relied upon the reasonable anticipation standard, "the court indeed would have applied a 'watered down' and incorrect standard." Ives, 102 S. Ct. at 2188 n. 13.6 Both the majority and concurrence agreed upon the proper legal test. Liability for passing off or contributory infringement must be established by proof that the manufacturer intentionally induced another to pass off or continued to supply the product, knowing, or having reason to know, that the pharmacist was engaging in illegal substitutions. See id. 102 S.Ct. at 2188, 2191-92 (White, J., concurring).

The passing off claim in Ives was brought under section 32 of the Lanham Act, while CIBA prosecutes the same claim in this case under section 43(a). The issue becomes whether the higher standard of intentional inducement should also be applied to section 43(a).

The Ives Court noted that "Sec. 43(a) prohibits a broader range of practices than does Sec. 32...." 102 S. Ct. at 2190 (emphasis added).7 A brief discussion of these two statutory provisions will illustrate that this simply means that section 43(a) proscribes a greater number of activities than does section 32. See generally SK&F, 625 F.2d at 1065 (discussing various types of behavior actionable under section 43(a)). On its face, section 32 is a very narrow provision, limited to action for infringement of a registered trademark.8 Although duplication of a trade dress is not, in itself, infringement of a trademark, the imitative appearance might induce a pharmacist to place a generic drug in a bottle bearing the brand name. Since the brand name is generally a registered trade-mark, the manufacturer can be held liable for contributory infringement of the inducement of passing off under this section. In contrast, section 43(a) does not require infringement of a mark, prohibiting instead any false designation of origin.9 The mere act of trade dress duplication is actionable under this section if it tends to misrepresent a product's origin. Passing off is also actionable, affording a plaintiff like CIBA a choice of statutory remedies.10 However, the nature of the activity, duplicating a drug's trade dress, and the potential for wrongdoing by an unscrupulous pharmacist, are the same regardless of which statutory section a plaintiff invokes. Consistency and logic dictate, therefore, that the same legal standard should be applied to the same behavior under both sections 32 and 43(a). To hold that a plaintiff need only prove reasonable anticipation under section 43(a) would render section 32 superfluous and eviscerate the intentional inducement standard of Ives. No plaintiff would voluntarily choose to proceed under section 32, given the higher burden under Ives, when they could invoke section 43(a) and be required to prove only reasonable anticipation. Moreover, the Ives court indicated that the standard for functionality is the same under both statutory sections. See 102 S. Ct. at 2190 n. 20; id. 102 S.Ct. at 2192-93 (White, J., concurring). It follows that the same must be true for "passing off." Thus, I would hold that Ives signaled the demise of the reasonable anticipation standard, implicitly overruling that portion of this court's decision in SK&F.

The district court did address the Ives intentional inducement standard, deciding in the alternative should this court hold that the SK & F reasonable anticipation standard was no longer viable. It found that Bolar copied the trade dress of APRESAZIDE with the intent to facilitate illegal substitutions. CIBA, 547 F. Supp. at 1108, 1116. In support of this finding, the court below explained that there was no legitimate reason for the duplication. Id. at 1108.11 The only other evidence relied upon by the court was the conviction of ninety-three (93) New York pharmacies and pharmacists for illegal substitutions. Id. at 1108 n. 9. This evidence was deemed sufficient by the majority in concluding that the district court's finding of intent to induce was not clearly erroneous. Majority Opinion at 855. I do not agree.

At the outset, I fail to see the relevance of the New York convictions. There is no evidence that these pharmacies had been passing off products produced by Bolar. Indeed, there was no suggestion that the convictions involved the illegal substitution of a generic for APRESAZIDE. These convictions may prove that illegal substitution is a problem, at least in New York, but, they prove nothing with respect to Bolar's actions or intent in duplicating the trade dress of APRESAZIDE.

The district court found that Bolar had the intent to induce illegal substitution. There was no finding of actual inducement. There is no claim that Bolar duplicated the packaging of its drug so that pharmacists might be innocently misled. Nor is there any evidence that Bolar said or did anything that could be reasonably construed as encouraging illegal substitution. According to the language of Ives, the manufacturer must intentionally induce another to pass off or continue to supply goods to one whom it knows or has reason to know is passing off. See 102 S. Ct. at 2188; id. 102 S.Ct. at 2191-92 (White, J., concurring) (emphasis added). This language requires some affirmative action apart from the act of copying. Such affirmative action is absent from this record.

Assuming that Ives punishes merely the intent to induce illegal substitution, the result remains unchanged. Proof of something in addition to the act of copying is required. It is beyond question that the Ives standard purports to have substantially more bite than the reasonable anticipation test. Under the latter, the act of copying proves the offense. By imitating the trade dress of a drug, it is virtually impossible to fail to anticipate that "it may be passed off as another product...." SK & F, 625 F.2d at 1062. Thus, to satisfy the higher standard of Ives, it would seem that proof of something more than the act of imitation is required. However, under the district court's analysis, the same quantum of evidence which would satisfy the reasonable anticipation standard would also suffice to prove intentional inducement. Under either, the act of copying would prove the offense.12 However, the rejection of the reasonable anticipation test by the Ives Court as " 'watered down' and incorrect....", 102 S. Ct. at 2188 n. 13, mandates that something in addition to the act of imitation is required to satisfy the intentional inducement standard. Since there is insufficient evidence to support the district court's finding of intent to induce, I conclude that this finding is clearly erroneous.

Relying upon SK & F, the majority assumes that the reasonable anticipation standard is the law of New Jersey. See Majority Opinion at 852. However, the court in SK & F sent out conflicting signals, first observing that New Jersey follows sections 711 and 713 of the Restatement (First) of Torts. Section 713 defines passing off, stating, "[o]ne fraudulently markets his goods as those of another if, though making no misrepresentation himself, he intentionally induces his purchasers to so market them." SK & F, 625 F.2d at 1062 (quoting Restatement (First) of Torts Sec. 713 (1938)) (emphasis added). This language sounds strikingly similar to that employed in Ives. The SK & F court quoted the illustration following section 713, which was derived from Warner & Co. v. Eli Lilly & Co., 265 U.S. 526, 44 S. Ct. 615, 68 L. Ed. 2d 1161 (1924). It must be noted that Warner was cited by both the majority and the concurrence in Ives as support for the intentional inducement standard. See Ives, 102 S. Ct. at 2188; id. 102 S.Ct. at 2191 (White, J., concurring). The SK & F court then stated that the reasonable anticipation test "has been held ... actionable ... under New Jersey law...." 625 F.2d at 1062 (citing Merrell-National Labs., Inc. v. Zenith Labs., Inc., 194 U.S.P.Q. 157, 159-60 (D.N.J.1977), aff'd on other grounds, 579 F.2d 786 (3d Cir.1978)). Other than citing cases from other jurisdictions, decided under both state and federal law, the court engaged in no further analysis of the Merrell decision or New Jersey state Law.13 See SK & F, 625 F.2d at 1062.

While Merrell was a look-alike drug case, it differs from this case in that it involved actual passing off by pharmacists, as well as a claim for contributory infringement on the part of the manufacturers. With respect to the latter, that court found that there was no New Jersey precedent on point. Merrill, 194 U.S.P.Q. at 160. Thus, it had to "predict" how the Supreme Court of New Jersey would rule, if presented with a claim of contributory infringement. Looking at "general tenants" of New Jersey unfair competition law, the court concluded that "fair play has become the watchword...." Id. The Court then discussed some rather dated federal cases which endorsed the reasonable anticipation standard. Based upon these cases and the "general tenants," the court concluded that plaintiff was likely to prevail against the manufacturers on a claim of unfair competition.14 Id.

Nothing in the "general tenants" of New Jersey unfair competition law, as articulated by the Merrell court, tends to prove that New Jersey would adopt the reasonable anticipation standard over the test of intentional inducement. Indeed, the SK & F court's acknowledgment that New Jersey follows sections 711 and 713 of the Restatement (First) of Torts would indicate the exact opposite. The only insight provided by New Jersey courts is found in dicta in the French American Reeds case. The Superior Court of New Jersey noted that the action did not present questions of passing off or "fraudulent marketing," citing the Restatement (First) of Torts Sec. 711 (1938). Quoting plaintiff, the court stated that the claim did "not involve any element of actual intent or any attempt on respondents' part to see their goods as those of appellant." French American Reeds, 20 N.J.Super. at 329, 90 A.2d at 52. If reasonable anticipation were the standard in New Jersey, the plaintiff in French American Reeds might not have limited the case to claims of unprivileged imitation.

Discounting the so-called "general tenants" of unfair competition, the Merrell decision, upon which the SK & F court relied, is based upon an analogy to federal law. This analogy is no longer justified in the wake of Ives. Assuming the reasonable anticipation standard has been overruled as a matter of federal law, it follows that it cannot be viable under New Jersey law in the absence of any endorsement of that standard by the courts of New Jersey. I would hold that, in accordance with the Restatement (First) of Torts Secs. 711, 713 (1938), and Warner & Co. v. Eli Lilly & Co., 265 U.S. 526, 44 S. Ct. 615, 68 L. Ed. 1161 (1924), New Jersey adheres to the traditional intentional inducement standard. For the reasons stated in the previous section, there is insufficient evidence to hold Bolar liable for passing off under this standard. Accordingly, I would reverse the decision below in its entirety.