Panduit appeals a preliminary injunction barring it from producing cable ties with certain features disclosed in an expired patent. Because the plaintiff has not established that the features are protectable trade dress, we reverse.

The combatants in this litigation are the nation's largest suppliers of cable ties, small nylon belts used to bundle wires. Cable ties, simple devices which cost pennies apiece to produce, are ubiquitous in industries which involve electronics, which is to say almost all industries; billions are sold each year.

Cable ties consist of a strap terminating at one end in a tapered tail and at the other in a head which incorporates a horizontal slit and locking mechanism. The strap is looped around the bundle of wires to be secured and the tail is inserted through the slit in the head and pulled snug--first by hand, then by a device which tightens the loop around the bundle and snips off the excess tail flush with the head of the tie. The result is a tight bundle with only the thickness of the head protruding above the strap and bundled wires. The head is made as small as possible to avoid snagging on other wires or components with which the bundle may come in contact.

Cable ties feature two types of locking mechanisms: one-piece or two-piece. In the one-piece system, a nylon pawl molded into the head of the cable tie protrudes into the horizontal slit through which the strap passes. The strap in turn incorporates, along essentially its entire length, small, closely spaced, transverse ridges or teeth. When the strap is inserted through the slit, the pawl engages the ridges on the strap like a one-way ratchet allowing the loop to be tightened around the bundle but not lengthened. Military size and strength requirements and the need to fit the nylon pawl mandate an essentially rectangular shape for the head of the one-piece cable tie. Defendant Panduit, the largest of a handful of producers of one-piece cable ties, sells its product under the trade-mark PAN-TY. Thomas & Betts ("T & B"), the plaintiff in this case, and several other producers also make one-piece cable ties, which like Panduit's meet military specifications and have rectangular heads [Pl.Ex. 233].1

In the two-piece tie, the nylon pawl is replaced by a metal barb which is inserted into the head of the tie and sits in a slot transverse to the slit for the strap. When the strap is inserted through the slit, the barb flexes into the transverse slot. When the strap is pulled taut, the tension on the strap causes the flexed barb to bite into the nylon strap and hold tight. Because the metal barb eliminates the need for teeth or ridges on the strap, the two-piece cable tie has a number of advantages, including infinite adjustability and a stronger strap. There are also disadvantages to this design, however, including the danger, real or perceived, that the metal barb may come loose and short out surrounding electronic equipment.

In 1965, T & B obtained a patent on the two-piece cable tie ("the Schwester patent"). That patent disclosed a two-piece cable tie with an oval head, metal barb and transverse slot. The top half of the slot allows for the insertion of the barb during manufacturing and the bottom half is required for the barb to flex and bite. The slot, barb and head portion are elements in each of the patent's claims. The oval shape of the head is not specifically claimed but is illustrated and described in the specifications. T & B currently markets a two-piece cable tie under the trademark TY-RAP that is essentially identical to the model disclosed in the Schwester patent. Though the Schwester patent expired in 1982 and a related patent also held by T & B expired in 1986, until 1993 T & B remained the sole producer of two-piece cable ties with annual sales of almost $100 million worldwide.

In 1993, Panduit entered the two-piece cable tie market with the BARB-TY, an oval-headed, metal-barbed cable tie essentially identical to T & B's TY-RAP. T & B promptly sued Panduit for trade dress infringement under Section 43(a) of the Lanham Act, 15 U.S.C. Sec. 1125(a), and the common law. T & B claimed that its trade dress "includes a rounded, low profile head configuration with a vertical slot which is aligned with the body of the cable tie and, in the case of Thomas & Betts' dominant product, includes a metal barb visible in the lower portion of the vertical slot." [R. 8 at 5]. T & B also brought an unfair competition claim under the same provision of the Lanham Act based on Panduit's use of the BARB-TY trademark and ancillary state law claims under the Illinois Consumer Fraud and Deceptive Business Practices Act, the Uniform Deceptive Trade Practices Act and the Illinois Anti-Dilution Act.

T & B moved for a preliminary injunction prohibiting Panduit from selling its BARB-TY. After an evidentiary hearing, the magistrate judge found in an order dated December 14, 1994, that "the transverse slot, the vertical slot, the steel barb and the tail or body" were all "functional parts" and thus not entitled to trade dress protection [Def.App. 48], but that the oval shape of the head was not functional and thus was protectable trade dress. The magistrate judge preliminarily enjoined Panduit "from further sales, advertising, promotion, or marketing of cable ties incorporating a steel barb locking mechanism with an oval shaped head and a transverse or vertical slot in the head." [Def.App. 56].

Panduit's request for a stay of the injunction was denied by both the magistrate judge and a panel of this Court. Panduit now appeals, contending that the magistrate judge erred in finding that T & B had established some likelihood of success on the merits. In reviewing the grant of a preliminary injunction we review factual findings for clear error and legal conclusions de novo. Schwinn Bicycle Co. v. Ross Bicycles, Inc., 870 F.2d 1176, 1181 (7th Cir.1989).

Trademark law aims to aid consumers in identifying the source of goods by allowing producers the exclusive right to particular identifying words or symbols which they may attach to their products as a designator of source. Allowing a particular producer to monopolize a symbol in this way is no burden on competition, the theory goes,2 because symbols are a dime a dozen. The only value of the initially arbitrary symbol comes from its association with the producer's products and the good or bad will consumers feel toward that producer based on the quality of those products. Because the symbol itself adds nothing to the product, consumer desire for products marketed with that symbol must derive solely from the belief that products bearing the mark originated with the producer for whom the consumers have developed goodwill. Therefore, the only reason a competitor would copy a mark would be to pass off his product as that of the original producer.

This simple theory becomes a bit more problematic when a trademark is descriptive or has some other value apart from its associated goodwill. In such a case, a court must determine whether the mark really signifies a particular source or the product itself. Still, the dispute concerns limitations on the language competitors can use to communicate with consumers about their products and the effect such limitations would have on competition. The products themselves remain unconstrained.

Trade dress protection in the configuration of the product itself opens up another can of worms. While it is true that features of a product may serve primarily to signify its source (e.g., the pink color of Owens-Corning fiberglass insulation, see In re Owens-Corning Fiberglas Corp., 774 F.2d 1116 (Fed.Cir.1985)), these features are still part of the product itself and may be valued by consumers as such ("I just could not sleep at night if the insulation above my head wasn't pink"). Trade dress protection only extends to the role of such features as signifier of source; when competitors are barred from duplicating features whose value to consumers is intrinsic and not exclusively as a signifier of source, competition is unduly hindered.

[I]mitation is the life blood of competition. It is the unimpeded availability of substantially equivalent units that permits the normal operation of supply and demand to yield the fair price society must pay for a given commodity. [citations omitted].

American Safety Table Co., Inc. v. Schreiber, 269 F.2d 255, 272 (2nd Cir.1959), certiorari denied, 361 U.S. 915, 80 S. Ct. 259, 4 L. Ed. 2d 185 (1959). Copying is not only good, it is a federal right--a necessary complement to the patent system's grant of limited monopolies.

Both the novelty and the nonobviousness requirements of federal patent law are grounded in the notion that concepts within the public grasp, or those so obvious that they readily could be, are the tools of creation available to all. They provide the baseline of competition upon which the patent system's incentive to creative effort depends.

Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 156, 109 S. Ct. 971, 980, 103 L. Ed. 2d 118 (1989).

The right to copy is even more robust when the copied product was previously patented but the patent has expired. In that case, the original producer has reaped his reward of a 17-year monopoly and the public has already "paid the congressionally mandated price for disclosure." Id. at 152, 109 S.Ct. at 978.

It is self evident that on the expiration of a patent the monopoly granted by it ceases to exist, and the right to make the thing formerly covered by the patent becomes public property. It is upon this condition that the patent is granted. It follows, as a matter of course, that on the termination of a patent there passes to the public the right to make the machine in the form in which it was constructed during the patent.

Kellogg Co. v. Nat. Biscuit Co., 305 U.S. 111, 120, 59 S. Ct. 109, 114, 83 L. Ed. 73 (1938), quoting Singer Mfg. Co. v. June Mfg. Co., 163 U.S. 169, 185, 16 S. Ct. 1002, 1008, 41 L. Ed. 118 (1896).

On the one hand then, trademark law allows a producer to prohibit the copying of a product feature which serves as a signifier of source in order to preserve his reputation and the goodwill consumers have for his brand. On the other hand, effective competition and the penumbra of the patent laws require that competitors be able to slavishly copy the design of a successful product.

[T]he plaintiff has the right not to lose his customers through false representations that those are his wares which in fact are not, but he may not monopolize any design or pattern, however trifling. The defendant, on the other hand, may copy plaintiff's good slavishly down to the minutest detail: but he may not represent himself as the plaintiff in their sale.

Bonito Boats, 489 U.S. at 157, 109 S.Ct. at 981, quoting Crescent Tool Co. v. Kilborn & Bishop Co., 247 F. 299, 301 (2d Cir.1917) (L. Hand, J.).

To police this distinction courts require as a prerequisite to protection that trade dress that is not inherently distinctive have acquired secondary meaning.3 " '[S]econdary meaning' is acquired when 'in the minds of the public, the primary significance of a product feature ... is to identify the source of the product rather than the product itself.' " Qualitex Co. v. Jacobson Products Co., Inc., --- U.S. ----, ----, 115 S. Ct. 1300, 1303, 131 L. Ed. 2d 248 (1995), quoting Inwood Laboratories, Inc. v. Ives Laboratories, Inc., 456 U.S. 844, 851 n. 11, 102 S. Ct. 2182, 2187 n. 11, 72 L. Ed. 2d 606 (1982). In Inwood Laboratories, for example, the secondary meaning question was whether the color of a pill signifies to most consumers a particular name-brand pharmaceutical or a chemical compound with particular medicinal properties.

The issue is slightly different in most trade dress product configuration cases, because product features, though they may identify the source of the product, do not really identify the product--they are the product. See Duraco Products, 40 F.3d at 1431, 1440-1441. The issue is not the primary meaning to consumers of a particular signifier, but is instead whether a product feature's primary significance to consumers is as an identifier of source or as an element which contributes to the inherent appeal of the product. See Kellogg, 305 U.S. at 120, 59 S.Ct. at 114 (denying trademark protection to "pillow shape" of shredded wheat: "a form in which the public has become accustomed to see the article and which, in the minds of the public, is primarily associated with the article rather than a particular producer.").

It is not enough that the consumers associate the form of the product with a particular producer. Such an association is inevitable when the first comer is the exclusive producer under a patent. Consumers must also care that the product comes from a particular producer (though they need not be able to identify him) and must desire the product with the particular feature because it signifies that producer. This Court articulated the requirement admirably in a state law unfair competition case brought by the first producer of jeweled "brodie knobs":4

Sinko created a desire on the part of the public for one of two things, either for knobs made by Sinko, above all other knob makers, or for knobs made in a particular manner regardless of who made them. If it is the first situation, the law of unfair competition gives Sinko the right to monopolize or to exclude other makers from copying the product. If it is the latter situation, Sinko receives no such right to monopolize even though he might have been the first to make the article in the particularly desirable manner.

Sinko v. Snow-Craggs Corp., 105 F.2d 450, 453 (7th Cir.1939).

Even if a product feature has acquired secondary meaning, as described above, the doctrine of functionality bars its protection as trade dress if the "feature is one that is costly to design around or do without, rather than one that is costly to have." Schwinn Bicycle Co. v. Ross Bicycles, Inc., 870 F.2d 1176 (7th Cir.1989).

Panduit contends that even if a product feature has acquired secondary meaning and is non-functional, it still cannot be protected trade dress if it was disclosed in an expired patent. Language in Scott Paper Co. v. Marcalus Co., 326 U.S. 249, 256, 66 S. Ct. 101, 104-05, 90 L. Ed. 47 (1945), supports its claim:

By the force of the patent laws not only is the invention of a patent dedicated to the public upon its expiration, but the public thereby becomes entitled to share in the good will which the patentee has built up in the patented article or product through the enjoyment of his patent monopoly. Hence we have held that the patentee may not exclude the public from participating in that good will or secure, to any extent, a continuation of his monopoly by resorting to the trademark law and registering as a trademark any particular descriptive matter appearing in the specifications, drawings or claims of the expired patent, whether or not such matter describes essential elements of the invention or claims.

Since the oval head was disclosed in the specification and drawing of the expired Schwester patent, Panduit wins, or so it says. Though references to trademark law in Scott Paper are dicta since that case involved assignor estoppel, not trademarks, Panduit may be right. A recent Tenth Circuit case comes close to adopting Panduit's proposed rule. Vornado Air Circulation Systems, Inc. v. Duracraft Corp., 58 F.3d 1498 (10th Cir.1995), held that "where a disputed product configuration is part of a claim in a utility patent, and the configuration is a described, significant inventive aspect of the invention, see 35 U.S.C. Section 112, so that without it the invention could not fairly be said to be the same invention, patent law prevents its protection as trade dress, even if the configuration is nonfunctional." Id. at 1510.

Vornado claimed trade dress protection for its spiral grill design. The spiral grill was also one element in a claim in a still valid utility patent. The patent specifications claimed that the spiral grill produced optimal air flow when in fact it made no real difference--as Vornado argued somewhat inconsistently in seeking trade dress protection. Because the air flow was not superior to a standard radial grill, the district court found the grill design to be non-functional and protectable trade dress. The Tenth Circuit reversed.

The present case is distinguishable from Vornado but it is not clear that the differences dictate a different result. The oval shape of the cable tie head was not specifically claimed in the Schwester patent, a point that T & B makes much of, while the spiral grill was a required element in at least one of the claims in the patent at issue in Vornado. Therefore, one could infringe T & B's patent without infringing its trade dress while any product which infringed the Vornado patent (at least the claim in which the spiral grill was an element) would also infringe its claimed trade dress.

Whether or not a feature is claimed is not necessarily, however, a good indicator of its relative importance to the invention as a whole. Additional elements in a claim narrow the scope of the patentee's monopoly since an infringing invention must contain each element of a claim or its equivalent. Lemelson v. United States, 752 F.2d 1538, 1551 (Fed.Cir.1985). Therefore, a patent applicant seeks to draw his claims as broadly as possible with as few specific elements as he can get away with. The fact that Vornado included the spiral grill as an element of a claim likely means that it could not have received a patent on the rest of the invention without including the spiral grill limitation--not that the spiral grill itself was a "significant inventive aspect." This is particularly true in Vornado where the spiral grill itself was not independently claimed (that is why Duracraft's fan did not infringe the still valid patent) because it was disclosed in the prior art.

Likewise, the fact that the Schwester patent includes the cable tie head as a required element without claiming a particular shape means that T & B was able to obtain a broader monopoly, i.e., they were able to exclude competitors from making metal barbed cable ties with any head shape. There is no reason, therefore, to infer from its absence as an element in a claim that the shape of the head is any less important to the two-piece cable tie than the spiral grill is to Vornado's fan. All one can infer is that because it was not separately claimed, the shape of the head, like the spiral grill, was not independently patentable.

The focus on what is or is not claimed is misguided for another more fundamental reason. In the patent "bargain," the claims define what the patentee receives, the "metes and bounds" from which he can exclude competitors. What the public receives is the entire invention as disclosed in the claims but primarily in the patent specifications which are required to "contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention." 35 U.S.C. Sec. 112.

The present case, however, does not require that we pass on the correctness of Vornado or Panduit's broader claim. On this record, T & B has failed to clear the initial hurdle of establishing that the TY-RAP's oval-shaped head is protectable trade dress.

Underlying the magistrate judge's analysis is the erroneous belief that Panduit's copying is in itself a wrong.5 In deciding unfair competition cases, "[o]ur natural inclination to disapprove of such conduct must give way to the public policy favoring competition, even by slavish copying." Keene Corp. v. Paraflex Industries, Inc., 653 F.2d 822 (3rd Cir.1981).

More important, the magistrate judge did not distinguish between goodwill toward the producer--all that trade dress law protects--and goodwill toward the article--"the attractive features, of whatever nature, that the product holds for consumers"--which is freely appropriable by second comers. Versa Products Co., Inc. v. Bifold Co. (Mfg.) Ltd., 50 F.3d 189, 215 (3rd Cir.1995). Although we have concerns about the judge's findings on nonfunctionality6 and likelihood of confusion,7 we will focus our analysis on secondary meaning.

In concluding that T & B "had established a strong likelihood of success in proving secondary meaning in spite of the fact that the trade dress consists largely of a simple geometric configuration" [Def.App. 34-35], the magistrate judge defined secondary meaning "as an association in the mind of the consumer between the trade dress of a product and a particular source." [Def.App. 33]. This is not the proper legal standard.8 As discussed above, in order to establish that its product feature possesses secondary meaning, T & B must prove that in the minds of consumers the primary significance of the oval head is to identify the TY-RAP as a T & B product. None of the evidence upon which the magistrate judge relied--T & B's advertising, T & B's consumer survey, Panduit's deliberate copying and Panduit's reference to the product's "classic design"--supports a finding of secondary meaning when the proper standard is considered. While we appreciate the magistrate judge's characteristically thorough and thoughtful work, we nevertheless hold that his conclusion was thus clearly erroneous.

The magistrate judge considered T & B's "substantial advertising with pictures of its product as well as large models used in trade shows" [Def.App. 33] as substantial evidence of secondary meaning even though "the advertising has not been directed specifically at calling the consumers attention to the oval shape of the TY-RAP heads." [Def.App. 33 n. 1].

Advertising which encourages consumers to identify the claimed trade dress with the particular producer ("look for the oval head") is some evidence of secondary meaning. First Brands Corp. v. Fred Meyer, Inc., 809 F.2d 1378, 1383 (9th Cir.1987). If a good deal of such advertising is done, the presumption is that at least some of it will take and some consumers will make the suggested connection. But as the magistrate judge noted, T & B's TY-RAP advertising never took this form. T & B's own product manager testified that to his knowledge the oval head shape was never advertised as a way to identify T & B TY-RAPS. [Tr. 149].9

Advertising that touts a product feature for its desirable qualities and not primarily as a way to distinguish the producer's brand is not only not evidence that the feature has acquired secondary meaning, it directly undermines such a finding. It supports instead the inference that consumers consider the claimed trade dress a desirable feature of the product and not primarily a signifier of source. To the extent T & B's advertising drew attention to the oval head shape at all, the proposed associations were of the latter type. T & B's merchandise rack includes a sign with an enlarged picture of TY-RAP's head next to which large letters proclaim "The Grip of Steel." Underneath, in smaller type, three bullets all draw attention to the functional advantages of the product: "Proprietary stainless steel barb molded into head giving superior strength and durability; Low profile head--won't snag looks good; Smooth body prevents breakage problems." [D. Ex. 77]. Similarly, T & B's trade show display includes the claim: "Low profile rounded head ... won't snag." [Pl.Ex. 192]. To the extent this advertising draws attention to the head shape at all, as opposed to the clearly functional barbed locking mechanism, it is only to tout claimed functional and aesthetic advantages.

T & B commissioned a survey to establish secondary meaning and likelihood of confusion. Posing as customers, the interviewers visited 154 electric supply retailers in 15 cities and asked the salespeople if they could identify the brand of a cable tie. Each salesperson was shown one of five different cable ties: 1) a T & B TY-RAP, 2) a T & B TY-RAP with the T & B logo removed, 3) a Panduit BARB-TY, 4) a Panduit BARB-TY with the Panduit logo removed, and 5) a one-piece cable tie manufactured by a company named GB.

Not surprisingly the survey concluded that the T & B TY-RAP had "developed a high level of secondary meaning" because 87% of the salespeople identified it as a T & B product and 43% of the salespeople identified it as a T & B product even with the T & B logo removed.10 The magistrate judge credited this survey as additional evidence that the T & B's oval head shape had acquired secondary meaning.

The fatal problem with the survey is that it did not separate the allegedly protectable trade dress, the oval head, from the clearly non-protectable elements, the metal barb and slot. When salespersons were queried as to how they identified the T & B TY-RAP, "as one would expect when salespeople are confronted with customers asking about a cable tie with a metal clip, their first response was to mention the metal clip." [Consumer survey at 11, Pl.App. 697]. A survey which asks consumers to identify the source of a product based on its overall configuration when most of the product's configuration is functional is worthless in determining whether a particular product feature has acquired secondary meaning. See Textron, Inc. v. U.S. Intern. Trade Comm'n, 753 F.2d 1019, 1027 (Fed.Cir.1985).

To bolster his conclusion that the oval head possessed secondary meaning, the magistrate judge found that "the evidence is undisputed that [Panduit] intentionally copied the entire line of plaintiff's TY-RAP cable ties because of the advantages this product enjoyed." [Def.App. 34]. Far from supporting a finding of secondary meaning, this conclusion explicitly undermines it. Copying is only evidence of secondary meaning if the defendant's intent in copying is to confuse consumers and pass off his product as the plaintiff's. In that situation, the defendant's belief that plaintiff's trade dress has acquired secondary meaning--so that his copying will indeed facilitate his passing off--is some evidence that the trade dress actually has acquired secondary meaning. See Blau Plumbing, Inc. v. S.O.S. Fix-it, Inc., 781 F.2d 604, 611 (7th Cir.1986). But "evidence of intent is often ambiguous," id., and this is particularly true where the product itself is copied. "[T]he copier may very well be exploiting a particularly desirable feature, rather than seeking to confuse consumers as to the source." Duraco, 40 F.3d at 1453. The magistrate judge found that Panduit copied T & B's TY-RAP knowing that it was a successful product and seeking a piece of the market for cable ties with particular features which consumers desire and have become accustomed to, i.e., oval heads and metal barbs. This copying of the "advantages that this product enjoyed" does not support an inference that any of the copied features possessed secondary meaning.11

Indeed, the record is devoid of evidence that Panduit intended to pass off the BARB-TY as T & B's product. Panduit's marketing of its BARB-TY does not suggest an attempt to pass off, but rather to distinguish the BARB-TY as a new entry in the two-piece cable tie market. Panduit's point of sale packages not only include the Panduit logo and green corporate color but also a bright yellow, black and red sticker prominently displayed on the clear plastic package that reads: "Using a nylon tie with a metal barb? Try the superior one ... BARB-TY from Panduit."

4. "Classic design"

The magistrate judge also considered probative Panduit's reference to the T & B TY-RAP as a "classic design" for the product. [Def.App. 34]. That Panduit copied the TY-RAP because the company considered it a classic design for a two-piece cable tie does not support a finding of secondary meaning, however. The purpose of trademark law is not to protect product designs, "classic" or otherwise. What the magistrate judge appears to have been intent on preventing was "not the loss of an identifying mark but the loss of a competitive advantage stemming from the exclusive possession of a popular design." Trademark law offers no such protection. W.T. Rogers Co., Inc. v. Keene, 778 F.2d 334, 348 (7th Cir.1985).

Considerable additional evidence in the record which the magistrate judge did not consider also weighs against a finding of secondary meaning in the oval shape of the TY-RAP's head. T & B's product manager testified that the oval shape of the TY-RAP's head was chosen not as a way to identify it as a T & B product but because the rounded head would help prevent snagging, would not cut installers' hands and looked better. [Tr. 129-130].

Also undermining T & B's claim of secondary meaning was testimony by T & B's product manager and its marketing director concerning the development of T & B's one-piece cable tie. Both testified that T & B chose a square head for its one-piece tie to distinguish it from the TY-RAP, because some customers prefer round heads and some square and the company wanted to meet both needs. [Tr. 54, 129, 132-133]. This testimony both undermines T & B's contention that the oval head shape is a universal T & B symbol and establishes that consumers value the cable tie's head shape apart from its possible role as a signifier of source.12

Finally, T & B's marketing director testified that all cable ties have either oval or square heads, with slight variations [Tr. 61]. This general uniformity makes it quite unlikely that consumers who care which brand of cable tie they are purchasing (and as we have said, they are the only ones at whom trademark and unfair competition law is aimed) would rely on the shape of the head in identifying the brand.

Because T & B has not established a reasonable likelihood of success on the merits, the preliminary injunction is reversed.13