dissenting:
The majority reaches the unprecedented conclusion that an entity’s use of its foreign trademark solely to sell services in a foreign country entitles it to trademark protection under United States law, even though the foreign mark holder has never used or registered its mark in the United States. In my view, the majority errs in holding that the protection of United States trademark law extends to a mark used exclusively in Monaco by a company incorporated there. For this rea-son and others set forth within, I respectfully dissent.
I.
Under United States law, the holder of an unregistered mark must demonstrate “use in commerce” of that mark in order to be eligible for trademark protection. The Lanham Act provides that “[a] mark shall be deemed to be in use in commerce ... on services when it is used or displayed in the sale or advertising of services and the services are rendered in commerce.” 15 U.S.C.A. § 1127 (West Supp.2003). Thus, there are two essential elements that must be present to constitute “use in commerce” for Lanham Act purposes: (1) advertising that employs the mark and (2) the rendering of services to which the mark attaches. Neither alone is sufficient. This two-pronged statutory meaning of “use in commerce” is what I refer to when I say that SBM did not “use” its mark in commerce because it did not “use” the mark in the United States.1 Prior to today’s holding, all existing authority, employing precisely this same two-pronged understanding of use, has similarly concluded that use in the United States is necessary to meet the Lanham Act’s use in commerce requirement. None of this authority has ever suggested that if one element of the use in commerce requirement — advertising— takes place in the United States while the other — the rendering of services — occurs outside the United States, there has been use in the United States. Both elements must occur in the United States in order to satisfy the use in commerce requirement.
Indisputably, SBM has satisfied the first element of the use in commerce requirement, for no one questions that SBM employed the Casino de Monte Carlo mark in the United States to advertise and promote the gambling services that it provides exclusively at its casino in Monaco. This is not and has never been a case about advertising. Rather, the question here is whether the second element of the use in commerce requirement has been satisfied, ie. have the services that give rise to the Casino de Monte Carlo mark been rendered in commerce of the United States. The majority holds that they have and thus accords SBM’s mark eligibility for trademark protection under United States law. I believe, however, that because SBM has not rendered its casino services in the United States, it has not satisfied the statutory use in commerce requirement in a manner sufficient to mer*384it protection under the Lanham Act. Subsection LA provides my analysis of this question; subsection I.B sets forth my response to the majority’s efforts to rebut that analysis.
A.
In this case, some United States citizens purchased “casino services” in Monaco from the Societe des Bains de Mer et du Cercle des Estrangers a Monaco (SBM), a “subject of a foreign nation.” The majority determines that those services “constitute trade with a foreign nation that Congress may regulate under the Commerce Clause.” Ante at 368. Based on this determination, the majority accords SBM, the holder of the foreign mark, “Casino de Monte Carlo,” eligibility for trademark protection under the Lanham Act for that mark even though SBM has never used the mark in the United States.
SBM has not argued, and the district court did not hold,2 that the mark was entitled to protection under the Lanham Act simply because SBM used it to sell gambling services in Monaco to United States citizens. Nor did SBM argue, or the trial court hold, that these sales of gambling services in Monaco to United States citizens constitute foreign trade that has been regulated by Congress under the Lanham Act. Indeed, prior to today, no court, administrative agency, or treatise has ever espoused or adopted this theory.
Rather, it has long been recognized that use of a foreign mark in a foreign country creates no trademark rights under United States law. As the foremost trademark authority has explained, “[f|or purposes of trademark rights in the United States, ‘use’ means use in the United States, not use in other nations.” 2 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 17.9 (4th ed. 2000) (“McCarthy”); see also 3 Rudolf Callman, The Law of Unfair Competition, Trademarks and Monopolies § 19.24 (4th ed. 1998) (“Callman”) (noting that “[t]he use required under United States law as the foundation of trademark rights must be use in this country and not abroad.... It is well settled that foreign use creates no trademark rights in the United States.”).3
*385Until today, every court to address this issue has held that use of a foreign trademark in connection with goods and services sold only in a foreign country by a foreign entity does not constitute “use of the mark” in United States commerce sufficient to merit protection under the Lan-ham Act. As the Federal Circuit explained in rejecting the contention that a Japanese mark holder’s use of a trademark solely in Japan in connection with goods sold to a United States citizen constituted use of the mark in commerce for Lanham Act purposes:
Such foreign use has no effect on U.S. commerce and cannot form the basis for a holding that appellant has priority here. The concept of territoriality is basic to trademark law; trademark rights exist in each country solely according to that country’s statutory scheme.
Person’s Co., Ltd. v. Christman, 900 F.2d 1565, 1568-69 (Fed.Cir.1990); accord Fuji Photo Film Co., Inc. v. Shinohara Shoji Kabushiki Kaisha, 754 F.2d 591, 599 (5th Cir.1985); Fin. Matters, Inc. v. PepsiCo, Inc., 806 F.Supp. 480, 484 (S.D.N.Y.1992); see also La Societe Anonyme des Parfwms Le Galion v. Jean Patou, 495 F.2d 1265, 1270 n. 4 (2d Cir.1974) (“[I]t is well settled that foreign use [of a trademark] is ineffectual to create trademark rights in the United States.”). Although we have never before directly addressed this question, we have “accept[ed] the assertion ... that ‘trade dress use in foreign countries does not create protectable trade-mark rights in the United States.’ ” CBS, Inc. v. Logical Games, 719 F.2d 1237, 1239 (4th Cir.1983); cf. Buti v. Perosa, 139 F.3d 98, 103 (2d Cir.1998) (noting that plaintiff conceded that “Congress has no constitutional authority to regulate the operation of the Fashion Cafe in Milan” perhaps in “recognition of the fact that ... registration and use of the Fashion Cafe name in Italy has not, given the territorial nature of trademark rights, secured it any rights under the Lanham Act”).
Aside from its singularly unpersuasive attempt to distinguish Buti,4, the majority ignores all of this authority. The majority also ignores the fact that courts have uniformly rejected its precise argument, i.e., use of a foreign mark in a foreign country somehow grants the foreign holder of the mark priority over one who uses the mark first in the United States. See, e.g., Per*386son’s, 900 F.2d at 1567-69; Fuji Photo Film, 754 F.2d at 599; Fin. Matters, 806 F.Supp. at 484-85; Johnson & Johnson v. Diaz, 339 F.Supp. 60, 63 (C.D.Cal.1971).
Like the courts, the Trademark Trial and Appellate Board (TTAB), whose decisions are entitled to “great weight,” Buti 139 F.3d at 105 (internal quotation marks omitted); see also In re Dr. Pepper Co., 836 F.2d 508, 510 (Fed.Cir.1987), has also long rejected this theory, concluding instead that “[pjriority of right in a trademark in the United States depends on priority of use in the United States and is not affected by priority of use in a foreign country.” Sterling Drug Inc. v. Knoll AG. Chemische Fabriken, 1968 WL 8198, 159 U.S.P.Q. 628, 630 (1968); see also Techex Ltd. v. Dvorkovitz, 1983 WL 51872, 220 U.S.P.Q. 81, 83 (1983); Mother’s Restaurants Inc. v. Mother’s Other Kitchen, Inc., 1983 WL 51992, 218 U.S.P.Q. 1046, 1048 (1983); Stagecoach Properties, Inc. v. Wells Fargo & Co., 1978 WL 21236, 199 U.S.P.Q. 341, 349 (1978); see also McCarthy § 29.3 (“Prior use of a trademark in a foreign country does not entitle its owner to claim exclusive trademark rights in the United States as against one who used a similar mark in the U.S. prior to entry of the foreigner into the domestic American market.”)5
Nor is the rule any different when, as here, the mark is used in a foreign country in connection with services or goods sold to United States citizens. See, e.g., Person’s, 900 F.2d at 1567-69 (rejecting argument that use of trademark on goods sold in Japan by Japanese company to a U.S. citizen could establish priority rights against person using mark first in the United States); Mother’s Restaurants, 218 U.S.P.Q. at 1047-48 (rejecting argument that use of trademark on restaurant services provided in Canada by Canadian entity to Americans created “priority rights in said mark in the United States”); Stagecoach Properties, 1978 WL 21236, 199 U.S.P.Q. at 349 (rejecting argument that use of trademark on hotel and restaurant services provided in Mexico to “people from the United States” created rights protectable under U.S. law in part because this “argument ignores the fundamental rule that activity outside the United States is ineffective to create rights in marks within the United States”); see also Sterling Drug, 1968 WL 8198, 159 U.S.P.Q. at 630-31.
I recognize that, consistent with the Paris Convention and other international agreements, owners of foreign trademarks can obtain United States trademark protection by registering their marks in the United States without having to show actual prior use in this country. See 15 U.S.C.A. § 1126(e) (West Supp.2002) (providing that a foreign applicant from a country with whom the United States maintains certain treaty rights can obtain a U.S. registration for a mark registered in its home country that it intends to use in U.S. commerce). But, as the majority acknowledges, SBM has not registered its mark in the United States. See ante at 361.6
*387Moreover, when the owner of a foreign trademark, like SBM, has properly registered its foreign mark under the Lanham Act, that registration merely entitles it to the “same national treatment as any other registrant.” Imperial Tobacco Ltd. v. Philip Morris, Inc., 899 F.2d 1575, 1578 (Fed.Cir.1990). Therefore, after it has registered its foreign mark, the holder of the mark can maintain United States protection of the mark only if it complies with the requirements of United States law. Id. at 1578-79. The Lanham Act provides that any mark owner — including the owner of a foreign mark — that discontinues use of the mark may be deemed to have “abandoned” the mark and that “nonuse” for a statutorily required period of time, now three consecutive years (formerly two consecutive years) “shall be prima facie evidence of abandonment.” 15 U.S.C.A. § 1127 (West 1998 & Supp.2002); see also Emergency One, Inc. v. American FireEagle, Ltd., 228 F.3d 531, 536 (4th Cir.2000). In this context too “ ‘use’ and ‘nonuse’ [of a trademark] mean use and nonuse in the United States.” Imperial Tobacco, 899 F.2d at 1579 (emphasis added); see also Rivard v. Linville, 133 F.3d 1446, 1449 (Fed.Cir.1998); Oland’s Breweries Ltd. v. Miller Brewing Co., 189 U.S.P.Q. 481, 487-88 (1976).
Thus, even when the owner of a foreign mark has registered its mark in the United States, it will be presumed to have abandoned the mark if it does not use the mark for the statutorily required period of time in the United States. Use, no matter how extensive, of the mark in a foreign country during this period, does not rebut the presumption of abandonment. See Rivard, 133 F.3d at 1448-49; Imperial Tobacco, 899 F.2d at 1579; Oland’s, 189 U.S.P.Q. at 487-88. In Oland’s, for example, although a Canadian company used its mark extensively on beer sold in Canada to United States citizens, because it did not sell any beer with this mark in the United States during the required statutory period, “a prima facie case of abandonment ... [was] shown.” Oland’s, 189 U.S.P.Q. at 488.
Indeed, in recent years one of our sister circuits has considered and rejected the majority’s theory in the context of facts far more favorable to the owner of the foreign trademark. See Rivard v. Linville, 133 F.3d 1446 (Fed.Cir.1998), aff'g, 41 U.S.P.Q.2d 1731 (1997). There, an owner of thirty-seven hair styling salons in Canada (the majority of which were within one hour of the U.S. border) maintained that although he had not used his mark in the United States itself, he had nonetheless “engaged in activities ... which constitute use of the mark in [United States] commerce.” 41 U.S.P.Q.2d at 1735.
Specifically, the owner maintained that far-reaching Canadian advertisement that extended into the United States; distribution of coupons, gift certificates, and bal*388loons in the United States; and, most importantly here, purchase by a “ ‘significant’ United States clientele” of the salon services in Canada constituted use of his mark in the United States sufficient to merit protection under the Lanham Act. The TTAB concluded that “the above activities neither individually or collectively constituted use of the ... mark in commerce,” and specifically rejected the owner’s “argument” that “travel of customers from the United States to [his] Canadian salons” meant that those services had been rendered “in commerce” because “activity outside the United States does not create rights in marks within the United States.” Id. at 1735-37. The Federal Circuit affirmed, explaining that “a prima facie case” had been established that the owner abandoned his mark “because he did not use it in connection with hairdressing and beauty salon services in the United States during the relevant time period.” 133 F.3d at 1449.
In sum, as the cases and commentary make clear, a mark cannot satisfy the use in commerce requirement sufficient to merit protection under the Lanham Act without a claimant showing use in the United States. Indeed, I have failed to find, and the majority fails to cite, a single case in which a court has accorded trademark rights to a foreign mark holder absent a showing of use in the United States. Moreover, as all of the existing authorities to address this issue have concluded, such use requires that both the advertising that employs the mark and the rendering of the services to which the mark attaches take place in the United States. Advertising in the United States alone, no matter how extensive, does not suffice to demonstrate the necessary use in the United States.
Before concluding, I must note the potential consequences of adoption of the majority’s rule. The rule announced by the majority today would mean that any entity that uses a foreign mark to advertise and sell its goods or services to United States citizens in a foreign country would be eligible for trademark protection under United States law.7 Such a rule threatens to wreak havoc over this country’s trademark law and would have a stifling effect on United States commercial interests generally. Before investing in a mark, firms and individuals would be forced to scour the globe to determine when and where American citizens had purchased goods or services from foreign subjects to determine whether there were trademarks involved that might be used against them in a priority contest or in an infringement action in the United States.8 On the other *389hand, SBM and companies like it would, under the majority’s rule, suddenly acquire a windfall of potential United States trademark rights for all of the goods and services advertised to and purchased by United States citizens while traveling in their countries. Like some sort of foreign influenza, these new entitlements would accompany American travelers on their return home, creating a vast array of new duties for individuals in the United States seeking to use the same or similar marks on goods or services sold in the United States.
Of course, if the law required us to permit this sort of reverse imperialism, whereby foreign subjects would be allowed to colonize American markets with their foreign trademarks based on sales conducted exclusively abroad, we would have no choice but to allow it. But the law does not compel this. Rather, the majority’s new theory is contrary to all extant authority. Applying this authority here leads to only one conclusion: SBM’s use in Monaco of its “Casino de Monte Carlo” mark does not constitute “use in commerce” of the United States sufficient to gain protection under the Lanham Act. Therefore, the grant of summary judgment to SBM should be reversed.9 On this ground, the plaintiff companies were entitled to judgment as a matter of law.
B.
My friends in the majority respond, at length but without effect, to the above analysis. See ante at 371-382. It is unclear to me precisely what they believe the cases, the TTAB decisions, and the treatises discussed above mean by their repeated directive that the statutory use in commerce requirement mandates use in the United States. See ante at 374-375 n. 7. What is clear is that the majority contends that this directive does not require that both elements of the statutory use in commerce requirement occur in the United States. For, according to the majority, existing authority, with only two exceptions, presents no obstacle to the majority’s own view that the second element of use in commerce — the rendering of services — need not occur in the United States. To bolster that argument, the majority valiantly tries to distinguish the wealth of authority discussed above. But its distinctions rest on a misreading of the cases. In truth, contrary to the majority’s assertions, the new rule it announces today does place our court squarely in conflict with at least two of our sister circuits, the TTAB, and all other existing authority on this issue.
First, the Federal Circuit has clearly and repeatedly embraced a rule that directly conflicts with the majority’s. See, e.g., Rivard v. Linville, 133 F.3d 1446 (Fed.Cir.1998), aff'g 41 U.S.P.Q.2d 1731 (1997); Person’s, 900 F.2d at 1567-69; see also Imperial Tobacco, 899 F.2d at 1579 (noting that “ ‘use’ and ‘nonuse’ [of a trademark] mean use and nonuse in the United States”). The majority mistakenly attempts to diminish Rivard’s precedential value, by explaining at length that this 1998 opinion was preceded by a prior unpublished opinion vacating an earlier TTAB decision. See ante at 375-377. In fact, the Federal Circuit’s 1998 opinion does indeed constitute recent precedent on *390this precise issue, and clearly holds that the TTAB correctly concluded that “activity outside of the United States does not create rights in marks within the United States,” 41 U.S.P.Q.2d at 1736; Rivard, 133 F.3d at 1449-50. The resulting rule is unambiguous: in order for a service mark to qualify for U.S. trademark protection, the services to which that mark attaches must be rendered in commerce in the United States.
The majority’s attempt to distinguish Person’s is equally misguided. For the majority contends that the Person’s court, in holding that a foreign mark on goods sold abroad to a United States consumer was not used in the United States and so not subject to U.S. trademark protection, was relying on the fact that there had been only “foreign advertising of [the] product to foreign consumers.” See ante at 374-375. But, in fact, Person’s involved a mark for goods, not services; under the Lanham Act, as long as a mark is attached to the goods it signifies, there is no advertising requirement akin to that pertaining to service marks. See 15 U.S.C.A. § 1127 (paragraph 16). This was clearly the case in Person’s, where the court specifically found that the Japanese clothing manufacturer had “applied a stylized logo bearing the name ‘PERSON’S’ to clothing in his native Japan,” and that a United States citizen “purchased several clothing items bearing the ‘PERSON’S’ logo and returned with them to the United States.” 900 F.2d at 1566-67. Thus, the Person’s court could not have been relying on the lack of U.S. advertising as the basis for its holding; instead, it must have grounded its conclusion that the mark had not been used in the United States, and for that reason did not qualify for U.S. trademark protection, on the fact that the goods had not been sold in the United States. Although the majority does not suggest that the Person’s holding was incorrect, applying the majority’s new rule to the facts of Person’s would lead to a holding exactly the opposite of that reached by the Federal Circuit. Because the mark owner in Person’s sold his goods (with mark attached) to a U.S. citizen, those goods were, under the majority’s new rule, sold in foreign commerce of the United States, thereby making the mark eligible for U.S. trademark protection.
The majority’s new rule also conflicts with the Second Circuit’s view on this question. In addition to rejecting my analysis of Buti, ante at 377, the majority dismisses the more recent Second Circuit opinion in Morningside, 182 F.3d 133, which eliminates any doubt as to the correctness of my analysis of Buti. Ante at 377 n. 10. In Momingside, the Second Circuit explained that a foreign financial services firm, unlike the Italian cafe owner in Buti, did have a valid mark under United States law because:
this is not a case like Buti v. Impressa Perosa, S.R.L., 139 F.3d 98 (2d Cir. 1998), in which the service mark claimant did not provide its actual business services (the conduct of a restaurant business) in the United States, but only advertised and promoted those services in the United States. Mere advertising and promotion of a mark in this country are not enough to constitute “use” of the mark “in commerce,” so as to bring the activity within the scope of the Lanham Act (id. at 105). By contrast, here Momingside Group does conduct material aspects of its investment services in the United States.
Morningside, 182 F.3d at 138.
The majority also ignores the fact that the Buti court itself, 139 F.3d at 103-05, after holding that the foreign mark holder’s promotional “activities in the United States were insufficient to establish ‘use in *391commerce’ of the Fashion Cafe name absent proof that [it] offered any restaurant services in United States commerce,” explicitly embraced the TTAB rule adopted in Rivard, 41 U.S.P.Q.2d at 1735-37, and Mother’s Restaurants, 1983 WL 51992, 218 U.S.P.Q. at 1047-48 — the two TTAB cases that the majority explicitly concedes are “on point” with this case but which reach a contrary conclusion. See ante at 374, 378-382; Buti, 139 F.3d at 105 (“We take particular comfort in our reliance on the TTAB decisions cited above [ie., Mother’s Restaurants and Rivard ] because the principles enunciated therein rest on solid footing.”). As the Buti Court explained, the basic principle animating these TTAB decisions has long been recognized by the Supreme Court. Id. at 105 (quoting United Drug Co. v. Theodore Rectanus, Co., 248 U.S. 90, 98, 39 S.Ct. 48, 63 L.Ed. 141 (1918) (“[T]hat a trademark right is not limited in its enjoyment by territorial bounds, is true only in the sense that wherever the trade goes attended by the use of the mark, the right of the [mark-holder] ... will be sustained.” (emphasis added by Buti court))).10
Furthermore, all of the other federal appellate decisions, which I have cited and the majority has rejected, do, in fact, explicitly embrace the general principle that United States trademark rights, or trade dress rights, cannot be based on “use” of the trademark or trade dress in a foreign country. See Fuji Photo Film, 754 F.2d at 599; Patou, 495 F.2d at 1270 n. 4; Logical Games, 719 F.2d at 1239; see also Rogers v. Ercona Camera Corp., 277 F.2d 94, 97 (D.C.Cir.1960) (“We take it as settled that ownership rights in a trademark in the United States, in the case of private persons, exist only as an appurtenance to a manufacturing or marketing business conducted in the United States in which the mark is used.” (footnote providing citation omitted)). The majority tries to distinguish these cases on the ground that although they admittedly involved exclusively “foreign use” such “ ‘foreign use’ ... is not what is before us today” given the purchase of casino services by United States citizens in Monaco. Ante at 377.11 But there is nothing in any of these cases (or any of the other cases cited in part I.A above) to suggest that the purchase of goods or services by American citizens traveling abroad would somehow render the use of trademarks to sell those goods or services no longer “foreign.” The crux of the matter, again, is whether the claimant can demonstrate use in the United States.
*392As for the numerous TTAB decisions holding that eligibility for trademark protection requires use in the United States, the majority dismisses most of them as inapposite and rejects the two (Rivard, 41 U.S.P.Q.2d at 1735-37, and Mother’s Restaurants, 218 U.S.P.Q. at 1047-48) that it concedes are “on point.” Ante at 378-381. In fact, the assertedly “inapposite” cases are not inapposite at all; each supports the basic rule that foreign “use,” including sales to United States citizens traveling abroad, does not entitle a foreign mark to eligibility for trademark protection under United States law.12
Moreover, the majority’s two reasons for rejecting Rivard and Mother’s Restaurants—that “there is no other precedent to support these two decisions,” ante at 378-379, and that their reasoning is in error because they “conflate” the two distinct elements of the use in commerce requirement, ante at 372 —do not withstand scrutiny. First, as demonstrated above, both the Federal Circuit and the Second Circuit have explicitly embraced the rule articulated in the TTAB’s Rivard and Mother’s Restaurants decisions, see Rivard, 133 F.3d 1446, and Buti, 139 F.3d at 103-105, and every other court has adopted an approach entirely consistent with those TTAB decisions. Second, in both Rivard and Mother’s Restaurants, the TTAB, in expressly refusing to embrace an interpretation of use in commerce like that of the majority, took great pains not to “conflate” the two elements of the use in commerce requirement.
Thus, in the TTAB’s 1997 decision in Rivard (not the earlier 1993 decision quoted by the majority), the Board canceled a United States mark registered by the owner of Canadian beauty salons because the Canadian owner had failed to demonstrate use in the United States, even though (1) he had engaged in extensive advertising that reached the United States; and (2) his beauty salon services had been sold to a “significant United States clientele.” 1996 WL 795315, 41 U.S.P.Q.2d at 1735-36. In doing so, the Board took care to separate the two elements of the use in commerce requirement:
*393We reject respondent’s argument that through his advertising, he has used the mark ULTRACUTS in the United States since 1986. Under Section 45 of the Trademark Act, a mark is deemed to be used in commerce on services
when it is used or displayed in the sale or advertising of services and the services are rendered in commerce .... [emphasis added by TTAB]
Similarly, we reject respondent’s argument that his services are rendered in commerce because commerce is affected by the travel of customers from the United States to respondent’s Canadian salons.... This case is readily distinguishable [from the single-location service provider cases cited by respondent] as respondent rendered no hair dressing and beauty salon services in the United States during the relevant time period.
It is well settled that activity outside the United States does not create rights in marks within the United States. See Stagecoach Properties, Inc. v. Wells Fargo & Company, 199 U.S.P.Q. 341 (1978). The concept of territoriality is basic to trademark law. See Person’s Co. Ltd. v. Christman, 900 F.2d 1565, 14 U.S.P.Q.2d 1477 (Fed.Cir.1990). Thus, the fact that United States residents have availed themselves of respondent’s services while in Canada does not constitute technical trademark use of respondent’s service mark which is sufficient to obtain or maintain a registration in the United States.
Id. at 1736-37.
There is no conflation of the two elements of use in commerce here. In fact, the Board’s conclusion is unambiguous: mere advertising that reaches the United States is not enough to constitute use in commerce; rather, such advertising must be combined with a rendering of services in the United States in order to “constitute technical trademark use of respondent’s service mark” and thereby “obtain or maintain a registration in the United States.” Id. at 1737. Thus, contrary to the majority’s assertion, ante at 380, the Board did in fact “reject a position similar to that present here” (indeed, the relevant facts are identical to those here: advertising that reaches the United States plus purchase of services by United States citizens abroad) to reach precisely the opposite conclusion — “that no qualifying commerce was present.” While the majority may disagree with the Board’s conclusion (and with the Federal Circuit’s published affirmance of this conclusion, 133 F.3d 1446), it can not fairly do so on the ground that the Board’s reasoning somehow conflated the two elements of use in commerce.
The same is true for Mother’s Restaurants, which presents a virtually identical set of facts in the context of a priority contest. In that case, the opposer argued that advertising of Canadian restaurant services that reached into the United States combined with purchase of those services by United States citizens traveling in Canada sufficed to establish priority of use in the United States. 218 U.S.P.Q. at 1047-48. In rejecting this claim, the Board certainly did not engage in anything approaching the elaborate “use in commerce” -analysis advanced by the majority today. But the Board, like every court to address this issue, clearly saw no reason to do so because it determined that advertising that reached American consumers could not by itself provide the basis for trademark protection in the United States absent a showing of “use of the mark on goods or services sold and/or offered in the United States.” 218 U.S.P.Q. at 1048. Thus, the Board recognized and correctly addressed both elements necessary to es*394tablish priority of use in disposing of oppo-ser’s claim. That it reached a contrary conclusion from the majority when confronted with the same set of relevant facts simply illustrates, once again, the extent to which the majority’s rule breaks with all prior authority on the matter.
II.
Even if my colleagues in the majority had correctly resolved the “use in commerce” issue, I would nonetheless have to dissent. This is so because, contrary to their holding, the district court erred in resolving disputed material facts when granting summary judgment to SBM on the secondary meaning issue and applied an improper secondary meaning analysis.
A.
The majority properly acknowledges both that the district court decided the case on cross motions for summary judgment and that in doing so the court engaged in factfinding. Ante at 362-363. However, the majority apparently believes that this was justified because the parties agreed to submit on the “voluminous record” to a bench trial and “did not contradict one another’s proffered facts, but only disputed the inferences that a fact-finder could draw from those underlying facts.” Id. at 362. In truth, the parties never agreed to submit on the “voluminous record” to a bench trial. Rather, the,, parties and the district court agreed that the court would review that record and, on the basis of it, grant summary judgment to the appropriate party, unless the court determined that material factual disputes required a trial, in which case the court would hold a trial.
Thus, the district court was not entitled to find facts and resolve conflicting inferences at this stage of the proceedings. Nor was it allowed to draw reasonable inferences from the facts before it as if it were a factfinder. Indeed, given the posture of the case, the court was required to resolve all inferences in favor of the non-moving party. Because the plaintiff companies opposed SBM’s motion for summary judgment on trademark infringement, they were entitled to all such favorable inferences in the context of the secondary meaning inquiry. See United States v. Diebold Inc., 369 U.S. 654, 655, 82 S.Ct. 993, 8 L.Ed.2d 176 (1962) (“On summary judgment the inferences to be drawn from the underlying facts ... must be viewed in the light most favorable to the party opposing the motion.”); Rossignol v. Voorhaar, 316 F.3d 516, 523 (4th Cir.2003) (“When faced with cross-motions for summary judgment, the court must review each motion separately on its own merits ‘to determine whether either of the parties deserves judgment as a matter of law.’... When considering each individual motion, the court must take care to ‘resolve all factual disputes and any competing, rational inferences in the light most favorable’ to the party opposing that motion.” (citations omitted)).
The small portion of transcript quoted by the majority, when considered in context, does not suggest that the parties ever agreed to anything other than the attempted resolution of the case via summary judgment. Both before and after the exchange quoted by the majority, which for clarity I emphasize in the material quoted below, the district court and the parties made their intentions clear:
THE COURT: ... I am going to proceed to scour the record, if you all submit it to the Court on summary judgment, and I will make my own determinations under Rule 56, Local Rule 56(b) and the like, and I will see what I ultimately come up with. If you *395have some concerns about that, we will go ahead and have a trial on Monday. [Counsel for plaintiff companies]: .... we would submit with the record to the Court for consideration without a trial....
THE COURT: So then it seems to me that the most efficient way to proceed is to decide now — is for me to decide whether there is a necessity for a trial. It seems to me that there are no experts. I have excluded the plaintiffs. That would have been an issue of dispute, presumably .... so there is really no reason why the Court should not dispose of this matter on the current record; isn’t that right?
[Counsel for plaintiff companies]: I believe that’s correct, your Honor.
[Counsel for SBM]: I think that sounds sensible, your Honor.
[Counsel for plaintiff companies]: That’s exactly what I would ask for, your Hon- or.
THE COURT: ... I will now proceed to attack this matter on the basis of the existing summary judgment record.
.... If there is a potential hearing, it will be in two weeks or three weeks or four weeks, when I get to the bottom. [Counsel for SBM]: Okay.
THE COURT: There is always the possibility — I don’t see it happening in this case, but I have had the experience of a voluminous record where counsel and I thought it was properly disposable on summary judgment, and when I finally got through working my way through this voluminous record, out popped a disputed issue of fact that I had to call to the parties’ attention. And that may happen, but I don’t think so. I don’t think so.
In any event, so unless I hear from counsel jointly on Monday on this issue, I will assume that I will proceed on the summary judgment issue — on the summary judgment record, and then I will ask for oral argument only if it’s needed. I will enter an order today indicating that....
J.A. 998-1005 (emphasis added to indicate portion of transcript quoted by majority).
That very day the district court did issue an order stating: “The parties, by counsel, agreed that the matter should be submitted to the Court on the existing summary judgment record. Should further oral argument be necessary, the parties will be advised.” Order, December 7, 2001 (emphasis added).
Moreover, after filing its initial memorandum opinion granting summary judgment to SBM, when the district court reopened the proceedings on the plaintiff companies’ motion for reconsideration, the court continued to recognize that the matter would be resolved either on summary judgment or after an actual trial, if the court determined that there were disputed issues of fact. For example, the court noted that “Rule 56(e) requires a party opposing summary judgment to submit affidavits setting forth specific facts showing that there is a genuine issue for trial.” J.A. 1020-21. In a later exchange with counsel for SBM, the court again remarked, “[y]our predecessor indicated in court that he thought it was appropriate to *396submit it on summary judgment.... So you need to have that in mind. Now — but I want to get these issues correct, and so if the record needs [to be] reopened for either side, I will do so, and try the matter if necessary.” J.A. 1023. The district court also recognized that “[t]he Celotex trilogy of cases ... govern in this case.” J.A. 1024-25. Finally, the court stated that in its view, “the defendants either prevail or don’t prevail on summary judgment on the record as it existed at the time the original memorandum opinion issued.” J.A. 1041.
Given all of these statements and the fact that the court ultimately granted SBM summary judgment, J.A. 1115, it is clear that the majority is mistaken in concluding that the parties “agreed to submit the voluminous record to the court” for a bench trial. Ante at 362. The parties agreed only to submit the voluminous record as it stood — without further evidence or witnesses — for determination on summary judgment, if that was possible, ie., only if, as the district court put it, no “disputed issue of fact ... popped out,” J.A. 1003, after the court reviewed the record. Therefore, the district court’s resolution of material factual disputes and conflicting inferences from the submitted facts when granting summary judgment constituted error, which the majority compounds with its attempted justification.
B.
The district court’s error in resolving material factual disputes infected its holding on secondary meaning. In addition, the district court erred in the manner in which it conducted its secondary meaning inquiry.
In order to receive trademark protection for a geographically descriptive mark like “Casino de Monte Carlo” the party seeking protection must prove not only that the mark was used in commerce but also that the mark has attained secondary meaning. Perini Corp. v. Perini Constr., Inc., 915 F.2d 121, 125 (4th Cir.1990).13 A party establishes secondary meaning “when, ‘in the minds of the public, the primary significance of [the mark] is to identify the source of the product rather than the product itself.’ ” Wal-Mart Stores, Inc. v. Samara Bros., Inc., 529 U.S. 205, 211, 120 S.Ct. 1339, 146 L.Ed.2d 182 (2000) (quoting Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844, 851 n. 11, 102 S.Ct. 2182, 72 L.Ed.2d 606 (1982)); see also Resorts of Pinehurst, Inc. v. Pinehurst Nat’l Corp., 148 F.3d 417, 421 (4th Cir.1998) (“Secondary meaning has been established in a geographically descriptive mark where the mark no longer causes the public to associate the goods with a particular place, but to associate the goods with a particular source.” (internal quotation marks, alteration, and citation omitted)).
In this circuit, “[p]roof of secondary meaning entails vigorous evidentiary requirements.” Perini, 915 F.2d at 125 (listing six factors “relevant to” but “not dis-positive of’ secondary meaning inquiry— “(1) advertising expenditures; (2) consumer studies linking the mark to a source; (3) sales success; (4) unsolicited media coverage of the product; (5) attempts to plagiarize the mark; and (6) the length and exclusivity of the mark’s use” (internal quotation marks and citation omitted)). Ordinarily, a party claiming trademark protection must meet its burden on secondary meaning by offering proof under the Perini factors. We held, in Larsen v. *397Terk Technologies Corp., 151 F.3d 140, 148-49 (4th Cir.1998), however, that evidence of “intentional, direct copying” of a mark, provides the party claiming protection with “a presumption of secondary meaning.” Under the Larsen rule, if the alleged infringer fails to rebut that presumption, the party claiming infringement is entitled to judgment. Id.
But even if the Larsen presumption did apply here,14 it must be properly applied. In Larsen, we expressly held that an alleged infringer could rebut the presumption. Id. at 148 (noting that evidence of intentional, direct copying creates a presumption entitling the party claiming infringement to judgment “in the absence of rebutting proof’). Because the alleged in-fringer in Larsen “failed to rebut the presumption,” we did not there address the question of the effect of such rebuttal evidence. However, the ultimate burden of proof always remains on the one asserting a claim. See St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502, 507, 118 S.Ct. 2742, 125 L.Ed.2d 407 (1993); Fed.R.Evid. 301. Therefore, Larsen must be read as consistent with Fed.R.Evid. 301, which provides that “a presumption imposes on the party against whom it is directed the burden of going forward with evidence to rebut or meet the presumption, but does not shift to such party the burden of proof in the sense of risk of nonpersuasion, which remains throughout the trial upon the party on whom it was originally cast.” Thus, evidence proffered by an alleged infringer that rebuts the Larsen presumption shifts the burden of establishing secondary meaning back to the party seeking protection, who must then make the requisite showing under the Perini factors that its mark had in fact attained secondary meaning.
Properly applied to the case at hand, this would mean that the plaintiff companies’ extensive evidence rebutting the presumption of secondary meaning, including abundant third-party use of the Casino de Monte Carlo mark by both land-based and web-based casinos located in the United States, at least some of which SBM conceded it knew of yet made no effort to eliminate, would shift the burden of proof on this issue back to SBM. At that point, SBM should have been required to prove secondary meaning under the Perini factors. The district court, however, did not interpret Larsen in this way and so did not require SBM, the party seeking protection, to prove that its mark had, in fact, attained secondary meaning.15
Instead, the district court found that the Larsen presumption applied, even though “there [w]as no direct evidence of intentional copying,” because “inferring intent of the plaintiff companies is proper given that the putative infringing term is identical to SBM’s mark and the infringing use is in the same relevant market: gambling services.” Assuming such an inference is permissible, surely the plaintiff companies’ evidence of similar extensive third-party use (known and undisturbed by SBM) should have been considered as rebutting the presumption of secondary meaning, *398thereby shifting the burden of proof back to SBM on that issue. But that did not happen. Upon concluding that the Larsen presumption applied, the district court then determined not whether SBM had ;proved secondary meaning but whether the plaintiff companies had “show[n] that the Perini factors weigh against secondary meaning.” (Emphasis added). Thus the district court placed the burden of proof on the wrong party—requiring the alleged in-fringer to prove no secondary meaning, rather than requiring SBM to prove secondary meaning.
The district court also made improper factual determinations in resolving the matter in SBM’s favor. For example, the district court determined that the extensive third-party use of the mark, constituted evidence of “plagiarism.” Yet a fact-finder could well infer from this evidence that SBM’s “Casino de Monte Carlo” mark was not distinctive enough to have achieved the requisite level of secondary meaning necessary to merit protection under United States law. See, e.g., Echo Travel, Inc. v. Travel Assoc., Inc., 870 F.2d 1264, 1269 (7th Cir.1989) (stating that evidence of “third party use of a substantially similar mark to promote the same goods or services to the same consumer class weighs against a finding that the consumer class associates the mark with one source” (emphasis in original)).16
Similarly, the district court ignored the evidence of SBM’s failure to police its mark in any systematic or vigorous way; yet a fact finder could conclude from this evidence that SBM’s mark had lost any distinctiveness that it may have had. See McCarthy, § 17:17 (“ ‘Without question, distinctiveness can be lost by failing to take action against infringers. If there are numerous products in the marketplace bearing the alleged mark, purchasers may learn to ignore the “mark” as a source identification. When that occurs, the conduct of the former owner, by failing to police its mark, can be said to have caused the mark to lose its significance as a mark.’ ” (quoting Wallpaper Mfrs., Ltd. v. Crown Wallcovering Corp., 680 F.2d 755, 766 (1982))).
For these reasons, I would conclude that, even if the plaintiff companies were not entitled to summary judgment on the use in commerce question (which I believe they were), they have produced sufficient evidence to raise genuine issues of material fact as to whether SBM shouldered its burden under the Perini factors. Accordingly, even if the majority’s holding on “use in commerce” were correct, I would reverse the district court’s grant of summary judgment to SBM.
. This two-pronged meaning of "usé” originated in the pre-Lanham Act trademark cases. See, e.g., Trade-Mark Cases, 100 U.S. 82, 94 — 95, 25 L.Ed. 550 (1879). The Lanham Act's “use in commerce” requirement draws directly on this earlier understanding. See Zazu Designs v. L'Oreal, S.A., 979 F.2d 499, 503 (7th Cir. 1992). In a case presenting facts almost identical to the case at hand, the Second Circuit clearly recognized this understanding of "use” and its grounding in earlier Supreme Court doctrine. See Buti v. Perosa, S.R.L., 139 F.3d 98, 103 (2d Cir.1998) ("Under this rule, therefore, Santambrogio's mere advertising of the Fashion Cafe mark, standing alone did not constitute 'use' of the mark within the meaning of the Lanham Act.”) (citing United Drug Co. v. Theodore Rectanus Co., 248 U.S. 90, 97, 39 S.Ct. 48, 63 L.Ed. 141 (1918)).
. The majority properly rejects the only rationale offered by SBM (and adopted by the district court) for finding the mark used in United States commerce for Lanham Act purposes, i.e. that advertising in the United States combined with the booking of reservations to SBM's various resorts through SBM's New York office sufficed to meet the use in commerce requirement. See ante at 364-365.
. This principle, that use in the United States provides the foundation for U.S. trademark rights, is a corollary of the well-established principle that trademark rights exist in each country solely as determined by that country’s law. See Ingenohl v. Olsen & Co., Inc., 273 U.S. 541, 544, 47 S.Ct. 451, 71 L.Ed. 762 (1927) ("A trademark started elsewhere would depend for its protection in Honkong upon the law prevailing in Honkong and would confer no rights except by the consent of that law.”). The United States has long committed itself to this territoriality principle by joining international agreements based on it. See, e.g., Paris Convention for the Protection of Industrial Property, March 20, 1883, art. 6(3), available at http:// www.wipo.int/clea/docs/en/wo/wo020en.htm (last visited March 20, 2003) — a convention that the United States joined in 1883. The majority builds its contrary thesis on the unremarkable proposition that for Lanham Act purposes "commerce” is defined as “all commerce which may lawfully be regulated by Congress.” 15 U.S.C.A. § 1127 (West 1998 and Supp.2002). What the majority overlooks is that to the extent Congress can regulate sales of goods and services by foreigners, bearing foreign marks, in foreign nations, it has chosen to do so by providing in the Lan-ham Act for compliance with international agreements based on the territoriality principle. See, e.g„ 15 U.S.C. § 1126(b) (West 1998).
. Contrary to the majority's suggestion, ante at 369-370, the plaintiff's concession in Buti that it did not sell cafe services in "foreign trade” with the United States, does not render Buti "unpersuasive" here. This is so because SBM has made — albeit implicitly' — exactly the same concession. Indeed, SBM does not argue that its sale of services abroad constitutes use in the "foreign trade” of the United States. Rather, like the plaintiff in Buti, SBM claims only that its U.S. advertisement and promotion activities constitute use in the United States. See Brief of Appellees at 19 (arguing that record provides "evidence of its mark's use in the United States ”); 22 (same); 23 (same); see also id. at 28 (arguing that application of famous marks doctrine entitles it to trademark protection "as an alternative to a showing of 'use' in the United States.”). Thus, in a context very similar to that at hand, the Second Circuit in Buti recently indicated its agreement with the rule that rendering services in a foreign country does not suffice to establish use in commerce for Lanham Act purposes. Moreover, Morningside Group Ltd. v. Morningside Capital Group, L.L.C., 182 F.3d 133 (2d Cir.1999), decided a year after Buti, makes clear the Second Circuit's intent. There, in the course of holding that a mark claimant met the "use in commerce” requirement by "conductpng] material aspects of its investment services in the United States,” the Second Circuit distinguished Buti as a case "in which the service mark claimant did not provide its actual business services (the conduct of a restaurant business) in the United States, but only advertised and promoted those services in the United States.” Id. at 138.
. Even if the plaintiff companies knew of SBM’s use of its foreign mark in connection with services rendered in Monaco, this does not render the plaintiff companies bad-faith users of the mark in the United States or preclude them from using the mark in the United States. See Person’s, 900 F.2d at 1570 ("Knowledge of a foreign use does not preclude good faith adoption and use in the United States.”); accord Buti, 139 F.3d at 106; Callman § 19.24 and cases cited therein ("[E]ven an intentional imitator may acquire domestic trademark rights if the mark he imitates is used in a foreign country.”).
. Of course, the Supreme Court has long held that when a United States trademark (i.e., a mark used or registered in this country) is at *387issue, no "rule of international law” prevents the United States " 'from governing the conduct of its own citizens ... in foreign countries when the rights of other nations or their nationals are not infringed.’ " Steele v. Bulova Watch Co., 344 U.S. 280, 285-86, 73 S.Ct. 252, 97 L.Ed. 319 (1952) (quoting Skiriotes v. Florida, 313 U.S. 69, 73, 61 S.Ct. 924, 85 L.Ed. 1193 (1941)). For this reason, Congress has the power, which it has exercised in the Lanham Act, to prohibit United States citizens from engaging in “deliberate acts” abroad which result in "[u]nlawful effects in this country,” i.e., infringement of a trademark registered in the United States. Bulova, 344 U.S. at 288, 73 S.Ct. 252. As the majority seems to recognize (by citing Bulova briefly only to distinguish it, see ante 369 n. 6), Bulova offers no support for its theory, which arises from a claim by a foreign entity to United States trademark protection for a foreign mark never used or registered in this country.
. It is worth noting here that, contrary to the majority’s suggestion, ante at 370, proof of distinctiveness by way of secondary meaning does not necessarily operate as a check on the massive expansion of protection that would be accorded to foreign marks used solely in foreign countries under the majority's "use in commerce” holding. This is so because if a mark is inherently distinctive, the mark holder need not offer any proof of secondary meaning in order to receive protection. See, e.g., Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992).
. Although I appreciate the majority’s concern with the policy objective of avoiding consumer confusion, ante at 381, I do not see how its rule would enhance the ability of trademark law to achieve this objective. Indeed, by allowing foreign mark holders to acquire trademark protection in the United States for goods and services sold exclusively abroad, the majority's rule would likely generate a whole new set of trademark disputes in the United States, exacerbating consumer confusion in the process. Moreover, for the reasons mentioned in the text, the "notice" function of trademarks vis-a-vis other firms would be severely undermined by the majority’s rule. See Zazu, 979 F.2d at 503 (discussing notice function of "use” requirement under U.S. trademark law).
. Nor does the "famous marks” doctrine provide SBM any refuge. That doctrine has been applied so seldom (never by a federal appellate court and only by a handful of district courts) that its viability is uncertain. Perhaps for this reason, SBM conceded in the district court that it could not prevail on a famous marks argument without showing "some use” of its mark in the United States. See J.A. 196. Since SBM has shown no use of its mark in the United States, this concession forecloses appellate pursuit of this argument.
. This principle — that the trade upon which the trademark is based (i.e. the goods or services to which the mark attaches) must reach those markets in which the trader seeks protection — inevitably leads to the conclusion that foreign conduct that does not reach United States markets does not suffice as a basis for according eligibility for U.S. trademark rights. See United Drug, 248 U.S. at 98, 39 S.Ct. 48; Hanover Star Milling Co. v. Metcalf, 240 U.S. 403, 416-17, 36 S.Ct. 357, 60 L.Ed. 713 (1916) ("But the mark, of itself, cannot travel to markets where there is no article to wear the badge and no trader to offer the article_ [T]he trade-mark right assigned” cannot be "greater in extent than the trade in which it [is] used.” (internal quotation marks omitted)). The majority’s rule turns this principle on its head. Under its rule, whenever United States citizens go into foreign markets to purchase goods or services, they render any foreign marks attached to those goods and services eligible for trademark protection in United States markets.
. The majority also suggests that some language in Logical Games "portend[s] something akin to the very case before us today.” Ante at 377. Actually, when read in context, it is clear that this language merely refers to situations in which an American importer purchases goods abroad and sells them in the United States.
. Thus, Techex cites the rule announced in Mother’s Restaurants, supporting the conclusion that, for the TTAB at least, there is no valid distinction between foreign use that involves the purchase of goods or services by U.S. citizens traveling abroad and foreign use that does not involve such purchases. 220 U.S.P.Q at 83. Stagecoach Properties cites the same principle to reject the argument that use of a mark to sell hotel and restaurant services to United States citizens in Mexico sufficed to create U.S. trademark rights. 199 U.S.P.Q. at 349.
In Oland’s, the Board concluded that the Canadian brewer had abandoned its American registration of the SCHOONER mark in the United States, even though the Canadian company had advertised with its mark extensively in the United States and even though United States citizens had purchased the company's beer in Canada. Oland's, 189 U.S.P.Q. at 482-83. Thus, although the Board concluded that a prima facie case of abandonment had been shown because the Canadian brewer had not sold any beer under the SCHOONER mark "in commerce which may be lawfully regulated by Congress,” id. at 488, it did so on facts identical to those in this case (U.S. advertising plus sales to U.S. citizens in a foreign country).
Finally, in Sterling Drug, a goods case, the TTAB concluded that even though the German manufacturer had provided its drug "denoted as TALUSIN” to doctors who then prescribed it to United States citizens in Germany, such use did not qualify as "use in commerce” sufficient to merit protection under the Lanham Act. 159 U.S.P.Q. at 630-31. In attempting to distinguish this case on the ground that the advertising there was insufficient to merit trademark protection, the majority makes the same mistake it made in trying to distinguish Person’s, i.e. importing the advertising requirement for service mark protection to trademark protection for goods.
. For those marks found to be inherently distinctive, the mark holder would not need to offer any proof of secondary meaning to merit protection (assuming, of course, that the mark also satisfies the use in commerce requirement). See, e.g., Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992).
. Of course, I do not believe the Larsen presumption applies here because copying a foreign trademark that has not been used or registered in the United States violates no law of this country. See, e.g., Buti, 139 F.3d at 106; Person’s, 900 F.2d at 1570.
. Contrary to the majority’s suggestion, the district court did not “employ[] two distinct analyses to assess the secondary meaning,” beginning with the "traditional Perini analysis” and then "furthermore" proceeding under Larsen. See ante at 370-371. Rather, the district court first applied the Larsen presumption and then considered the Perini factors in light of this presumption.
. The majority's recognition, ante at 363 n. 4, 371, that conflicting inferences could be drawn from the plaintiff companies’ evidence of third-party use simply confirms, once again, that the district court erred in drawing the inferences it did given that it decided the case on summary judgment. Because the plaintiff companies were the non-moving party on the issue of secondary meaning, any such inferences should have been viewed and resolved in the light most favorable to them. See Diebold, 369 U.S. at 655, 82 S.Ct. 993; Harley-Davidson Motor Co., Inc. v. Powers-ports, Inc., 319 F.3d 973, 989 (7th Cir.2003) ("The phrase in the light most favorable to the nonmoving party ... simply means that summary judgment is not appropriate if the court must make a choice of inferences.” (internal quotation marks and citation omitted)).