Pharmacia Corp. v. Alcon Laboratories, Inc.

201 F. Supp. 2d 335 | D.N.J. | 2002

201 F. Supp. 2d 335 (2002)

PHARMACIA CORP., Pharmacia AB, Pharmacia Enterprises S.A. and Pharmacia & Upjohn Co., Plaintiffs,
v.
ALCON LABORATORIES, INC., Defendant.

Civ. No. 01-1539(WGB).

United States District Court, D. New Jersey.

May 14, 2002.

William W. Robertson, Esq., Jeffrey A. Cohen, Esq., Robertson, Freilich, Bruno & Cohen, LLC, Newark, NJ, Bruce P. Keller, Esq., Michael R. Potenza, Esq., Debevoise & Plimpton, New York City, for Defendant.

Andrew T. Berry, Esq., Sherilyn Pastor, Esq., McCarter & English, LLP, Newark, NJ, Edward E. Vassallo, Esq., Timothy Kelley, Esq., James Gibson, Esq., Tila Duhaime, Esq., Fitzpatrick, Cella, Harper & Scinto, New York City, for Plaintiffs.

OPINION

(Amended)

BASSLER, District Judge.

This is an action arising under the Lanham Act, 15 U.S.C. § 1051, et seq. and the laws of the State of New Jersey for trademark infringement and trademark dilution. The Court has subject matter jurisdiction pursuant to 15 U.S.C. § 1121(a) and 28 U.S.C. §§ 1331 and 1338. This Court has supplemental subject matter jurisdiction over all other claims under 28 U.S.C. § 1367(a). Venue is proper in this District. 28 U.S.C. § 1391(b).

Plaintiff Pharmacia Corporation and its related companies filed a complaint on March 30, 2001 against Alcon Laboratories Inc., seeking a preliminary injunction against the use of an existing trademark. The application for a preliminary injunction was submitted to the Court for determination on the basis of five days of testimony of witnesses at an evidentiary hearing, as well as affidavits, exhibits and the transcripts of testimony given upon oral depositions.

FINDINGS OF FACT

The Court makes the following findings of fact^[1] and conclusions of law pursuant to Rule 52 of the Federal Rules of Civil Procedure. To the extent that any of the findings of fact might constitute conclusions of law, they are adopted as such. Conversely, to the extent that any conclusions of law constitute findings of fact, they are adopted as such.

I. THE PARTIES AND THE NATURE OF THE ACTION

A. THE PARTIES

1. Plaintiff Pharmacia Corporation is a Delaware corporation with its principal place of business located at 100 Route 206 North, Peapack, New Jersey 07977. Pharmacia Corporation is the parent corporation of the other plaintiffs: Pharmacia AB (formerly known as Pharmacia & Upjohn AB), a company organized under the laws of Sweden; Pharmacia Enterprises S.A. (formerly known as Pharmacia & Upjohn S.A.), a company organized under the laws of Luxembourg; and Pharmacia & Upjohn Company, a company organized under the laws of Delaware (collectively, "Pharmacia").

2. Defendant Alcon Laboratories, Inc. ("Alcon") is a Delaware corporation and a wholly-owned subsidiary of Alcon Holdings, which is a wholly-owned subsidiary of Alcon Universal Limited, which is a wholly owned subsidiary of Nestle S.A. Alcon's principal place of business is 6201 South Freeway, Fort Worth, Texas 76134.

3. Pharmacia and Alcon are pharmaceutical companies that manufacture and market, inter alia, ophthalmic preparations for the treatment of the eye disease glaucoma.

B. THE NATURE OF THE ACTION

4. Pharmacia initiated this civil action on March 30, 2001 under the Lanham (Trademark) Act of 1946, 15 U.S.C. § 1051, et seq. (the "Lanham Act") and New Jersey statutory and common law.

5. On April 16, 2001, Pharmacia moved for a preliminary injunction to enjoin defendant Alcon from infringing and diluting Pharmacia's federally registered Xalatan trademark in connection with Alcon's use of the name Travatan for the sale of a prostaglandin-based product used for the treatment of the eye disease glaucoma.

II. THE EYE DISEASE GLAUCOMA AND THE GLAUCOMA MARKET

6. There is no dispute about the nature of the eye disease glaucoma and the glaucoma market. This case is about the medications used in the treatment of glaucoma, a disease of the eye marked by increased intraocular pressure ("IOP") within the eye. Elevated IOP is caused by excess fluid within the eye that can result in damage to the optic disk and cause gradual loss of vision. Left untreated, glaucoma can result in blindness.

7. There is no known cure for glaucoma. The goal in treating patients with glaucoma and ocular hypertension is to decrease and maintain the IOP to within normal limits.

8. There million Americans, primarily those age 40 and over, now suffer from the disease. One million are African-Americans, for whom the disease strikes with greater frequency, is more severe and is more difficult to treat.

9. The newest class of prescription drugs used to treat elevated IOP associated with glaucoma are "prostaglandin analogues." They reduce IOP by increasing the outflow of fluid from the eye. The three largest-selling prostaglandin analogue products are Xalatan, Lumigan and Travatan. Pharmacia markets Xalatan; the generic name of its active ingredient is "latanoprost." Alcon markets Travatan; the generic name of its active ingredient is "travoprost." Allergan, Inc. ("Allergan") markets Lumigan; the generic name of its active ingredient is "bimatoprost."

10. Travatan, Xalatan and Lumigan can be dispensed only with a doctor's prescription.

11. Surveys of ophthalmologists conducted by Pharmacia in May and August of 2001 show over 95% awareness of the brand names of all major glaucoma medications, including Travatan and Xalatan.

12. The Xalatan mark as a whole is strong and is a "famous" mark among ophthalmologists.

III. PHARMACIA AND THE DEVELOPMENT OF THE XALATAN PRODUCT AND NAME

13. Alcon does not dispute Pharmacia's sales figures, advertising expenditures or that Xalatan has received media coverage.

14. Prior to Pharmacia's introduction of Xalatan, no prostaglandin product for glaucoma was available on the market in this country and many doctors were uncertain about its use because the product was widely believed to cause inflammation, pain and other detrimental effects in eyes.

15. Xalatan was a commercial success from the beginning; in its first year, sales in this country were approximately $26 million.

16. Since 1996, sales in this country have continued to increase: In 1997 sales were $126 million; in 1998 sales were $197 million; in 1999 sales were $247 million and in 2000 they were $295 million.

17. In those years, worldwide sales of Xalatan were 139, 282, 452 and 615 million dollars.

18. From its launch of Xalatan in 1996, Pharmacia undertook a massive marketing and educational campaign designed to educate the medical field about its new product and its benefits, and to encourage practitioners to switch their patients to Xalatan. This marketing and educational campaign was particularly important because of the general unfamiliarity with the use of prostaglandins for ocular indications.

19. Pharmacia has spent over $100 million educating doctors about the benefits of using prostaglandin products in the eye, and demonstrating the effectiveness of Xalatan solution. Such efforts have included seminars, publications and face-to-face meetings.

20. By 1998, Xalatan solution had become the leading branded drug for the topical treatment of glaucoma and ocular hypertension, capturing 22.5% of dispensed glaucoma prescriptions in the United States.

21. In the thirteen month period from January 2000 to January 2001, the Xalatan solution was the leader in its class of drugs dispensed.

22. Pharmacia spent large sums of money for market research for Xalatan related research in 2001. Gurreri Tr., at 368-70; Def. Ex. W (DX02498).

23. The Xalatan and Travatan products will be prescribed, dispensed and used by the same groups: physicians, pharmacists and other health care professionals and glaucoma patients.

IV. SELECTION OF THE XALATAN NAME

24. The Xalatan trademark was derived by combining the "LATAN" prefix from its generic name, latanoprost, with the "XA" portion of the code name PHXA41 used by Pharmacia during the product's development.

25. In 1997, Pharmacia obtained Federal trademark registration No. 2,118,188 for its Xalatan trademark.

26. The following paragraphs (a) thru (g) support the finding that what identifies Pharmacia's product in the marketplace and is the dominant part of the Xalatan mark is the prefix "XAL" and not the suffix "ATAN.".

(a) When Pharmacia adopted the Xalatan mark, it was aware of seven pharmaceutical marks with the suffix "ATAN" and ten more with the suffix "TAN." Def. Ex. I (DX00751-52) (Battle Dep. Ex. 1); 12/21 Tr., at 14-16 (Battle).

(b) At least three other pharmaceutical trademarks currently in the market bear trademarks with the suffix "ATAN." Def. Ex. H (DX00614) (Mintz Decl., ¶ 4); id. (DX00624-40) (Mintz Decl., Ex. 3-5) (Rynatan, Phenatan, Germ-tan).

(c) At least 11 other pharmaceutical trademarks currently in the market bear trademarks with the suffix "TAN." Id. (DX00641-72, DX00679-729) (Mintz Decl., Exs. 6-16).

(d) "AN" is the most common suffix among generic names for pharmaceutical products and the sixth most common suffix among registered trademarks in U.S. Category 005 (Pharmaceuticals). Def. Ex. G (DX00580-89) (Bruce L. Lambert, et at., "Descriptive Analysis of the Drug Name Lexicon," 56 Drug Information J.163, 170 Tbl. 3 & 171 Tbl. 4 (2001)); 12/18 Tr., at 189-91 (Lambert).

(e) Timolol is the generic name of the active ingredient of many products in a class of drugs called "beta-blockers," which traditionally have been the "first-line" topical treatment for lowering elevated IOP. Def. Ex. O (DX02073) (Harfstrand Decl., ¶ 14). For its proposed timolol-latanoprost combination product, all of the marks Pharmacia filed with the Food and Drug Administration ("FDA") and the United States Patent and Trademark Office ("PTO") begin with "Xal," none contain either the "ATAN" or "TAN" suffixes. Def. Ex. I (DX01203-28) (PTO filings for Xalaplus, Xalaplus T, Xaltimol, Xalacom, Xalacomb, Xalitol and Xalcom); Garanzini Tr., at 62,99-100; Def. Ex. I (DX01259)(internal Pharmacia notes referring to "Xal family"); Garanzini Tr., at 59-69, 77-96, 108-109, 112-19, 122-29 (describing "Xal" names submitted to FDA); Linvill-Neal Tr., at 80-81 (only "XALA" part of Xalatan name considered for inclusion in new combination product name); Garanzini Tr., at 132-33 ("Xalatan family" trademarks begin with "Xal"); 12/20 Tr., at 226-27 (Gurreri) (Pharmacia refers to Xalatan and Xalcom products as the "Xalatan family" or "Xalatan franchise").

(f) In 1993 when Eli Lilly, a major pharmaceutical company that monitored trademarks approved by the PTO saw that the Xalatan mark had been approved, it opposed that mark as confusingly similar to Lilly's mark, Oxalactam. Pharmacia responded by emphasizing that Xalatan "starts with the letter X, which makes it special." Def. Ex. I (DX01260) (facsimile from Gunnell Nilsson to Doreen Lute, dated December 8, 1993).

(g) In 2000, Pharmacia defended its "Xalcom" mark against an opposition filed by Alcon in Spain by again arguing that "[t]he letter X at the beginning of the word ... demands the most attention from the consumer." Id. (DX01082)(translation of Pharmacia brief in opposition).

V. ALCON AND THE DEVELOPMENT OF THE TRAVATAN PRODUCT

27. Alcon invested more than ten years and over $100 million to bring the Travatan product to market. Def. Ex. E (DX00330) (Kreuger Decl. ¶ 6).

28. Because prior to FDA approval of a drug, companies are permitted to promote new drugs only on a limited basis, Alcon's marketing investment in the Travatan product, prior to FDA approval, was just over $1 million. Barton 10/25 Tr., at 65-66; Gurreri Tr., at 213, 282-86.

29. Since approval by the FDA, Alcon has spent over $15 million on the sales and marketing of Travatan. Def. Ex.B (DX00024) (Barton Decl., ¶ 26); 12/17 Tr., at 150 (Barton). In 2002, Alcon's planned marketing expenditures are $38 million. Def. Ex. B (DX00024) (Barton Decl., ¶ 26.).

30. This investment has resulted in the recognition of the Travatan name; Pharmacia's marketing studies measuring ophthalmologists' familiarity with the Travatan name show a "total awareness" of over 95%. Gurreri Tr., at 226, 232-36; Def. Ex. I (DX00938, DX00945) (Wave II Awareness/Usage Study)(excerpts); Def. Ex. QQQ (DX04757, DX04767)(same)(complete); 12/17 Tr., at 150 (Barton).

VI. ALCON'S SELECTION OF THE TRAVATAN MARK

31. Although Alcon was aware of Pharmacia's mark Xalatan, the record does not support Pharmacia's contention that Alcon copied Pharmacia's mark with an intent to confuse consumers or that Alcon chose the "ATAN" suffix in order to trade off the success and name recognition of Xalatan. Alcon has credibly explained the origin and marketing goals behind the adoption of the Travatan mark.

32. Alcon's Trademark Committee selected the name Travatan after a global search and clearance process, that started in 1994 when Alcon's Research and Development team was early in the process of formulating the drug. Def. Ex. B (DX00018)(Barton Decl., ¶ 4).

33. In July 1997, the United States Adopted Names Council ("USAN") proposed the generic name "travoprost" as an alternative to Alcon's suggested generic names. Def. Ex. B (DX00031) (July 7, 1997 USAN Letter). In September 1997, Alcon consented to the name "travoprost," id. (DX00032) (Sept.12, 1997 Letter of Charles L. Schmidt), and, in March 1998, USAN formally adopted it. Def. Ex. GGG (DX04210-12) (Mar.25, 1998 USAN Letter).

34. Before Alcon learned of the travoprost name, it was considering trademark names for the product such as "alcoprost," "prostalon" and "prostal." See Def. Ex. B (DX00030)(Aug.8, 1994 Memo); id. (DX00033)(Jan.8, 1996 Memo). In connection with the name travoprost, Alcon began to consider potential trademarks that integrated the prefix "trav." Id. (DX00035-36) (Barton Decl., Ex.4); 12/17 Tr. 161-62 (Barton). Other possible names included "Travodin" and "Travopress." Most of these names appended common drug suffixes to a prefix taken from or evocative of the generic "travoprost." Def. Ex. B (DX00019) (Barton Decl., ¶¶ 7-8); id. (DX00035) (Barton Decl., Ex. 4).

35. There is sufficient credible evidence for this court to find that Alcon's selection and adoption of the name Travatan was done in "good faith."

36. Travatan was created by taking the prefix "travo" of travoprost, changing the "o" to an "a," and combining the resulting prefix "trava" with what is a common pharmaceutical suffix-"tan" in order to create a coined name. Def. Ex. B (DX00020-21) (Barton Decl., ¶¶ 11,14); 12/17 Tr., at 154-56, 161-62, 171 (Barton); see also Def. Ex. H (DX000614)(Mintz Decl., ¶ 4 (listing drug names ending in "a-t-a-n")).

37. The court finds credible the testimony of Barton and Schira at the evidentiary hearing that Alcon did not create the Travatan trademark by copying the "ATAN" ending of Xalatan. 12/17 Tr., at 171 (Barton); 12/19 Tr., at 184-85 (Schira). Rather, suffixes other than "TAN" were rejected because they did not sound or look as good as Travatan. Def. Ex. B (DX00020-21) (Barton Decl., ¶¶ 10-14); id. (DX00036) (Barton Decl., Ex. 4); id. (DX00042-43) (Barton Decl., Ex. 6).

38. Travoprost, bimatoprost and latanoprost all are pronounced in practice as if they ended in "aprost." Barton 7/18 Tr., at 27-28; Nesi Tr., at 46-47, 92-93; Di Domizio Tr., at 226-27; Def. Ex. B (DX00020) (Barton Decl., ¶ 13); 12/17 Tr., at 142, 170-171 (Barton). The prefix "TRAVA" was selected over "TRAVO" because: (1) the assonance of the repeated A vowel was desirable; and (2) the FDA generally does not approve of brand names that contain too much of the generic name. 12/17 Tr., at 156 (Barton); Def. Ex. K (DX01432) (Di Domizo Decl., Ex.2); Def. Ex. B (DX00020) (Barton Decl., ¶ 12).

39. Alcon considered several other marks and selected Travatan but only after its first choice "PRAVATIN," was considered to be too close to "PREVATAIN," a mark for a dry-eye product marketed by Allergan, Inc. 12/17 Tr., at 135-36, 140, 166-69, 172-73, 177 (Barton); 12/18 Tr., at 25-26, 47-48, 50-51 (Gill); Def. Ex. DDD (DX04145-65)(PRAVATIN trademark search report).

40. Because of the importance of the product, even after the Travatan name was tentatively selected, Tim Sear, Alcon's CEO, repeatedly urged the generation of additional potential names throughout the first half of 1998. Def. Ex. B (DX00022) (Barton Decl., ¶ 20); id. (DX00046-47) (Barton Decl., Ex. 8).

41. The minutes of the meeting of Alcon's Trademark Committee on November 19, 1997 contain the statement that: "The proposed trademark is Travatan which should compete well with Pharmacia's Xalatan's trademark." PX-25 at DP 002210. The Court does not accept the sinister connotation urged by Pharmacia. Alcon ultimately selected the Travatan mark as a trademark because it is suggestive of the generic ingredient "travoprost," Barton 7/18 Tr., at 27; Gill Tr., at 51, 72; Def. Ex. B (DX00018, DX00020) (Barton Decl., ¶¶ 5,11), and because it is easy to spell, pronounce, write and remember, because it is different from the names of its competitors' products, and because it creates a brand identity unique to Alcon. Barton 7/18 Tr., at 29, 64-65; Def. Ex. B (DX00019-22) (Barton Decl., ¶¶ 8,11,14,18). Because of these characteristics Alcon concluded the Travatan name "competes well" against the Xalatan name. Def. Ex. B (DX00022) (Barton Decl., ¶ 18); 12/17 Tr., at 151, 163-64, 172, 179 (Barton); 12/18 Tr., at 35(Gill); 12/19 Tr., at 126-28 (Schira).

42. To determine whether the Travatan name was "clear," Alcon conducted both its own database trademark search and ordered an external one, conducted by Thompson & Thompson. 12/17 Tr., at 172 (Barton); 12/18 Tr., at 6-8, 12-13, 22-23, 29-30 (Gill); 12/19 Tr., at 138-40, 154-56 (Schira); Def. Ex. B (DX00055-100) (Barton Decl., Ex. 9). The internal search disclosed the existence of Xalatan, Def. Ex. B (DX00059), but the Thompson & Thompson search did not list Xalatan as a potential conflict. (DX00101-31) (Thompson and Thompson Search) 12/18 Tr., at 12.

43. The internal trademark search and the Thompson and Thompson trademark search revealed eleven other marks that end in the suffix "TAN," out of which three marks end in "ATAN".

44. The "ATAN" portion of the Xalatan mark is weak.

45. At the evidentiary hearing Alcon's in-house intellectual property counsel, Jeffrey Schira Esq. testified, as did Alcon's manager of trademark services, Anna Gill. The Court found their testimony very credible. Both Mr. Schira and Ms. Gill participated in the trademark clearance of the Travatan name. Schira Tr., at 73-75, 103, 107, 124; Gill Tr., at 12-13, 76, 124. Mr. Schira is an experienced trademark attorney. 12/19 Tr., at 141, 195; Schira Tr., at 8-10, 17-18. His clearance of the Travatan name was part of his usual practice of reviewing and providing his "final approval" for all prospective marks. Schira Tr., at 15-16, 69-70, 76, 102-103, 110-12; Gill Tr., at 12-13.

46. Ms. Gill has had over twenty years experience in trademark reviewing and clearance. As part of the clearance process, Ms. Gill would investigate potential marks and provisionally "clear" them before bringing the file to Mr. Schira for review. Until this litigation, none of the marks Ms. Gill had cleared during her long career ever resulted in a trademark dispute. 12/18 Tr., at 27-28, 48 (Gill). But only if Mr. Schira also agreed with Ms. Gill that the mark was available would the Trademark Committee be allowed to adopt it. 12/19 Tr., at 134-35, 138-39, 155-56, 164, 192-93 (Schira); Schira Tr., at 59-62; 12/18 Tr., at 4-5, 21, 23 (Gill); 12/17 Tr., at 169-70 (Barton).

47. Although Mr. Schira was aware of the Xalatan mark at the time he reviewed the Travatan mark, he compared both the front end of the mark and the back end of the mark, and determined that there was not a likelihood of confusion. Schira Tr., at 79-92; 12/19 Tr., at 161.

48. Both Ms. Gill and Mr. Schira also concluded that Pharmacia did not have exclusive rights to the "ATAN" suffix because of the number of other pharmaceutical products that also end in the suffixes "ATAN" and "TAN," and the number of other glaucoma products that end in "AN." 12/19 Tr., at 171-73 (Schira); Schira Tr., at 74, 83, 91-93, 127-28; 12/18 Tr., at 13, 31, 38 (Gill).

49. It is worth noting that the PTO, the FDA and various other foreign regulatory authorities have all approved the Travatan mark. 12/19 Tr., at 185-91 (Schira). And when Alcon, during the clearance process, identified "TRIVITAN," a mark for an ophthalmic product marketed by Merck & Co. ("Merck") in Argentina, it successfully negotiated a co-existence agreement with Merck. 12/18 Tr., at 47 (Gill).

50. Alcon filed an intent-to-use application for Travatan with the PTO on June 3, 1998. Def. Ex. B (DX00023)(Barton Decl., ¶ 23).

51. The PTO determined that the Travatan mark was not confusingly similar to any other existing mark and published the mark for opposition on May 18, 1999. Id. (DX00023) (Barton Decl., ¶ 23); id. (DX00163-64)(Barton Decl., Ex.12).

52. In addition, the brand names for all prescription drugs, such as Travatan, must be approved by the FDA before they can be used in the United States. The FDA has the authority to reject any proposed name that "because of similarity in spelling or pronunciation, may be confused with the proprietary name or the established name of a different drug or ingredient." 21 C.F.R. § 201.10(c)(5); see also 12/17 Tr., at 173 (Barton); Garanzini Tr., at 62-64.

53. Prior to October 1999, the FDA name review process was conducted under the auspices of the FDA's Labeling and Nomenclature Committee ("LCN"), a division of the FDA's Center for Drug Evaluation and Research ("CDER"). The LCN was composed of nine health professionals from different divisions within the FDA. The LCN reviewed the submitted name for potential confusion with other drug names. If it found no such confusion likely, the LCN recommended that the name be approved by the CDER Medical Reviewer, an FDA Division Director who had ultimate approval authority over the drug. Def. Ex. E (DX00335) (Kreuger Decl., ¶ 24); id. (DX00339-43) (Kreuger Decl., Ex. 2).

54. In 1999 Alcon submitted the Travatan mark to the LCN, which after conducting its review recommended that Travatan was "acceptable for use." Def. Ex. E (DX00331-35)(Kreuger Decl., ¶¶ 13-17); Id. (DX00345)(Kreuger Decl., Ex. 3)(OPDRA Report).

55. The FDA implemented a new and more rigorous screening process designed to eliminate potentially confusing names. That process is now overseen by the Office of Post-Marketing Drug Risk Assessment ("OPDRA"), which tests proposed names on outpatient and inpatient prescriptions in a simulated clinical setting. Def. Ex. C (DX00173-75) (De Santis Decl., ¶¶ 20, 24). The purpose of this process is to reduce the risk of approving a new drug name that might result in medication errors. Di Domizio Tr., at 241.

56. OPDRA conducted a study, employing ninety-four healthcare professionals, to determine under simulated real-world conditions whether Travatan would be confused with other U.S. drug names due to similarity in visual appearance with handwritten prescriptions or verbal pronunciation of the name." Def. Ex. E (DX00346) (OPDRA Report).

57. OPDRA tested two handwritten prescriptions, one of which read "D/C Travatan today" with no reference to concentration, or strength of the drug. Id. (DX00346-47) (OPDRA Report at 3-4). It also tested two verbal prescriptions transmitted by telephone. Id.

58. The OPDRA study found "there were no erroneous interpretations of this proprietary name with other U.S. marketed drug products" in any of the four simulated prescriptions tests it conducted, including the handwritten test where no concentration of Travatan was specified. Id. (DX00334) (Krueger Decl., ¶ 21); id. (DX 00347) (OPDRA Report). As a result, OPDRA recommended that the FDA approve the Travatan name.

59. At the time OPDRA initially tested the Travatan name, Alcon had submitted data on two concentrations of travoprost: 0.0015% and 0.004%

60. In December 2000, prior to approving the Travatan product, the FDA recommended to Alcon, and Alcon agreed, that Travatan be marketed only in one concentration: 0.004%. Krueger 6/20 Tr., at 28, 30-31; Krueger 10/30 Tr., at 12, 33-34. As a result, the FDA was aware that Travatan would be marketed in one concentration (travoprost ophthalmic solution 0.004%) well before it issued its final approval of the Travatan product in March 2001. Def. Ex. E (DX00335) (Krueger Decl., ¶ 25).

61. On March 16, 2002, the FDA's CDER Medical Reviewer and Division Director, Robert DeLap, approved the Travatan name as part of the FDA approval to market Travatan. Def. Ex. E (DX00331, DX00335)(Krueger Decl., ¶¶ 11, 24).

62. In addition to approving drug names, the FDA also must approve the way the name is presented on all product packaging. The FDA approved the packaging for Travatan and did not find it to be confusingly similar to any other packaging, including Xalatan. Def. Ex. E (DX00335-36) (Krueger Decl., ¶¶ 27-28).

63. Although Pharmacia could have petitioned the FDA at any time to reconsider its approval of the Travatan name and/or packaging, it did not. See 21 C.F.R. § 10.30 Pharmacia never contacted the FDA with its concerns about the Travatan name, Gurreri Tr. 204-07, nor did it ask its paid consultant, Mr. George Di Domizio, to use his experience with and connections with the FDA to intercede on Pharmacia's behalf. Tr. 97.

64. The packaging for both Travatan and Xalatan prominently displays their manufacturers' respective house marks. Def. Ex. E (DX00336)(Krueger Decl., ¶ 28): Def. Ex. III (Xalatan package); Def. Ex. JJJ (Travatan package). In addition, Alcon emphasizes its well-known house mark in promoting its products. Guerri Tr., at 165-66; Def. Ex. V (DX03305) (Agenda for May 6, 1998 Trademark Committee Meeting).

65. Alcon has also submitted the Travatan name for regulatory review in more than fifty countries. None has raised any concern with the name. Moreover, despite opportunities to do so, Pharmacia did not oppose any of these submissions. Def. Ex. B (DX00023-24)(Barton Decl., ¶ 24); 12/19 Tr.186-91 (Schira); 12/18 Tr. 29 (Gill).

VII. PHARMACIA'S KNOWLEDGE OF THE TRAVATAN MARK

66. Pharmacia did not move for preliminary injunctive relief until April 2001 although it had constructive knowledge as early as the Spring of 1999 and actual knowledge commencing in June of that year and going forward. Although the Court does not find that Pharmacia timed the lawsuit to cause maximum disruption of the Alcon launch of Travatan, such delay must mean that Pharmacia did not initially consider Alcon's use of Travatan to infringe on its trade mark.

67. Pharmacia had constructive (if not actual) knowledge of the Travatan name as early as May 18, 1999, when Travatan was published in the Official Gazette. Pharmacia certainly had actual knowledge of the Travatan name through (1) its internal and external competitive intelligence reporting from the summer of 1999 forward and (2) its extensive testing of the Travatan name in internal market research beginning in the summer and fall of 2000 when it designed strategies to counter the anticipated Travatan launch. Gurreri Tr. 154-55.

68. Pharmacia failed to diligently monitor its trademarks.

69. The PTO approved the Travatan mark for publication on April 16, 1999 and then published the Travatan mark for opposition in the Official Gazette on May 18, 1999. Def. Ex. SS (DX03875-76) (Travatan Prosecution History); Def. Ex. B (DX00163-64) (Barton Decl., Ex. 12) (Official Gazette excerpts).

70. Of course publication in the Official Gazette is "notice to the world that Alcon" planned to use the mark Travatan. 12/21 Tr. 62 (Battle). The purpose of such publication is to provide "any person who believes that he would be damaged by the registration of a mark" the opportunity to object to registration by the PTO. 15 U.S.C. § 1063(a).

71. Although Pharmacia's head intellectual property lawyer, Mr. Battle testified that when he joined Pharmacia in October 1999 "there was nobody to monitor" the Official Gazette, 12/21 Tr. 63 (Battle), in fact Pharmacia did monitor trademark filings in 1999 in several ways:

(a) It had a "strong in-house intellectual property component" that monitored trademark filings and formally opposed at least seven trademarks that Pharmacia believed were likely to infringe or dilute its marks. Jacobs-Meadway Tr., at 13-14, 25-27, 41-44; Def. Ex. I (DX01160-1202) (Pharmacia oppositions filed in or about 1999).

(b) Pharmacia also engaged multiple outside law firms to protect its trademarks, including one that acted as a trademark watch service and monitored the Official Gazette. Jacobs Meadway Tr., at 25-31, 36-38, 40-44; 12/21 Tr. 63 (Battle).

(c) Pharmacia subscribed to a number of trade journals and other publications that reported on the approval of the Travatan mark, including a well-known trade publication, Pharmaceutical Approvals Monthly, and the PTO's Official Gazette. Def. Ex. I (DX01253-54) (2d Int. Ans.Nos.2-3).

(d) Mr. Battle's predecessor at Pharmacia was Ray Arner. He was head of Global Intellectual Property and prior to October 1999 (and Battle's arrival at Pharmacia), had responsibility for Pharmacia's trademarks. Ray Arner monitored the PTO's Official Gazette for trademark filings in May of 1999, when the Travatan mark was published for opposition. Jacobs-Meadway Tr., at 22, 25; 12/21 Tr., at 20 (Battle).

(e) As previously noted, Pharmacia received the Pharmaceutical Approvals Monthly. In its June 1999 issue, which featured articles about Pharmacia's products, it reported Alcon's intent to use the Travatan mark for its travoprost product. Def. Ex. B (DX00165-66) (Barton Decl., Ex. 13) (excerpts).

72. Despite the fact that Xalatan in 1999 "had the biggest sales worldwide in the entire corporation," and that the Xalatan product and brand name are "extremely important" to Pharmacia, 12/17 Tr., at 79-80 (Harfstrand), and although Pharmacia opposed the registration of at least seven marks that it believed infringed on various marks, it did not oppose Travatan.

73. Pharmacia has an entire department devoted to Global Competitive Intelligence ("GCI"). It is inconceivable to this Court that information about Travatan did not reach executives of Pharmacia well in advance of Alcon's launch of Travatan on March 16, 2001.

74. Information about Travatan was collected by GCI and circulated periodically to senior marketing executives in what was called "Competitor Outlook & Review" reports. 12/20 Tr., at 154-55 (Gurreri); Gurreri Tr., at 43-51.

75. One source of Global Competitive Intelligence information that is of particular significance was an ophthalmic congress regularly held in Europe by the European Society of Ophthalmology ("SOE"). Nesi Tr., at 28-42.

76. The 1999 SOE Congress was a "major convention[]" in the industry. 12/17 Tr., at 91 (Harfstrand).

77. Pharmacia was a "Main Sponsor" of the 1999 SOE Congress, which that year was held in Stockholm, Sweden from June 27-July 1, 1999. Harfstrand Tr., at 79-80; 12/17 Tr., at 92-93.

78. At this June 1999 SOE Congress, Alcon's Vice President of Pharmaceutical Products, Research and Development, Dr. Stella Robertson, made a public presentation on the clinical results achieved with Travatan. Def. Ex. J (DX01261) (Robertson Decl., Ex. 2).

79. The first slide in Dr. Robertson's presentation read, "Dose Response Evaluation Travoprost Ophthalmic Solution Travatan." Def. Ex. J (DX01269) (Robertson Decl., Ex 2). It is important to note that an abstract of the presentation was published in the conference's "Abstract Book," which was distributed to all conference attendees. Def. Ex. I (DX 01025)(Abstract Book excerpt); Def. Ex. J. (DX01262)(Robertson Decl., ¶ 4); 12/17 Tr., at 94-95 (Harfstrand). The title of the presentation actually included the designation "Travatan Trade mark" and was listed both in the Abstract Book and in the "Final Program" that contained the schedule of events at the Congress. Def. Ex. I (DX01025) (Abstract Book excerpt); Def. Ex. O (DX02002) (Final Program excerpt); Def. Ex. J (DX01262) (Robertson Decl., ¶¶ 3-6); 12/17 Tr., at 94-97 (Harfstrand).

80. Alcon's Travatan presentation was made on June 30, 1999 at 2:00 o'clock as part of a one-and-one half hour session on glaucoma therapies, in the same room as, and along with a presentation on the use of latanoprost, the active ingredient in Xalatan. Def. Ex. OO (DX02001-02)(1999 SOE Final Program excerpts).

81. Pharmacia claims that it was unaware or not made aware of Dr. Robertson's presentation. That claim defies belief for the following reasons:

(a) Dr. Anders Harfstrand, then Pharmacia's Vice President for the Global Ophthalmology Business, the highest ranking marketing executive responsible for Xalatan, attended the 1999 SOE Congress. 12/17 Tr., at 93 (Harfstrand).

(b) Also in attendance at the Congress were at least seven Pharmacia employees or agents including Paul Chaney, Pharmacia's current Vice President for the Global Ophthalmology Business, who was then primarily responsible for marketing Xalatan in the United States; Michael Garanzini, Pharmacia's Senior Global Product Manager for Xalatan and Dr. Stephen Obstbaum, a Pharmacia declarant and outside consultant paid to attend such meetings. 12/17 Tr. 93 (Harfstrand); Harfstrand Tr., at 79-81; 99; Garanzini Tr., at 15, 137-38; Def. Ex. I (DX01024) (SOE Abstract Book Excerpt).

(c) Pharmacia engaged outside private investigators to monitor Alcon's activities. Pharmacia retained TJN ("TJN") at $15,000-$30,000 per meeting, to attend ophthalmic conferences, including the 1999 SOE Congress, and keep tabs on Pharmacia's prostaglandin analogue competitors. Nesi Tr., at 18, 32-33, 36; 12/17 Tr., at 90-92 (Harfstrand); 12/20 Tr., at 155-56 (Gurreri).

(d) At the time of the 1999 SOE Congress, Pharmacia was involved in a European patent opposition proceeding with Alcon over Xalatan (which Pharmacia lost on July 2, 1999, just two days after Dr. Robertson's presentation). Def. Ex. O (DX02006-08) (Pharmacia European Patent Register File). Dr. Harfstrand retained TJN to, among other things, gather information on Travatan in connection with that patent opposition and its eventual appeal. Harfstrand Tr., at 63-64; 12/17 Tr. 90 (Harfstrand).

(e) The "Final Program" and a blownup reproduction of the page reprinting the abstract of Dr. Robertson's presentation resided in Pharmacia's files and was produced to Alcon during discovery. Def. Ex. CCCC (PH072589) (1999 SOE Final Program); Def. Ex. O (DX02001-02) (Final Program excerpts); Def. Ex I (DX01110) (blown-up Robertson abstract); 12/17 Tr., at 126-27 (Harfstrand).

(f) Given that twenty pages of the Final Program were devoted to glaucoma, that the one page summary in the Abstract Book identified the three panels presenting papers on glaucoma, and that the two page list of symposia in the Abstract Book identified the glaucoma symposia held at the Congress, Dr. Harfstrand's explanation that he missed Dr. Robertson's presentation and the references to Travatan in the 1999 SOE Abstract Book and Final Program because they were difficult to find rings a little hollow. 12/17 Tr., at 118-19 (Harfstrand); Def. Ex. O (DX02001)(Final Program table of contents); Def. Ex. I (DX01023) (Abstract Book listing free papers by topic); Def. Ex. I (DX01022) (Abstract Book listing symposia by topic).

82. Moreover, Pharmacia also had constructive knowledge of the Travatan presentation. Dr. Harfstrand acknowledged that he engaged TJN, 12/17 Tr., at 90, until the patent litigators took over the TJN contracts. Dr. Harfstrand's investigative agent attended the Travatan presentation on his behalf. 12/17 Tr., at 90-92 (Harfstrand); Def. Ex. I (DX01112-18) (TJN Report.)

83. Dr. Robertson's presentation on the clinical results achieved with Travatan at the June 1999 SOE Congress was reported in Global Competitive Intelligences's October 1999 Competitor Outlook & Review which described Alcon's product as "the first potential competitor ... identified ... in late-stage head-to-head studies v. Xalatan." Def. Ex. I (DX01048) (Harfstrand Def. Ex. 15) (excerpts). The Review referred to Travatan by name (misspelling it "Travatan") and contained information relating to Alcon's launch plans. (DX01048-49); 12/17 Tr., at 100-04 (Harfstrand); Harfstrand Tr., at 138-42. Top Pharmacia executives received copies of the Global Competitive Intelligence Competitor Outlook & Review reports. 12/20 Tr., at 154-55 (Gurreri); 12/17 Tr. 97-98, 100-102 (Harfstrand); Gurreri Tr., at 47 Garanzini Tr., at 142-43; Battle Tr., at 71.

84. The October 1999 Competitor Outlook & Review was the first Global Competitive Intelligence document distributed after the 1999 SOE Congress.

85. Six months later, in April 2000 (one year before Pharmacia filed its ex parte application for immediate injunctive relief), Global Competitor Outlook & Review again referred to "Travatan" (spelled correctly), reporting that Alcon would be filing for approval "this quarter." Def. Ex. I (DX00846-52) (same); Gurreri Tr., at 43-45; 12/17 Tr., at 98-99 (Harfstrand).

86. Although Dr. Harfstrand testified that he had not read any of the Competitor Outlook & Review Reports in Pharmacia's files discussing the Travatan mark because they "were boring and usually lack of news in them," 12/17 Tr., at 77, preferring instead to throw them away, id. at 98, other Pharmacia executives paid more attention to them. Mr. Chaney requested to have someone "make [a] file for me of these reports" and "keep them centrally." Def. Ex. I (DX01042)(October 1999 Report with handwritten notes). Carl Battle, Esq. Pharmacia's vice president for intellectual property, testified that although he could not remember exactly how he first learned of the Travatan "in approximately May" of 2000 it was from "some document" distributed within Pharmacia. 12/21 Tr., at 7, 10, 23, 38 (Battle).

87. In any event "based on this report ... [ by] April of 2000 Pharmacia knew that Alcon had been testing their product called Travatan ..." Gurreri Tr., at 51; 12/20 Tr., at 157-61 (Gurreri).

88. In July of 2000, Dr. Harfstrand, Mr. Chaney, Mr. Gurreri and others with marketing responsibility for Xalatan participated in a "Therapeutic Area Review" for Pharmacia's ophthalmic business. To help prepare them, Global Competitive Intelligence distributed a memo to Dr. Harfstrand, Mr. Chaney and others. The memo included a chart that summarized the status of competitive ophthalmic products and which listed Travatan by name, its anticipated launch date, its perceived strengths, its perceived weaknesses, and relevant commercial considerations. While the chart does list as an important "commercial consideration" the connection of Travatan to Alcon and its sales force, it does not indicate any concern about the Travatan name. 12/17 Tr., at 104 (Harfstrand); 12/20 Tr., at 162 (Gurreri); Def. Ex. QQ (DX03453-59) (Therapeutic Area Review Memo); Id.; Def. Ex. RR (DX03859T-X) (Therapeutic Area Review Chart); Def. Ex. QQ at DX03456; Def. Ex. RR at DX03859; 12/20 Tr., at 162-66 (Gurreri).

89. Throughout the summer of 2000, Pharmacia tested the appeal to ophthalmologists of both the Travatan product and the Travatan name against Xalatan by conducting "War Games" and other market research. As a part of the preparation for these "War Games," certain Pharmacia executives were dubbed "Team Travatan" and instructed to think like Alcon to develop a message that would anticipate how Alcon would likely promote Travatan. 12/20 Tr., at 166-78(Gurreri); Guerri Tr., at 81-86; Def. Ex. W (DX0239-56) (Xalatan Games Kick-Off Meeting); 12/20 Tr., at 167-69 (Gurreri); Garanzini Tr., at 187-93.

90. At no point during this process did any Pharmacia marketing executive give any thought to the possibility that the Travatan mark was likely to cause confusion with or dilute the Xalatan mark, or that Alcon was attempting to do so by naming its product Travatan. # 12/20 Tr., at 174-75 (Gurreri); Gurreri Tr., at 109, 117, 129-30, 153; Garanzini Tr., at 145; Linvill-Neal Tr., at 122.

91. By October 2000, Pharmacia's internal marketing department repeatedly had evaluated the Travatan name. According to Pharmacia, Travatan was a "mini-van" trademark — a "weakness," and not a threat. 12/20 Tr., at 178-81 (Gurreri); Gurreri Tr., at 109-18, 166-68; Def. Ex. I (DX00881) (Global Commercial Team Workshop Notes & Presentation Slides).

92. In the various analyses by Pharmacia, none identified the Travatan trademark as an issue. Def. Ex. U (DX02271) (Xalatan U.S. Product Plan 2002 listing "vigorous defense of three main patents" as a "strategic initiative"); 12/20 Tr., at 179-80 (Gurreri).

93. While document after document from 1999 through early 2001 reflects that Pharmacia analyzed what it knew of Alcon's clinical studies, launch dates and marketing concepts for Travatan, and repeatedly tested the Travatan name and drug profile with ophthalmologists, none identify the Travatan mark as a problem. Def. Ex. I (DX01029-39); Id. (DX01111-18); Id. (DX01119-40); Id. (DX00846-52); Id. (DX011410-50); Def. Ex. (DX03859L-R); Def. Ex. QQQ(DX04719-800A); Def. Ex. QQ (DX04719-800A); Id. (DX03603-57); Id. (DX03723-76); Id. (DX03509-83).

94. Throughout almost all of this period, Pharmacia perceived Lumigan, and not Travatan, as the prime competitive threat until late January or early February 2001, when Pharmacia received its first indication that Travatan would be the greater competitive threat. CDC, a marketing research firm hired by Pharmacia, distributed a report dated early February 2001 stating that, "[i]t appears Travatan will provide the greatest threat to Xalatan." Gurreri Tr., at 106, 141-42; Def. Ex. QQQ (DX04884)(CDC Report); 12/20 Tr., at 208-12 (Gurreri); Def. Ex. I (DX00904); Def. Ex. QQQ (DX05006).

95. In connection with negotiations over patent issues Alcon told Pharmacia about its plans to market Travatan as early as June 30, 2000. Def. Ex. A (DX0002)(Arno Decl., ¶ 3); Id. (DX005-06) (Arno Decl., Ex. A); 12/21 Tr., at 10-12 (Battle).

96. In subsequent monthly correspondence and in a face-to-face meeting in December, 2000, both parties consistently used the name "Travatan." Def. Ex. A (DX0005-16) (Arno Decl., Exs. A-J) (negotiation letters); Def. E. KKK (DX04646-64) (Battle Dep. Exs.2-9) (same); 12/21 Tr., at 24, 40-42, 51-57 (Battle).

97. The correspondence and negotiations were an attempt to resolve patent issues, but even before Alcon initiated these discussions, Mr. Battle had regular discussions with Dr. Anders Harfstrand concerning the potential patent threat of Alcon's prostaglandin analogue product and the negotiations with Alcon. 12/21 Tr., at 29-31, 36-37, 40-42 (Battle).

98. Despite the fact that Mr. Battle knew of the Travatan mark no later than May 2000, Mr. Battle testified that he and Dr. Harfstrand "never" discussed the Travatan name before October 2000. The Court finds this inconceivable in view of Mr. Battle's consistent use of the Travatan name in his correspondence with Alcon. 12/21 Tr., at 48-49 (Battle); Def. Ex. A (DX0005-16) (Arno Decl.Exs. A-J); Def. Ex KKK (DX0464-64) (Battle Dep. Exs.2-9). Mr. Battle claims that he did not recognize, until Dr. Harfstrand brought it to his attention in October 2000, that Pharmacia might have trademark concerns, as well as patent concerns, about Travatan. 12/21 Tr., at 12-13, 37-38, 44-45, 48-49 (Battle).

99. Yet even after October 2000, Mr. Battle never suggested the possibility of a trademark conflict to Alcon by letter, telephone, or in their face-to-face meeting in December 2000. Battle Tr., at 58-59;12/21 Tr., at 44-45, 50, 53-54 (Battle). Nothing was said to Alcon even after Dr. Harfstrand raised trademark issues with Mr. Battle in October 2000. 12/21 Tr., at 44 (Battle); 12/20 Tr., at 177, 184.

100. At the evidentiary hearing, Dr. Harfstrand testified that he was "very upset" upon learning of the Travatan mark in connection with the October 2000 conference of the American Academy of Ophthalmology. 12/17 Tr., at 75 (Harfstrand). In an October meeting, he told Mr. Battle the Travatan trademark "was a problem" and "absolutely" infringed Pharmacia's trademark," and "diluted Pharmacia's trademark." 12/17 Tr., at 111-12 (Harfstrand). According to Dr. Harfstrand, Mr. Battle promised to "take care of it and work on it" and thereafter provided Dr. Harfstrand with status updates on an "almost weekly" basis. 12/17 Tr., at 75-76 (Harfstrand).

101. Mr. Battle has "ultimate responsibility" for Pharmacia's trademarks, 12/21 Tr., at 6 (Battle), and has been familiar with the Xalatan product since October 1999. Id. at 11. No trademark issue crossed his mind when he first saw the name Travatan in May 2000, although he specifically noted at that time that the suffix for Travatan "was similar to the A-T-A-N in Xalatan." Id. at 10, 12-13. Rather, Mr. Battle focused exclusively on a patent case as "the way to go" in dealing with the competitive threat posed by Travatan. Id. at 38, 49.

102. Mr. Battle testified that even after Dr. Harfstrand brought his concern over the Travatan trademark to Mr. Battle's attention in October 2000, he waited "a couple of weeks" to have Pharmacia's in-house trademark counsel, Ms. Cynthia Summerfield, "look at the matter for me." Id. at 33-45. Ms. Summerfield, along with the help of outside counsel, spent from October 2000 to February 2001 reviewing whether Pharmacia would have a "meritorious claim" with respect to Travatan. Id. at 52, 54, 58-59.

103. Mr. Battle also testified, in contrast to Dr. Harfstrand, that in their subsequent meetings, they "primarily" discussed "patent matters" and he only told Dr. Harfstrand that "we were looking into the trademark issue" rather than providing any specific updates. Id. at 48.

104. On February 5, 2001, after the patent negotiations reached an impasse, Pharmacia advised Alcon in writing that it had commenced a patent infringement case over the Travatan product in Delaware. Def. Ex A (DX0003) (Arno Decl ¶.10); Id. (DX00016) (Arno Decl., Ex. J). On February 7, 2001, Pharmacia first notified Alcon that the Travatan trademark was a "concern" to Pharmacia. 12/21 Tr., at 64 (Battle); Def. Ex. A (DX0004) (Arno Decl., ¶ 12).

105. It was not until March 9, 2001, or seven days before the launch of Travatan, that Pharmacia formally demanded that Alcon abandon the Travatan mark. 12/21 Tr., at 66-67 (Battle).

VIII. THE ABSENCE OF CONFUSION IN THE MARKET PLACE

106. Confusion between the Xalatan and Travatan products is unlikely, due to the different prefixes, the differences in packaging and marketing methods, as well as the products' nine-month coexistence with no actual confusion.

107. First is the fact that the dominant prefixes of Travatan and Xalatan are distinctive and very different from one another.

108. Moreover, the product packages for Xalatan and Travatan are different from one another. Alcon's Travatan packaging displays the Alcon mark and uses a distinct all-capital font, type size and purple coloring, whereas Pharmacia's Xalatan packaging reinforces the "Xal" prefix with an oversized, turquoise capital and stylized "X" juxtaposed with the lowercase and plain font "alatan." Def. Ex. E(Kureger Decl., Ex.5); Def. Ex. III (Xalatan Package); Def. Ex. JJJ (Travatan package). The Travatan box also is significantly larger that the Xalatan box and contains the unique foil-overwrap packaging used only by Alcon. Def. Ex. E(DX00336)(Krueger Decl., ¶ 28); Def. Ex. III (Xalatan Package); Def. Ex. JJJ (Travatan Package); 12/17 Tr., at 224-25 (Di Domizio); 12/18 Tr., at 178 (Lambert); Gurreri Tr., at 360-61.

109. The promotion of the products also serves to distinguish them. Detailing visits by product representatives from different manufacturers is a primary method of promoting ophthalmic prescription medicines, including Xalatan and Travatan. Garanzini Tr., at 39; Obstbaum Tr., at 25-26; Gurreri Tr., at 12-16. During such visits, Alcon representatives promote the entire line of Alcon products, including Travatan, whereas Pharmacia representatives promote only Xalatan. Def. Ex. B (DX00021) (Barton Decl., ¶ 16); Garanzini Tr., at 38-39.

110. Sales representatives often make comparisons to competing products during these detailing visits. Eisenberg Tr., at 65-66; Ex. I (DX00807-809); Id. (DX00978-92).

111. Advertising and articles in medical journals also identify the manufacturers of glaucoma products. 12/19 Tr., at 85, 88-89, 114-15 (McCullough); 12/18 Tr., at 54, 76-78 (Eisenberg); Obstbaum Tr., at 13-14, 25, 34; Garanzini Tr., at 20-21, 39, 190.

112. A published chart of sound alike names also "serves to decrease the likelihood of any confusion between Xalatan and Travatan." 12/17 Tr., at 216 (Di Domizio).

113. These various sources of information about glaucoma products would help to make ophthalmologists and optometrists aware that the different medications each company's representatives promote are manufactured by different companies. Obstbaum Tr., at 25-26.

114. There is no evidence of actual confusion in the market place. Alcon began shipping Travatan on March 16, 2001, the day it received FDA approval. Through the end of October, 2001, more than 115,000 prescriptions for Travatan had been filed. Def. Ex. TT (DX03877-82), and more than 200,000 units of Travatan had been shipped to pharmacies. Def. Ex. E (DX00331) (Krueger Decl., ¶ 11). There have been no reported instances of confusion, nor is there any evidence of medication errors involving the two drugs. Pharmacia has acknowledged that it knows of no instances of actual confusion. Def. Ex. I (DX01236) (1st Int. Ans. No 10); Gurreri Tr., at 78-80; Obstbaum Tr., at 61-65; Eisenberg Tr., at 80-81; 12/17 Tr., at 113 (Harfstrand); Harfstrand Tr., at 159-60; Garanzini Tr., at 38-41.

115. In a comprehensive survey conducted by Pharmacia from July through September 2001, ophthalmologists consistently demonstrated recognition of various characteristics of both Xalatan and Travatan. These studies show no indication that ophthalmologists surveyed believed Xalatan to be an Alcon product or Travatan to be a Pharmacia product. 12/20 Tr., at 221-24 (Gurreri); Gurreri Tr., at 225-32, 374: Def. Ex. QQQ (DX04747-800A) (Wave Awareness Usage Study) (complete): Def. Ex. I (DX00930-54) (same) (excerpts); Id. (DX00966-77) (Wave III research).

116. The United States Pharmacopeia ("USP") collects and reports instances (both "actual" and "potential") of misprescription and medication error. It has not reported any actual instances of misprescription or error between Travatan and Xalatan. Def. Ex. G (X00605-08) (McCabe Decl., Ex. 7) (USP Quality Review, May 1999); Def. Ex. K (DX01438-43) (Di Domizio Decl., Ex. 3) (USP Quality Review, March 2001); Def. Ex. C (DX00229) (DeSantis Decl., Ex. 6) (USP Medication Errors Reporting Program form); 12/17 Tr., at 210-13 (Di Domizio); 12/18 Tr. 168-71 (Lambert); Di Domizio Tr. 182-83.

117. By FDA regulation, 21 C.F.R. § 314.80(i), Pharmacia is required to maintain a database of all adverse events involving Xalatan. Gurreri Tr., at 187-88; 12/17 Tr., at 112 (Harfstrand); 12/20 Tr., at 227 (Gurreri). Pharmacia has not produced any evidence of any confusion between Xalatan and Travatan.

118. Alcon also maintains a similar database. It has received no reports of any confusion between Xalatan and Travatan. Krueger 10/30 Tr., at 52-54; Def. Ex. E (D00335) (Krueger Decl., § 26).

119. Pharmacia did not conduct a confusion survey.

120. There is no credible evidence in the record that patients are likely to confuse Xalatan and Travatan.

IX. EXPERT TESTIMONY

121. Pharmacia submitted opinion evidence to support its contention that confusion between the Travatan and Xalatan marks is likely. It relied on four expert witnesses: Dr. Stephen A. Obstbaum, Dr. Dan L. Eisenberg, Mr. George Di Domizio and Dr. Bruce L. Lambert. In addition, Pharmacia relied on a survey conducted by Walter McCullough to support its contention that the Travatan mark dilutes the Xalatan mark. The Court, for reasons that will be discussed, has accorded these opinions little weight, and in some case, none. Alcon, on the other hand, submitted a survey of ophthalmologists and optometrists conducted by Professor Shari Diamond (who also testified). Relying on a net confusion rate of 1.5%, Dr. Diamond concluded that there is "no evidence that the name of Alcon's product Travatan is likely to cause confusion with the name of Pharmacia's product Xalatan in the relevant population," i.e. ophthalmologists and optometrists. The Court finds her opinion persuasive and her testimony credible.

A. Pharmacia's Expert Testimony

(1) Dr. Obstbaum

122. In his declaration, Dr. Obstbaum opined that "the Travatan name is very similar to the Xalatan name," and that "there is a likelihood that medication substitution could occur, because of the substantial similarity" of the names. It was his opinion that "a medication error by which a pregnant woman patient received Travatan rather than Xalatan can pose a health risk to that patient." Def. Ex. K (DX01351) (Obstbaum Decl., ¶ 14); Id. (DX01352) (Obstbaum Decl ¶¶ 15, 16). Dr. Obstbaum's declaration did not disclose that he has been a Pharmacia consultant for fourteen years and is currently paid $190,000 per year for his services. Obstbaum Tr., at 43-44. In any event, for the reasons set forth in the Court's Opinion and Order dated November 28, 2001, the Court held these opinions inadmissible.

(2) Dr. Eisenberg

123. Dr. Eisenberg was of the opinion that the FDA's approval of the Travatan name was based on "reasoning" that "no longer holds." Def. Ex. K (DX01402-03) (Eisenberg Decl., ¶¶ 5-6). However, the Court in its November 28, 2001 Opinion and Order, held that this opinion regarding the FDA's reasoning was inadmissible.

124. Dr. Eisenberg also was of the view that "it is not unreasonable to anticipate a confusion between Xalatan and Travatan." Def. Ex. K (DX01402-03) (Eisenberg Decl., ¶¶ 5-6); 12/18 Tr., at 55 (Eisenberg). In his opinion, confusion would occur "when the patient transfers from one physician to another and is asked to give a list of their medications." 12/18 Tr., at 56. The only specific grounds for his opinion regarding Travatan and Xalatan are two experiences he had where prescriptions he wrote for Beltagan and Betimol were filled with Betimol and Betoptic S, respectively. Def. Ex. K (DX01403) (Eisenberg Decl., ¶ 6).

125. Dr. Eisenberg's testimony must be evaluated in the context of his business relationship with Pharmacia (not disclosed in his declaration). Since 1998, he has been hired to speak for an organization that markets Xalatan for Pharmacia, charging $1500 per afternoon or $5,000 per day. 12/18 Tr., at 55, 100 (Eisenberg). Pharmacia had also provided Dr. Eisenberg with "material support" such as Baerveldt shunt devices and a Helon GB surgical device. 12/18 Tr., at 100-01; Eisenberg 9/5 Tr., at 35-36. At the time he testified, Dr. Eisenberg also was being considered for a Pharmacia-sponsored study that would pay over $150,000. 12/18 Tr., at 101-02 (Eisenberg).

126. Dr. Eisenberg testified that he has no knowledge concerning whether the beginnings of words are more likely to cause confusion than the endings. Eisenberg Tr., at 88. Nor was his opinion supported by any survey or other empirical research. Id. at 78-79; 12/18 Tr., at 96-97.

127. Any confusion generated by patient transfers is just as likely to be off set by the differences in packaging and storage between the two products. See Def. Ex. E (DX00353) (Krueger Decl., Ex. 5); 12/17 Tr., at 223-27 (Di Domizio); 12/18 Tr., at 117 (Eisenberg).

128. Dr. Eisenberg's testimony as to confusion among pharmacists and their assistants was less than convincing: such confusion is "possible" not "likely," 12/18 Tr., at 56-57 (Eisenberg), moreover it occurs "completely apart from any confusion with regard to names." Id. at 112-113.

129. The Court finds the "not unreasonable to anticipate" confusion testimony of little value, even with Eisenberg's subsequent attempt at clarification that this phraseology is "statistical terminology" for "likely." This results from the absence of any statistical analysis from Dr. Eisenberg, any experience in trademark-related matters, or in the FDA clearance process for pharmaceutical trademarks. Eisenberg 9/5 Tr. XX-XXX-XX. Moreover, he has no knowledge of the number of pharmaceutical trademarks that end with "TAN" or "ATAN." Id. at 74.

(3) Mr. Di Domizio

130. George Di Domizio, retired Chairman of the Merck Trademark Committee, Di Domizio Tr. 8-11, offered the opinion on behalf of Pharmacia that "[t]he name Travatan is similar to Xalatan," Def. Ex. K(DX01418) (Di Domizio Decl., ¶ 14), that "[t]his name similarity ... leads me to conclude that medication errors between the two products are likely to occur," Id.; 12/17 Tr., at 184-85 (Di Domizio), that health professionals do not recognize shared terminal syllables as denoting treatment class similarity, Def. Ex. C (DX00171) (DeSantis Decl. ¶¶ 14-16), and OPDRA'S recommended approval of the Travatan name was "based on erroneous information," 12/17 Tr., at 188-89 (Di Domizio).

131. The Court did not find his testimony persuasive for the following reasons:

(a) Most of the time he testified about the "possibility" of confusion, not the "likelihood" of confusion, which is irrelevant to the governing legal standard.

(b) He claimed no knowledge of the prevalence of "TAN" or "AN" in the pharmaceutical market, even though he knew that some suffixes are very common in the industry. Di Domizio Tr., at 93-94.

(c) In his own practice, Mr. Di Domizio did not have any hesitation about approving pharmaceuticals whose trademarks used suffixes identical to those already in use on other drugs. Id. at 116-24, 208-209.

(d) His assessment of the confusing similarity of the two marks conflicts with the FDA's testing of the Travatan mark. Di Domizio Tr., at 105.

(e) Although he worked on a ten-factor model (the "1010 Model"), purportedly designed to analyze trademarks for confusing similarity, Mr. Di Domizio testified that he did not apply that model, or any other model, in reaching his opinion in this matter. 12/17 Tr., at 216-17; Di Domizio Tr., at 31.

132. Mr. Di Domizio also testified that healthcare practitioners are not trained to recognize similar suffixes in pharmaceutical brand names as indicating that the drugs share a "relevant product category." Di Domizio Reply Decl., ¶¶ 11, 14. This statement is not only inconsistent with his admission that many pharmaceutical products in the same class share common suffixes, even when those suffixes are not suggestive of the product's attributes, Di Domizio Tr., at 208, but also with other evidence demonstrating that healthcare professionals are trained to make suffix class associations. Def. Ex. C (DX00171-72) (De Santis Dec. ¶¶ 14-17).

133. Mr. Di Domizio was also of the opinion that the report of the FDA's OPDRA department, Def. Ex. E (DX00344-48)(Krueger Decl., Ex. 3), is "based on erroneous information," # 12/17 Tr., at 188-89 (Di Domizio), but other than the report itself, Mr. Di Domizio had no personal or direct knowledge of any considerations that went into the OPDRA review of the Travatan name and product. 12/17 Tr., at 191 (Di Domizio).

134. An acquaintance of Mr. Di Domizio, Dr. Neil Davis, published on the eve of the hearing in this case, a list of potential "look-alike," "sound alike" names that included Xalatan and Travatan. 12/17 Tr., at 186, 206-07; PX 287 (Di Domizio). Mr. Di Domizio's own testimony seriously devalues his opinions on this chart: he did not have "any idea" how Travatan and Xalatan got onto the list. 12/17 Tr., at 207. He acknowledged the names did not get on Dr. Davis's list through the U.S. Pharmacopeia Institute for Safe Medication Practices medication error reporting program (reporting the name), and that a single unknown individual could have Xalatan and Travatan appear on the list simply by suggesting their inclusion even if "they didn't look alike or sound alike." Id. at 214-215. He couldn't rule out the possibility that the lawsuit itself could have caused Dr. Davis or someone else to include the name on his list. Id. at 208.

(4) Dr. Lambert's Statistical Model

135. Pharmacia relies on the opinions of Dr. Bruce Lambert, a professor of pharmacy with degrees in speech communication, as the proponent of a statistical model purportedly capable of predicting likelihood of confusion between drug names by measuring the similarity in spelling and pronunciation of pairs of drug names. Ph. Supp. Br., at 29-31; Def. Ex. K (DX01536-38) (Lambert Decl., ¶ 7).

136. For purposes of automated regulatory screening, Pharmacia claims that Dr. Lambert's research regarding objective measures of similarity between drug names has been relied on by the United States Adopted Names Council and the Med-Errs Division of the Institute for Safe Medication Practices. Lambert, 12/18/01 Tr., at 121:2-22; Lambert Decl., ¶¶ 1, 2 [PX-2].^[2]

137. Dr. Lambert's research purportedly reveals that similarity in drug names is highly predictive of the probability of confusion between those drugs, and that such similarity can be used to distinguish between confusing name pairs and pairs of names that are not confusing. Lambert 12/18/01 Tr., at 122:20 to 123:6; 129:15 to 132:8.

138. Dr. Lambert has evaluated a number of methods for measuring similarities between drug names and has developed a statistical model to analyze the degree of similarity between the Xalatan and Travatan names. Lambert 12/18/01 Tr., at 123:7 to 125:8; 128:24 to 130:23; Lambert Decl., Ex. 2 [PX-2]. Although Dr. Lambert's research may have some theoretical value, his models are far too malleable for them to have any reliable predictive value in real-world situations. Accordingly, for the following reasons the Court will accord Dr. Lambert's conclusions little weight in determining whether the names Xalatan and Travatan are confusingly similar.

139. Based on his analyses employing mathematical modeling designed to predict name confusion, Dr. Lambert concluded that the Xalatan and Travatan names are substantially similar, and classified the pair as likely to be involved in medication error based upon common attributes with 1127 known error pairs. Lambert 12/18/01 Tr., at 122:11-19; 139:16 to 140:163; 141:21 to 142:6; 146:9 to 147:13; 149:3-24; 151:20 to 152:8; Lambert Decl., ¶ 7, Ex. 2, at pp. 4-5 [PX-2].

140. Dr. Lambert posits that "the similarities between Xalatan® and Travatan are substantial, particularly when compared with those names recently rejected by the FDA, suggesting a strong likelihood of confusion." In addition, but not as a function of any statistical or other empirical analysis, Dr. Lambert opines that "the products' non-name attributes are nearly identical, again suggesting a strong likelihood of confusion." Def. Ex. K (DX01537-38) (Lambert Decl., ¶¶ 7-8). These opinions are entitled to little weight for the following reasons.

141. Dr. Lambert's initial declaration did not disclose that in a pre-litigation publication, he described his method as having "poor positive predictive value." Def. Ex. G (DX00577) (Bruce L. Lambert, et al., "Similarity as a Risk Factor in Drug-Name Confusion Errors," 37 Medical Care 1214 at 1223 (1999)). In that publication, Dr. Lambert described the very same methods and numerical assumptions he employed in reaching his initial opinion in this case as having error rates that are "too high" to be relied on "as the sole basis for regulatory decisions." Id.; 12/18 Tr., at 193-194 (Lambert); Lambert Tr., at 225. He also acknowledged at the hearing that whatever results his statistical method produced, ultimately those results should be submitted to a "panel of human experts" to make actual determinations as to the likelihood of medication errors based on "other factors." 12/18 Tr., at 160-61, 193-95 (Lambert).

142. The first step in Dr. Lambert's methodology was to calculate spelling and pronunciation similarity scores between Travatan® and Xalatan® using certain simple mathematical formulas. Def. Ex. K (DX01537-38) (Lambert Decl., ¶¶ 7-8).

143. Although Dr. Lambert's methodology in computing similarity is not in question, it was made apparent to the Court that his computations can lead to spurious results. The limited real world value of this analysis was demonstrated at the hearing. Analyzing the words "Mexican" and "American," Dr. Lambert reached scores either higher than (supposedly more confusing) or identical to that of Travatan® and Xalatan® (and, therefore, equally or more confusing). 12/18 Tr., at 179-83 (Lambert); Def. Ex. PPP (DX04717A) (Mexican/ American Trigram-2b and Normalized Edit Distance demonstrative chart). The scores obtained in this basic exercise illustrated the abstract nature of his model and why it sheds no light on the central legal issue: the extent to which two words are likely to be confused in the real world.

144. The second step in Dr. Lambert's methodology was to compare the similarity scores for Travatan and Xalatan to similarity scores derived from two groups of branded drug name pairs. The first group consisted of "Known Error Pairs," or "pairs of names that have appeared in published lists of confusing drug names." Def. Ex. K (DX01556) (Lambert Decl., Ex. 2 at 4) (hereinafter "Lambert Report"); Id. (DX01537-38) (Lambert Decl., ¶¶ 7-8). According to Dr. Lambert's initial declaration, there are 1,127 "Known Error Pairs." Id. (DX01556) (Lambert Report at 4); see also id. (DX01437-43) (Di Domizio Decl., Ex. 3) (U.S. Pharmacopeia list of branded drug name pairs).

145. The second group was an artificially created "control" group of 1,127 "Non-Error Pairs," or drug name pairs that have not appeared in published lists of confusing drug names.^[3] Adding the "Known Error Pairs" to the "control" group of "Non-Error Pairs" results in a total universe of 2,254 drug name pairs to which Dr. Lambert applied his methodology. In his initial report, however, Dr. Lambert identified 235,152,141 names pairs as the universe of actual possible branded drug name pairs in the real world. Id. (DX01556) (Lambert Report at 4).

146. Dr. Lambert claimed in his initial declaration that, in his controlled universe of 2,254 names, he was able to use his method to distinguish the Known Error Pairs from the Non-Error Pairs with "93% accuracy." Id. (DX01559) (Lambert Report at 7). Compare id. (DX01576) (Bruce L. Lambert et al., "Similarity as a Risk Factor in Drug-Name Confusion Errors," 37 Medical Care 1214 at 1219 tbl. 2, 1223 (1999)) (Dr. Lambert admits that his model has a 6.3% rate of generating false positives).

147. Dr. Lambert, however, did not compute how well his method would work in the real world universe he identified as consisting of over 235,152,141 branded drug names. As was made clear by the testimony of Professor George P. McCabe,^[4] in that setting, Dr. Lambert's method generates an error rate of over 99%. This is because his method's "positive predictive value" — the likelihood that a name pair predicted to be an error pair actually is an error pair — is dictated by its "false positive rate" (here 6.3%) and by the actual number of error pairs relative to non-error pairs in the total population of name pairs (also known as "prevalence"). Because there are so many more non-confusing name pairs (235,152,141-1,127) than confusing ones (1,127), when the model's conceded 6.3% false positive rate is applied to the 235,151,014 Non-Error Pairs, it incorrectly identifies over 14.8 million pairs (6.3% of 235,151,014) as being confusing. Because only 1,127 "error pairs" actually have been reported, that means the Lambert method is incorrect more than 99% of the time.^[5] 12/19 Tr. 218-19 (McCabe); see also 12/18 Tr., at 165, 196 (Dr. Lambert concedes that the 99% number was calculated using the "right equation"). Stated differently, out of the 14.8 million names pairs that the Lambert method identifies as "Error Pairs," only about 1,127 actually were "Error Pairs." The vast majority — 14,798,873 name pairs — were mistakenly identified as "Error Pairs" by Dr. Lambert's model.

148. Simply put, as applied to the current issue, it is much more likely (over 99% likely) that Dr. Lambert's model lumped Travatan® and Xalatan® together with the over 14,798,873 other false positives, rather than having accurately identified it as one of the roughly one thousand true "Error Pairs." See 12/19 Tr. 210-18 (McCabe); Def. Ex. G (DX00524-25) (McCabe Decl., ¶ 10); Def. Ex. SSS (DX05025) (Positive Predictive Value Assuming 1,127 "Error Pairs" in a Universe of 235 Million Total Name Pairs Demonstrative); 12/20 Tr. 45-47 (McCabe).

149. When confronted with this analysis of his numbers and his work, Dr. Lambert repudiated and changed the numbers he had provided in his pre-litigation publications and initial declaration. See 12/18 Tr., at 167 (conceding that the 1,127 figure for known error pairs comes "from one of my earlier papers"), 196-97 (same); 208 (numbers used in court not submitted as part of Dr. Lambert's reply declaration).

150. At the hearing, Dr. Lambert refused to commit to what number should be used in place of the number 1,127 for known or actual error pairs, 12/18 Tr., at 168, but at varying times he suggested numbers ranging from 1,250, to 794,000, to seven or eight million, all the way up to 300 million error pairs. Lambert Tr., at 311-13, 394-95, 398; 12/18 Tr., at 174, 196, 200-07 (Lambert).^[6]

151. Dr. Lambert's inability to settle on a particular number (or even range) not only creates great uncertainty, but undermines the statistical foundation and reliability of his model. Dr. Lambert's model was built using 1,127 non-randomly selected error pairs. The sensitivity (false negative rate) and specificity (false positive rate) measures he calculated are derived from the characteristics of these 1,127 "Known Error Pairs." Because these 1,127 Known Error Pairs are not a random sample, it is scientifically improper simply to assume that the characteristics of those 1,127 (which form the basis for Dr. Lambert's measures of specificity and sensitivity) necessarily apply to the much larger universe of names. 12/20 Tr., at 44-47, 52-54 (McCabe). That remains true whether that universe consists of seven to eight million name pairs or the 300 million name pairs that Dr. Lambert initially adopted during his deposition, but later recanted. In other words, the reliability of Dr. Lambert's entire model is thrown open to question if one assumes there are so many more error pairs than the 1,127 Known Error Pairs that form the basis of his statistical construct. Id.

152. Even giving this model every benefit of the doubt (and assuming the sensitivity and specificity of Dr. Lambert's model remains constant notwithstanding a huge increase in error pair assumptions) it still has too high an error rate. For example, if one arbitrarily increases the number of "Known Error Pairs" to 10,000 (instead of 1,127), and dramatically limits the universe of drug name pairs by assuming there are only two million pairs of drug names (instead of 235 million), Dr. Lambert's model would still be wrong more than 90% of the time, as Professor McCabe demonstrated during his deposition. McCabe Tr. 70-74; see also 12/19 Tr., at 225-29 (McCabe).

153. Dr. Lambert also abandoned his earlier use of 235 million as the number of total name pairs of drug trademarks. 12/18 Tr., at 196-99. Rather, he suggests a possible replacement figure of 87 million, Lambert Tr., at 393-94, but could not explain why that number is more accurate other than to say that the 235 million figure used in his peer-reviewed publication is "inflated." 12/18 Tr., at 171-73.

154. Because Dr. Lambert was readily willing to change the numbers that he used to support his model, and because his model's predictive value turns entirely on initial assumptions about how many pairs should be included in the model, Dr. Lambert's entire methodology is suspect. In fact, it is no longer the same model that has been evaluated and peer-reviewed. Compare, 12/18 Tr., at 125, 159, 219, 220 (pre-litigation model evaluated in peer-reviewed journals), with id. at 230 (model used to support Dr. Lambert's last declaration has not been published or peer-reviewed).

155. Alternatively, Dr. Lambert offered an opinion, not based on specific numbers, that assumed that up to one in ten pairs of names is an error pair (that is, that "prevalence" is somewhere between 0 and 10%). He reached that assumption based on one study that concluded 8% of prescriptions led to some type of a medication error (including overdoses, missed doses, pill mix-ups, etc.). From that 8% figure, he extrapolated and concluded that one out of about every ten name pairs is an error pair. 12/18 Tr., at 175, 177, 200-04, 206-08 (Lambert). On its face, this extrapolation compares apples to oranges and otherwise has no scientific basis. 12/20 Tr., at 39-40, 44-45, 50-51, 70-71 (McCabe).

156. Even assuming 10% of name pairs are error pairs, however, the positive predictive value of Dr. Lambert's model is only 62%. 12/19 Tr., at 223 (McCabe); 12/20 Tr., at 68-69 (McCabe). This means there is a 38%, or a more than one in three chance, that his model has incorrectly identified Travatan and Xalatan as an error pair, even making all possible assumptions in Dr. Lambert's favor.^[7] 12/19 Tr., at 223 (McCabe).

157. The 10% figure, however, is not credible. By manipulating the universe of drug name pairs and the universe of error pairs, Dr. Lambert can theoretically achieve positive predictive values that are sufficient to make his model appear valid. Ultimately, however, Dr. Lambert's model lacks predictive value because Dr. Lambert is unable to substantiate any of the numbers that he has used in the model. Even more troubling, it appears to the Court that in order for Dr. Lambert's model to reach sufficiently high predictive values, for the following reasons it must make assumptions about the number of relevant pairs that are sufficiently unbelievable so as to lack any credibility.

158. The lack of real world reliability of Dr. Lambert's model was illustrated by the Court's question C1 to both parties' experts. The Court noted first that, if there were 8,000,000 error pairs (the number Dr. Lambert posited in order to reach a prevalence of 10%, given his total universe of drug name pairs), mathematically, the minimum number of drug names required to produce 8 million error pairs must be 4,000.^[8] 12/20 Tr., at 72-73 (McCabe). This minimum number, moreover, can be achieved only if one assumes that each one of the 4,000 drug names is confusingly similar to every other one of the other drug names in the group — i.e., each of the 3,999 other names. Id. Further, the total number of drug names, assuming 87 million pairs of names, is around 13,000. 12/20 Tr., at 77 (Lambert). This means that, in order to render Dr. Lambert's model capable of making a prediction with even 62% accuracy, his model condemns at a minimum almost one-third of all drug names as confusingly similar to another drug name. Id. In the Court's estimation, absent evidence to the contrary, such an assumption about the realworld state of drug names is simply not credible.

159. Dr. Lambert responded to the issues raised by the Court by testifying that his model nonetheless identifies Travatan as falling into the one-third of all drug names that he believes to be similar to another drug name, conceding that the only way to tell which of these one-third of names is actually likely to cause errors is to analyze other factors for which his model does not account. 12/20 Tr., at 75-77 (Lambert). Not only does this expose the inability of Dr. Lambert's statistical model (and methodology) to assist traditional trademark infringement analysis, it vastly understates the problem. Four thousand is the bare minimum number of names mathematically required to produce 8,000,000 unique pairs of names. For that number to be the actual number, each one of the 4,000 names would have to be confusingly similar to every one of the other 3,999 names. As Professor McCabe explained, it is extremely unlikely — in the real world — that there would be 4,000 names each confusingly similar to every one of the 3,999 other names in that group. Id. at 73 (McCabe) ("that's not real world"). In reality, because almost all of the names will be confusingly similar to only one or at most several other names, it would take significantly more than 4,000 names to produce 8,000,000 error pairs. Id. at 72-73. This means that the "fix" Dr. Lambert proposes for his model is even more at odds with reality than the Court's conservative 4,000 name estimate suggests.

160. Furthermore, the "Known Error Pairs" Dr. Lambert has identified from "published lists" include not only pairs of names that have resulted in error, but also an undetermined number of pairs of names that various authors subjectively determined looked or sounded alike without actually having resulted in any errors. Lambert Tr., at 203, 205-06. Professor Lambert could not tell from the lists he examined how many of the 1,127 "Known Error Pairs" actually had been confused or resulted in errors, and how many simply reflect someone's subjective view that two names might be possibly confused. Id. As a result, that a pair of names fits Professor Lambert's profile as a "Known Error Pair" is even less useful as a guide to the actual chance of error: What Dr. Lambert has constructed is a test that arguably might indicate the existence of some spelling similarity between two names, but his methodology is not a reliable predictor of actual medical substitution, let alone trademark confusion.

161. Dr. Lambert's other opinion, that there is a 99% chance that the Travatan name was not selected by chance, 12/18 Tr. 184, is entitled to no weight at all. It is not informed by any analysis of or investigation into the facts and circumstances under which the Travatan name was selected, is not the subject of any peer review or analysis of its actual and potential error rate, cannot be independently verified, and is not probative of the ultimate issue that it seems to try to prove — that the Travatan name was intentionally picked to cause confusion as to the source of the Xalatan® product. Id. at 184-86.

(5) The McCullough Survey

162. With respect to the application for a preliminary injunction on its claim that the Travatan mark dilutes the Xalatan mark, Pharmacia relies on the survey conducted by Walter McCullough. The survey did not test for trademark confusion. McCullough Tr., at 28. According to Pharmacia, McCullough's survey demonstrated a 14% dilution of the Xalatan mark. Because the following mechanical and conceptual problems with the survey limit its value, however, the Court finds that the true level of dilution of the Xalatan mark was even less than the 14% proposed by Pharmacia.

163. McCullough's survey contained a test cell and a control cell. In the test cell, 199 ophthalmologists were shown a box of Travatan and asked what other "products," if any, were brought to mind. Def. Ex. K (DX01593-94) (McCullough Report at 5, 6). In the control cell, 112 ophthalmologists were shown a box of Lumigan, a competing drug marketed by Allergan, Inc., and asked the same questions. Id. (DX01594) (McCullough Report at 6).^[9]

164. If a test cell respondent answered Xalatan, they were asked four times to specify "why" Xalatan came to mind. Id. (DX01593) (McCullough Report at 5).

165. Any answers that indicated "name similarity" in the test cell — including several that appeared to be referring to similarity of the generic names — were coded as dilution responses. Id. at 6.

166. After netting out the dilution or "noise" results of the control cell, Mr. McCullough opined that 15% of respondents thought Travatan brought Xalatan to mind because of similarities in the trade names, Def. Ex. K (DX01594) (McCullough Report at 6); 12/19 Tr., at 29-30 (McCullough); McCullough Tr., at 134.

167. At the hearing in this case, Mr. McCullough admitted to having miscoded at least one response. 12/19 Tr., at 7-8, 26-27; McCullough Tr., at 292-93. Simply by omitting that one, concededly wrongly coded response, the dilution percentage dropped to, at best, 14%. 12/19 Tr., at 7-8 (McCullough); see also Def. Ex. K (DX01595-1639) (McCullough Report at Ex. A) (36 minus 1 "dilution" responses of 199 in test cell (17.6%) and 4 of 112 "dilution" responses (3.6%) in control cell).

168. Even the 14% figure advanced by Pharmacia is suspect, as the survey conducted by Mr. McCullough contained numerous flaws and limitations that affected the reliability of his results. First, Mr. McCullough did not survey a sample of the entire universe of consumers of glaucoma medications; he surveyed ophthalmologists only, despite that optometrists in all but three states nationwide prescribe glaucoma medications. McCullough Tr., at 23-26. In its internal marketing studies, on the other hand, Pharmacia surveys optometrists, Def. Ex. U (DX02213) ("Xalatan Hot Issues"); Def. Ex. GG (DX02586) ("Xalatan Product Information" excerpt), and Pharmacia budgeted over $500,000 in the year 2001 alone for marketing directly to optometrists. Def. Ex. U (DX02289) (Xalatan U.S. Product Plan 2001).

169. Second, having already limited the value of the survey by sampling only the universe of ophthalmologists, Mr. McCullough further undercut the value of his survey by failing to take any steps to ensure that the sample drawn from this universe was even representative of the marketplace of ophthalmologists. He did nothing to screen for respondents' years of experience or the size of their practices, McCullough Tr., at 87, a flaw Pharmacia itself characterized as the antithesis of good market research practices. Gurreri Tr., at 219.

170. Even assuming away these problems, there are other mechanical flaws in the survey that cause the purported 14% dilution level to be exaggerated.

171. The hearing highlighted two specific mechanical problems with Mr. McCullough's survey. First, Mr. McCullough did not apply the same standard in coding the test and control cells. In the control cell, Mr. McCullough did not code the following as a dilution answer: "because it's in a similar class of product, it has similar endings." Def. Ex. JJ (DX03007-13) ("Doctor 393"); 12/19 Tr., at 67. In the test cell, however, at least three substantively identical answers were coded as dilution. 12/19 Tr., at 66-69, 71, 74, 76 (McCullough). Compare Def. Ex. JJ (DX 03023-29) ("Doctor 193") ("Because first, it is in the same class of drugs. The name itself is reminiscent."); 12/19 Tr., at 69-72; Def. Ex. JJ (DX02893-99) ("Doctor 282") ("The endings [sic] the same; same medication; same type drug" and "same type medication in same family"); 12/19 Tr., at 72-75; Def. Ex. JJ (DX02933-39) ("Doctor 348") ("Similar product" and "Similar name. Does the same thing. Both medications work almost the same."); 12/19 Tr., at 75-76. This inconsistent coding has the effect of exaggerating the net level of reported dilution. Mr. McCullough admitted that if he "changed the coding of these responses ... [he] could do something which would drive the dilution level down." 12/19 Tr., at 75. In fact, when the three test cell responses are recoded as nondilution answers so that they are consistent with Mr. McCullough's coding of the nearly identical control cell response, the net dilution level falls to 12.5%. See Appendix A (chart setting forth dilution shown by the McCullough Survey after accounting for flawed responses).

172. The second problem that undermines the reliability of Mr. McCullough's 14% figure is that it depends upon the inclusion of a set of responses obtained by one interviewer (Susan Boyd) in Los Angeles, California, under suspect circumstances. These responses stand out from those obtained by the rest of the 55 interviewers nationwide in the following ways:

(a) Ms. Boyd was one of only two interviewers nationwide to learn of the lawsuit during her work. McCullough Tr., at 223-24.

(b) She was her own supervisor in the field. 12/19 Tr., at 50 (McCullough); McCullough Tr., at 247.

(c) She did more interviews on her own than Mr. McCullough "would have liked her to do," 12/19 Tr., at 51 (McCullough), and more than any other interviewer nationwide. Id. at 53-54.

(d) She alone was responsible for 9, or 25%, of the answers coded as dilution. Id. at 54; McCullough Tr., at 239.

(e) Half of the test cell interviews she conducted were coded as dilution responses, a "success" rate greater than any other interviewer nationwide. 12/19 Tr., at 54 (McCullough); McCullough Tr., at 240.

(f) She obtained 7 out of the 10 most "unusually detailed" answers recorded nationally in the test and control cells — answers in which a respondent purportedly said that either "TAN," "ATAN" or the "last 3 letters" were the reason for the association. McCullough Tr., at 329-31; Def. Ex. JJ (DX02900-07, DX02998-3005) (McCullough Dep. Ex. 29, 42); 12/19 Tr., at 55-61, 66 (McCullough); McCullough Tr., at 263-65.

(g) Two of Ms. Boyd's other interviews included the unusually specific (and curious) phrase "puts me in mind of," a phrase that does not appear in any other questionnaire. 12/19 Tr., at 62-65 (McCullough).

173. Mr. McCullough has testified that it would be appropriate to "toss" interviews if they are suspicious. Id. at 48-49; McCullough Tr., at 168-69; see also Diamond Tr., at 121 (if an interviewer becomes aware of the purpose of the survey, it is necessary to evaluate whether that person "was more or less likely to produce confusion responses"). Eliminating the Boyd interviews as anomalous or, at best, unexplained, drops the net dilution rate from 14% to 11.5%. See Appendix A to Alcon's Post-Hearing Proposed Findings of Fact and Conclusions of Law. When combined with the recoding of the three miscoded test cell responses, the dilution level is reduced even further, to 10.4%. Id.

174. Pharmacia has defended Ms. Boyd's work by contending that the Los Angeles market itself is anomalous. See, e.g., McCullough Tr. 275-78; 12/19 Tr. 19, 61, 66; 12/20 Tr. 150-51 (Diamond). Other than that, it offers no explanation for why this market produced such an unusually large percentage — nearly 25% — of all dilution answers. 12/19 Tr. 57, 61, 65-66 (McCullough); McCullough Tr. 341-42. If the Los Angeles interviews are eliminated because of this unexplained anomaly, then, with the three recoded test cell responses discussed above, see supra Findings ¶ 219, the dilution level drops to 9.8%. See Appendix A to Alcon's Post-Hearing Proposed Findings of Fact and Conclusions of Law.

175. In light of this finding that the dilution level is 9.8% the Court declines to further evaluate the McCullough survey to determine whether or not it is so conceptually flawed that it would justify Alcon's position that the cumulative net dilution should be further reduced to 5.9%. Alcon's objection to the questions in McCullough's survey asking what other "products" rather than what names or brands come to mind appear to be the kinds of questions appropriate in survey evidence to show actual dilution. See Patrick M. Bible, Defining and Quantifying Dilution Under The Federal Trademark Dilution Act of 1995: Using Survey Evidence To Show Actual Dilution, 70 U. Col. L.Rev. 295 (1999).

(6) The Diamond Survey

176. Alcon produced evidence in the form of a survey of the relevant population, that the marketplace is not confused. A well-designed survey of ophthalmologists and optometrists conducted by Prof. Shari Seidman Diamond shows a net confusion rate of 1.5%. Def. Ex. D (DX00239-40, DX00246-47, DX00250) (Diamond Decl., ¶¶ 6, 26, 30) (there is "no evidence that the name of Alcon's product Travatan is likely to cause confusion with the name of Pharmacia's product Xalatan in the relevant population," i.e., ophthalmologists and optometrists) Dr. Diamond also testified at the evidentiary hearing. The Court found her to be a very credible witness.

177. Professor Diamond is a professor of law and psychology at Northwestern University School of Law and a senior research fellow at the American Bar Foundation. She is the author of the Reference Guide to Survey Research contained in the Federal Judicial Center's Reference Guide on Scientific Evidence (1994, 2000), as well as a similar chapter in West Publishing's four volume treatise, Modern Scientific Evidence (1997, 2000), and several other well-respected articles in the field. 12/20 Tr., at 83-84 (Diamond); Def. Ex. D (DX00238-39) (Diamond Decl., ¶¶ 1-3); id. (DX00251-64) (Shari S. Diamond)

178. Professor Diamond chose to survey a nationwide population of ophthalmologists and optometrists because they are the market gatekeepers with respect to the prescription glaucoma medications at issue in this case. 12/20 Tr., at 85-86, 114-15 (Diamond). Ophthalmologists and optometrists thus make the ultimate determination as to which medications pharmacists will dispense and end-users —patients—will receive. Id.; see also McCullough Tr., at 25-26. Pharmacia's survey expert, Mr. Walter McCullough, also chose to survey ophthalmologists, McCullough Tr., at 23-26, and Pharmacia routinely surveys only doctors, not pharmacists or patients, in its internal marketing research. Gurreri Tr., at 70, 72, 74-75, 120-21, 201, 218-19, 227-28, 235, 241-42, 277-78, 331, 364.

179. In order to ensure that the population surveyed was representative of glaucoma prescribers, Professor Diamond considered several lists of doctors from which to draw interviewees, including the membership list for the American Ophthalmology Association. Professor Diamond ultimately used a list assembled by the National Data Center because it is widely used for research purposes within the pharmaceutical industry and because it provided enough information about potential respondents' practice size and years of experience to allow for a stratified random sample. 12/20 Tr., at 86-87 (Diamond).

180. Because doctors are a particularly busy population of potential respondents, id. at 86, Professor Diamond arranged for each one to receive an advance letter, as well as follow-up phone calls, to schedule interviews at a convenient time. Id. at 88.

181. The Diamond Survey tested for product source confusion. It asked a series of questions based on a stimulus that was mailed to respondents in a sealed envelope. Def. Ex. D (DX00243) (Diamond Decl., ¶ 16); 12/20 Tr., at 84, 88-89 (Diamond). There were two cells in the survey: (1) a test cell of 200 prescribers that used a Travatan advertisement as the stimulus; and (2) a control cell, also of 200 prescribers, that used an identical Lumigan® advertisement as the stimulus. Def. Ex. D (DX00243) (Diamond Decl., ¶ 16); 12/20 Tr., at 89 (Diamond). The stimulus was designed to focus the respondent's attention primarily on the name of the product. 12/20 Tr., at 89-91 (Diamond).

182. The questions that Professor Diamond selected for the survey ("What company do you believe puts out the product whose advertisement you just saw?;" "Please tell me whether or not you believe the company whose advertisement you just saw puts out any other medications for the eye;" and "Please tell me whether or not you believe that the company whose advertisement you just saw needs authorization, permission or approval from some other company in order to put out the product advertised."), id. at 100-02 (Diamond); Def. Ex D (DX00278-85) (sample questionnaire), are a standard type and format of questions used to gauge confusion in trademark cases. 12/19 Tr., at 91-94 (McCullough). Pharmacia's survey expert, Walter McCullough, used virtually identical questions in a trademark confusion survey he designed in a prior infringement case. Def. Ex. LL (DX 03216-85) (Declaration of Walter McCullough in Westchester Media Co. v. PRL USA Holdings, Inc.); see also 12/19 Tr., at 96-97 (McCullough).

183. Five percent of the test cell respondents confused the source of Xalatan® and Travatan and 3.5% of control cell respondents confused the source of Xalatan® and Lumigan®, yielding net confusion of 1.5%. 12/20 Tr., at 108-09 (Diamond); Def. Ex. D (DX00239-40) (Diamond Decl., ¶ 6).

184. Professor Diamond selected the survey firm D.S. Howard & Associates ("D.S. Howard") to conduct the survey because it specializes in medical surveys and medical marketing research, fields surveys for pharmaceutical companies for approximately 60% of its work and is staffed by surveyors with an academic background and a reputation for attention to detail. 12/20 Tr., at 91-92 (Diamond).

185. D.S. Howard took numerous measures to ensure that respondents did not view the stimulus prior to the interview. First, during the screening interview, respondents were explicitly instructed not to open the survey materials until they were contacted for the interview. Corrigan Tr., at 111; 12/20 Tr., at 89 (Diamond). Second, the materials were sent to respondents in a sealed envelope, with a sticker reminding respondents not to open the envelope before the interview. Id.; Diamond Tr., at 43, 46. Finally, interviewers were specifically instructed to report any indication that respondents had previously opened the envelope to Karen Corrigan, who carefully supervised the interview process. Ms. Corrigan credibly testified that she specifically received numerous reports from the interviewers that they could hear the respondents opening the envelopes. She also listened to tapes of the interviews (as did the court), which confirmed that respondents did not view the stimulus prior to the interview. Corrigan Tr., at 112-13; Diamond Tr., at 43-44.^[10] Moreover, none of the doctors volunteered any statement that would even lead to suspicion that they had opened the stimulus ahead of time. 12/20 Tr., at 110 (Diamond).

186. Interviewers were instructed to record absolute and accurate verbatim responses and capture every word and sound uttered by respondents, including the "humms" and "ahhs," even if doing so meant simultaneously transcribing the responses in less than perfect penmanship. 12/20 Tr., at 93-94 (Diamond); Diamond Tr., at 94; Corrigan Tr., at 92. In order to record some of the lengthier responses, interviewers quickly recorded verbatim all of the statements of doctors in ways that ensured they did not miss anything. Immediately following the interview, they transcribed those responses, when necessary, in a more legible fashion on a replacement or additional page. 12/20 Tr., at 96, 124-25 (Diamond); Diamond Tr., at 94; Corrigan Tr. 95-96.

187. A review of all the questionnaires, as well as monitoring reports and tapes of the interviews, confirms that the responses were recorded accurately and completely. 12/20 Tr. 96-97 (Diamond); Corrigan Tr. 99; Def. Ex. VV (DX03888-4110) (transcripts of taped interviews and accompanying recordings); Def. Ex. AAAA (SD0001-4601) (complete set of Diamond Survey verbatims). There are no paraphrases, gaps or any other indications that they do not represent what the respondents actually said. Corrigan Tr., at 90, 99-100. The raw data collected for this survey, in the form of the verbatim responses of doctors, was preserved and was made available to Pharmacia and the Court for review. Diamond Tr., at 17, 20, 97; Corrigan Tr., at 11, 98.

188. After deposing Prof. Diamond, Pharmacia also deposed Ms. Corrigan. She testified at length regarding the steps she personally took to ensure that Prof. Diamond's admonition, to capture all comments made by the respondents to the substantive questions in the survey, was followed. For example, Ms. Corrigan prepared a separate interviewer training manual for the Diamond Survey, which included the explicit instruction that interviewers record each doctor's comments verbatim. Corrigan Tr., at 22-24. Ms. Corrigan personally trained the interviewers, during which time she emphasized to the interviewers that their "primary task" was to record every verbatim comment made by the doctors accurately and completely. She also conducted mock interviews in which she herself acted as respondent. Id. at 79, 92, 95. Finally, Ms. Corrigan carefully reviewed each questionnaire, as well as the monitoring reports and tapes. She concluded that the verbatim recording completed for this survey was "exceptionally accurate." Id. at 100.

189. Pharmacia has accused Professor Diamond and Ms. Corrigan of failing to control for the respondents' exposure to the stimuli and destroying evidence. McCullough Reply Decl., ¶ 9; 12/19 Tr., at 77-78 (McCullough); 12/20 Tr., at 110, 121 (Diamond). The Court finds the accusation to be without merit.

190. Walter McCullough, who leveled the charges on behalf of Pharmacia, testified that one of the reasons that the Diamond survey was entitled to no weight was the lack of stimuli control. 12/19 Tr., at 78 (McCullough). By that he meant that respondents in the Diamond survey had opened up the envelope they were sent that contained the test stimulus ahead of time. Id. He had no evidence that happened but assumed it had "based on good research practice I think." Id. The court finds from its listening of the tapes that the respondents did not view the stimulus prior to the interview. See Dx. Ex. VV; W.

191. McCullough admitted that he had no factual basis to support his "assumption" that respondents "could have" opened the stimulus envelopes ahead of time, 12/19 Tr., at 78 (McCullough), never listened to any of the taped interviews to verify their accuracy, id. at 78-79, reviewed only 10-20 (out of 400) verbatims, id. at 80, and did not see any statements indicating any concern that the envelopes were opened. Id. at 79.

192. Similarly, Mr. McCullough has no evidence that any of the verbatim transcriptions are anything but the complete contemporaneous recordings of Professor Diamond's interviews. Id. at 80, 86-87. To the contrary, the evidence on record clearly demonstrates that Professor Diamond and Ms. Corrigan made concerted efforts, above and beyond that found in many other surveys, to ensure that the verbatim recording of each response was absolutely complete and accurate. 12/20 Tr., at 96-97 (Diamond); see also 12/19 Tr., at 94-95 (Mr. McCullough admitting that his choice of words in criticizing Professor Diamond may have been "too strong.")

193. Indeed, the ophthalmologists and optometrists surveyed proved to be so verbose that without the precautions taken by Ms. Corrigan and Professor Diamond, they would likely have missed parts of the interviewees' substantive responses. 12/10 Tr., at 96, 103, 105-106, 135-137, 147 (Diamond); Corrigan Tr., at 98-99.

194. The extensive verbatim responses demonstrate that ophthalmologists and optometrists are highly sophisticated and can readily distinguish Travatan and Xalatan®. Def. Ex. II (DX02602) (Diamond Survey test cell response 8746) (identifying Alcon as company that puts out Travatan by "process of elimination" because "Pharmacia makes Xalatan."); id. (DX02609) (test cell response 8736) (same); id. (DX02616) (test cell response 22903) ("I know, I prescribe [Travatan®]. I know about it. I use it in place of Xalatan."); id. (DX02626) (control cell response 171) ("[P]rostaglandins have been popular like with Travatan and Xalatan lately. Those are two prostaglandins."); id. (DX02630) (control cell response 402) ("Xalatan is made by Pharmacia ... there are two on the market, another one on the market, Travatan."); id. (DX02637) (control cell response 8669) ("I know Alcon is Travatan and Xalatan's a competitor, it's Pharmacia.").

195. Mr. McCullough's other expressed concern about Professor Diamond's survey — that it is likely that doctors were exposed to information (such as an Internet website) during the interview allowing them to associate Travatan with Alcon, 12/19 Tr., at 85, 89 (McCullough) — has no basis in fact. Id. at 89-90. Moreover, the Diamond Survey verbatims often make it clear that the doctors were speaking from memory. 12/20 Tr., at 109-110 (Diamond). Many doctors made statements to the interviewers that confirm they were opening the materials for the first time during the interview. Corrigan Tr., at 112-13; Diamond Tr., at 43-44; see also supra, Findings ¶ 146 n. 21. In the one instance, where a doctor apparently was looking at a pre-printed prescription pad, Professor Diamond removed the verbatim from consideration. 12/20 Tr., at 110 (Diamond).

196. Because physicians and pharmacists are a sophisticated population in their area of expertise the differences between the Xalatan and Travatan, their presentation and the products themselves are more significant than any partial similarities.

X. THE RELATIVE RISKS OF TRAVATAN AND XALATAN

197. There is no evidence in this record to conclude that either Xalatan or Travatan is more or less of a risk to pregnancy. Pharmacia asks the court to find that there are potential risks to pregnant patients and pregnant caretakers who use or handle Travatan, believing it to be Xalatan, relying on Eisenberg Reply Decl., at § 21 (PX-11). Pharmacia argues in support of its argument that Travatan has risks that Xalatan doesn't by noting that the FDA approved label (package insert) for Travatan solution displays a significant precaution that women who are pregnant or trying to become pregnant should not use the solution, and should even "avoid direct exposure to the contents of the bottle." See Harfstrand 12/17/01 Ex.(PX-1). However, Pharmacia notes, that no such precaution is included on the potential risks to pregnant patients and pregnant caretakers who use or handle Travatan believing it to be Xalatan. Eisenberg Reply Decl., at § 21 (PX-11). In addition Pharmacia points to the package insert for Travatan which indicates significantly higher hyperemia (red eye) rates than Xalatan-a problem that can lead to noncompliance and potential detrimental effects. The court rejects Pharmacia's proposed finding because:

(a) The testimony of record is that Alcon voluntarily included a clearer pregnancy warning on its package insert, not that it was required to do so.

(b) Pharmacia uses similar language in other countries.

(c) As opposed to Pharmacia's experts' speculation, Pharmacia has no empirical evidence or other data to suggest the products pose different risks to pregnancy.

(d) Both drugs are FDA Pregnancy Category C and have been approved as safe and effective by the FDA.

XI. COMPARATIVE EFFICACY OF XALATAN AND TRAVATAN

198. There is some small medical advantage of Travatan over Xalatan, particularly for African Americans.

199. There is at least one known study published in a recognized ophthalmological journal evaluating the safety and IOP-lowering efficacy of two concentrations of travoprost (0.0015% and 0.004%) compared with latanoprost 0.005% and timolol 0.5% in patients with open-angle glaucoma or ocular hypertension. Netland, et al., "Travoprost Compared With Latanoprost and Timolol in Patients With Open-angle Glaucoma or Ocular Hypertension," 132 Am. J. Ophthalmology 472, 478 (2001) (Eisenberg Reply Decl., Ex. 2). The study reported a response (i.e., there was a 30% or more reduction in IOP) rate of 54.7% to travoprost .004%, 49.3% to travoprost .0015% and 49.6% to latanoprost.

XII. IRREPARABLE HARM TO ALCON

200. Pharmacia has contended that when it first objected to Alcon's mark on February 7, 2001, five weeks before Alcon received FDA approval for Travatan, Alcon should have changed the name of its product. By that time, such a course was neither commercially feasible, nor reasonable given Pharmacia's lengthy period of inaction.

201. The need to obtain both FDA and PTO approval (as well as the approvals of health and trademark regulators worldwide) for a new name would have resulted in the product being off-market for months while those approvals were being obtained. Def. Ex. B (DX00024) (Barton Decl., ¶ 25); 12/17 Tr., at 173 (Barton); 12/19 Tr., at 187-88 (Schira).

202. The Court rejects the opinion of Pharmacia's expert, Mr. George Di Domizio, that "it's never too late to change a trademark," 12/17 Tr., at 196 (Di Domizio), and that a new mark could be approved in as little as 30-60 days. Id. at 198-99. That opinion is contradicted by Mr. Di Domizio's own testimony and other facts.

203. Mr. Di Domizio's testimony about the amount of time that the FDA takes to approve a new name for a pharmaceutical product was inconclusive at best. Although he initially testified that a new name could be approved "in a week or two," id. at 199 (Di Domizio), he conceded at the hearing that the approval process is "situational" and could take "over a year" or longer. Id. at 203-05. Mr. Di Domizio gave inconsistent answers when asked how many actual pharmaceutical products he had handled where the FDA had approved a new name in the time frame he had suggested was possible for Travatan. See id. at 199 (two such products), 200 (one such product), 201 (two such products), 202 (three or four such products). Moreover, these incidents involved name pairs that were not only far more similar than Travatan and Xalatan, but also, as Mr. Di Domizio conceded, were identified as confusing by the FDA itself. Id. at 202.

204. Pharmacia's own experience with the launch of its still-unapproved glaucoma latanoprost-timolol combination product shows that the process of selecting a new name for a prescription drug product is time consuming, particularly when, as is true of both parties here, the name is intended to be used on a worldwide basis.

205. Nor would it have been a reasonable business alternative for Alcon to market Travatan under its generic name. Not only would Alcon incur substantial costs in educating the medical community about the name change, but none of that investment would ever translate into goodwill, because no company can claim exclusive rights in a generic name. Di Domizio Tr., at 230, 234-35. As Pharmacia's own expert put it, it is not realistic to expect a company to "gamble[ ] away all future brand equity" by marketing its product under its generic name. Di Domizio Reply Decl., Ex. 4.

CONCLUSIONS OF LAW

I. Jurisdiction and Venue

1. Pharmacia's claims against Alcon are for unfair competition, including trademark infringement under the Lanham Act, 15 U.S.C. 1114(1)(a), false designation of origin and false representation under Section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a), violation of the federal dilution law, § 1125(c), and for violation of state unfair competition and anti-dilution laws under the common law and N.J. Stat. § 56:3-13.20.

2. This Court has subject matter jurisdiction under 15 U.S.C. § 1121(a) and 28 U.S.C. §§ 1331, 1367 and 1338. This Court has supplemental subject matter jurisdiction over all other claims under 28 U.S.C. § 1367(a).

3. A substantial portion of the events giving rise to plaintiffs' claims and Alcon's defenses occurred in this District. Venue is proper, therefore, in this District. 28 U.S.C. § 1391(b).

II. The Preliminary Injunction Standard

4. Preliminary injunctive relief is an extraordinary and drastic remedy that should not be granted unless the party seeking the injunction can, upon a clear showing, demonstrate: (1) that there is a likelihood of eventual success on the merits; (2) that there is a probability of irreparable injury if relief is not granted; (3) that the non-moving party will not suffer irreparable harm if the preliminary injunction is issued; and (4) that preliminary injunctive relief is in the public interest. Times Mirror Magazines, Inc. v. Las Vegas Sports News, LLC, 212 F.3d 157, 160 (3d Cir.2000), cert. den'd, 531 U.S. 1071, 121 S. Ct. 760, 148 L. Ed. 2d 662 (2001).

III. Likelihood of Success On The Merits

5. To satisfy its burden on the merits, Pharmacia must prove that confusion from Alcon's use of Travatan is likely or probable among consumers. Checkpoint Sys., Inc. v. Check Point Software Tech., Inc., 269 F.3d 270, 279-80 (3d Cir. 2001); Versa Prods. Co. v. Bifold Co., 50 F.3d 189, 200 (3d Cir.), cert. den'd, 516 U.S. 808, 116 S. Ct. 54, 133 L. Ed. 2d 19 (1995). Pharmacia has failed to meet its burden of proof.

6. In A & H Sportswear Inc. v. Victoria's Secret Stores, Inc., 166 F.3d 197, 205-06 (3d Cir.1999) (en banc) (Victoria's Secret I), the Third Circuit unanimously adopted a uniform "likelihood of confusion" standard for all trademark infringement cases. Therefore the Court rejects Pharmacia's argument that the Court should apply the "possibility of confusion" standard in cases involving prescription drugs. It is this Court's conclusion that Morgenstern Chemical Co. v. G.D. Searle & Co., 253 F.2d 390, 394 (3d Cir.1958), is no longer good law in this Circuit for the following reasons. First, as the Third Circuit noted, "`likelihood of confusion' is the language used in the Lanham Act." Victoria's Secret I, 166 F.3d at 205. There is no basis in the Lanham Act to draw any distinction in the substantive standard based on the nature of the products at issue. Second, the Third Circuit cited favorably and relied upon numerous decisions, from this Circuit and elsewhere, applying the "likelihood of confusion" standard in trademark cases involving pharmaceutical and other medical products. See id. at 205-06, 211 citing, Barre-National, Inc. v. Barr Laboratories, Inc., 773 F. Supp. 735 (D.N.J.1991); Vitek Sys. Inc. v. Abbott Labs., 675 F.2d 190, 192 (8th Cir.1982) (involving medical equipment) G.D. Searle & Co. v. Chas. Pfizer & Co., 265 F.2d 385 (7th Cir.1959) (drugs); U.S. Surgical Corp. v. Orris, Inc., 5 F. Supp. 2d 1201, 1209 (D.Kan.1998) (rejecting "possibility of confusion" in case involving surgical products), aff'd, 185 F.3d 885, 1999 WL 130688 (Fed.Cir.1999); see also Reedco, Inc. v. Hoffman-La Roche, Inc., 667 F. Supp. 1072, 1080 (D.N.J.1987) (refusing to apply Morgenstern even before Victoria's Secret I).^[11] Third, Morgenstern was not a federal trademark case but a diversity action based on New Jersey common law. The only New Jersey case that the court relied on was Rubber & Celluloid Harness T. Co. v. Rubber-Bound Brush Co., 81 N.J.Eq. 419, 519, 88 A. 210 (N.J.Err. & App.1913), which is not a trademark case, or one dealing with medical products, but a passing off case involving brushes. Its vitality on the issue of whether a likelihood of confusion standard or a possibility of confusion applies to trademarks of medical products would appear to be negligible. This is particularly so in light of the provision in N.J.S.A. 56:3-13a that federal law trademark law is persuasive authority for the interpretation and construction of New Jersey's trademark act.

7. Pharmacia has not proved that there is a likelihood of confusion between Alcon's use of the Travatan mark and Pharmacia's use of the Xalatan mark. In determining whether confusion is likely, the following non-exclusive list of factors are relevant: (1) the degree of similarity between the marks; (2) the strength of the owner's mark; (3) the price of the goods; (4) the length of time the defendant has used the mark without evidence of actual confusion; (5) the intent of the defendant; (6) the evidence of actual confusion; (7) whether the goods are marketed through the same channels of trade; (8) the similarity of the sales targets; (9) the relationship of the goods in the mind of consumers; and (10) other facts suggesting that consumers might expect the prior owner to manufacture both products. A & H Sportswear, Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198, 215 (3d Cir. 2000) (Victoria's Secret II). Not all the factors need be considered if some are dispositive. Checkpoint Sys., 269 F.3d at 280. Moreover, the weight to be given any particular factor is a fact sensitive determination left to the discretion of the trier of fact. Id.

8. Pharmacia has not proved that there is a likelihood of confusion between Alcon's use of the Travatan mark and Pharmacia's use of the Xalatan mark.

9. Here, at least five factors — the length of time of concurrent use without confusion, the sophistication of the market, the channels of trade, Alcon's good faith in adopting the mark and the dissimilarities in the dominant features of the marks — dispositively weigh in Alcon's favor and indicate that confusion is not likely.

a. The Absence Of Confusion In The Marketplace

10. Both Travatan and Xalatan have coexisted for over nine months. Thousands of prescriptions for each have been written and filled. Thousands of "detailing" visits have been made during this time period by representatives of both companies. Despite this, and despite the existence of voluntary and mandatory reporting systems that catalogue mix-ups between prescription products, there is no evidence in the record of any confusion. Although evidence of actual confusion is not necessary to prove likelihood of confusion, Ford Motor Co. v. Summit Motor Products, Inc., 930 F.2d 277, 292 (3d Cir.1991), such concurrent use for a substantial period of time with no confusion may create a "presumption that there is little likelihood of confusion," and weighs heavily against a finding of likely confusion. Barre-National, 773 F.Supp. at 744; see also, Pignons S.A. de Mecanique de Precision v. Polaroid Corp., 657 F.2d 482 (1st Cir.1981); Aktiebolaget Electrolux v. Armatron Int'l, Inc., 999 F.2d 1 (1st Cir.1993) ("absence of actual confusion, or a negligible amount of it, between two products after a long period of coexistence on the market is highly probative in showing that likelihood of confusion exists"); Gruner + Jahr USA Publish. v. Meredith Corp., 793 F. Supp. 1222, 1232-33 (S.D.N.Y.1992) (where both parties products have been on the market for six months, "we find that the absence of evidence of actual confusion weighs heavily against a finding of likelihood of confusion"), aff'd, 991 F.2d 1072 (2d Cir.1993); Mars, Inc. v. H.P. Mayer Corp., 1988 WL 86314, *2 (D.N.J. Aug.17, 1988) (finding that where "defendants have been distributing [the product in question] for approximately six months without any evidence of actual confusion occurring, that there is little to no likelihood" of confusion).

11. The FDA's approval of the Travatan mark, which is based on its own pre-approval testing for confusing similarity, is entitled to some deference. See Zeneca Inc. v. Eli Lilly & Co., 1999 WL 509471, at *34 (S.D.N.Y. July 19, 1999) ("as a recognized expert in evaluating data from clinical trials, the FDA's conclusion ... is persuasive evidence" in assessing Lanham Act claims); Rhone-Poulenc Rorer Pharms., Inc. v. Marion Merrell Dow, Inc., No. 93-0144-CV-W-1, 1994 U.S. Dist. LEXIS 20782, at *13 (W.D.Mo. Sept. 30, 1994) ("[In a Lanham Act case,] ... [i]f the tests were conducted in a manner approved by the FDA and the results are accepted as valid by the FDA, then they are deemed valid by this Court."), aff'd in relevant part, 93 F.3d 511 (8th Cir.1996); Brandenfels v. Heckler, 716 F.2d 553, 555 (9th Cir.1983) ("Determining reliable scientific data is not the judicial function. Congress vested that responsibility in the FDA.")

12. The Court rejects Pharmacia's argument that OPDRA, a lower-level subdivision of the FDA, was somehow misled by earlier plans to market Travatan in two concentrations. This ignores the fact that the final decision of the FDA's Division Director to approve Travatan was made with full knowledge that it would be marketed in one concentration. The decision of the Division Director is entitled to the Court's deference. See Serono Labs., Inc. v. Shalala, 158 F.3d 1313, 1320-21 (D.C.Cir.1998) (district court erred in giving weight to lower-level FDA staff reports; only Division Director's decision was entitled to deference).

13. The Court is aware that Pharmacia is not legally required to conduct a confusion survey. But under the circumstances of this case, Pharmacia's failure to conduct any confusion survey weighs against its request for a preliminary injunction. Such a failure, particularly when the trademark owner is financially able, justifies an inference "that the plaintiff believes the results of the survey will be unfavorable." Charles Jacquin et Cie, Inc. v. Destileria Serralles, Inc., 921 F.2d 467, 475 (3d Cir.1990); see also Eagle Snacks, Inc. v. Nabisco Brands, Inc., 625 F. Supp. 571, 583 (D.N.J.1985); Tyco Indus. Inc. v. Lego Sys. Inc., 5 U.S.P.Q.2d 1023, 1033, 1987 WL 44363 (D.N.J.1987), aff'd, 853 F.2d 921 (3d Cir.1988).

14. Pharmacia's attempt to explain away its decision not to conduct a confusion survey (economics) further supports the appropriateness of such an inference, in light of the fact that from July 2000 forward, Pharmacia conducted numerous surveys concerning the Travatan mark, but never attempted to assess whether Travatan was likely to cause confusion with Xalatan. Moreover, the results of that research show that over 95% of ophthalmologists were aware of Travatan and Xalatan as distinct products. The enormous sums Pharmacia invested in market research to prepare for the Travatan launch stand in stark contrast to its failure to conduct a confusion survey here, and fully warrant a negative inference regarding likelihood of confusion.

15. Finally, Professor Diamond's well-designed survey of ophthalmologists and optometrists, showing a net confusion rate of 1.5%, Def. Ex. D (DX00239-40, DX00246-47, DX00250) (Diamond Decl. ¶¶ 6, 26, 30), provides further evidence to support the Court's conclusion that no confusion is likely. Henri's Food Prods. Co. v. Kraft, Inc., 717 F.2d 352, 358-59 (7th Cir.1983) (holding that 7.6% confusion finding is "a factor weighing against infringement"); Pfizer Inc. v. Astra Pharm. Prods., Inc., 858 F. Supp. 1305, 1326 (S.D.N.Y.1994) (adjusted confusion rate of 7.27% in a survey is "low" and insufficient evidence of likelihood of confusion); see also 5 McCarthy § 32:188, at 32-311 (4th ed.2001) (generally, "figures below 20 percent become problematic because they can only be viewed against the background of other evidence ...").

16. For all these reasons, the actual confusion factor strongly favors Alcon.

b. The Sophistication Of The Relevant Market

17. The relevant market for the products is a significant factor, particularly when the market consists of sophisticated professionals. Victoria's Secret II, 237 F.3d at 215, 225; Ford Motor Co., 930 F.2d at 293; Restatement (Third) of Unfair Competition, § 25, at 271 (1995) (hereinafter "Restatement") ("Discerning purchasers are more likely to recognize elements that distinguish two similar marks and are thus more likely to retain separate mental associations for each of the designations.").

18. In trademark cases involving competing prescription drugs, the relevant consumers are physicians because patients do not choose their own prescription drugs. Smithkline Beckman Corp. v. Pennex Prods. Co., 605 F. Supp. 746, 752-53 (E.D.Pa.1985) ("Prescription drugs are a unique commodity. It is the physician, not the consumer, who selects the prescription."); Pennwalt Corp. v. Zenith Labs., Inc., 472 F. Supp. 413, 422 (E.D.Mich.1979) ("[T]he dispensing physician is in fact the individual who truly exercises the consumer-patient's freedom of choice in the marketplace when issuing a prescription.") Physicians are capable of fine distinctions between marks, making confusion even less likely. Doral Pharmamedics v. Pharmaceutical Generic Developers, Inc., 148 F. Supp. 2d 127, 138-39 (D.P.R.2001) (confusion between prescription drugs "Exotic-HC" and "Genexotic-HC" unlikely); Pfizer Inc. v. Astra Pharm. Prods., 858 F.Supp. at 1309, 1328 (identical XL suffix unlikely to confuse physicians, "as sophisticated a group as one could imagine"); Barre-National, 773 F.Supp. at 742, 745 (sophistication made confusion between "Barre" and "Barr" unlikely); Schering Corp. v. Thompson Med. Co., 209 U.S.P.Q. 72, 74 (S.D.N.Y.1979) (same; confusion between "Polaramine" and "Prolamine" unlikely).

19. Both parties' surveys in this case confirm that doctors are a sophisticated market.

20. Although Pharmacia argues about possible medication errors later in the prescription chain, even assuming that such errors occur, they do not establish trademark confusion. Medication errors include missed doses, drugs given to the wrong patient, prescriptions filled with wrong dosages and many other kinds of errors caused by any number of reasons, including poor handwriting, verbal miscommunications or other breakdowns in the dispensing chain. Misreading or mishearing a prescription is not actionable under the Lanham Act, because trademark confusion is limited to consumer confusion in the context of a purchasing decision. See Harlem Wizards Entm't Basketball, Inc. v. NBA Properties, Inc., 952 F. Supp. 1084, 1098 (D.N.J.1997) ("Actual confusion is not the same as clear mistake or misidentification on the part of consumers.... Moreover, there is no evidence that these purported instances of actual confusion could have any effect on consumer purchasing decisions"); Lang v. Ret. Living Publ'g Co., 949 F.2d 576, 582-83 (2d Cir. 1991) (rejecting mistaken communications as evidence of confusion because there was no link to any purchasing decision).

21. When a consumer knows the identity of the product or service he or she desires, but is mistakenly directed to another source (for example, because of a mistake by directory assistance), the error is not evidence of trademark confusion. Checkpoint Sys., 104 F.Supp.2d at 463 (misdirected communications based on directory assistance "not actionable under the Lanham Act"), aff'd, 269 F.3d 270, 298 (3d Cir.2001) (distinguishing consumer confusion from error); citing, Duluth News-Tribune v. Mesabi Pub'g Co., 84 F.3d 1093, 1098 (8th Cir.1996); see also Prime Media, Inc. v. Primedia, Inc., 33 F. Supp. 2d 932, 939-40 (D.Kan.1998) (calls misdirected to plaintiff by directory assistance is not "the type of confusion the Lanham Act guards against"); Major League Baseball Props., Inc. v. Sed Non Olet Denarius, Ltd., 817 F. Supp. 1103, 1122 (S.D.N.Y.1993) ("typographical errors" do not constitute trademark confusion).

22. Because the evidence of record demonstrates that doctors (who play the role of the ultimate consumer in prescription drug cases) are not confused, to the extent that pharmacists make mistakes, they are analogous to mistaken directory assistance operators. More importantly, however, Pharmacia has no evidence of any misprescriptions between Xalatan and Travatan.

23. The sophistication of consumers strongly favors Alcon.

c. Channels Of Trade And The Relevant Marketing Practices

24. Pharmacia's marketing executives specifically anticipated that Alcon would market Travatan by expressly comparing and contrasting it to Xalatan. The channels of trade — particularly the myriad inperson detailing visits by both companies — also reduce the likelihood of confusion. Pfizer Inc. v. Astra Pharm. Prods., 858 F.Supp. at 1324 (doctors know they cannot receive information for competing drugs from the same sales representative, "dispelling any issue concerning likelihood of confusion"); American Home Prods. v. Barr Labs., Inc., 656 F. Supp. 1058, 1069 (D.N.J.1987), aff'd, 834 F.2d 368 (3d cir. 1987) (no likelihood of confusion where comparative advertisements "essentially beg consumers to distinguish [the defendant's] generic ibuprofen tablets from Advil"); Warner Lambert Co. v. McCrory's Corp., 718 F. Supp. 389, 398-99 (D.N.J. 1989) (comparative statements "distinguish[] the two products"); Pfizer Inc. v. Perrigo Co., 988 F. Supp. 686, 698 (S.D.N.Y.1997) (same).

25. The relevant marketing practices and the channels of trade strongly favor Alcon.

d. Standards For Finding Bad Faith

26. The credible evidence does not support Pharmacia's assertion that Alcon intentionally copied the "ATAN" part of the Xalatan mark. There is no factual support for that proposition.

27. Even accepting Pharmacia's argument that Travatan was named by copying the ending of Xalatan (which the Court rejects), such copying is legally irrelevant for at least three reasons.

28. First, because the marks as a whole are not identical, no presumption of intentional copying arises and Pharmacia retains the burden of establishing likelihood of confusion. Barnes Group Inc. v. Connell L.P., 793 F. Supp. 1277, 1302 (D.Del.1992); see also 20th Century Wear, Inc. v. Sanmark-Stardust, Inc., 815 F.2d 8, 10-11 (2d Cir.1987); 3 McCarthy § 23:107, at 23-250.

29. Second, the reasoned judgment by counsel that Travatan was available precludes a finding of bad faith.

30. Third, intentional copying is irrelevant when other factors indicate that there is not a likelihood of confusion, American Home Prods. Corp. v. Barr Labs., Inc., 834 F.2d 368, 371 (3d Cir.1987), or when the facts do not show that the defendant intended to deceive consumers as to the product's source. Versa Prods., 50 F.3d at 205-06; Victoria's Secret II, 237 F.3d at 226 ("[D]efendant's intent will indicate a likelihood of confusion only if an intent to confuse consumers is demonstrated via purposeful manipulation of the junior mark to resemble the senior's.")

31. There is no evidence that Alcon intended doctors to be confused as to the source or nature of Travatan.

32. Because there is no evidence that Alcon intended to deceive consumers as to the source of the Travatan product (or even that it intentionally copied the "ATAN" portion of the Xalatan mark), Alcon's "adoption of a trademark with actual knowledge" of Xalatan may be entirely "consistent with good faith." Lang, 949 F.2d at 583-84. "Indeed, even if [Mr. Schira's] professional advice had been wrong, it does not follow that [Alcon's] reliance on that advice would have constituted bad faith." Mead Data Central, Inc. v. Toyota Motor Sales, U.S.A., 875 F.2d 1026, 1028 (2d Cir.1989); see also A & H Sportswear Co. v. Victoria's Secret Stores, Inc., 926 F. Supp. 1233, 1268 (E.D.Pa.1996), aff'd, 166 F.3d 191 (3d Cir.1999) ("[W]here Defendants relied on advice of counsel, even if wrong, record insufficient to find Defendants liable for `that species of bad faith which constitutes independent and substantive proof of infringement.'") (emphasis added) (quoting, Information Clearing House v. Find Magazine, 492 F. Supp. 147, 161 (S.D.N.Y.1980)), aff'd, 166 F.3d 191 (3d Cir.1999); Parenting Unlimited, Inc. v. Columbia Pictures Television Inc., 743 F. Supp. 221, 230 (S.D.N.Y.1990) ("[D]efendants relied on the advice of counsel from the outset and should not be deemed to have acted in bad faith purely on the basis of plaintiff's contention that their assessment of the mark's availability was incorrect.").

33. Accordingly, the lack of bad faith favors Alcon.

e. Differences In The Marks And Packaging

34. Pharmacia bases its infringement and dilution claims on the shared "ATAN" suffixes of the two names, ignoring the many dissimilarities in the spelling, pronunciation and presentation of the marks. However, "marks should be viewed in their entirety," rather than dissected. Victoria's Secret II, 237 F.3d at 216.

35. Dissection of marks is particularly inappropriate in the pharmaceutical context because suffix similarity is not uncommon and, for that very reason, not likely to confuse highly trained doctors. See Upjohn Co. v. Schwartz, 246 F.2d 254, 262 (2d Cir.1957) ("`Syrocol' [and] `Cheracol' ... do not look or sound enough alike to justify a holding of trademark infringement. The only similarity is in the last syllable, and that is not uncommon in the names given drug compounds.") (emphasis added).

36. Any claim by Pharmacia to exclusive rights to the suffix "ATAN" is weak because of the number of other pharmaceutical products that use the suffix (e.g., Rynatan, Phenatan, Germa-tan), and because of the existence of similar marks in the glaucoma market (e.g., Alphagan, Betagan, Lumigan). 2 McCarthy § 11:85, at 11-163 ("[I]n a `crowded' field of similar marks, each member of the crowd is relatively `weak' in its ability to prevent use by others in the crowd."); see also Victoria's Secret II, 237 F.3d at 223-24 (third-party use weighs against finding mark strength and likely confusion); Accu Personnel, Inc. v. AccuStaff, Inc., 823 F. Supp. 1161, 1166 (D.Del.1993) (numerous similar marks enable consumers to distinguish "on the basis of minute distinctions") (citation omitted); Restatement § 13, at 110.

37. No particular number of other products or users are necessary to show that concurrent use weakens claims to exclusivity. Castle Oil Corp. v. Castle Energy Corp., Civ. A. No. 90-6544, 1992 WL 394932, at *12 (E.D.Pa. Dec.29, 1992).

38. The relevant market regarding third party use is not limited to ophthalmic solutions but is the overall pharmaceutical market. Barre-National, 773 F.Supp. at 741 (liquid pharmaceuticals market too narrow; market was all pharmaceutical products); A & H Sportswear Co., 926 F.Supp. at 1267 (entire market within which goods were registered); Hershey Foods Corp. v. Mars, Inc., 998 F. Supp. 500, 517 (M.D.Pa.1998) ("food industry," not just candy market, is relevant); see also Victoria's Secret II, 237 F.3d at 223 ("[T]he extensive use of the term in other [related] markets may also have a weakening effect on the strength of the mark.").

39. The prominent display of the Pharmacia and Alcon house marks, along with the distinctive packaging used by each company, further weigh against likely confusion. Victoria's Secret II, 237 F.3d at 218-19 (house marks are significant in determining overall impression); Nabisco, Inc. v. Warner-Lambert Co., 220 F.3d 43, 46 (2d Cir.2000) (house mark "significantly reduces, if not altogether eliminates," any likelihood of consumer confusion); Bristol-Myers Squibb Co. v. McNeil-P.P.C., Inc., 973 F.2d 1033, 1046 (2d Cir.1992) (same); American Cyanamid, 800 F.2d at 309 (2d Cir.1986) (same); Richards v. Cable News Network, Inc., 15 F. Supp. 2d 683, 690-91 (E.D.Pa.1998) ("World Beat" and "World Beat" not confusingly similar because of differences in presentation); Harlem Wizards Entm't Basketball, 952 F.Supp. at 1096 (same; "Harlem Wizards" and "Washington Wizards"); Pfizer, Inc. v. Astra Pharm. Prods., 858 F.Supp. at 1324 (same); Genovese Drug Stores, Inc. v. TGC Stores, Inc., 939 F. Supp. 340, 346 (D.N.J.1996) (house mark and other visual differences); Barre-National, 773 F.Supp. at 742 (stylized "b" distinguished "Barre" and "Barr"); Schmid Labs. v. Youngs Drug Prods. Corp., 482 F. Supp. 14, 18 (D.N.J.1979) (same); 3 McCarthy § 23:52, at 23-141 to 23-142. Alcon made its mark even more distinct by using ALL CAPS. Barnes, 793 F.Supp. at 1302.

40. This factor strongly favors Alcon as well.

f. Pharmacia's Opinions Regarding Likelihood of Confusion

41. Because Pharmacia has no marketplace evidence of likely or actual confusion, it emphasizes evidence in the form of expert opinion testimony. This evidence is insufficient to overcome the marketplace evidence that confusion is unlikely. Apart from consumer surveys, "[L]ay or even expert opinion about the likelihood of confusion is inadmissible or entitled to little weight." Richard L. Kirkpatrick, Likelihood of Confusion in trademark Law, § 1.8.c, at 1-45 (PLI 1995); Barnes Group, 793 F.Supp. at 1293, 1301-02 (expert opinion that confusion was likely, offered without any supporting empirical data or marketplace observations, could not be relied on).

42. The opinions of Dr. Eisenberg and Mr. Di Domizio with regard to the likelihood of confusion between Travatan and Xalatan are based primarily on a subjective evaluation of the marks in light of their experience in the pharmaceutical industry (Di Domizio) or as an ophthalmologist (Eisenberg). There are no reported trademark cases in which a court has based its findings of a likelihood of confusion or dilution on the types of "opinions" on which Pharmacia relies. The bases for these opinions stand in stark contrast to the survey conducted by Professor Shari S. Diamond, J.D., Ph. D., which demonstrates persuasively a confusion rate of 1.5%. See John H. Harland Co. v. Clarke Checks, Inc., 711 F.2d 966, 979 n. 23 (11th Cir.1983) (lay opinion of bank president and opinion of attorney with no experience in relevant industry and who conducted no confusion survey entitled to little weight).

43. Dr. Lambert's opinion on the likelihood of confusion similarly is entitled to little weight. His statistical model cannot predict with meaningful reliability in the real world whether Xalatan and Travatan are accurately characterized as an "Error Pair," or if instead they represent a "false positive" — that is, a pair of drug names improperly classified by Dr. Lambert as confusingly similar. Because in the real world there are many, many more non-confusing drug name pairs than confusingly similar ones, it follows from the application of basic statistical principles that the Travatan-Xalatan pair is likely to be a false positive. In any event, Dr. Lambert was unable to opine with any degree of certainty what numbers should be used to determine his model's reliability, much less whether Travatan and Xalatan ever will be confused in the real world. Lambert Tr., at 144-46.

44. Because the Court must assess confusion based on a well-defined set of factors intended to measure what is likely to happen in the real world, Checkpoint Sys., 269 F.3d at 279-80, Dr. Lambert's model lacks significant probative value for the purposes for which it is offered. Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 594, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993) (courts should consider error rate of scientific technique); Kumho Tire Co. v. Carmichael, 526 U.S. 137, 153-54, 119 S. Ct. 1167, 143 L. Ed. 2d 238 (1999) (question is not whether method is useful in general, but whether method is reasonable in reaching conclusion about specific event at issue); General Electric Co. v. Joiner, 522 U.S. 136, 146, 118 S. Ct. 512, 139 L. Ed. 2d 508 (1997) ("A court may conclude that there is simply too great an analytical gap between the data and the opinion proffered."). The model is entitled to very little weight in a Lanham Act case and cannot serve as the basis for a finding of likely confusion, especially in light of contrary marketplace evidence and the FDA's regulatory determination.

45. None of the expert opinions remaining after the Court's November 28 in limine ruling, therefore, is sufficiently compelling to overcome the marketplace evidence demonstrating that confusion is not likely.

g. The Dilution Claim

46. Pharmacia also has not shown a likelihood of success on the merits of its trademark dilution claims. In order to succeed on the merits of its trademark dilution claim, Pharmacia must prove, inter alia, that: (1) the Xalatan mark is famous,^[12] and (2) Alcon's use of the Travatan mark causes dilution of the capacity of Xalatan to identify and distinguish its latanoprost product. Times Mirror, 212 F.3d at 162-63.

47. The Court recognizes that Pharmacia's claim for dilution is separate and distinct from its claim for confusion.^[13] Nevertheless, dilution is not a "fallback protection" for invalid infringement claims. I.P. Lund Trading v. Kohler Co., 163 F.3d 27, 48 (1st Cir.1998) (affirming denial of preliminary injunction and cautioning against crediting dilution claims involving competitive products where there is no confusion). It is "up to the judiciary to apply ... potent [anti-dilution] laws with care and common sense lest they damage the competitive system they are designed to enhance." 4 McCarthy § 24:114, at 24-250.

48. Pharmacia has failed to meet its burden on its dilution claim for several reasons. First, a dilution plaintiff must be threatened by a very similar, if not identical, mark. Courts have repeatedly rejected dilution claims unless the marks are essentially the same, which they are not here. American Cyanamid Co. v. Nutraceutical Corp., 54 F. Supp. 2d 379, 392-93 (D.N.J.1999) (differently designed rainbows); Luigino's, Inc. v. Stouffer Corp., 170 F.3d 827, 832 (8th Cir.1999) ("Lean `N Tasty" mark did not dilute "Lean Cuisine" mark); I.P. Lund, 163 F.3d at 50; Mead Data, 875 F.2d at 1028-29 (finding "LEXIS" and "LEXUS" not substantially similar for dilution purposes); Ringling Bros.-Barnum & Bailey Combined Shows, Inc. v. B.E.Windows Corp., 937 F. Supp. 204, 211-12 (S.D.N.Y.1996).

49. Indeed, differences such as the use of house marks "alone" can defeat dilution claims. Astra Pharm. Prods., Inc. v. Beckman Instruments, Inc., 718 F.2d 1201, 1209-10 (1st Cir.1983) ("Astra" and "Astra"); Hormel Foods Corp. v. Jim Henson Prods., Inc., 73 F.3d 497, 506 (2d Cir.1996) (display of "Muppet Treasure Island" next to "Spa`am" character "alone" could prevent dilution of Hormel's "Spam").

50. In the pharmaceutical context, because prescription drug packaging is required to display generic names, the likelihood of dilution is further diminished when, as is the case here, those generic names differ. Pfizer, 858 F.Supp. at 1324.

51. Pharmacia's dilution claim suffers from two other obvious infirmities. First, there are many other drug names that end in "AN," "TAN" and "ATAN." The existence of: (1) Alphagan, Betagan and Lumigan; (2) the eleven other pharmaceutical products that end in "TAN," (3) the seven other pharmaceutical products ending in "ATAN" that preceded Xalatan,^[14] disposes of Pharmacia's claim to exclusivity in the "ATAN" suffix, the only part of the Travatan name that Pharmacia claims dilutes Xalatan.^[15]

52. Second, in the pharmaceutical field, professionals have been trained to distinguish generic and proprietary names within a class of drugs. This means that the sophistication and training of these professionals must be taken into account. Times Mirror, 212 F.3d at 168; Astra Pharm. Prods., 718 F.2d at 1209-10; Restatement § 25 cmt. f, at 271 ("Discerning purchasers are more likely to recognize elements that distinguish two similar marks and are thus more likely to retain separate mental associations for each of the designations.").

53. Moreover, courts should be wary of applying a dilution remedy when it is invoked in trademark cases between competing products, especially when there is no accompanying evidence of a likelihood of confusion. I.P. Lund, 163 F.3d at 45 ("the archetypical problems [addressed by the Federal Trademark Dilution Act] involved non-competing products as to which there could, by definition, be no confusion"); Fed. Express Corp. v. Fed. Espresso, Inc., 201 F.3d 168, 177-78 (2d Cir.2000) (plaintiff's inability to show likely confusion supported denial of preliminary injunction on dilution claim); 4 McCarthy § 24:72, at 24-130.1 to 130.2 ("It is difficult to understand why an anti-dilution law is invoked when parties operate in competitive or closely-related product or service lines.").

54. In light of the above, dilution is not likely.

55. The Court reaches this conclusion notwithstanding the McCullough survey. In fact, if anything, the McCullough survey supports the court's conclusion that Pharmacia has not met its burden of showing a likelihood of dilution.

56. In claims for likelihood of confusion, case law teaches that a finding of infringement requires a 15 to 20 percent range of confusion. Gerald L. Ford, Dilution Surveys Update 1998 Practicing Law Institute, 537 PLI/Pat 551, 562 (1998). See also Xuan-Thao N. Nguyen, The New Wild West: Measuring And Proving Fame And Dilution Under The Federal Trademark Dilution Act, 63 Alb. L.Rev. 201 (1999) ("Courts in trademark infringement cases have consistently relied on survey evidence in their findings of likelihood of confusion in violation of the Lanham Act. Survey evidence within a range of 25% to 50% is well accepted as solid support for a finding of a likelihood of confusion." See fn. 270.)

57. A dilution claim may require a slightly higher number "because the nature of the underlying cause of action is different. It is harder to establish that consumers believe that a defendant's product is put out by the plaintiff. That is harder to establish than it is to establish dilution which simply requires a showing that defendant's product calls to mind the plaintiff's brand name. [A plaintiff] doesn't have to make the source association. [A plaintiff] just [has] to show that the plaintiff's brand is called to mind or that looking at the defendant's product makes them think of plaintiff ..." See Ford, Dilution Surveys Update (1998). See also Gerald L. Ford, "Dilution Surveys New Challenges," Litigating Copyright, Trademark and Unfair Competition Cases (1997) (suggestion that 15-20% should not be sufficient in dilution cases); Xuan-Thao N. Nguyen, The New Wild West: Measuring and Proving Fame and Dilution Under the Federal Trademark Dilution Act, 63 Albany L.Rev. 201, 237 (1999)("To establish proof of dilution, survey evidence must meet a minimum threshold of at least 20% of respondents making a mental association between the famous mark and junior mark.").

58. Even fully crediting Mr. McCullough's survey, it is insufficient on its face to support a preliminary injunction for trademark dilution. That survey shows at best a dilution level of 14%.^[16] That is less than half of the lowest percentage of dilution in any reported decision basing a finding of dilution on a survey. WAWA Dairy Farms v. Haaf, 40 U.S.P.Q.2d 1629, 1996 WL 460083 (E.D.Pa. 1996) (29%), aff'd, 116 F.3d 471 (3d Cir. 1997). Other courts have demanded a much higher percentage. Kellogg Co. v. Exxon Mobil Corp., 192 F. Supp. 2d 790, 806-808 (W.D.Tenn.2001) (70% association sufficient proof of blurring). Without dictating that any particular level of dilution is necessary at a minimum, it is plain that 14% is not adequate to sustain the extraordinary remedy sought here — a preliminary injunction against an FDA-approved prescription drug. Jerre B. Swann, Dilution Redefined for the Year 2000, 90 Trademark Rep. 823, 860 n. 237 (2000) (required association is "substantial," not merely "appreciable"); 5 McCarthy § 32:188, at 32-311.^[17]

59. That 14% figure, however, is entitled to little weight because of the flaws in methodology and execution of the survey which led to court to find that the dilution rate is 9.8%. Johnson & Johnson-Merck Consumer Pharms. Co. v. Rhone-Poulenc Rorer Pharms., Inc., 19 F.3d 125, 134 (3d Cir.1994) ("The probative value of a consumer survey is a highly fact-specific determination and a court may place such weight on survey evidence as it deems appropriate.") (citation omitted); Barre-National, 773 F.Supp. at 745 (according "little weight" to survey with "several major flaws which minimize its probative value"); Transfer Print Foils, Inc. v. Transfer Print America, Inc., 720 F. Supp. 425, 432-33 (D.N.J.1989) (same).

60. Once it became evident that two of the interviewers learned of the lawsuit, precautions should have been taken to ensure that any further responses they elicited were scrutinized for any anomalies and, if any appeared, were excluded. Paco Sport, Ltd. v. Paco Rabanne Parfums, 86 F. Supp. 2d 305, 323 (S.D.N.Y.2000) ("[T]he experts on both sides agree that for survey results to be reliable, interviewers should not be aware of the purpose of the survey ... or the answers the sponsor wants to receive. Interviewers occasionally fabricate answers instead of completing interviews, and knowledge of the desired responses makes cheating easier."), aff'd, 234 F.3d 1262, 2000 WL 1721126 (2d Cir.2000); Pittsburgh Press Club v. United States, 579 F.2d 751, 758 (3d Cir.1978). Because Ms. Boyd's interviews (as well as the two other Los Angeles interviews) sharply contrast with those obtained nationwide, they should not be counted.

61. When taken in the context of the previous findings of fact and conclusions of law, the McCullough survey with a dilution level of 9.8 % (or 14 % for that matter) is inadequate to meet Pharmacia's burden of proving likely success on its dilution claim.

62. The McCullough survey, accordingly, is inadequate to meet Pharmacia's burden of proving likely success on its dilution claims.

63. The court rejects Pharmacia's contention that as Alcon continues its marketing efforts to position Travatan to take market share from Xalatan, the percentage of ophthalmologists who think Xalatan when they see Travatan will increase with time. Pharamacia essentially argues that dilution is a slowly progressing tort, see Times Mirror, 212 F.3d at 163 ("gradual attenuation of the value of a famous mark ... gives rise to independent commercial tort for trademark dilution."); Liquid Glass Enterprises, Inc. v. Dr. Ing, h.c.F. Porsche AG, 8 F. Supp. 2d 398, 405 (D.N.J. 1998) (unauthorized use of trademark and trade dress likely to slowly whittle away the distinctiveness of Porsche's marks); Nabisco v. PF Brands, Inc., 50 F. Sup.2d 188, 210 (S.D.N.Y.1999) ("over time, the presence of Nabisco's goldfish-shaped cracker ... likely to weaken."). Here, Pharmacia's own dilution survey indicates (as modified by the Court) a 9.8% dilution rate. There is no evidence that the dilution rate will increase over time. Moreover, the Court has already concluded that there is no likelihood of dilution. See, supra, Conclusions, at ¶ 54.

IV. IRREPARABLE HARM

64. Plaintiff contends that it has demonstrated irreparable harm because it has made a strong showing of a likelihood of confusion. Jews For Jesus v. Brodsky, 993 F. Supp. 282, 311.(D.N.J.1998) ("[W]here a plaintiff makes a strong showing of a likelihood of confusion, irreparable injury follows as a matter of course"). Again, Pharmacia argues that irreparable harm is generally presumed in a case of trademark dilution. Eli Lilly & Co. v. Natural Answers, Inc., 233 F.3d 456, 469 (7th Cir.2000). This line of law is unavailing to Plaintiff, however, because Pharmacia has not made a showing of likelihood of confusion, let alone a strong one, and the court has concluded that Pharmacia has not met its burden of proving likely success on its dilution claims. Any claim to irreparable injury is undercut by its lengthy and inexcusable delay in bringing this action for preliminary injunctive relief. This delay is a separate but equally dispositive basis, apart from the weakness of Pharmacia's case on the merits, for denying Pharmacia's motion for a preliminary injunction.

65. Actual or constructive notice is the governing standard for measuring delay in moving for preliminary injunctive relief. Thus, courts consider both a trademark plaintiff's actual knowledge of infringement and its constructive knowledge. See, e.g., Commerce Nat'l Ins. Servs. v. Commerce Ins. Agency, Inc., 995 F. Supp. 490, 502 (D.N.J.1998) (applying constructive knowledge standard), rev'd on other grounds, 214 F.3d 432 (3d Cir.2000); Warner Lambert Co. v. McCrory's Corp., 718 F. Supp. 389, 395 (D.N.J.1989) (same); Reedco, 667 F.Supp. at 1081-82 (same); Calvin Klein Co. v. Farah Mfg. Co., 229 U.S.P.Q. 795, 797, 801, 1985 WL 3953 (S.D.N.Y.1985) (same); Mego Corp. v. Mattel, Inc., 203 U.S.P.Q. 377, 383, 1978 WL 21347 (S.D.N.Y.1978) (same).

66. A trademark plaintiff has a "duty to police" for allegedly infringing uses, Commerce Nat'l, 995 F.Supp. at 502 (relief denied; delay period for purposes of denying preliminary relief begins to run when plaintiff "could have known ... and should have known" of mark's use), before seeking the extraordinary remedy of a preliminary injunction. See also Reedco, 667 F.Supp. at 1081-82 (relief denied; applying constructive knowledge standard based on plaintiff's trademark-monitoring activities); Warner Lambert, 718 F.Supp. at 395 (relief denied; plaintiff's claim it had no actual knowledge rejected given its practice of monitoring the marketplace); Calvin Klein, 229 U.S.P.Q. at 801, 1985 WL 3953 at *10 (relief denied; plaintiff was "chargeable" with knowledge given its practice of monitoring the marketplace); Mego Corp., 203 U.S.P.Q. at 383, 1978 WL 21347 (relief denied; plaintiff held to have waited "until the Eleventh Hour" when it "should have known ... and could have known" about defendant's plans earlier from trade journals).

67. The "should have known" standard is satisfied by contrasting the full scope of a party's monitoring practices and the way it addressed marks it truly regarded as violating its rights, with its inactivity with respect to the mark it belatedly seeks to enjoin. Warner Lambert, 718 F.Supp. at 395 (relying on evidence of "Warner-Lambert's practice" of monitoring); Citibank N.A. v. Citytrust, 756 F.2d 273, 277 (2d Cir.1985) (evidence of plaintiff's practices in monitoring and opposing trademark applications and use of other "resources and information sources"); Calvin Klein, 229 U.S.P.Q. at 798, 1985 WL 3953 at *6 (evidence of "the extensive investigative forces Calvin Klein employed throughout the relevant years"); see also Andersen Consulting, LLP v. American Mgmt. Sys., Inc., 95 Civ. 5428, 1995 WL 510042, at *4 (S.D.N.Y. Aug.28, 1995) (finding "incredible" that if intellectual property was so valuable, plaintiff failed to thoroughly read competitor's business proposal containing alleged trade secrets). There is ample evidence in the record that during the relevant time period, Pharmacia monitored for infringing uses of its trademarks, and filed oppositions before the PTO to marks that it felt infringed its intellectual property rights.

68. Pharmacia has had constructive or actual notice of Alcon's intent to use the Travatan mark for its glaucoma product since at least May 1999, when the mark was published for opposition in the Official Gazette. See, e.g., Commerce Nat'l, 995 F.Supp. at 502 (applying "could have known ... and should have known" standard); Reedco, 667 F.Supp. at 1081-82 (publication in Official Gazette provided constructive knowledge); Warner Lambert, 718 F.Supp. at 395 (rejecting no actual knowledge defense given marketplace monitoring); Calvin Klein, 229 U.S.P.Q. at 801, 1985 WL 3953 at *10 (plaintiff "chargeable" with knowledge because of marketplace monitoring); Mego Corp., 203 U.S.P.Q. at 383, 1978 WL 21347 (plaintiff "should have known ... and could have known"); Citibank N.A., 756 F.2d at 277 ("It strains one's credulity to argue that a major financial institution such as Citibank, with all its resources and information sources, could not establish" earlier infringing use). Indeed, the very purpose of publication for opposition in the Official Gazette is to allow "[a]ny person who believes that he would be damaged by the registration of a mark" to "file an opposition in the Patent and Trademark Office ... within thirty days after the publication ..." 15 U.S.C. § 1063(a); see also id. § 1062(a); 37 C.F.R. § 2.101.

69. By the summer and fall of 1999, Pharmacia had to have known of the Travatan name, even if Pharmacia's delay is measured only from April 2000, when it concedes it knew of the Travatan mark. Such a delay — one full year — knocks the bottom out of any claim of immediate and irreparable harm. New Dana Perfumes v. The Disney Store, Inc., 131 F. Supp. 2d 616, 630 (M.D.Pa.2001) (delay of two months in sending demand letter and five months in moving for relief precludes preliminary injunction); Warner Lambert Co., 718 F.Supp. at 393-95 (seven-month delay and failure to move against other potential infringers "conclusively refutes" claim of irreparable harm); Nabisco, Inc. v. PF Brands, Inc., 191 F.3d 208, 222 (2d Cir. 1999) (delay is "perhaps more" important in dilution cases than in infringement cases); Tough Traveler, Ltd. v. Outbound Prods., 60 F.3d 964, 968 (2d Cir.1995) ("[D]elay alone may justify denial of a preliminary injunction.") (citation omitted).^[18]

70. Here, however, it is not simply the passage of time that refutes Pharmacia's claims of irreparable harm. In this case, Pharmacia not only knew about Alcon's Travatan mark, but also exhaustively market-tested and analyzed that mark for at least eight months before seeking relief. It also concluded, during those analyses, that the mark was not a competitive threat, but a weakness.

71. Neither the parties' negotiations over patent issues, nor the (remote) possibility that the FDA would not approve Travatan, excuses Pharmacia's delay. First, nothing about either of those potential developments prevented Pharmacia from asserting its rights. Harley-Davidson, Inc. v. Estate of O'Connell, 13 F. Supp. 2d 271, 280-81 (N.D.N.Y.1998) (delay not excused by possibility that defendant would never get zoning approvals); see also Gidatex, 13 F.Supp.2d at 419-20 (negotiations over contract validity did not affect plaintiff's ability to bring trademark suit). Second, the "FDA might not approve it" argument has been rejected under virtually identical circumstances. Reedco, 667 F.Supp. at 1082. The policy excusing delays attributable to settlement negotiations does not cover subjects not negotiated, particularly when one party was less than candid about the range of topics it wished to negotiate.

72. Moreover, the negotiations over the patent issues, if anything, undercut Pharmacia's position. The record shows that from at least July 2000 through February 2001, Pharmacia engaged in negotiations with Alcon over the possibility of resolving those patent issues by way of a license. Pharmacia negotiated, presumably in good faith, on patent issues alone, repeatedly referring to Travatan by name in those negotiations. This justifiably created an expectation on Alcon's part that any objections Pharmacia had to the Travatan product were limited to the patent and not its trademark. This expectation then was reinforced by Pharmacia's failure to oppose the Travatan mark anywhere in the world, either before or after the commencement of the patent negotiations.

73. Pharmacia is also not excused by its claim that it needed from April or May 2000 to February 2001 to investigate whether it had a valid trademark claim. Investigation is no excuse when the plaintiff is in possession of all of the necessary facts constituting the alleged infringement. See Gidatex, 13 F.Supp.2d at 420 (rejecting investigation excuse "particularly as the `extent and nature' of [defendant's] allegedly offending use was already known"); Richard Feiner & Co. v. Turner Entm't Co., 98 F.3d 33, 35 (2d Cir.1996); Markowitz Jewelry Co. v. Chapal/Zenray, Inc., 988 F. Supp. 404, 407-08 (S.D.N.Y.1997); Bear U.S.A., Inc. v. A.J. Sheepskin & Leather Outerwear, Inc., 909 F. Supp. 896, 910 (S.D.N.Y.1995) (plaintiff's failure to act sooner, particularly where it previously faced no competition in that market, "undercut the need for and urgency of a preliminary injunction").

74. Nor is Pharmacia's delay excused because of Alcon's asserted bad faith in adopting the Travatan mark. First of all, the Court has determined that there is no merit to Pharmacia's claim that Alcon acted in bad faith. Furthermore, in order to excuse Pharmacia's delay, Pharmacia must demonstrate that Alcon committed "fraudulent infringement," Anheuser-Busch, Inc. v. Du Bois Brewing Co., 175 F.2d 370, 374 (3d Cir.1949), as part of an effort to "foist upon the public the (products) of the defendant as those of the plaintiff." Cuban Cigar Brands N.V. v. Upmann Int'l, Inc., 457 F. Supp. 1090, 1098-99 (S.D.N.Y.1978) (Weinfeld, J.), aff'd, 607 F.2d 995 (2d Cir. 1979); see also Joint Stock Soc'y v. UDV North Am., Inc., 53 F. Supp. 2d 692, 721 (D.Del.1999) (using the "fraudulent infringe[ment]" language); Swanson v. Georgetown Collection, Inc., No. Civ. A No. 94-CV-1283, 1995 WL 72717, at *9 (N.D.N.Y. Feb.14, 1995). For the reasons set forth earlier, no inference of bad faith, much less fraud, is warranted.

75. Because Pharmacia inexcusably delayed raising any objection to the mark until February of 2001, Alcon would be seriously prejudiced by the granting of the present application for a preliminary injunction. Indeed, that Pharmacia sat silently on its rights with full knowledge of Alcon's intent to use the Travatan name gives rise to a separate laches defense, see Conopco, Inc. v. Campbell Soup Co., 95 F.3d 187, 192-93 (2d Cir.1996); Reedco, 667 F.Supp. at 1082; New Dana Perfumes, 131 F.Supp.2d at 627 n. 10; Warner Lambert, 718 F.Supp. at 395; Barton Tr. 48-49. Alcon has changed trade names when timely objections were raised, further weakening Pharmacia's claim on the merits.

V. BALANCING FACTORS

76. In determining whether a preliminary injunction should issue the Court, of course, is required to balance the potential hardships that either side will endure as a result of the injunction. Opticians Ass'n of Am. v. Ind. Opticians of Am., 920 F.2d 187, 197 (3rd Cir.1990). Where, as here, a plaintiff's trademark case involving FDA-approved medicines otherwise is weak, the public interest factor is paramount because an injunction would deprive the public of an important medical benefit. See Reedco, 667 F.Supp. at 1080-81 (refusing to enjoin advance in psoriasis treatment); Hybritech Inc. v. Abbott Labs., 4 U.S.P.Q.2d 1001, 1003-04, 1014-15 (C.D.Cal.1987) (relief denied on public interest grounds despite other factors in plaintiff's favor), aff'd, 849 F.2d 1446 (Fed.Cir.1988); Neuromedical Sys., Inc. v. Neopath, Inc., No. 96 Civ. 5245 (JFK), 1998 WL 264845, at *16 (S.D.N.Y. May 26, 1998) (same); Ethicon Endo-Surgery v. U.S. Surgical Corp., 855 F. Supp. 1500, 1517 (S.D.Ohio 1994) (same).

77. Pharmacia could not substantiate its claim, made at the outset of this case, that Travatan should be enjoined because it poses a greater risk to pregnant women than does Xalatan. Neither pregnancy-associated, nor any other, risks warrant enjoining Travatan, which the FDA has approved as safe and effective.

78. Furthermore, Alcon would be harmed by any injunction, not only in the cost of getting a new name and packaging approved by the FDA, but also in lost sales during the time that its product would be off the market. The harm Alcon would suffer would bear no relation to any benefit it might have derived from the Travatan name, much less from any similarity of Travatan to Xalatan. See Continental Group, Inc. v. Amoco Chem. Corp., 614 F.2d 351, 357 (3d Cir.1980); Times Mirror, 212 F.3d at 168 (harm to defendant relevant to dilution claim). Alcon's legitimate interest in providing an important new drug to the market, under whatever name, would be substantially impaired.

VI. PREVAILING PUBLIC INTEREST

79. In a trademark case, this factor requires the court to consider the right of the consuming public not to be deceived or confused by an infringing mark. Opticians Ass'n, 920 F.2d at 197. Since the plaintiff is unlikely to succeed on its claims of confusion or dilution, and since the FDA has approved the Travatan mark, the court concludes that the public interest would not be served by issuing a preliminary injunction enjoining the use of the mark.

VII. PHARMACIA'S CLAIMS UNDER NEW JERSEY LAW

80. Pharmacia is not entitled to injunctive relief under the laws of New Jersey and the common law. Both parties agree that in the Third Circuit the test for common law infringement and unfair competition is identical to the test for federal infringement and unfair competition. Apollo Distrib. v. Jerry Kurtz Carpet Co., 696 F. Supp. 140, 143 (D.N.J.1988) Similarly, New Jersey's statutory unfair competition law, N.J.S.A. 56:4-1, is equivalent to Section 43(a) of the Lanham Act. Id. at 143; see also J & J Snack Foods Corp. v. Nestle USA, Inc., 149 F. Supp. 2d 136, 157 (D.N.J.2001).

81. N.J.S.A. 56:3-13a also provides that "[t]he interpretation and construction of the federal Trademark Act of 1946 shall be examined as persuasive authority for the interpretation and construction of this 1995 amendatory and supplementary act."

82. Therefore, since the Court has concluded that the plaintiff is unlikely to succeed on the merits of its federal claims, it also concludes that the plaintiff will be unlikely to succeed on the remaining state statutory and common law claims.

83. Pharmacia's dilution claim under N.J.S.A. 56:3-13.20 is now barred as a matter of law because the PTO has issued Travatan a federal registration. 15 U.S.C. § 1125(c)(3); see also Viacom Inc. v. Ingram Enters., Inc., 141 F.3d 886, 888 (8th Cir.1998) ("[T]he FTDA also provides a complete defense to state law dilution claims to owners of valid, federally registered marks."); Westchester Media Co. v. PRL USA Holdings, Inc., 103 F. Supp. 2d 935, 976-77 (S.D.Tex.1999), aff'd on relevant grounds, 214 F.3d 658 (5th Cir.2000) (applying statute to preclude state law dilution claim because mark at issue is federally registered).

CONCLUSION

For the reasons discussed, the plaintiff's motion for a preliminary injunction is denied.

NOTES

[1] In assessing the credibility of each witness in this case, the Court has taken into consideration how well each witness was able to recall and describe the things testified to, the manner of the witness while testifying, whether the witness had an interest in the outcome of the case or any bias or prejudice concerning any party or matter involved in the case, how reasonable the witness's testimony was considered in light of all the evidence in the case. 9A Charles A. Wright & Arthur R. Miller, Federal Practice and Procedure, § 2585 (2d. ed. 1995); see also Miller v. Mercy Hospital, Inc., 720 F.2d 356, 365 (4th Cir.), 470 U.S. 1083, 105 S. Ct. 1841, 85 L. Ed. 2d 141 (1985). ("Credibility involves more than a witness's demeanor and comprehends an overall evaluation of testimony in light of its rationality or internal consistency and the manner in which it hangs together with other evidence" and whether the witness's testimony was contradicted by what that witness had said or done at another time, by the testimony of other witnesses, or by other evidence.)

[2] OPDRA "evaluated" Dr. Lambert's model and decided not to use it as part of their separate review and screening processes. 12/18 Tr., at 192 (Lambert); Lambert Tr., at 259-60. Professor Lambert was "unaware" of any instance in which OPDRA used his mathematical model in reaching a decision as to confusing similarity. Lambert Tr., at 95-96. The Institute for Safe Medicine Practices, which had used it, stopped doing so because of a business dispute. 12/18 Tr., at 229 (Lambert).

[3] In fact, the second group was created by Dr. Lambert by simply taking names from the first group of Known Error Pairs and shuffling them to create 1,127 other pairs. Def. Ex. K (DX01573) (Bruce L. Lambert et al., "Similarity as a Risk Factor in Drug Name Confusion Errors," 37 Medical Care 1214 at 1216 (1999)).

[4] Professor George P. McCabe is head of the Statistical Consulting Service and Professor of Statistics at Purdue University. Def. Ex. G (DX00521) (McCabe Decl., ¶ 1); Id. (DX00531-48) (curriculum vitae of Professor McCabe). Professor McCabe has served as the Chair of the Statistical Consulting Section of the American Statistical Association and as a consultant on statistical issues to the FDA and the Department of Veterans Affairs. Id.; 12/20 Tr., at 69 (McCabe).

[5] (14,800,000 - 1,127) ÷ 14,800,000 = 0.XXXXXXXXX, or more than 99%.

[6] Dr. Lambert recanted the 300 million figure at the hearing. 12/18 Tr. 173-74, 200-02.

[7] Dr. Lambert suggests that at a prevalence of 10%, the positive predictive value of his model should be 70%. 12/18 Tr. 175 (Lambert). 70% is not the correct figure, however. The 70% figure is based on a "three-variable model" that Dr. Lambert did not use in this case. 12/19-221, 223-24 (McCabe). Even if the positive predictive value is 70%, there remains almost a one in three chance that his model is wrong. 12/18 Tr., at 177 (Lambert).

[8] (4,000 × 3,999) ÷ 2 = 7,998,000, using Dr. Lambert's formula. Def. Ex. K (DX01556) (Lambert 1st Del. Ex. 2).

[9] The Lumigan control cell was designed to measure "noise" — that is, the percentage of respondents who associated Travatan with Xalatan for reasons other than the trademark. To determine the true trademark dilution percentage, the percentage of responses coded as dilution in the control cell is subtracted from the percentage of dilution responses in the test cell. Def. Ex. S (DX02141-42) (Diamond Dep. Ex. 192, at 250-51) (Federal Judicial Center, Reference Manual on Scientific Research).

[10] See also Def. Ex. VV (taped Diamond Survey test cell response 8800) (respondent comments "Secret ballot, right?" while opening envelope); id. (taped control cell response 8864) ("Can I open the envelope?"); id. (test cell response 17008) ("Have not opened them yet per the instructions."); id. (control cell response 8724) ("This is like mission impossible ... you open it up on the spot ... does it self-destruct in the end? ... I'm opening it now."); id. (test cell response 8861) ("I'm ripping it open."); id. (test cell response 22878) ("I hadn't opened anything yet."); id. (control cell response 8632) ("Please do not open until we contact you, right? ... Can I open it?").

[11] To the extent the possibility of confusion standard has any continued application (outside the Third Circuit), it is limited to situations, unlike this one, where confusion between products could have "disastrous" or "dire" consequences. 3 McCarthy § 23:32, at 23-96 (4th ed. 2001) ("If preparation XYLON and CILAN are made to be used for different types of medical treatment, the results could be catastrophic"); Morgenstern, 253 F.2d at 393 (different type of drugs); American Cyanamid Corp. v. Connaught Labs., Inc., 800 F.2d 306, 309 (2d Cir.1986) (refusing to apply Morgenstern where drug products were for same condition).

[12] A court may consider the following eight non-exclusive factors in determining the fame of a mark: (1) the degree of inherent or acquired distinctiveness of the mark; (2) the duration and extent of use of the mark in connection with the goods or services with which the mark is used; (3) the duration and extent of advertising and publicity of the mark; (4) the geographical extent of the trading area in which the mark is used; (5) the channels of trade for the goods and services with which the mark is used; (6) the degree of recognition of the mark in the trading areas and channels of trade used by the marks' owner and the person against whom the injunction is sought; (7) the nature and extent of use of the same or similar marks by third parties; and (8) whether the mark was registered. Times Mirror, 212 F.3d at 163. In light of Alcon's acknowledgment that the Xalatan mark is famous among opthamologists, the court does not address these factors to determine whether the mark is famous. ("Alcon does not dispute that Xalatan is a famous mark among opthalmologists." See Alcon's Response to Plaintiffs' Post-Hearing Proposed Findings of Fact and Conclusions of Law at p. 398.)

[13] See Melanie M. Routh, Trademark Dilution and the Effect of the Federal Trademark Dilution Act, 50 Rutgers L.Rev. 253 (1997) and David S. Welkowitz, Reexamining Trademark Dilution, 44 Vand. L.Rev. 531 (1991).

[14] STEMATAN, SARATAN, PENATAN, PHENIRATAN, SAVATAN, RYNATAN, and GERMA-TAN.

[15] Moreover, Pharmacia's use of Lumigan as a control in its dilution survey confirms that it claims no exclusive rights in the "AN" portion of the suffix.

[16] Although Mr. McCullough reported 15% dilution in his initial report, Def. Ex. K (DX01591) (McCullough Report at 3); McCullough Tr. 134, he admitted that the 15% figure was rounded up from 14.52%, id., and later conceded that one of the dilution responses was miscoded, id. at 292-93; 12/19 Tr., at 7-8, 26-27 (McCullough), which brought the dilution level he reports down to 14%.

[17] Dilution does not implicate any public interest against consumer deception because, by definition, it protects only a trademark owner's private interest. TCPIP Holding Co. v. Haar Communications Inc., 244 F.3d 88, 95 (2d Cir.2001); I.P. Lund, 163 F.3d at 53 ("The traditional interest in trademark protection is stretched very thin in dilution cases where confusion is absent and a professed aim is to protect the maker's investment in the trademark."); Edward E. Vassallo & Maryanne Dickey, Protection in the United States for "Famous Marks," 9 Fordham Intell. Prop. Media & Ent. L.J. 503, 505 (1999), citing, Genovese Drug Stores, Inc., 939 F.Supp. at 340 and I.P. Lund Trading v. Kohler Co., 11 F. Supp. 2d 112 (D.Mass.1998).

[18] See also Nina Ricci, S.A.R.L. v. Gemcraft Ltd., 612 F. Supp. 1520, 1530-31 (S.D.N.Y. 1985) (12 weeks); Bear U.S.A., Inc. v. A.J. Sheepskin & Leather Outerwear, Inc., 909 F. Supp. 896, 910 (S.D.N.Y.1995) (3 months); Gidatex, S.R.L. v. Campaniello Imports, Ltd., 13 F. Supp. 2d 417, 420 (S.D.N.Y.1998) (at least 4½ months); Krueger Int'l, Inc. v. Nightingale Inc., 915 F. Supp. 595, 613 (S.D.N.Y. 1996) (9 months); Ivy Mar Co., Inc. v. C.R. Seasons Ltd., 907 F. Supp. 547, 563-64 (E.D.N.Y.1995) (10 months).