LEVAL, Circuit Judge:

Defendants Haar Communications Inc.' and Richard Haar appeal from the order of the United States District Court for the Southern District of New York (Richard Conway Casey, District Judge), preliminarily enjoining them from using 81 internet domain names consisting of variations of the words “The Children’s Place.” The district court held that these names were likely to dilute plaintiff TCPIP Holding Company’s trademark and service mark, “The Children’s Place,” in violation of the Federal Trademark Anti Dilution Act (“Dilution Act” or “FTDA”), 15 U.S.C. § 1125(c), and to infringe plaintiffs mark in violation of Section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a).

To the extent the court’s ruling was on the basis of the Dilution Act, it is vacated. Because plaintiffs mark, “The Children’s Place,” for a store selling children’s clothes, is descriptive and lacks inherent distinctiveness, it does not qualify for the protection of the Act; in addition, plaintiff failed to show that its mark is “famous,” as required by the Act. To the extent the injunction was granted under the Lanham Act on the basis of likelihood of confusion, we affirm as to a number of the defendants’ domain names that are confusingly similar to plaintiffs mark. However, some of the defendants’ domain names are sufficiently different from plaintiffs mark, especially given its weak, descriptive character, that no infringement is properly found.

BACKGROUND

1. Facts

Plaintiff, TCPIP Holding Company, Inc. (“TCPIP”), operates a chain of stores, primarily in the eastern half of the United States, selling children’s clothing and accessories. Plaintiffs stores operate under the registered mark, “The Children’s Place.” The articles sold in the stores are also labeled “The Children’s Place.” 1 Plaintiff has conducted this activity under its mark for about thirty years. Its has grown from 87 stores in 1994 to 228 in 1998, and its sales volume has increased during this same period from approximately $100 million to $280 million.

In August 1996, TCPIP registered two internet domain names, “tcpkids.com” and “childrensplace.com,” which can be used to gain access to its web site, though which it sells goods bearing the mark, “The Children’s Place.”

Defendants, Haar Communications Inc., a New York corporation, and Richard Haar, its president and sole employee (collectively “Haar” or “defendant”), offer consulting and networking services in the area of telecommunications. Haar Communications was incorporated on April 24, 1998 and is located in a Manhattan apartment. Around the fall of 1998, Richard Haar developed the idea of creating a “portal” on the internet for children. This portal would be a webpage that would facilitate “surfing” the web for materials concerning children by providing links to a broad array of child-related products, services and information.

On November 9, 1998, Haar registered the domain name, “thechildrensplace.com,” with the domain name registrar, Network Solutions, Inc. (“NSI”). After obtaining this domain name, Haar posted the following on the domain’s website:

*91 “THIS IS THE FUTURE HOME OF THE CHILDRENS PLACE THE PLACE FOR YOUR CHILDREN

For more information write haarcom@yahoo.com”

In late January 1999, TCPIP discovered that Haar had registered the domain name, “thechildrensplace.com.” Counsel for TCPIP sent a letter demanding that Haar cease and desist from using plaintiffs mark and that Haar transfer to it the domain name registration for “thechil-drensplace.com.”

Haar then purchased and registered at least 66 more domain names containing variations of the words “children” and “place” in combination, at the price of $119 per name. 2 Approximately half of the names are identical to the other half, but for the fact that half end with the top-level domain 3 of “.com,” while half end with “.net” — as in “thechildrensplaces.com” and “thechildrensplaces.net,” or “thechild-plaee.com” and “thechildplace.net.” They also vary in that they employ combinations of the singular and plural forms of “child” and “place,” as well as of the possessive form of “child/children.” In some cases, the word “the” is either omitted or is replaced by “a,” “your,” or “our.” Examples include “childplaces.com/.net,” “achild-places.com/.net,” and “ourchilds-place.com/.net.”

By letter dated February 10, 1999, Haar refused to cease using the domain name “thechildrensplace.com.” Around this time, Richard Haar contacted Ezrah Dabah, the Chief Executive Officer of TCPIP, and requested a meeting to discuss the possibility of a joint venture. Dabah agreed to meet because TCPIP wished to acquire the domain name, “thechildrensplaee.com.”

Haar and Dabah met on February 22, 1999. At this meeting, Haar presented to Dabah a “business plan” for creating a portal on the internet and invited Dabah to enter into a joint venture to that end. Dabah declined this offer and, instead, asked Haar to name a price for “thechil-drensplace.com.”

By letter dated March 3, 1999, Haar offered to sell TCPIP a “Domain Matrix Package,” including “theehildrens-place.com” and 35 of the other domain names, for $570,000. By letter of March 8, 1999, Haar offered another package of 44 names, for $697,000.

Dabah rejected the offers and proposed to purchase “thechildrensplace.com” alone for $30,000. Haar responded by offering to sell a package of sixteen domain names, including “thechildrensplace.com” and “thechildrensplace.net” at the price of $30,000 per name, or $480,000, plus TCPIP’s transfer of its registration for “childrensplace.com” to Haar. Dabah refused, and TCPIP filed suit.

*92 2. Proceedings Below

TCPIP’s complaint alleged, inter alia, that Haar was liable for trademark infringement, unfair competition and dilution, and sought declaratory judgment, injunction, and punitive damages. The complaint listed 44 of Haar’s registered domain names. 4

On May 21, 1999, the district court heard oral argument on TCPIP’s motion for preliminary injunction and granted the motion. Finding a “high degree of similarity” between TCPIP’s mark and Plaar’s domain names, the court ruled that TCPIP was likely to succeed on the merits of its claims under the Dilution Act and for infringement under the Lanham Act. The court directed TCPIP to submit a proposed order listing the domain names to be enjoined.

TCPIP submitted a proposed order listing 67 domain names to be enjoined, consisting of the 44 names listed in footnote 4, plus an additional 23 similar names that begin with “your” or “our,” rather than “the” or “a.” 5 Haar argued that “the domain names ... which commence with the word ‘your’, or ‘our’, or ‘a’, or ‘child’, or ‘the child’, or ‘children’ are not confusingly similar or likely to cause dilution of the mark The Children’s Place, and defendants should not be enjoined from use of the same.” Haar proposed a counter-order that would enjoin only eight names.

On May 27, 1999, the court signed TCPIP’s proposed order, preliminarily enjoining Haar from “using, as part of a domain name or otherwise” TCPIP’s mark, “The Children’s Place,” or any “colorable imitation thereof,” including the 67 names proposed by TCPIP.

On December 9, 1999, the district court granted TCPIP’s motion to expand the scope of the injunction by adding an additional fourteen domain names Haar had registered at some point that year, 6 bringing the total number of enjoined domain names to 81.

DISCUSSION

To obtain a preliminary injunction, TCPIP must establish: “(1) the likelihood of irreparable injury in the absence of such an injunction, and (2) either (a) likelihood of success on the merits or (b) sufficiently serious questions going to the merits to make them a fair ground for litigation plus a balance of hardships tipping decidedly” in its favor. Federal Express Corp. v. Federal Espresso, Inc., 201 F.3d 168, 173 (2d Cir.2000). The standard of appellate review of a grant or denial of a preliminary injunction is said to be abuse of discretion. See, e.g., North Atl. Instruments, Inc. v. Haber, 188 F.3d 38, 43 (2d Cir.1999); Otokoyama Co. v. Wine of Japan Import, Inc., 175 F.3d 266, 270 (2d Cir.1999). However, that standard is generally considered satisfied if the district court relied on incorrect law. See North Atl. Instruments, 188 F.3d at 43.

*93 1. The Dilution Claim

TCPIP employs the mark “The Children’s Place” to designate stores that are devoted to the sale of children’s clothing and accessories, as well as the children’s clothes and accessories sold in these stores. A store that specializes in children’s merchandise is a children’s place—a place that sells children’s merchandise, a place to buy what children need. The use of the words “children’s place” as a mark for such a store is thus highly descriptive. Descriptive marks are considered non-distinctive and weak under trademark law. The first question we face is whether the special broad protection of the Dilution Act is available to a descriptive mark.

In considering the precise terms of the Dilution Act, its fit with the general trademark law as set forth in the Lanham Act, the policies underlying trademark law, and the legislative history of the Dilution Act, we conclude that a descriptive mark does not come within the protection of the Dilution Act.

a) The requirement of distinctiveness

The Dilution Act provides that “[t]he owner of a famous mark shall be entitled, subject to the principles of equity ..., to an injunction against another person’s commercial use in commerce of a mark or trade name, if such use begins after the mark has become famous and causes dilution of the distinctive quality of the mark.” 15 U.S.C. § 1125(c)(1) (emphasis added). Because the Act protects against the dilution of the mark’s “distinctive quality,” trademark owners seeking protection under the Act must establish that their marks possess a “distinctive quality” in order to state a claim for dilution. See Nabisco, Inc. v. PF Brands, Inc., 191 F.3d 208, 216 (2d Cir.1999) (“A mark that, notwithstanding its fame, has no distinctiveness is lacking the very attribute that the antidilution statute seeks to protect.”). The Act furthermore invites courts to consider “the degree of inherent or acquired distinctiveness of the mark,” 15 U.S.C. § 1125(c)(1)(A), among other listed factors.

The relevance of the “degree” of distinctiveness suggests that a bare minimum of distinctiveness is not adequate. If it were, the degree of distinctiveness would have no relevance. We turn first to explore the meaning and significance of the term “distinctive quality” or distinctiveness in the trademark law.

The inherent distinctive quality of a mark has long been an important standard in trademark law. The law has always disfavored marks that lack distinctiveness and either denied them protection outright or grudgingly granted protection under special conditions. Distinctiveness is traditionally measured on a scale memorably described by Judge Friendly in Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4, 9-11 (2d Cir.1976). The scale, progressing from least to most distinctive, is described in terms of marks that are: (1) generic; (2) descriptive; (3) suggestive; and (4) arbitrary or fanciful. Judge Friendly noted that the ascending order reflected not only “eligibility to trademark status,” but also “the degree of protection accorded.” Id. at 9.

Generic marks, consisting of words that identify the type or species of goods or services to which they apply, are totally lacking in distinctive quality; they are not entitled to any protection against infringement, even if they have become famous as marks, because according such protection would deprive competitors of the right to refer to their goods by name. See id. at 10; see also Nabisco, 191 F.3d at 215.

Descriptive marks are marks that describe a product or its attributes. Because they are descriptions, they also lack distinctive quality as marks. The same policy considerations that justify barring protection for generic marks disfavors the grant of protection to descriptive marks, albeit to a lesser degree. According trademark exclusivity to a descriptive mark would inhibit competitors from using descriptions of *94 their competing products. Thus, at common law, and under the Trademark Act of 1905, “neither those terms which were generic nor those which were merely descriptive could become valid trademarks.” Abercrombie, 537 F.2d at 9. Over time, this rule was softened in deference to the merchant’s and the public’s shared interest in maintaining the reliability of marks that had developed consumer recognition through use in commerce. In contrast to prior law, the Lanham Act accorded regis-trability to a descriptive mark if the public had come to associate the mark with the goods or sendees of the user — in trademark parlance, if the mark had acquired “secondary meaning.” In permitting the registration of descriptive marks that have acquired secondary meaning, Judge Friendly explained, the Lanham Act “strikes the balance ... between the hardships to a competitor in hampering the use of an appropriate word and those to the owner who, having invested money and energy to endow a word with the good will adhering to his enterprise, would be deprived of the fruits of his efforts.” Id. at 10.

The compromise of the Lanham Act that made such descriptive marks registrable is set forth in 15 U.S.C. § 1052(e) and (f). Clause (e) denies registrability to a mark which “when used on or in connection with the goods of the applicant is merely descriptive.” 7 Id. § 1052(e). Clause (f), however, provides that nothing in the Chapter shall prevent registration of a mark “which has become distinctive of the applicant’s goods in commerce.” Id. § 1052(f). Thus, secondary meaning, or acquired distinctiveness saves a mark from unregistrability and non-enforceability. Nonetheless, it is the policy of the Act to grant broader protection to those marks with a higher degree of inherent distinctiveness. See Abercrombie, 587 F.2d at 9 (the “ascending order ... [of the four categories of marks] reflects ... the degree of protection accorded”); Arrow Fastener Co. v. Stanley Works, 59 F.3d 384, 391 (2d Cir.1995) (holding that “the strength or distinctiveness of a mark determines ... the degree of protection it will be accorded”) (internal quotation marks omitted). In short, descriptive marks are traditionally disfavored under the law and have been granted a minimum of protection under the Lanham Act when secondary meaning has been shown, due to equitable concerns about the unfairness of depriving those who have invested in a mark of the goodwill they have thereby developed and depriving the public of the ability to rely on a mark it has come to recognize as an emblem of quality.

Marks that do not directly describe goods or services or their attributes, but rather are suggestive of them, have some distinctive quality and are thus protected without need to prove secondary meaning. See Abercrombie, 537 F.2d at 10-11; Nabisco, 191 F.3d at 215-16. The most distinctive marks, and thus the strongest, are marks that are arbitrary or fanciful. Arbitrary or fanciful marks enjoy the broadest protection of the trademark laws. See Abercrombie, 537 F.2d at 11; Nabisco, 191 F.3d at 216.

In creating a cause of action for dilution, see Federal Trademark Dilution Act of 1995, Pub.L. No. 104-98, 109 Stat. 985 (1996), Congress greatly expanded the scope of protection available to owners of famous trademarks. Under the traditional action for infringement, the owner of a mark may bar another from using a mark in a manner that creates a likelihood of consumer confusion as to the source of goods. See 15 U.S.C. § 1125(a). Thus, as a general proposition, under traditional trademark law, a mark is enforceable within the area of commerce in which the mark has been established. Its establishment in *95 one segment of commerce generally does not prevent others from using the same or a similar mark in a different, non-competing area; ordinarily, little confusion will result when the junior use is in an area of commerce that is outside the senior owner’s area.

In contrast, the Dilution Act gives the owner a far greater scope of exclusivity. It permits the owner of a qualified, famous mark to enjoin junior uses throughout commerce, regardless of the absence of competition or confusion. See 15 U.S.C. § 1127 (providing that dilution may be found “regardless of the presence or absence of ... (1) competition between the owner of the famous mark and other parties, or (2) likelihood of confusion, mistake, or deception”); H.R.Rep. No. 104-374, at 3 (1995), reprinted in 1995 U.S.C.C.A.N. 1029, 1030 (“Dilution is an injury that differs materially from that arising out of the orthodox confusion.”). One circuit has characterized the Dilution Act as coming “very close to granting rights in gross in a trademark.” Avery Dennison Corp. v. Sumpton, 189 F.3d 868, 875 (9th Cir.1999) (internal quotation marks omitted).

The Dilution Act further differs from traditional trademark law in that the class of entities for whose benefit the law was created is far narrower. The action for infringement under the Lanham Act serves the interests of consumers, as well as sellers, in having trademarks functions as source-identifiers. By preventing competitors from marketing their goods under the same or a confusingly similar mark, the action for infringement “helps assure a producer that it (and not an imitating competitor) will reap the financial, reputation-related rewards associated with a desirable product.” Qualitex Co. v. Jacobson Prods. Co., 514 U.S. 159, 164, 115 S.Ct. 1300, 131 L.Ed.2d 248 (1995). At the same time, it “reducefs] the customer’s costs of shopping and making purchasing decisions, for it quickly and easily assures a potential customer that ... the item with this mark ... is made by the same producer as other similarly marked items that he or she liked (or disliked) in the past.” Id. at 163-64, 115 S.Ct. 1300 (internal quotation marks and citation omitted). In contrast, the Dilution Act is designed solely for the benefit of sellers. Its purpose is to protect the owners of famous marks from the kind of dilution that is permitted by the trademark laws when a junior user uses the same mark in a non-confusing way in an unrelated area of commerce. The Dilution Act offers no benefit to the consumer public — only to the owner.

Against a background of policies that strongly disfavor marks lacking inherent distinctiveness, according them only narrow protection, we think it highly unlikely that Congress intended to extend to such marks the expanded rights conferred by the Dilution Act. See I.P. Lund Trading v. Kohler Co., 163 F.3d 27, 47 (1st Cir.1998) (“[T]he standard for fame and distinctiveness required to obtain anti-dilution protection is more rigorous than that required to seek infringement protection.”) As we observed in Nabisco, “the degree of distinctiveness of the senior mark has a considerable bearing on the question whether a junior use will have a diluting effect.” 191 F.3d at 217. In our view, the language of the Act, specifying the need for a showing of dilution of the “distinctive quality” of the plaintiffs mark, and inviting the court to assess its “degree” of distinctiveness, is entirely consistent with the policies rooted in traditional trademark law that have always disfavored non-distinctive, descriptive marks, because their enforcement tends to deprive others of the opportunity to use proper descriptions of their products. We conclude that the Dilution Act accords its special, broad protection only to marks that have a significant degree of distinctiveness.

The tenuous compromise wrought by the Lanham Act, according descriptive marks the minimum degree of protection, would be seriously distorted if such marks now became entitled to the much broader scope of exclusivity accorded by the Dilution Act. *96 Descriptive marks, often asserting geographical identity or nation-wide prominence, or claiming merit or strength, abound in the U.S. marketplace. A few well-known examples are American, National, Continental, Metropolitan, Pacific, Southern, Texas, Chicago, Federated, United, Consolidated, Allied, First National, Acme, Merit, and so forth. Some of the holders of these inherently weak marks are huge companies; as a function of their commercial dominance, their marks have become famous. 8 It seems unlikely that Congress could have intended that the holders of such non-distinctive marks would be entitled to claim exclusivity for them throughout all areas of commerce. Innumerable good-faith junior users of the same weak marks, who have developed goodwill in these marks, would be denied further use of their marks to their detriment and that of their customers. And nation-wide, throughout all areas of commerce, the use of ordinary, descriptive marks like American would be restricted to one famous user (and others whose use pre-dated the plaintiff user’s achievement of fame).

Some confirmatory indication of our assessment of the meaning of the statute may be drawn from the House Report. See H.R.Rep. No. 104-374 (1995), reprinted in 1995 U.S.C.C.A.N. 1029. In explaining the operation of the new statute and the need for protection against dilution, Congress focused on three famous marks as to which a junior use would be “actionable under this legislation.” H.R.Rep. No. 104-374, at 3, reprinted in 1995 U.S.C.C.A.N. 1029, 1030. The three marks cited as possible beneficiaries of the Act were Dupont, Buick, and Kodak — all highly distinctive, arbitrary or fanciful marks. See id. We recognize that one cannot read examples in a legislative report as indicating the limits of a statute’s coverage. We therefore attach little importance to this passage in the legislative report, but note only that it tends to confirm the indications from the text of the statute, the history and theory of trademark doctrine, and the inherent improbability that Congress would have wished to grant broad-scale exclusivity throughout all areas of commerce to weak descriptive marks.

All of these reasons lead us to the conclusion that descriptive marks, which possess no distinctive quality, or at best a minimal degree, do not qualify for the Act’s protection. “The Children’s Place,” for stores selling children’s merchandise, is such a mark.

b) The relevance of acquired distinctiveness

The next question we face in considering the applicability of the Dilution Act to a non-distinctive, descriptive mark is whether the mark’s lack of inherent distinctiveness can be compensated by the fact that the mark has achieved “secondary meaning,” or “acquired distinctiveness.”

Although the district court never expressly stated that TCPIP’s mark has acquired secondary meaning, we believe it implicitly made such a finding in holding that the mark is “strong, given the duration of its existence and its prominence in the relevant market.” We have no doubt, furthermore, that TCPIP can establish that its mark has acquired secondaiy meaning, given the evidence that it (1) operates about 230 retail stores under its mark in twenty-seven states, (2) sold $280 million worth of goods in 1998, and (3) has expended “tens of millions of dollars” advertising its mark in the last decade. See Yarmuth-Dion, Inc. v. Dion Furs, Inc., 835 F.2d 990, 993 (2d Cir.1987) (holding that evidence relevant to a finding of secondary meaning includes “the senior user’s advertising expenditures, ... [and] sales success”).

*97 The argument might be made that the Dilution Act’s requirement of distinctiveness may be satisfied by either inherent distinctiveness or acquired distinctiveness. There is ambiguous language in the statute that could be read to support that argument. The statute states: “In determining whether a mark is distinctive and famous, a court may consider ... (A) the degree of inherent or acquired distinctiveness of the mark,” among other listed factors. Id. § 1125(c)(1)(A).

On the basis of this language, two arguments might be advanced: (1) Clause (A) treats “inherent” and “acquired” distinctiveness as equivalents; and (2) if acquired distinctiveness does not satisfy the statute’s requirement of distinctiveness, then the reference to acquired distinctiveness is meaningless and superfluous; and such an interpretation of a statute — one that renders a portion of the statute superfluous— should be avoided.

Both arguments are ill-conceived. As to the first argument, the fact that the statute invites courts to consider both inherent and acquired distinctiveness in no way implies that they are to be deemed equivalent to one another. In fact, they refer to quite different phenomena. Inherent distinctiveness refers to a mark’s theoretical capacity to serve forcefully as an identifier of its owner’s goods, regardless whether the mark has fulfilled those expectations. Acquired distinctiveness refers to a different phenomenon, one measured solely by the extent of recognition that the mark in fact has earned in the marketplace as a designator of its owner’s goods or services, regardless of the mark’s innate suitability to serve as a mark. Not only are the two phenomena different, but they serve very different purposes under the statute, as described below. The mention of each as a relevant factor neither states nor implies that they should be deemed equivalent.

As to the second argument, it is incorrect that the concept of acquired distinctiveness has no discernable function in the statute, unless as a substitute for inherent distinctiveness. The list of factors set forth in § 1125(e)(l)(A)-(H) is relevant to two separate questions. The factors are listed as pertinent to the court’s “deter-min[ation] whether a mark is distinctive and famous.” 15 U.S.C. § 1125(c)(1) (emphasis added). The “degree of ... acquired distinctiveness of the [plaintiffs] mark” is directly relevant to the determination whether the mark is “famous,” as the Act requires. Acquired distinctiveness is the essential ingredient in the determination of fame, within the meaning of the statute. The statute’s requirement of fame is not satisfied by any kind of fame. The mark must have become famous as the designator of the plaintiffs goods or services. A merchant’s taking a famous name — Shakespeare or Zeus — as the mark for its product would not thereby satisfy the statute’s requirement of fame. It is true, such a mark would be famous in the sense that universal recognition would attach to the name Shakespeare or Zeus. To satisfy the statute, however, the mark must be famous in its capacity as a mark designating the plaintiffs goods. 9 In other words, to be famous within the meaning of the statute, the mark must have achieved a high “degree of ... acquired distinctiveness,” meaning that it must have become very widely recognized by the U.S. consumer public as the designator of the plaintiffs goods.

*98 Because a high “degree of ... acquired distinctiveness” is crucial to a finding of fame, there is no merit to the proposition that the statute’s reference to acquired distinctiveness would be meaningless unless the element of distinctiveness could be satisfied by acquired distinctiveness. Indeed, it is just the other way around. If the statute’s requirement of distinctiveness could be satisfied by “acquired distinctiveness,” as opposed to “inherent distinctiveness,” the reference to inherent distinctiveness would become superfluous. Why is this so? In order to qualify for the Act’s protection, the mark must be famous. By definition, every mark that is famous, in the sense intended by the Act, has a high degree of acquired distinctiveness. Thus, no mark can qualify for the Act’s protection without acquired distinctiveness. If that acquired distinctiveness satisfies not only the fame requirement, but also the distinctiveness requirement, then there will never be a case when a court needs to consider whether the mark has inherent distinctiveness. The statute’s invitation to courts to consider the mark’s degree of inherent distinctiveness would serve no function.

We therefore understand Clause (A) of § 1125(c)(1) to invite two inquiries: (1) Has the plaintiffs mark achieved a sufficient degree of consumer recognition (“acquired distinctiveness”) to satisfy the Act’s requirement of fame? (2) Does the mark possess a sufficient degree of “inherent distinctiveness” to satisfy the Act’s requirement of “distinctive quality.” The latter requirement cannot be satisfied by the mere fact that the public has come to associate the mark with the source. Thus, weak, non-distinctive, descriptive marks do not qualify for the Act’s protection, even if famous.

Because TCPIP’s mark, “The Children’s Place,” as a designator of stores for children’s clothing and accessories, is descriptive, and thus, lacks inherent distinctiveness, it cannot qualify for the protection of the Dilution Act. The mark’s deficiency in inherent distinctiveness is not compensated by the fact that TCPIP’s mark has achieved a significant degree of consumer recognition.

The question next arises whether the distinctiveness requirement is satisfied in view of the fact that TCPIP uses it mark, “The Children’s Place,” not only to designate its stores, but also to designate the merchandise sold in the stores. Merchandise is not a “place.” Accordingly, it may be argued that, even if “The Children’s Place,” designating stores for children’s merchandise, is insufficiently distinctive to qualify for the protection of the Dilution Act, the same mark as a designation for children’s merchandise is not.

Arguments to the contrary are: (1) because the mark is used only on merchandise sold at a store or website bearing the mark, “The Children’s Place,” it remains descriptive; and (2) the Dilution Act requires a significant degree of inherent distinctiveness; a bare minimum will not suffice; even if the mark is less descriptive (hence, more distinctive) when used on children’s merchandise than when used on a store, it is nonetheless too low in distinctiveness to qualify under the Dilution Act.

We need not resolve the controversy because there is another reason why the plaintiff has not shown entitlement to the preliminary injunction premised on the Dilution Act: it has failed to show compliance with the statutory requirement of fame.

c) The requirement that the mark be famous

As noted above, the benefits of the Dilution Act are available to owners of a “famous” mark. The Act does not tell how famous a mark must be. Nor does it provide any direct guidance as to how courts should answer the question. The word “famous” is susceptible to many widely different understandings. If a hypothetical Grendel’s Coffee Shop in Smalltown, U.S.A. has for years been the *99 favorite hangout of Smalltown high school students, Grendel’s may well be famous among the students and graduates of Smalltown High. Or another mark for a catalogue selling rare plant specimens may be famous among 100,000 collectors scattered throughout the country. Are those then “famous” marks within the meaning of the statute? The argument might be made that if the plaintiffs mark is “famous” in any sense coming within a dictionary definition, it qualifies for the statute’s protection.

The Merriam Webster Third New International Dictionary (1993) defines “famous,” in the relevant entries, as “la. much talked about: well-known_ b. honored for achievement: celebrated.... ” Does this mean that the hypothetical Grendel’s — famous among the students of Smalltown High — and the catalogue mark — famous among 100,000 collectors of rare plant specimens — must be found to qualify as “famous” marks under the Act?

In our view, this is a question of statutory interpretation for which dictionaries provide no answer. Legislatures do not in all instances undertake to specify the precise standards courts must follow. They do hot necessarily regard courts as adversaries to be rigidly controlled, lest they take unreasonable liberties of interpretation with the meaning of a statute. In some enactments, legislatures look upon the courts as their partners. Recognizing the difficulty of answering all questions prospectively, legislatures at times deliberately use a vague term, like “combination in restraint of trade,” or “famous,” knowing full well that its meaning is not precisely communicated, but counting on courts to understand the legislature’s intentions and to interpret the word or phrase in a sensible manner to carry out those intentions.

It seems most unlikely that Congress intended to confer on marks that have enjoyed only brief fame in a small part of the country, or among a small segment of the population, the power to enjoin all other users throughout the nation in all realms of commerce. The examples of eligible “famous marks” given in the House Report — Dupont, Buick, and Kodak, see H.R.Rep. No. 104-374, at 3 (1995), reprinted in 1995 U.S.C.C.A.N. 1029, 1030 — are marks that for the major part of the century have been household words throughout the United States. They are representative of the best known marks in commerce. Once again, we recognize that examples in a legislative report cannot be taken as defining the limits of a statute’s coverage. Putting together the extraordinary power the Act confers on a “famous” mark and the improbability that Congress intended to grant such outright exclusivity to marks that are famous in only a small area or segment of the nation, with the hints to be gleaned from the House Report, we think Congress envisioned that marks would qualify as “famous” only if they carried a substantial degree of fame.

We do not believe TCPIP’s submissions to the district court demonstrated the degree of fame necessary to qualify. The submissions were quite sketchy. According to TCPIP’s affidavit, in 1999 it operated 228 retail stores in 27 states under the mark “The Children’s Place,” and achieved sales in 1998 of $280 million, having grown from $100 million in sales though 87 stores in 1994. Over the past decade, the affidavit asserts, TCPIP spent “tens of millions of dollars” advertising its mark, but does not tell how many millions, when expended, or how effectively. Nor did TCPIP submit consumer surveys, press accounts, or other evidence of fame. The affidavit gives no statistics of any kind pertaining to any year earlier than 1994. While it asserts that the plaintiff (and its. predecessors) have used the mark continuously and exclusively for thirty years, it gives no further information — apart from unsubstantiated conelusory phrases like “the mark The Children’s Place has been widely recognized by American consumers”— that would enable a court to determine the extent or dimension of public recognition *100 of the mark and whether it has fame sufficient to meet the requirement of the Act. While the information given undoubtedly shows considerable commercial success and growth, the aggregate sales under the mark since it originated (which have not been furnished to us) may well not equal the sales of Dupont, Buick, or Kodak in any given month.

We express no view whether TCPIP may be capable of showing at trial that its mark is “famous” within the meaning of the statute, if it submits appropriate evidence. But in its motion for preliminary injunction, it did not make an adequate showing on that question. For that reason, as well as others, the preliminary injunction must be vacated to the extent it was premised on the Dilution Act.

2. The Lanham Act Infringement Claim

To establish a likelihood of success on its claim for trademark infringement under the Lanham Act, TCPIP was required to show that Haar’s use of its 81 domain names containing variations of the words “child” and “place” was likely to cause consumer confusion. See 15 U.S.C. § 1125(a)(1) (stating that “[a]ny person who, on or in connection with any goods or services, ... uses in commerce any word, term, name, symbol, or device, or any combination thereof, or any false designation of origin, false or misleading description of fact, or false or misleading representation of fact, which ... is likely to cause confusion” shall be liable for trademark infringement).

In this circuit, courts generally begin the inquiry by considering whichever may be relevant of the non-exclusive list of factors set forth by Judge Friendly in Polaroid Corp. v. Polarad Elecs. Corp., 287 F.2d 492, 495 (2d Cir.1961): (a) the strength of the plaintiff’s mark; (b) the degree of similarity between the plaintiffs and the defendant’s marks; (c) the proximity of the products; (d) the likelihood that either owner will bridge the gap, using the mark on products closer to the other’s area of commerce; (e) the sophistication of the buyers; (f) the quality of the defendant’s product; (g) actual confusion; and (h) good or bad faith. See Time, Inc. v. Petersen Publ’g Co. L.L.C., 173 F.3d 113, 117 (2d Cir.1999).

(a) Strength of the senior mark. discussed above, a mark’s strength may measured in two very different ways: (1) inherent strength, resulting from the mark’s degree of inherent distinctiveness, usually measured on the ladder ranging from unprotectable generic marks to arbitrary, fanciful marks that enjoy the broadest protection, see Abercrombie, 537 F.2d at 9-11; and (2) acquired strength, reflecting the degree of consumer recognition the mark has achieved. The policies of trademark law disfavor according exclusive rights to marks that are descriptive. See id. at 10. The Lanham Act’s compromise allowed descriptive marks registration and minimal scope of protection, upon establishing that the mark had acquired a secondary meaning as a designator of source.

The narrower scope of protection commanded by weak, descriptive marks has two explanations. The first is that granting one merchant a monopoly over the use of descriptive marks tends unfairly to prevent competing merchants from employing appropriate descriptive terms in their marketing. The second has to do with the character of descriptive marks in relation to the likelihood of consumer confusion. The more arbitrary and fanciful the mark in relation to the goods on which it is used, the more the consumer is likely to assume that a similar mark designates the owner of the first as the source of the goods. The arbitrariness of the mark in relation to those goods makes it unlikely that an unrelated merchant would select a similar mark for closely related goods. Conversely, the more descriptive the mark is of the goods, the less likely a consumer is to assume that a similar mark used on related goods came from the same source. The similarity between the two marks is as *101 likely explained by the fact that each seeks to describe the goods, as by their coming from the same merchant.

To illustrate the point, if a car is marked “Buieke,” or film in a yellow box is labeled “Kodok,” the unsophisticated observer may well fail to notice the slight differences and assume that they come from the makers of Buick and Kodak. Conversely, if a consumer knows crunchy candy bars coming from Confectioner A under the mark “CRUNCHIES” and later encounters cookies branded “CRUNCHY delights,” the consumer is as likely to attribute the similarity of name to two unrelated confectioners independently selecting names that describe and vaunt the crunchy texture of their products. Likewise, a person walking down Broadway who sees first a “Broadway Theater,” and later a “Broadway Deli” is unlikely to assume that they belong to the same owner. 10 The descriptive nature of the mark suggests that unrelated merchants may have chosen their marks for the same descriptive reason. Thus, similarity in descriptive marks is less likely to cause confusion than similarity in arbitrary or fanciful marks.

Likelihood of confusion depends on consumer expectations. If the similarity between two marks would lead consumers to expect that they identify the same source, that similarity is likely to cause confusion. If the similarity between the two marks would lead consumers to assume the similar marks were chosen because they describe similar, desirable attributes of the products, the similarity is less likely to invite a mistaken assumption that the products must come from the same source.

As noted above, “The Children’s Place” is a weak, descriptive mark, especially as used on a store for children’s merchandise, and also (although to a lesser extent) as used on children’s merchandise sold in a store of the same name specializing in children’s merchandise. Because the plaintiffs mark has achieved consumer recognition, it is not denied the protection of the trademark law, but it commands a lesser degree of protection than arbitrary, fanciful marks.

(b) Similarity of the marks. All of the defendant’s domain names are not equally similar to TCPIP’s mark. The closest of them are different only in that the separate words are run together without separating spaces or punctuation, and the resulting word is followed by a top-level domain identifier of “.com,” or “.net.” Those differences, however, are of little or no significance. It is necessary in the registration of an internet address to eliminate spaces and possessive punctuation.. See Tami K. Lefko, The West Just Got Tamer: Appeals Court Considers Use of Trademarks in Domain Names and Meta-tags, Computer Law., June-July 1999, at 27. It is necessary, furthermore, to add a top-level domain at the end of the address. Thus, consumers would see the domain name “thechildrensplace.com/.net” as employing functionally the same name as “The Children’s Place.”

Others of Haar’s disputed domain names, such as “childrensplace,” “achil-drensplaee,” and “thechildrenplace,” are so close to plaintiffs mark that the differences are scarcely noticeable. 11 However, as the differences separating the defen *102 dant’s names from the plaintiffs mark become more noticeable, the similarities are likely to appear to consumers as more reflective of the fact that both are being used to designate something having to do with children and places, rather than indicating a common source. Second-level domain names like “achildplace,” “achild-places,” “yourchildsplace,” “yourchild-places,” “mychildsplaee,” “mychildrens-place,” and “ourchildrensplaces” seem to us to be sufficiently dissimilar to plaintiffs descriptive mark that the similarities to plaintiffs mark will be attributed to descriptive goals, rather than to common source.

(c) and (d) Proximity of the products and likelihood of bridging the gap. Because Haar’s plans are as yet only vaguely formulated, it is not easy to discern with assurance the degree of proximity between plaintiffs and defendant’s areas of commerce. Nonetheless, it appears that Haar’s plans for a portal will provide links for the purchase of children’s merchandise. The district court so found. Thus, there is sufficient proximity of goods and services covered that confusion can easily arise if the similarity between the two marks is sufficient. Because the defendant is expected from the start to use its mark in the same area of commerce as plaintiffs use, the issue of the likelihood of bridging the gap is of no relevance. There is no gap to be bridged.

(e)Sophistication of consumers. The more sophisticated the consumers, the less likely they are to be misled by similarity in marks. See Cadbury Beverages, Inc. v. Cott Corp., 73 F.3d 474, 480 (2d Cir.1996) (“The sophistication factor recognizes that the likelihood of confusion between the products at issue depends in part on the sophistication of the relevant purchasers.”) (internal quotation marks omitted). In this case, there is no reason to expect that plaintiffs consumers and users of defendant’s portal will be pai’ticularly sophisticated. Plaintiff solicits a general clientele, of purchasers of children’s clothing and accessories. Defendant expects to solicit users generally exploring matters of interest to children, including purchasers of children’s merchandise. There is no reason to expect that a high degree of sophistication will protect consumers from confusion resulting from confusingly similar marks.

(f) The quality of Haar’s product. The quality of Haar’s product cannot as yet be discerned because it has scarcely taken first steps to launch a portal. If Haar in fact follows through with the establishment of a portal under names confusingly similar to plaintiffs mark and the site functions badly, or provokes consumer dissatisfaction in any way, this could harm plaintiffs mark to the extent that consumers believe there is an association between the two.

(g) Actual confusion. When two marks have existed side by side in commerce, the evidence of actual confusion among consumers, or of absence of confusion, can be very helpful in determining wdiether a likelihood of confusion results from the junior’s use of a mark similar to the senior mark. Here, because Haar has not yet launched its portal in a serious way, there has been little or no opportunity for actual confusion to be manifested. As a result, the absence of evidence of actual confusion sheds no light whatever on the problem. See Nabisco, 191 F.3d at 228.

(h) Bad faith. Bad faith on the part of a party can influence the court in at least two ways. First, where a substantive issue such as irreparable harm or likelihood of confusion is a close question that could reasonably be called either way, a party’s bad faith could cause it to lose the benefit of the doubt. Second, if prospective entitlement to relief has been established, the good or bad faith with which the parties had conducted themselves could influence the court in the fashioning of appropriate equitable relief, or even cause it to deny equitable relief to a party that had conducted itself without clean hands. A preliminary injunction can have drastic consequences — potentially putting a party out of *103 business prior to trial on the merits. A court may be less concerned about imposing such drastic consequences on a party that had conducted itself in bad faith.

Haar registered at least 66 domain names after receiving plaintiffs initial demand letter. It thereafter made exorbitant demands. It refused plaintiffs request to name a price for “thechildrens-place.com,” offering instead packages of sixteen or more names at prices around a half million dollars, and demanding plaintiffs domain name, “childrens-place.com,” in return. On the basis of this conduct, the district court found that Haar had acted in bad faith, and we cannot say the finding was unreasonable.

There was no suggestion of bad faith in any of TCPIP’s conduct toward Haar.

* * * * * *

On consideration of those factors, we conclude that all of Haar’s domain names are not equally vulnerable to preliminary injunction. We have no doubt that some of them, listed in footnote 11, are sufficiently close to TCPIP’s mark, especially when used in so closely related a field of enterprise, that consumers are likely to assume a common origin. Those domain names are likely to cause confusion.

On the other hand, others of Haar’s domain names, as outlined above, are less similar to the plaintiffs mark. Given the highly descriptive nature of the plaintiffs mark, the scope of protection it receives under the trademark law is less broad than for arbitrary or fanciful marks. Furthermore, by reason of its descriptive nature, the similarities in such cases are less likely to cause consumer confusion than in the case of arbitrary or fanciful marks, because consumers are likely to attribute some degree of similarity to the common descriptive objectives of the plaintiff and the defendant, rather than to common origin.

We therefore' conclude that the likelihood of confusion, and plaintiffs likelihood of success on the merits, differ significantly as between various of Haar’s domain names. It is clear to us that the injunction was properly granted as to the domain names listed in footnote 11, all of which so closely resemble the plaintiffs mark as to give rise to a high likelihood of confusion. It is also clear to us that, because of the weak, descriptive nature of the plaintiffs mark, the injunction should not stand as to the defendant’s domain names which are sufficiently dissimilar to the plaintiffs mark.

In granting the injunction, the district court did not consider the narrower scope of protection afforded by the Lanham Act to descriptive marks or the lesser likelihood of confusion that arises in the case of descriptive marks. On remand, the court should review all the domain names in question — other than those listed in footnote 11, as to which we affirm the injunction — to determine on the basis of the principles articulated in this opinion whether to issue a preliminary injunction.

8. Fair Use

Haar claims it is protected from liability by the doctrine of fair use. Under this doctrine, one party’s exclusive right to use a mark “will not prevent others from using the word or image [constituting the mark] in good faith in its descriptive sense, and not as a trademark.” Car-Freshner Corp. v. S.C. Johnson & Son, Inc., 70 F.3d 267, 269 (2d Cir.1995) (citing Abercrombie, 537 F.2d at 12-13); see also Restatement (Third) of Unfair Competition § 28 (1995) (“In an action for infringement of a trademark, ... it is a defense that the term used by the actor is descriptive ... of the actor’s goods, services, or business, ... and the actor has used the term fairly and in good faith solely to describe the actor’s goods, services, or business....”). The principle is codified in the provision of the Lanham Act that fair use is established where “the [defendant’s] use of the name ... charged to be an infringement is a use, otherwise than as a mark, ... which is descriptive of and used fairly and in good *104 faith only to describe the [defendant’s] goods or services, ... or their geographic origin.” 15 U.S.C. § 1115(b)(4). 12

The simple answer to Haar’s claim is that the privilege specified in the statute applies only where the name is used descriptively “otherwise than as a mark.” Haar uses the name “thechildrens-place.com” as the address, or name, of its website, and has similarly reserved all the other names in contention. This is not simply an adjectival use, as might be the case if Haar named his website otherwise, but referred to it in publicity materials as “a children’s place.” Haar’s use is as a mark. It therefore cannot qualify for the protection of 15 U.S.C. § 1115(b)(4). We need not consider whether, in view of the statute’s requirement that the name must be “used fairly and in good faith,” a finding of fair use is also excluded by the court’s finding of bad faith.

CONCLUSION

The preliminary injunction issued against Haar under the Lanham Act is affirmed as to its use of the domain names listed in footnote 11—names that are sufficiently similar to “The Children’s Place” as to cause a genuine likelihood of confusion among consumers. As to Haar’s other domain names, the injunction is vacated and the matter remanded for further consideration in light of the principles discussed in this opinion. To the extent the preliminary injunction was predicated on the Dilution Act, it is hereby vacated.

. As TCPIP’s trademarks and service marks are both "The Children's Place,” we henceforth refer, for ease of expression, to its "trademark” or "mark.” Likewise, we sometimes use "goods” to refer to both goods and services.

. These 66 domain names consisted of (1) the following 30 second-level domains, with the top-level domains ".com” and “.net” added: thechildrenplace, thechildrenplaces, thechil-drensplaces, thechildplace, thechildsplace, thechildplaces, thechildsplaces, childrenplace, childrenplaces, childrensplaces, childplaces, childsplaces, achildrenplace, achildrensplace, achildrenplaces, achildrensplaces, achild-place, achildplaces, achildsplaces, yourchil-drenplace, yourchildrensplace, yourchildrens-places, yourchildrenplaces, yourchildplace, yourchildsplace, yourchildsplaces, yourchild-places, ourchildrensplaces, ourchildsplace, ourchildsplaces; and (2) in addition: thechil-drensplace.net, childrensplace.net, child-place.net, childsplace.net, achildsplace.net, and ourchildrensplace.net.

. A domain name consists of a second-level domain, freely selected by the individual registering the name, followed by a top-level domain, such as ".com,” ".net,” or ".org.” See Sporty’s Farm L.L.C. v. Sportsman’s Market, Inc., 202 F.3d 489, 492-93 (2d Cir.2000); Katherine C. Spelman & Jason D. Firth, Internet Legal Forms for Business Domain Names, in Understanding the Intellectual Property License, at 75, 77 (PLI Patents, Copyrights, and Literary Property Course Handbook Series No. 496, 1997).

. These 44 domain names listed in the complaint consisted of (1) the following 20 second-level domains, with the top-level domains “.com” and ".net" added: thechildrensplace, thechildrenplace, thechildrenplaces, thechil-drensplaces, thechildplace, thechildsplace, thechildplaces, thechildsplaces, childrenplace, childrenplaces, childrensplaces, childplaces, childsplaces, achildrenplace, achildrensplacc, achildrenplaces, achildrensplaces, achild-place, achildplaces, achildsplaces; and (2) in addition: childrensplace.net, childplace.net, childsplace.net, and achildsplace.net.

. The additional 23 domain names consisted of (1) the following 11 second-level domains, with the top-level domains “.com” and “.net” added: yourchildrenplace, yourchil-drensplace, yourchildrensplaces, yourchil-drenplaces, yourchildplace, yourchildsplace, yourchildsplaces, yourchildplaccs, ourchil-drensplaces, ourchildsplace, ourchildsplaces; and (2) ourchildrensplace.net.

.These additional fourteen domain names consisted of (1) the following six second-level domains, with the top-level domains ".com” and “.net” added: mychildrensplace, mychildsplace, the-childrens-place, the-chil-drensplace, thechildrens-place, theplaceior-children; and (2) in addition: thechildrens-place.org, and mychildrensplaces.com.

. "No trademark by which the goods of the applicant may be distinguished from the goods of others shall be refused registration ... on account of its nature unless it — (e) Consists of a mark which (1) when used on or in connection with the goods of the applicant is merely descriptive ... of them.” Id. § 1052(e).

. Examples are American, United, and Continental Airlines, Metropolitan Life, Allied Chemical, and First National Bank of Wherever.

. Nor would the fame requirement be satisfied where a plaintiff used as its mark a mark that was famous as another entity’s mark. For example, Pacific Gas and Electric has long been famous as the name of a huge electric power company serving a significant portion of the nation’s population. When a rock band takes the name Pacific Gas and Electric as its mark, the rock band is not deemed the “owner of a famous mark” merely by virtue of having taken another user’s famous mark as its own. Within the meaning of the statute, the band’s mark would not be a “famous mark,” unless and until it became famous as the designator of the band, or in the words of the statute, unless and until it achieved a high "degree of ... acquired distinctiveness” as the designator of the band.

. The Business Section of the current Manhattan telephone directory carries approximately 250 listings of business entities using the trade name "Broadway” (the vast majority of which are located either on or close to Broadway), including: Broadway Accounting Service, Broadway Auto Driving School, Broadway Carpet, Broadway Deli, Broadway Diner, Broadway Jewelry, Broadway Nails, Broadway Theatre, Broadway Video, and Broadway Vision Center. It is most unlikely that consumers assume there is a connection, other than geographic, between the 250 establishments.

. Haar's names that are so clearly similar to "The Children’s Place” that the differences are hardly noticeable are: "thechildrens-place,” "childrensplace,” "thechildrenplace,” ‘ 'thechildrenplaces,' ’ ' 'thechildrensplaces, ’ ’ ‘ ‘the-childrens-place,' ’ ' 'the-childrensplace,' ’ "thechildrens-place,” and "achildrensplace.”

. The district court believed fair use has no application to an action under the Dilution Act. We disagree. While the statute, passed prior to the Dilution Act, refers only to a charge of ''infringement,” we observed in Nabisco that we "see no reason why it should have any less application to actions for dilution.” 191 F.3d at 221. If "Sweet” were a famous brand of automobile, its owner presumably could not use the Dilution Act to bar a seller of candy from describing its candies as "sweet.”