Stuart Hall Company, Inc., a Missouri Corporation v. Ampad Corp., a Massachusetts Corporation
51 F.3d 780
8th Cir.
1995
Check Treatment
MAGILL, Circuit Judge.

Stuart Hall appeals the district court’s denial of its motion for a preliminary injunction against Ampad in this trade dress action brought under Lanham Act § 43(a). Stuart Hall argues that the district court applied an unrecognized legal standard to determine whether its trade dress is inherently distinctive, and that the court failed to consider or to grant appropriate weight to the evidence presented on the other elements of the trade *783 dress infringement standard. We reverse and remand.

I. BACKGROUND

Stuart Hall Company produces and markets a line of specialized pre-bound pads and notebooks designed for mass market sales under the names “Executive” and “Executive Suite.” The products at issue in this appeal are the Executive Project Planner and Notebook and the Executive Suite Personal Notebook.

The products were designed to provide the calendars and specialized forms available in expensive personal organizer lines at a price comparable to that of an ordinary legal pad. They are generally displayed in the area of stores dedicated to legal pads, pads of graph paper, and other standard office supply items. Stuart Hall's specialized pads are packaged with a “false cover” that displays the design of the inside forms, differentiating the pads from ordinary lined pads. A substantial advertising and promotional budget is allocated to the Executive and Executive Suite lines: Stuart Hall has spent almost $1.7 million on advertising alone for the two lines. The advertisements and promotional displays feature the design of the products as an element that makes Stuart Hall pads different from other office supply pads. The Executive and Executive Suite products have been very successful in terms of sales volume.

The only other major company producing a mass-market personal planner for some time was Mead Corporation, whose Cambridge line of office supply products includes a Planner Pad. 1 In the autumn of 1989, however, Ampad Corporation, another large-scale producer of office supply products, introduced its Gold Fibre and Gold Fibre Designer lines. The Gold Fibre line included a Legal Pad, Project Planner and Personal Notebook; the Gold Fibre Designer line included a Project Planner. These items are in direct competition with Stuart Hall’s Executive and Executive Suite products. As do the Stuart Hall pads, the Ampad pads include a false cover with a copy of the inside pages on the front. The graphics and text of the Ampad pages are almost identical to that of the Stuart Hall pages, and, according to the testimony of Ampad’s president, were copied directly from Stuart Hall’s design. Ampad successfully marketed its pads to Wal-Mart, which had previously been a major retailer of Stuart Hall’s pads, as a replacement for Stuart Hall products at a lower price.

Stuart Hall brought an action in the district court for a preliminary injunction against Ampad’s sales of the Gold Fibre and Gold Fibre Designer pads, alleging trade dress infringement under § 43(a) of the Lanham Act, 15 U.S.C. § 1125(a) (1994). After a three-day hearing, the court denied the motion, finding that Stuart Hall had not shown a likelihood of success on the merits. 2

In reaching this conclusion, the district court first stated the standard for a trade dress infringement recently confirmed by the Supreme Court: to obtain relief, plaintiff must show that (1) its trade dress is either inherently distinctive or has acquired secondary meaning; (2) that the trade dress is nonfunctional; and (3) that defendant’s imitation of the trade dress “creates a likelihood of confusion in consumers’ minds as to the origin of the services.” Prufrock Ltd., Inc. v. Lasater, 781 F.2d 129, 132 (8th Cir.1986); see Two Pesos, Inc. v. Taco Cabana, Inc., — U.S. -, -, 112 S.Ct. 2753, 2758, 120 L.Ed.2d 615 (1992). The court proceeded to find that Stuart Hall was not likely to succeed on the merits on several grounds. The court found that the trade dress was not inherently distinctive because it was not “striking in appearance, or at least memorable,” and found that the trade dress *784 had not acquired a secondary meaning because there had been no showing that consumers would “think Stuart Hall” when seeing the trade dress. The court further found that the trade dress was “not purely functional,” although it did not address functionality at length because its denial of Stuart Hall’s motion rested on the finding that the trade dress was not inherently distinctive and had not acquired a secondary meaning. Finally, the court stated that it did not find Stuart Hall’s consumer survey convincing. Because the survey was intended to show likelihood of confusion, this finding merged the likelihood of confusion prong of the three-part standard into the inherent distinctiveness analysis, making it unclear whether a finding was made as to likelihood of confusion.

II. DISCUSSION

The district court has broad discretion in deciding whether to grant or deny a motion for preliminary relief, and we will reverse only for clear error, an error of law, or an abuse of discretion. Dakota Indus., Inc. v. Dakota Sportswear, Inc., 988 F.2d 61, 63 (8th Cir.1993) (citing General Mills, Inc. v. Kellogg Co., 824 F.2d 622, 625 (8th Cir.1987)). Whether each of the three prongs of the trade dress infringement standard is satisfied is a question of fact. Aromatique, Inc. v. Gold Seal, Inc., 28 F.3d 863, 868 (8th Cir.1994).

Stuart Hall appeals the district court’s denial of its motion for a preliminary injunction, arguing that: (1) the court erred as a matter of law in requiring that Stuart Hall show that its trade dress is “memorable or striking” to meet the inherently distinctive standard; (2) the court erred as a matter of law in requiring that Stuart Hall show that consumers “think Stuart Hall” on seeing the products to meet the secondary meaning standard, and in failing to address additional evidence of secondary meaning; (3) the court erred in failing to provide a reason or cite evidence supporting its limited findings on functionality; and (4) the court was clearly erroneous when it completely discounted Stuart Hall’s survey evidence designed to show likelihood of confusion. We address each of these issues in turn.

A. Inherent Distinctiveness

The district court stated that its judgment of the merits turned on the requirement of inherent distinctiveness, and found that Stuart Hall’s trade dress was not inherently distinctive. The only test of inherent distinctiveness the district court applied was that “[t]he trade dress must ... be striking in appearance, or at least memorable,” noting examples of such striking and memorable items as “Durante’s nose, Holmes’ mustache, Chief Judge Arnold’s bow ties and Congressman Leach’s sweaters.” Mem. and Order at 4 n. 1 (May 23,1994). The court treated this standard as a threshold test for acquired distinctiveness, because “it seems evident that a trade dress must have sufficient strength of visual impact to impress itself on the mind of a consumer, and thus be subject to recall when the consumer next enters the market” if the trade dress is to be “inherently distinctive in a manner to serve as a source-identifier.” Id. A single unreported case from an Illinois district court was cited in support of this “striking or memorable” test. See Hanig & Co. v. Fisher & Co., No. 92 C 1779, 1994 WL 97758 (N.D.Ill. Mar. 24, 1994). The district court, applying this test, found that “it is entirely unlikely that any ordinary consumer would be so fascinated with the inside page layout to remember it and associate it with a known or unknown source.” Mem. and Order at 5-6 (May 23, 1994). Finally, the court rejected the widely-used test for inherent distinctiveness derived from Chevron Chemical Co. v. Voluntary Purchasing Groups, Inc., 659 F.2d 695, 702 (5th Cir. Unit A Oct. 1981), cert. denied, 457 U.S. 1126, 102 S.Ct. 2947, 73 L.Ed.2d 1342 (1982), as not creating a “comprehensive and usable definition” of inherent distinctiveness, id., and rejected Stuart Hall’s survey evidence as providing no showing of memorableness.

Stuart Hall argues that this “striking or memorable” test is contrary to established law in this circuit and in others, and we agree.

*785 The Supreme Court recently resolved a split in the circuit courts, holding that when a trade dress is inherently distinctive, no additional showing of secondary meaning 3 is required for the trade dress to qualify for protection. Two Pesos, — U.S. at-, 112 S.Ct. at 2760. In Abercrombie & Fitch Co. v. Hunting World, Inc., Judge Friendly for the Second Circuit set forth a series of classifications for trademark protections: (1) generic; (2) descriptive; (3) suggestive; (4) arbitrary or fanciful, and held that if a mark is suggestive or arbitrary or fanciful, it is a registera-ble trademark. 587 F.2d 4, 9 (2d Cir.1976). In Chevron, the Fifth Circuit applied Aber-crombie classifications to trade dress, distinguishing between descriptive and arbitrary trademarks, and held that “if the features of the trade dress sought to be protected are arbitrary and serve no function either to describe the product or assist in its effective packaging, there is no reason to require a plaintiff to show consumer connotations associated with such arbitrarily selected features.” 659 F.2d at 702. In Two Pesos, the Court set out the “classic” Abercrombie test and approved the Fifth Circuit’s application of its Chevron test to the restaurant trade dress at issue; although Chevron does not set out each of the individual Abercrombie classifications, the Court stated that “[t]he Fifth Circuit was quite right in Chevron, and in this ease, to follow the Abercrombie classifications consistently and to inquire whether trade dress for which protection is claimed under § 43(a) is inherently distinctive.” — U.S. at-, 112 S.Ct. at 2760.

The Court further noted that the Chevron test for inherent distinctiveness, or variations on it, is followed by the Seventh, Ninth and Eleventh Circuits as well as the Fifth. Id. The Lanham Act discusses secondary meaning only in the context of trademarks that are “merely descriptive” under the Aber-crombie classifications, and because the bases for legal protection of trade dress are not analytically distinct from protection of trademarks, the Court found that requiring a showing of secondary meaning in addition to inherent distinctiveness for trade dress, when such a showing is not required for trademark protection, would be inconsistent with congressional intent. Id.

This circuit has consistently applied the Abercrombie classifications when addressing the inherent distinctiveness of unregistered trademarks: In Co-Rect Products v. Marvy! Advertising Photography, 780 F.2d 1324 (8th Cir.1985), we followed Abercrombie, and held that if an unregistered trademark is arbitrary or fanciful, it is inherently distinctive and no showing of secondary meaning is required, id. at 1329. If it is suggestive, again no showing of secondary meaning is required. Id. If it is descriptive of the product itself, a showing of secondary meaning is required to obtain trade dress protection, and if it is generic, it will not be protected. Id.; see also Cellular Sales, Inc. v. Mackay, 942 F.2d 483 (8th Cir.1991) (applying same test to trade name registered in Missouri as fictitious name, but unregistered as to trademark protection). This test is equally applicable to trade dress under § 43(a), because “§ 43(a) provides no basis for distinguishing between trademark and trade dress.” Two Pesos, — U.S. at-, 112 S.Ct. at 2760; see Aromatique, 28 F.3d at 868 (“The difference between trade dress and trademark is no longer of importance in determining whether trade dress is protected by federal law.”).

The classification of trade dress as arbitrary or fanciful or suggestive, and thus inherently distinctive, requires no showing that the trade dress is memorable or striking. The distinction between suggestive and descriptive trademarks, and thus between inherently distinctive and nondistinctive trademarks, was delineated in Abercrombie: “A term is suggestive if it requires imagination, thought and perception to reach a conclusion as to the nature of goods. A term is descriptive if it forthwith conveys an immediate idea of the ingredients, qualities or characteristics *786 of the goods.” Abercrombie, 537 F.2d at 11 (quoting Stix Prods., Inc. v. United Merchants & Mfrs., Inc., 295 F.Supp. 479, 488 (S.D.N.Y.1968)). These definitions address the relation between the product and the trade dress, not the relation between the trade dress and the consumer. The question they present is whether, and how much, the trade dress is dictated by the nature of the product, not whether consumers remember or are struck by the design, or whether consumers associate the design with its source. If the specific design of the trade dress is only tenuously connected with the nature of the product, then it is inherently distinctive, and secondary meaning, which requires a showing that consumers attach significance to the trade dress as a source-signifier, need not be proven. If the design of the trade dress is dictated by the nature of the product, then secondary meaning must be proven.

Abercrombie’s emphasis on the relation of the product’s trademark or trade dress to the product, rather than on the relation of the consumer to the product, is echoed in Two Pesos, where the Supreme Court stated that “[sjuggestive marks are eligible for protection without any proof of secondary meaning, since the connection between the mark and the source is presumed,” — U.S. at-, 112 S.Ct. at 2759 (emphasis added) (quoting Thompson Med. Co. v. Pfizer Inc., 753 F.2d 208, 216 (2d Cir.1985)), and quoted from the Fifth Circuit’s “sound” disposition below:

While the necessarily imperfect (and often prohibitively difficult) methods for assessing secondary meaning address the empirical question of current consumer association, the legal recognition of an inherently distinctive trademark or trade dress acknowledges the owner’s legitimate proprietary interest in its unique and valuable informational device, regardless of whether substantial consumer recognition yet bestows the additional empirical protection of secondary meaning.

Id. at -, 112 S.Ct. at 2758 (emphasis added). Similarly, in Chevron, the Fifth Circuit stated that “there is no reason to require a plaintiff to show consumer connotations associated with” inherently distinctive trade dress. 659 F.2d at 702.

The district court ignored this presumption of source identification inherent in the classification-based tests for inherent distinctiveness when it applied a memorableness requirement to Stuart Hall’s trade dress; the court justified its memorableness requirement by stating that if consumers do not remember the trade dress, they will not later identify it with its source, but, as discussed in Two Pesos and Chevron, source-identification is presumed when trade dress is suggestive or arbitrary or fanciful. The analysis in these cases, as in Abercrombie, focuses on the connection between the trade dress and the product, not the trade dress and the consumer.

Another widely-adopted standard, based on the Abercrombie classifications, for identifying an inherently distinctive trade dress was crafted by the Court of Customs and Patent Appeals: “[i]n determining whether a design is arbitrary or distinctive this court has looked to whether it was a ‘common’ basic shape or design, whether it was unique or unusual in a particular field, [or] whether it was a mere refinement of a commonly-adopted and well-known form of ornamentation for a particular class of goods viewed by the public as a dress or ornamentation for the goods.” Seabrook Foods, Inc. v. Bar-Well Foods Ltd., 568 F.2d 1342, 1344 (C.C.P.A.1977); see 1 McCarthy, Trademarks and Unfair Competition § 8.02[4]. McCarthy describes this test as “ways to ask whether the design, shape or combination of elements is so unique, unusual or unexpected in this market that one can assume without proof that it will automatically be perceived by customers as an indicia of origin.” Id. (emphasis added). Thus, under the Seabrook test (also referred to as the AmBrit test, Ambrit, Inc. v. Kraft, Inc., 812 F.2d 1531, 1536 (11th Cir.1986), cert. denied, 481 U.S. 1041, 107 S.Ct. 1983, 95 L.Ed.2d 822 (1987)), as under the Chevron test, no proof of source-identification is required to show that trade dress is inherently distinctive. The Seabrook test focuses on a comparison of plaintiffs trade dress to others in the same *787 class of goods, not on the trade dress’s impact on consumers.

In summary, the Chevron and Abercrom-bie tests focus on the arbitrariness of the trade dress and its relevance to the product, whereas the Seabrook test focuses on a comparison of the trade dress with others in the same field. None of these tests in any combination examines the trade dress’s impact on consumers. A particular trade dress can be arbitrary and different from others in the field, without being particularly “striking or memorable” in the sense in which the district court used these words.

The single relevant case the district court cited as support for its contention that to be inherently distinctive a trade dress must be striking or memorable is not only an unreported decision by an Illinois district court with no precedential value in this circuit, but also does not strongly support the district court’s standard. See Hanig & Co. v. Fisher & Co., No. 92 C 1779, 1994 WL 97758 (N.D.Ill. Mar. 24, 1994). The Hanig court defined an inherently distinctive trade dress as “memorable enough for us to conclude that it would probably be retained in a viewer’s mind and recognized when seen again,” id. at 5, which does not demand a showing of memorableness in the sense of “striking” or “fascinating.” Moreover, although the Han-ig court stated a memorableness standard for inherent distinctiveness, it applied the Chevron test to the trade dress, and its analysis focused not on memorableness, but on the arbitrariness of the features at issue. Id. at 6.

The requirement of source-identification applies not to whether a trade dress is inherently distinctive, but to whether it has a secondary meaning. The three widely-applied tests for inherent distinctiveness, none of which were applied by the district court, focus not on the trade dress’s impact on consumers, but on the relation between plaintiff’s trade dress and his product or plaintiffs trade dress and other trade dresses in the same field. The district court thus focused on the wrong aspect of the trade dress in finding that it was not inherently distinctive because it was not memorable. By replacing the Abercrombie test for inherent distinctiveness with a version of the test for secondary meaning, the court effectively ruled only on secondary meaning, never addressing the substance of Stuart Hall’s inherent distinctiveness claim.

Ampad nevertheless argues that the Third Circuit’s recent decision, Duraco Products, Inc. v. Joy Plastic Enterprises, Ltd., 40 F.3d 1431 (3d Cir.1994), supports the district court’s test, and that we should accept the striking or memorable test on the basis of Duraco.

In the wake of Two Pesos, which overruled the Third Circuit’s previous requirement that trade dress must always have secondary meaning to be protected, the Third Circuit crafted a new test for inherent distinctiveness applicable to trade dress comprising product features or shape as opposed to packaging. Rejecting the Abercrombie test as inappropriate for “product configuration” trade dress the court held that, to be inherently distinctive, product configuration must be: (1) unusual and memorable; (2) conceptually separable from the product; and (3) likely to serve primarily as a designator of origin of the product. Id. at 1434. We decline to adopt the Third Circuit’s standard.

Stating that Two Pesos did not address the question of whether product configuration, as opposed to packaging, can ever be inherently distinctive, the Duraco court decided that configuration could be inherently distinctive, but that the existing tests for inherent distinctiveness should apply only to packaging. See id. at 1440. We, however, read Two Pesos as resting on a presumption that “trade dress” is a single concept that encompasses both product configuration and packaging, and find that its holding applies to trade dress as a whole, not merely to packaging. Two Pesos concerned the trade dress of a Mexican restaurant: The trade dress at issue was the restaurant’s decor. The Duraco opinion views a restaurant’s decor as more like' packaging than product configuration, and finds that therefore Two Pesos bears no implications regarding product configuration. Id. at 1442. We perceive a restaurant’s decor as being as akin to product configuration as to packaging. A *788 restaurant is composed of a room containing furnishings appropriate to dining in which food is served. The room and its furnishings are not packaging, which is separate from a product and to be thrown away once the product is removed from its packaging, but are part of the restaurant experience itself— the product itself is the room and its furnishings as well as whatever food is served. Thus, we view Two Pesos as a case that involves product configuration, and, although the Court did not review the finding below of inherent distinctiveness, its holding leads directly to the conclusion that product configuration, like packaging, can be inherently distinctive, and if it is, no showing of secondary meaning is required.

Further, restaurant decor is an ideal illustration of the fact that there is no distinct line between packaging and product configuration; some trade dresses, like restaurant decor, will be somewhere in between or perhaps could be seen as both and to create a distinction between the two artificially treats things that are alike differently. The language the Supreme Court used in Two Pesos suggests a similar view. Neither the term “packaging” nor the term “product configuration” is used in Two Pesos: the term used is “trade dress,” and both trademarks and trade dress are defined as “a verbal or symbolic mark or the features of a product design.” Two Pesos, — U.S. at -, 112 S.Ct. at 2759 (emphasis added). Further, the Court stated that “Suggestions that under the Fifth Circuit’s law, the initial user of any shape or design would cut off competition from products of like design and shape are not persuasive.” Id. at-, 112 S.Ct. at 2760 (emphasis added). Clearly, the Court envisions trade dress as including product features, design and shape, and intends its holding to apply to such trade dress. Therefore, we decline to create a distinction between protection of packaging and protection of product configuration, as such a distinction would run contrary to the holding of Two Pesos.

In addition, the Supreme Court approved the Abercrombie and Chevron tests for inherent distinctiveness in Two Pesos. See id. at -, 112 S.Ct. at 2760. This circuit has followed Abercrombie in the past, and, in light of Two Pesos’ approval, we see no reason to abandon the classic test now. A new test based on source-identification and mem-orableness would bear a very close resemblance to the standard for secondary meaning: A showing of secondary meaning requires empirical proof of consumer recognition and identification, but the new test would amount merely to a shift from actual consumer recognition to likely consumer recognition. See Duraco, 40 F.3d at 1450. Thus, a new test would undermine Two Pesos by requiring that, in order to be inherently distinctive, a trade dress be shown to meet a version of the secondary meaning test with a lesser burden of production.

The Third Circuit suggests that, because Abercrombie examines the relation between the trade dress and the product, it cannot apply to product configuration because the trade dress then is the product. Id. at 1440-41. We do not perceive this problem with Abercrombie because a product’s trade dress is not the entire product itself, but specific features of the product; we find no difficulty in looking at a specific feature of a product and determining whether and to what degree that feature is dictated by the nature of the product.

Finally, the district court criticized Stuart Hall’s survey as not adequate to show memorableness. Because source-identification and consumer recognition need not be proven to determine whether a trade dress is inherently distinctive, the quality of the survey is not pertinent to this issue.

We find that the district court erred as a matter of law by requiring that Stuart Hall’s trade dress be striking, or at least memorable, to be inherently distinctive. This requirement is inconsistent with the established law of this circuit, the majority of other circuits, and the Supreme Court, and we confirm our adherence to the Abercrom-bie test for inherent distinctiveness of unregistered trademarks, including trade dress, as applied in our previous decisions. See, e.g., Co-Rect, 780 F.2d at 1324.

*789 B. Secondary Meaning

The district court discussed secondary meaning only briefly, but nevertheless found that Stuart Hall had not shown that its trade dress had a secondary meaning:

There is no showing that plaintiffs alleged trade dress ... has acquired secondary meaning, so that any perceptible segment of the Wal-Mart consuming public would think “Stuart Hall” when seeing either the pads as a whole or the inside pages in particular. Plaintiff has introduced evidence of advertising, but there is no evidence that the advertising has impressed any part of the consuming public in the manner necessary to find secondary meaning.

Mem. and Order at 4 (May 23, 1994).

Stuart Hall argues first that requiring that consumers “think Stuart Hall” when seeing the pads was an error of law, and second that the district court erred in finding that there was no showing of secondary meaning.

It is well settled that secondary meaning, or “acquired distinctiveness,” requires that “the user of ... a [trade] dress ... show that by long and exclusive use in the sale of the user’s goods, the [dress] has become so associated in the public mind with such goods that the [dress] serves to identify the source of the goods and to distinguish them from those of others.” Aromatique, 28 F.3d at 870; see Co-Rect, 780 F.2d at 1330. “[T]he ultimate inquiry is whether in the consumer’s mind the mark denotes a ‘single thing coming from a single source.’” Co-Rect, 780 F.2d at 1330 (internal quotations omitted). That single source, however, need not be known by name by consumers. See Two Pesos, — U.S. at-, 112 S.Ct. at 2757; Centaur Communications v. A/S/M Communications, 830 F.2d 1217, 1221 (2d Cir.1987). The court’s phrase “think Stuart Hall” is ambiguous: It may imply a requirement that consumers think of the specific brand name of the trade dress, although it may merely be a careless locution not intended to imply such a requirement. We remind the district court that Stuart Hall need not show that consumers know their name; the showing required is that consumers associate the trade dress with a single source.

The district court relied entirely on its finding that consumers do not “think Stuart Hall” in deciding that the trade dress had not acquired secondary meaning. There are, however, a range of factors that can contribute to a determination of secondary meaning including consumer surveys, deliberate copying of the mark, advertising featuring the trade dress at issue, expenditure on such advertising, and anecdotal evidence showing consumer identification of the trade dress with its source. See Aromatique, 28 F.3d at 870. The district court did not examine any of these factors in making its determination regarding secondary meaning, although evidence on each factor was presented.

In particular, Stuart Hall presented survey evidence that may pertain to a showing of secondary meaning. See Cellular Sales, 942 F.2d at 486 (survey showing that public has associated a trade dress with the source may support claim of secondary meaning). The district court discussed the survey in the context of inherent distinctiveness, where survey evidence is not relevant. Stuart Hall’s survey is relevant to likelihood of confusion, discussed in part D, infra, and the court should also examine whether the survey evidence contributes to a showing of secondary meaning.

Stuart Hall also argues that its continuous and exclusive use of the trade dress for approximately five years creates an inference of secondary meaning. 15 U.S.C. § 1052(f) states that the trademark commissioner may accept proof of five years’ exclusive and continuous use of a mark as prima facie evidence of secondary meaning. This suggests that five years’ use is a strong factor in favor of secondary meaning: “[T]he general principles qualifying a mark for registration under § 2 of the Lanham Act are for the most part applicable in determining whether an unregistered mark is entitled to protection under § 43(a).” Two Pesos, — U.S. at-, 112 S.Ct. at 2757. Although “no absolute time span can be posited as a yardstick in cases involving secondary meaning,” Centaur Communications, 830 F.2d at 1225, length and exclusivity of continuous use *790 is a factor bearing on secondary meaning, id. at 1222. Stuart Hall’s exclusive and continuous use of its trade dress for the period of time that creates a statutory presumption of secondary meaning for trademark registration weighs in favor of a finding of secondary meaning, and the district court, on remand, should consider this factor in addition to the other factors bearing on secondary meaning.

C. Functionality

The only finding the district court made as to functionality was that:

although defendant has a persuasive contention that the page layout may be characterized as 75% to 90% functional in nature, and that the nonfunctional feature was more of a sales device than a source-identifier, I am not sufficiently satisfied with my grasp of functionality law to make the case turn on this point. The alleged trade dress is not purely functional.

Mem. and Order at 4 (May 23, 1994).

Ampad argues that forms designed to record rather than convey information cannot be protected by copyright or trade dress because they are intrinsically functional. Although the strictly functional aspects of forms are not protected, nonfunctional design elements may be. Ampad rather disingenuously cites John H. Harland Co. v. Clarke Checks, Inc., where, although the lines on a check stub for recording the date, amount, payee, and purpose of a check were held to be functional, the court also held that “[a] reasonable jury could conclude, however, that some features of the trade dress of Harland’s product — such as the decorative box around the lines for recording information on the stub, and the design and color of the carry-around case — are primarily nonfunctional.” 711 F.2d 966, 983-84 (11th Cir. 1983). We recently explained the distinction between functional and nonfunctional elements of trade dress, applicable to the trade dress of forms as to other products:

If the particular feature is an important ingredient in the commercial success of the product, the interests in free competition permits [sic] its imitation in the absence of a patent or copyright. On the other hand, where the feature or, more aptly, design, is a mere arbitrary embellishment, a form of dress for the goods primarily adopted for purposes of identification and individuality and, hence, unrelated to basic consumer demand in connection with the product, imitation may be forbidden where the requisite showing of secondary meaning is made. Under such circumstances, since effective competition may be undertaken without imitation, the law grants protection.

Aromatique, 28 F.3d at 873 (quoting Prufrock Ltd. v. Lasater, 781 F.2d 129, 133 (8th Cir.1986)). Therefore, we remind the district court, if it reaches the issue of nonfunctionality on remand, that individual format and design elements of Stuart HaU’s planner pages may be primarily nonfunctional under the standard explained in Aromatique, even if such elements as lines on which to write appointments throughout the day are not.

D. Likelihood of Confusion

The district court characterized the survey commissioned by Stuart Hall regarding the likelihood of confusion between its product and Ampad’s as “misleading,” and accorded it no weight regarding inherent distinctiveness or likelihood of confusion. 4 As we state supra, the survey is not relevant to the issue of inherent distinctiveness.

In Mutual of Omaha Insurance Co. v. Novak, this court held that surveys are an appropriate form in which to produce evidence of actual confusion. 836 F.2d 397, 400 (8th Cir.1987), cert. denied, 488 U.S. 933, 109 S.Ct. 326, 102 L.Ed.2d 344 (1988). The court went on to say that “[b]ecause manifestations of actual confusion serve as strong evidence of a likelihood of confusion, and may, in fact, be the best such evidence, this survey should be given substantial weight unless seriously flawed.” Id. at 400 (internal citations omitted); see also Woodsmith Pub. Co. v. Meredith Corp., 904 F.2d 1244, 1249 (8th Cir.1990) (stating that, although not required to show *791 likelihood of confusion, surveys are probably the most accurate evidence of actual confusion).

The district court’s complaints about the survey are not particularly significant, nor do they necessarily address actual flaws in the survey as performed. The court stated that it would have preferred a survey designed to enforce only casual and limited attention to the products and a time-delayed resurvey. The more limited and casual the attention to the products, however, the more likely consumers are to indicate confusion between them. The court verified this hypothesis itself when it also complained that some people surveyed may have been so casual that elements not at issue in the case may have been enough to cause an inference of identical source. The court is splitting hairs here: it would prefer a survey of people who are casual, but not too casual. The court also expressed concern that some people surveyed may have skewed their answers due to their views on political economy, but the testimony of John Bunge, Stuart Hall’s survey expert, indicated that not only were the people surveyed not told that the survey concerned issues of trade dress protection and thus potentially market competition, the interviewers and their employers were also not told for the specific purpose of avoiding such bias.

No evidence was presented to indicate that the survey was flawed. Stuart Hall’s expert testified that he had performed approximately 200 trademark-related consumer surveys with his current company, and had performed over 1,000 consumer surveys throughout his marketing career. His testimony indicates strict adherence to and extensive knowledge of standard methodology for trademark-related consumer surveys, and attention to the requirements of trademark law when designing surveys in general and this survey in particular. The testimony of Am-pad’s expert is, as the district court itself points out, unimpressive, and does nothing to indicate the presence of real flaws in the survey.

On remand, the district court must reconsider the evidence provided by the survey on the likelihood of confusion between Stuart Hall and Ampad products if it finds that Stuart Hall has shown that its trade dress is inherently distinctive or has secondary meaning and that it is primarily nonfunctional.

III. CONCLUSION

We reverse the district court’s denial of Stuart Hall’s motion for a preliminary injunction because the court erred in applying a novel and unrecognized test for inherent distinctiveness, failed to consider the factors that can properly support a finding of secondary meaning, failed to make a complete finding on nonfunctionality, and accorded no weight to Stuart Hall’s survey on likelihood of confusion. We remand the case to the district court for reconsideration of Stuart Hall’s motion for preliminary relief analyzing Stuart Hall’s likelihood of success on the merits in accordance with the law as outlined in this opinion.

Notes

1

. Stuart Hall has never challenged Mead's products as infringing on the trade dress of Stuart Hall products.

2

. A preliminary injunction may be granted when the movant has demonstrated: (1) the threat of irreparable harm to it; (2) the state of the balance between this harm and the injury that granting the injunction will inflict on other parties; (3) the probability that it will succeed on the merits; and (4) the public interest. Calvin Klein Cosmetics Corp. v. Lenox Lab., 815 F.2d 500, 503 (8th Cir.1987) (citing Dataphase Sys., Inc. v. C L Sys., Inc., 640 F.2d 109, 113-14 (8th Cir.1981) (en banc)).

3

. Secondary meaning is sometimes referred to as “acquired distinctiveness," and is established when "the user of ... a [trade] dress [has shown] that by long and exclusive use in the sale of the user’s goods, the [dress] has become so associated in the public mind with such goods that the [dress] serves to identify the source of the goods and to distinguish them from those of others.” Aromatique, 28 F.3d at 870.

4

. It is not clear from the district court's memorandum and order whether it made any finding on the likelihood of confusion element. We presume, given the court’s treatment of the survey, that if there was a finding, it was that there was no likelihood of confusion.