Arthur H. Seidel, Stephen J. Meyers (argued), Nancy Rubner-Frandsen, Seidel, Gonda, Lavorgna & Monaco, Philadelphia, Pa., for appellant/cross appellee.

Martin F. Savitzky, William H. Elliott, Jr., Synnestvedt & Lechner, Philadelphia, Pa., Albert Robin (argued), Robin, Blecker, Daley & Driscoll, New York City, for appellees/cross appellants.

Before HUTCHINSON, NYGAARD and ROSENN, Circuit Judges

In this Lanham Act case, appellant Charles Jacquin Et Cie, Inc. ("Jacquin") alleged that Destileria Serralles, Inc. ("DSI") and Crown Marketing International ("Crown") infringed on its products' trade dress in violation of 15 U.S.C. Sec. 1125(a) and state common law. The district court directed a verdict in favor of DSI and Crown on Jacquin's compensatory and punitive damages claim 730 F. Supp. 662. The jury found that Jacquin's trade dress had acquired secondary meaning and that there was a likelihood of confusion with DSI's trade dress, and the district court crafted injunctive relief to protect Jacquin. Jacquin appeals the scope of that injunction and the district court's directed verdict on punitive damages. We will affirm in part and reverse and remand in part.I.

Jacquin, a Pennsylvania corporation, produces alcoholic beverages, including cordials. DSI, a Puerto Rican corporation, produces rum and rum schnapps. Crown was a Florida partnership which distributed DSI's products in the continental United States. At the time of trial, Crown was no longer in business.1

In 1968 Jacquin developed a bottle of a particular shape for its line of cordials. The bottle is 10 and 3/4 inches high, with a beveled or tapered bottom. Jacquin has consistently used this same shape of bottle for all the cordials in its line. Jacquin promotes its cordials through billboards, print ads, and other materials. The district court found that approximately 75 per cent of Jacquin's promotional materials include the bottle as part of the advertisement.

In 1985, representatives of Peter Harvey Wines ("PHW") suggested to DSI that it produce a rum-based schnapps which PHW would market in the United States. The representatives of PHW suggested that the rum schnapps be sold in a bottle similar to the bottle used for Blackstone whiskey in Mexico. The Blackstone whiskey bottle has a beveled bottom. In 1986, DSI submitted a Blackstone whiskey bottle to Owens-Illinois, Inc., a bottle manufacturer, to use as a sample. DSI instructed Owens-Illinois to increase the height of the bottle to 10 inches and to make a few other minor adjustments.

The bottle developed by Owens-Illinois became the bottle DSI uses for Don Juan rum schnapps in the United States. The Don Juan bottle is shorter than Jacquin's, and Jacquin's bottle has a longer neck. The Don Juan bottle has an eight sided cross-section while the Jacquin's bottle has a four sided cross-section. However, the Don Juan bottle has a similar appearance to the Jacquin's bottle when viewed from the front, primarily because both have beveled bottoms.

In the fall of 1987, DSI sold 2,700 cases of Don Juan schnapps to Crown for distribution in the United States. Crown sold Don Juan in New York, Michigan, New Jersey, Massachusetts, Florida, Virginia, New Hampshire, Pennsylvania, Vermont and Maine. In February 1988, Jacquin sent a cease and desist letter to Crown, alleging that the Don Juan bottle infringed on Jacquin's distinctive trade dress. Later that year, DSI repurchased 1,421 unsold cases of Don Juan from Crown.

Jacquin filed suit alleging that the Don Juan bottle infringed on its trade dress in violation of section 43(a) of the Lanham Act, 15 U.S.C. Sec. 1125(a)2, and common law. Jacquin sought compensatory and punitive damages, as well as injunctive relief. At trial and following Jacquin's case in chief, the district court directed a verdict for DSI on Jacquin's compensatory and punitive damage claims. The district court concluded that Jacquin had failed to demonstrate actual consumer confusion and thus compensatory damages under section 43(a) of the Lanham Act were inappropriate. On Jacquin's claim for punitive damages under Pennsylvania common law, the district court concluded that Jacquin's evidence that DSI had intentionally copied Jacquin's bottle was insufficient.

On Jacquin's request for injunctive relief, the district court concluded that Jacquin could obtain injunctive relief if it demonstrated secondary meaning in its trade dress and a likelihood of consumer confusion. The district court denied DSI's motion for a directed verdict on this claim, holding that secondary meaning in the bottle design and a likelihood of consumer confusion were issues best left to the jury. After DSI presented its case, the jury was asked to answer two special interrogatories:

1) Has the shape of the plaintiff's bottle acquired secondary meaning?

(If your answer to this question is "yes," please go to the next question. If it is "no," your deliberations are over; please return to the courtroom.)

2) Is defendants' bottle as it appears in the market-place likely to lead consumers to think that the defendant's product was produced by the plaintiff or some entity related to the plaintiff?

The jury answered "yes" to both questions.

In response to the parties' post-verdict motions, the district court considered whether it was bound by the jury's verdict, since the only claim remaining when the case went to the jury was for equitable relief. The court concluded that it would not have been bound, had it notified the parties that the jury would be advisory only. Since it did not notify the parties, the court decided that it was bound by the jury's verdict. Neither party contests this decision.

The court then considered the extent of injunctive relief available to Jacquin. Although the jury found that Jacquin's bottle had acquired secondary meaning, it had not determined in which markets the bottle had acquired this status. The district court, after a thorough review of Jacquin's sales records, concluded that the bottle had acquired secondary meaning only in Pennsylvania, and issued an injunction prohibiting DSI from using its bottle for cordial or specialty beverages in Pennsylvania. Jacquin appeals.

The first issue is whether the district court erred by directing a verdict in DSI's favor on the issue of punitive damages. Our review is plenary. A verdict may only be directed if, viewing the evidence in the light most favorable to the non-moving party, there is no issue of material fact and only the verdict directed would be correct under the governing law. Macleary v. Hines, 817 F.2d 1081, 1083 (3d Cir.1987).

Jacquin's claim for punitive damages is governed by Pennsylvania law. Pennsylvania follows section 908(2) of the Restatement (Second) of Torts and allows punitive damages only for "conduct that is outrageous, because of the defendant's evil motive or his reckless indifference to the rights of others." Kirkbride v. Lisbon Contractors, Inc., 521 Pa. 97, 555 A.2d 800, 803 (1989) (quoting Restatement (Second) of Torts Sec. 908(2)). The district court found no evidence to support an award of punitive damages.3

Jacquin conceded at oral argument before us that it had offered no direct evidence of DSI's outrageous conduct, evil motive or reckless indifference in designing its Don Juan bottle. Jacquin contends, nonetheless, that bad intent can be inferred from the striking similarity between its bottle and the Don Juan bottle, the prominence of Jacquin's bottle in the marketplace, and the fact that DSI continued to use the Don Juan bottle after Jacquin had notified it that the bottle might infringe Jacquin's trade dress.

Assuming that covert copying of a non-trademarked package is outrageous conduct, there was simply no evidence that DSI willfully set out to copy Jacquin's bottle. The only testimony about the origin of the Don Juan bottle was elicited by Jacquin itself on direct examination--the unequivocal statements by Irwin Goldberg of Crown and Alberto Torruella of DSI that they used the Blackstone whiskey bottle as a model for DSI's rum schnapps bottle.4 Jacquin's bald assertion that its bottle design was so well known that DSI must have been intentionally copying does not make a prima facie case for punitive damages.

Similarly, the fact that DSI continued to market Don Juan after receiving Jacquin's cease-and-desist letter does not support a finding of outrageousness. See e.g. Andy Warhol Enterprise, Inc. v. Time Inc., 700 F. Supp. 760, 766 (S.D.N.Y.1988) (continuing use of trademark after notification not sufficient for finding of bad faith). Jacquin's argues that continued use of a mark demonstrates the defendant's willfulness. We find this argument unpersuasive in a case where there is no registered mark, but only an asserted right to protection of a distinctive trade dress. Since there was no evidence of outrageous conduct or willful disregard for Jacquin's rights, the district court did not err by granting a directed verdict in favor of DSI.

The district court fashioned an injunction to protect Jacquin's bottle in markets where it had established such market penetration that DSI's bottle would present a real likelihood of confusion among consumers. The court concluded that Jacquin only had significant market penetration in Pennsylvania. Jacquin argues that it was entitled to protection in other markets.

We review the district court's injunction for abuse of discretion. A district court has abused its discretion if it has rested its decision on "a clearly erroneous finding of fact, an errant conclusion of law, or an improper application of law to fact." International Union, UAW v. Mack Trucks, Inc., 820 F.2d 91, 95 (3d Cir.1987). Jacquin argues that, since a jury determined that secondary meaning and a likelihood of confusion existed in this case, we owe less deference to the district court. We disagree. Although the district court was bound by the jury's finding on likelihood of confusion, the special interrogatories did not direct the jury to consider specific markets. Rule 49(a) of the Federal Rules of Civil Procedure states that in a case decided by special verdict, if the district court fails to charge the jury on a fact issue raised by the pleadings or evidence the parties waive their right to a jury trial on that issue unless they object. In crafting injunctive relief, the district court properly supplied its own factual findings to supplement the jury's special verdict. Hence, when reviewing its execution of this task, we will only reverse the district court if it abused its discretion.

The proper geographic scope of an injunction in a trademark infringement case is determined by examining the market penetration of the mark. An injunction is appropriate only if the market penetration is "significant enough to pose the real likelihood of confusion among the consumers in that area."5 Natural Footwear Ltd. v. Hart, Schaffner & Marx, 760 F.2d 1383, 1397 (3d Cir.), cert. denied, 474 U.S. 920, 106 S. Ct. 249, 88 L. Ed. 2d 257 (1985) (quoting Sweetarts v. Sunline, Inc., 380 F.2d 923, 929 (8th Cir.1967)). Market penetration of a trademark is determined by: "(1) the volume of sales of the trademarked product; (2) growth trends (both positive and negative) in the area; (3) the number of persons actually purchasing the product in relation to the potential number of customers; and (4) the amount of product advertising in the area." Natural Footwear, 760 F.2d at 1398-99.

We examine the market, as the district court did, by reference to state boundaries. Although state boundaries are not always appropriate delineators of markets, see Natural Footwear, 760 F.2d at 1398 n. 34, alcoholic beverages are distributed and regulated on a state-by-state basis, and state boundaries are uniquely proper for defining their markets. State markets are generally divided into "control" states, where the sale and distribution of alcohol is strictly regulated by the state, and "open" states, where state control is minimal. Jacquin sells its cordials in both "open" and "control" states.

In Natural Footwear, we reviewed the market penetration of a New Jersey based clothing manufacturer. We determined that we need only consider market penetration in two states outside of New Jersey, since the clothing manufacturer had failed to establish sales above $5,000 per year or a total of over 50 customers for any one year in any other states. Jacquin argues that, since its sales of cordials exceeds $5,000 per year in Alabama, Delaware, Iowa, Maryland, New Hampshire, New Jersey, North Carolina, Rhode Island, Vermont, Virginia, West Virginia, its market penetration is more than de minimis and it is entitled to protection in those states. Jacquin misunderstands our holding in Natural Footwear. Surpassing the de minimis threshold did not result in automatic protection for the clothing manufacturer in Natural Footwear. In fact, we held that the market penetration in the two states surpassing the de minimis threshold was insufficient to warrant protection. Natural Footwear, 760 F.2d at 1403.

Jacquin argues that it has also surpassed the threshold set out in Wrist-Rocket Manufacturing Co., Inc. v. Saunders Archery Co., 578 F.2d 727, 733 (8th Cir.1978). In Wrist-Rocket, the court set a ratio of one sale of the product in question per 20,000 persons as a threshold for establishing secondary meaning. Jacquin's efforts to treat these cases as establishing bright line tests is misguided. Whether a volume of sales is significant will vary with the product and the market. The numbers that result in injunctive relief in one case may not be significant in another.

In determining that Jacquin had only established secondary meaning and likelihood of confusion in Pennsylvania, the district court relied primarily on the volume of sales prong of the Natural Footwear test. The district court summarized Jacquin's sales volume in the following chart:

STATE             JACQUIN SALES      TOTAL SALES      PERCENTAGE
Alabama              371               67,622             5.4
Iowa                 311               213,095             .14
New Hampshire        7,921             306,044            2.58
North Carolina       2,971             175,680            1.69
Vermont              315               82,469              .38
Virginia             12,112            225,034             .53
West Virginia        5,771             59,705             3.8
Pennsylvania         219,918           945,223           23.26