New York Times Co. v. Roxbury Data Interface, Inc., 434 F. Supp. 217 (D.N.J. 1977)
May 3, 1977
ROXBURY DATA INTERFACE, INC., Byron A. Falk, Jr., and Valerie R. Falk, Defendants.
United States District Court, D. New Jersey.
*218 Lankenau, Kovner & Bickford by John C. Lankenau, Robert C. Bickford, New York City, Schwartz, Steinberg, Tobia & Stanziale by Janet J. Burak, East Orange, N.J., for plaintiffs.
Cowan, Liebowitz & Latman by Carol Faye Simkin, New York City, Edward R. Weingram, Newark, N.J., for defendants.
MEANOR, District Judge.
Plaintiffs, holders of certain copyrights, move under Fed.R.Civ.P. 65(a) for a preliminary injunction pursuant to § 112 of the Copyright Act of 1909, 17 U.S.C. § 112 (1970), to halt alleged infringement of their copyrights. The suit presents an interesting and seemingly novel question regarding the extent of the proprietary rights inhering in a copyright. The issue is whether a copyrighted work may be indexed by an outsider without the permission of the holders of the copyrights to the original work.
Plaintiffs publish the New York Times Index annually. As indicated by its name, the Times Index essentially correlates data which has appeared in The New York Times in a given year with citations to the pages and columns of The New York Times on which the data appears. The data is denominated by various headings in the Times Index. These headings include the names of people, places, institutions, events, and the general headings which cover broad subject matters, e. g., "fires." The Times Index has been published annually, with a few hiatuses, since 1863, and covers the years from 1851 to the present.
In addition to correlating data with page citations to The New York Times, the Times Index contains abstracts of certain Times articles. The Times Index is a publication of substantial size. For example, the 1972 Index consists of over 2500 small-margined pages of fine print. The external dimensions of the volume are 12 inches by 9½ inches by 3½ inches. Similarly, volume 1 of the two-volume 1933 Index contains almost 1500 pages of closely packed fine print. The external dimensions of the first volume of the 1933 Index are 10 inches by 7 inches by 2¾ inches.
In addition to their annual index, plaintiffs publish four cumulative indexes to The New York Times. The New York Times Obituaries Index, published in 1970, lists the names of persons whose obituary has appeared in The New York Times between 1858 and 1968. The New York Times Book *219 Review Index is a cumulative index to material published in the weekly New York Times Book Review since 1896. The Book Review Index contains the names of the authors of reviewed books and of the reviewers. The New York Times Film Review Index is a cumulative index to film reviews published in The New York Times since 1913. This index contains the names of many directors, actors, screenwriters, producers and others involved in filmmaking. The New York Times Theater Review Index is a multivolume, cumulative index to theater reviews published in The New York Times since 1870. The Theater Review Index contains the names of playwrights, actors, directors, producers and others associated with the stage.
Plaintiffs also operate a computer data bank which consists of a subject index to The New York Times from 1969 to the present. Plaintiffs state that access to the data in the computer is obtained through certain "key words," of which approximately 350,000 are personal names. Plaintiffs sell subscriptions to the computer data bank and collect a fee based on use.
Plaintiffs allege that they hold valid copyrights on all of the annual Times Indexes and the cumulative indexes, and on the contents of the computer data bank. Plaintiffs assert that the Times Indexes are protected both as books and as compilations. See 17 U.S.C. §§ 5(a) & 7 (1970). Although defendants have indicated that they may challenge the validity of at least some of these copyrights, the court will assume the validity of these copyrights for the purposes of this motion.
Defendants are in the process of publishing a 22-volume personal name index to the annual New York Times Index. Two of the 22 volumes have been published, covering names beginning with the letter "A" through "Blo." Defendants compiled their index by perusing all of the annual New York Times Indexes covering the years 1851 to 1974 and culling every personal name that appeared as a heading. The culled names appear in defendants' index in alphabetical order. After each name appears citations to pages of the New York Times Index on which the name appears. Some names are followed by the year of death, certain titles held by the person, and years of birth and death; most, if not all, of this information was taken from the Times Index. Physically, each of the first two volumes of the set is approximately ten inches long, seven inches wide, and an inch thick. Each volume is covered with brown cloth and has gold lettering on the front and binding. The title of defendants' index, as printed on the cover, reads, "PERSONAL/NAME INDEX/TO `THE NEW YORK TIMES INDEX'/1851-1974." The title is divided into four lines, as indicated by the slashes above. The words "PERSONAL NAME INDEX" appear in letters 3/8 of an inch high; "TO `THE NEW YORK TIMES INDEX'" is written in 3/16 inch letters; and the years appear in 3/32 inch type. Approximately 4¾ inches below the title on the cover of all but the first volume will appear the following phrase in 1/8 inch gold lettering: "AN INDEPENDENT WORK NOT PUBLISHED OR APPROVED BY THE NEW YORK TIMES." On the binding, an abbreviated title appears: PERSONAL/NAME/INDEX/1851/-/1974. The binding also contains the volume number, an indication of the names appearing in the volume, and the logotype of Roxbury Data Interface, Inc.
Defendants compiled the first volume of their index in the following manner. Defendants hired several people to peruse each volume of the Times Index for the years 1851 to 1974, and to extract every personal name beginning with the letter "A." In addition to the name itself, the employees were instructed to copy any pseudonyms, nicknames, titles, years of birth and death, and to note the page and year of the Times Index on which the name appeared. Each extract was recorded on a three by five inch card. Defendants state that because the Times Index does not always list as a separate heading certain personal names which appear under certain subject matter headings, e. g., murders, defendants' employees were instructed to peruse certain subject matter headings for names beginning with *220 the letter "A," as well as the "A" section of each Times Index. Defendants then sorted into alphabetical order the names taken from each Times Index. The resulting 120 stacks of cards then were sorted into five alphabetized stacks, which finally were sorted by the two named authors into the lists that were published. During each of these sortings, defendants made some effort to standardize variant spellings of names where no confusion would result, list separately different people with the same names, and consolidate listings of single individuals appearing under different names. The extent of these efforts is disputed and unclear; it is clear that defendants' efforts failed in a number of cases and an appropriate warning to users appears in the Introduction to the second volume. Defendants also transformed the personal names which appear in all capital letters in the Times Indexes to capital and small letters in their index. The first two volumes contain approximately 240,000 names. The complete set is expected to list over 3,000,000 names. Editorial costs for compilation of the entire set will be $213,000.00, according to defendants' projection. This projection does not include the cost of typesetting and publishing the index. Defendants plan to sell the 22 volumes at a price of $600.00 to $650.00 per set.
Plaintiffs contend that the copying of these names constitutes a prima facie case of copyright infringement, entitling plaintiffs to a preliminary injunction. Defendants point out that it would be impossible for them to publish a personal name index to the New York Times Index without copying the names from the Times Index. In order to obtain a preliminary injunction, plaintiffs must show "a strong likelihood" of prevailing on the merits. Kontes Glass Co. v. Lab Glass, Inc., 373 F.2d 319, 321 (3d Cir. 1967). While the general rule in alleged infringement of creative works seems to presume irreparable harm from a copyright holder's showing of a clear probability of success on the merits, see American Metropolitan Enterprises, Inc. v. Warner Brothers Records, Inc., 389 F.2d 903, 905 (2d Cir. 1968), the most recent Third Circuit opinion indicates that plaintiffs may be required to make some showing of irreparable harm. Kontes Glass Co., supra, at 320.
The primary function of the Times Index is the correlation of data with citations to the pages and columns of The New York Times on which the data appears. This correlation constitutes the substance of plaintiffs' copyrights. See Kane v. Pennsylvania Broadcasting Co., 73 F. Supp. 307, 307 (E.D.Pa.1947). Cf. Affiliated Hospital Products, Inc. v. Merdel Game Mfg. Co., 513 F.2d 1183, 1188 (2d Cir. 1975) (copyright on rule book only protects arrangement of rules); Benny v. Loew's Inc., 239 F.2d 532, 536 (9th Cir. 1956), aff'd by an equally divided court, 356 U.S. 43, 78 S. Ct. 667, 2 L. Ed. 2d 583 (1958) (copyright on screenplay protects arrangement of events); Triangle Publications, Inc. v. Sports Eye, Inc., 415 F. Supp. 682, 686 (E.D.Pa.1976) (only form for expressing horse racing data is copyrightable). If defendants had copied both the personal names and the correlated citations to the pages of The New York Times from the Times Index, plaintiffs would appear to have a strong claim to infringement of their copyrights. See, e. g., Leon v. Pacific Tel. & Tel. Co., 91 F.2d 484 (9th Cir. 1937); Southwestern Bell Tel. Co. v. Nationwide Independent Directory Serv., Inc., 371 F. Supp. 900 (W.D.Ark.1974). Defendants, however, copied only the names appearing in the Times Index and correlated those names to citations of the multivolume New York Times Index on which the names appear.
Plaintiffs take the position that their copyrights extend beyond the correlation of data to page citations to The New York Times and encompass the effort entailed in extracting the personal names and other data from The New York Times. While plaintiffs could not copyright the personal names, they clearly could obtain a copyright on a compilation of names appearing in The New York Times. A question arises, however, whether or not millions of names scattered over more than one hundred volumes and integrated with a great mass of other data can qualify as a compilation. In addition, *221 while it is clear to this court that each personal name and its correlated citation to The New York Times would qualify as a unit for copyright protection under 17 U.S.C. § 3 (1970), which protects component parts of copyrighted works, it is not clear that either the personal name or the citation to The New York Times, standing alone, would qualify for such protection.
Assuming without deciding that plaintiffs' copyrights on their annual indexes encompass the use of the indexed personal names, the court must consider whether or not defendants' admitted copying from plaintiffs' indexes constitutes a "fair use" of that material. The "fair use" doctrine is a judicially created limitation on the exclusive right of copyright holders to use their copyrighted materials. In determining whether a use constitutes a fair use or a copyright infringement, courts have considered four factors:
(1) the purpose and character of the use;
(2) the nature of the copyrighted work;
(3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
(4) the effect of the use upon the potential market for or value of the copyrighted work.
The fair use doctrine was codified in section 107 of Congress' recent general revision of the copyright law, which will become effective January 1, 1978. Section 107 of the new statute contains the four factors listed above. Although the statute is not effective until next January, the guidelines set out therein are useful since they represent a codification of the case law. H.R.Rep. No. 1476, 94th Cong., 2d Sess. 65 (1976), U.S.Code Cong. & Admin. News 1976, p. 5659. See, e. g., Marvin Worth Products v. Superior Films Corp., 319 F. Supp. 1269 (S.D.N.Y.1970); Annot., 23 A.L.R.3d 139 (1969).
(1) The purpose and character of the use is at least twofold: (1) to make money for the defendants, and (2) to facilitate access by the public to information concerning certain individuals. While whether the alleged infringer seeks financial gain from its use is a factor to be considered, the fact that defendants seek to profit financially will not preclude their use from being a fair use. Rosemont Enterprises, Inc. v. Random House, Inc., 366 F.2d 303, 307-08 (2d Cir. 1966), cert. denied, 385 U.S. 1009, 87 S. Ct. 714, 17 L. Ed. 2d 546 (1967). It seems likely that defendants' index will serve the public interest in the dissemination of information. Without defendants' index, an individual seeking to find articles which appeared in The New York Times on a certain person whose career spanned, say, forty years, would be compelled to search through forty volumes of the Times Index. Using defendants' index, the researcher would discover immediately the pages and volumes of the Times Index on which the name of his subject appears. Armed with this information, the researcher then can proceed to a few of the forty potentially relevant volumes of the Times Index, from which he will be directed to the pages and columns of The New York Times itself. On its face, defendants' index appears to have the potential to save researchers a considerable amount of time and, thus, facilitate the public interest in the dissemination of information. Defendants have placed in the record affidavits from a librarian and a professor of library science in support of this observation.
(2) The nature of the Times Index does not require extensive comment, except to note that it is rather in the nature of a collection of facts than in the nature of a creative or imaginative work. Since the Times Index is a work more of diligence than of originality or inventiveness, defendants have greater license to use portions of the Times Index under the fair use doctrine than they would have if a creative work had been involved. See Benny v. Loew's Inc., supra, at 536; Gardner v. Nizer, 391 F. Supp. 940, 943 modified on other grounds, 396 F. Supp. 63 (S.D.N.Y.1975).
(3) Plaintiffs assert that defendants' index quantitatively uses approximately fifty percent of the headings in the Times Index *222 while defendants prefer to categorize the use as constituting eight percent of the words in the Times Index. This semantic difference refers to substantially the same quantum of use. Although the quantum of material taken from the copyrighted work must be considered, a bald quantitative approach is not determinative. While the taking of as little as three sentences has been held to constitute an infringement, Henry Holt & Co. v. Liggett & Myers Tobacco Co., 23 F. Supp. 302, 303 (E.D.Pa.1938), the copying of an entire work has been held to be a fair use, Williams & Wilkens Co. v. United States, 487 F.2d 1345, 1348 (Ct.Cl.1973), aff'd by an equally divided court, 420 U.S. 376, 95 S. Ct. 1344, 43 L. Ed. 2d 264 (1975). The inadequacy of a pure quantitative approach obviously results from the fact that the quantum of appropriated material must be interrelated with a number of other factors, such as the quality and substantiality of the appropriated material, the need to use the appropriated material in the manufacture of defendants' product, and the other factors discussed herein.
As stated above, the substance of plaintiffs' copyrights, as far as they pertain to personal names, covers the correlation of personal names with citations to the pages and columns of The New York Times. Defendants have not copied any of these correlations, but, rather, have taken only the personal names appearing in the Times Indexes; defendants, of course, also have copied the volume and page numbers of the Times Indexes where the personal names appear. Plaintiffs place much reliance on cases involving the appropriation of telephone and other directory listings. E.g., Leon v. Pacific Tel. & Tel. Co., supra; Cincinnati & Suburban Bell Tel. Co. v. Brown, 44 F.2d 631 (S.D.Ohio 1930); Southern Bell Tel. & Tel. Co. v. Donnelly, 35 F. Supp. 425 (S.D.Fla.1940); Sammons v. Larkin, 38 F. Supp. 649 (D.Mass.1940), modified on other grounds, 126 F.2d 341 (1st Cir. 1942); American Travel & Hotel Directory Co. v. Gehrig Publishing Co., 4 F.2d 415 (S.D.N.Y. 1925); Hartford Printing Co. v. Hartford Directory & Publishing Co., 146 F. 332, damages awarded, 148 F. 470 (C.C.D.Conn. 1906); Social Register Association v. Murphy, 129 F. 148 (C.C.D.R.I.1904). Each of these cases involved a defendant who copied both the name and the correlated data which formed the essence of the plaintiff's work and then sought to publish this copied material in its entirety as a new and virtually identical or very similar directory in direct competition with the plaintiff's directory. An analogous situation might exist in the instant suit if defendants had copied the personal names and the correlated citations to The New York Times. Defendants in the instant suit neither have copied the correlated data along with the personal names nor seek to publish a virtually identical index which would be in direct competition with the Times Index. In addition, all of the cases relied on by plaintiffs involved one-volume works devoted exclusively to the listing of the copied names and in which names were arranged in an order, usually alphabetical; defendants copied not only the names and correlated data, but also, by and large, the order in which the names were listed. The instant case involves well *223 over one hundred volumes in which millions of personal names are listed among millions of other listings. Defendants did not copy a list, as defendants in the above cited cases had done. Defendants rather extracted the names from the multivolume Times Index and compiled their own alphabetical listing. While this distinction is not determinative, it must be noted. The above cited cases, therefore, do not compel a finding of a prima facie infringement of plaintiffs' copyrights in the instant case.
Under this third factor, the court also must consider the extent to which defendants' index requires the copying of the names from the Times Index. See Marvin Worth Products v. Superior Films Corp., supra, at 1274. For all practical purposes, defendants could not publish their personal name index without copying the names from the Times Index. Plaintiffs indicate that defendants theoretically could have obtained the same information directly from The New York Times, as had the compilers of the Times Index. If defendants had copied from the Times Index both the personal names and the correlated citations to The New York Times, plaintiffs' argument would have merit because defendants would have produced an abridgement or other version of the Times Index without expending efforts equal to the compilers of the Times Index. See, e.g., 17 U.S.C. § 1(b) (1970) and cases cited at 222, supra. Since defendants have not produced an abbreviated index to The New York Times in direct competition with the Times Index, however, plaintiffs' position is not convincing.
Plaintiffs finally argue that defendants' use cannot be a fair use because defendants have failed to exercise any independent effort aside from the effort entailed in "copying" the personal names. Plaintiffs point out that defendants' index consists only of names extracted from the Times Index, with virtually no information added by defendants. In light of the nature of the work, an index, this circumstance is not surprising. As outlined supra at 219, defendants are expending substantial effort and money in culling the millions of personal names from over 100 volumes of the Times Index, alphabetizing those names, and editing the entries into their final form. This independent work is precisely the labor that an original indexer must undertake.
(4) The discussion now turns to the effect of the personal name index on the potential market for the Times Index. In the area of copyright law dealing with compilations and directories, this factor of competitive effect looms large.
The decisions [involving directories and similar compilations] recognize or suggest that the competitive or noncompetitive character of the publication making the claimed infringing use may be of particular importance. Thus, the courts have indicated that the uses which a noncompetitor, or even one who is not in direct competition with the copyright proprietor or whose work will not injure the potential market for the copyrighted material, may make of the copyrighted directory are greater or less limited than the uses allowed to a direct competitor.
Annot., 23 A.L.R.3d 139, 252 (1969), citing Sampson & Murdock Co. v. Seaver-Radford Co., 140 F. 539 (1st Cir. 1905); Sub-Contractors Register, Inc. v. McGovern's Contractors & Builders Manual, Inc., 69 F. Supp. 507 (S.D.N.Y.1946); Consumers Union v. Hobart Mfg. Co., 189 F. Supp. 275 (S.D.N.Y. 1960); Social Register Association v. Murphy, 128 F. 116 (C.C.D.R.I.1904). See Kane v. Pennsylvania Broadcasting Co., supra, at 307.
The effect of defendants' personal name index on the market for the Times Index appears slight. Since defendants' index carries citations only to the New York Times Index, defendants' index is useless unless its user has access to the Times Index, from which he will be directed to the articles appearing in The New York Times. *224 Despite the absolute dependence of defendants' index on the Times Index, plaintiffs allege potential and actual harm to the sales of the Times Index flowing from the publication of the personal name index. Plaintiffs appear to concede that the market for defendants' personal name index will be, for the most part, libraries which have many volumes of the Times Index. Plaintiffs point out, however, that many of the potential purchasers of defendants' index will own many volumes, but not complete sets, of the Times Index. Plaintiffs argue that libraries, given their limited funds allotted to the purchase of new books and reference materials, may choose to procure defendants' index in preference to purchasing back issues of the Times Index. The principal weakness of this argument derives from the essential difference in function of the two works. If a librarian wishes to obtain a more complete reference to The New York Times, he will purchase back issues of the Times Index. On the other hand, if the librarian wants to obtain a handy reference to the personal names appearing in the Times Index, he will purchase defendants' index. The two works clearly do not compete directly in the sense that telephone, city or social directories do. That is to say, purchase of defendants' index in no way supersedes the need for plaintiffs' index; indeed, both the viability and the utility of the personal name index depends entirely on the Times Index. Defendants' index competes for library dollars with the Times Index no more directly than does an encyclopedia, a dictionary, or any other publication. The publication of every new book no doubt increases competition for library dollars. This competition, however, does not amount to the direct competitive effect which would preclude defendants' use from being a fair use.
Plaintiffs also allege that their computer data bank and some of their other cumulative indexes, particularly the Obituaries Index, will suffer competitive harm from the publication of defendants' index. Plaintiffs, however, do not allege infringement of any of their copyrights on the cumulative indexes or the computer data bank; plaintiffs claim infringements only of their copyrights on the annual Times Indexes. Plaintiffs cite no authority for the proposition that potential competitive injury to noninfringed publications of the copyright holder must be taken into account in determining whether an allegedly infringing use is a fair use. The newly enacted general revision of the copyright law, as well as the case law which it codifies, instruct the court to weigh "the effect of the use upon the potential market for or value of the copyrighted work" from which the material was taken. Since none of the personal names were extracted from the computer data bank or from any of the cumulative indexes, this court cannot consider the possible competitive injury to these works in determining whether defendants' use of the annual Times Index is a fair use.
Plaintiffs also argue that publication of defendants' index deprives them of the right to exploit their copyrights. Since nothing stands in the way of plaintiffs' publishing a work to rival defendants' index, plaintiffs actually are arguing that publication of defendants' index will have an adverse effect on any personal name index to the Times Index which plaintiffs may wish to publish in the future. This argument is merely a restatement of plaintiffs' argument that a copyrighted work cannot be indexed without permission of the holders of the copyright to the original work. Plaintiffs seek to analogize themselves to a writer of a play or novel and defendants to a filmmaker seeking to make a movie from the copyrighted play or novel. This analogy is not convincing. First, defendants have not copied the substance of *225 plaintiffs' copyrights the correlation of personal names and other data with citations to the pages of The New York Times. If defendants had published microfilm of the Times Index, defendants would have infringed plaintiffs' copyrights and plaintiffs' analogy would be well taken. Second, defendants' index is not another version of plaintiffs' index, but a work with a different function and form. Third, no creative idea of plaintiffs' has been appropriated by defendants. Finally, no list comparable to defendants' index can be found in the Times Index; the personal names are scattered over more than one hundred volumes.
Plaintiffs place much reliance on Leon v. Pacific Tel. & Tel. Co., supra, and dicta in Reed v. Holliday, 19 F. 325 (C.C.W.D.Pa. 1884). In Leon, defendants copied the names and telephone numbers from two 1935 telephone directories for San Francisco and other communities in the Bay area. Defendants then inverted the correlation of name and telephone number and published a telephone directory arranged in order of the telephone number. The court held that defendants' directories infringed plaintiff's copyrights because defendants' works were mere rearrangements of plaintiff's directories. The apparent rationale behind the court's decision was that defendants' work, like plaintiff's, was a telephone directory to the Bay area and defendants had published their directory, use of which could supersede use of plaintiff's directory, without expending the same degree of work as had plaintiff. See id., 91 F.2d at 484-85. While the opinion mentions a contention by defendants that their directory served a purpose different from plaintiffs, it is obvious that the function of the two directories, correlation of names of telephone subscribers with their telephone numbers, was overlapping to a large degree, if not virtually identical. See id. at 486.
In Reed, plaintiffs published an English language textbook. The textbook instructed readers to analyze sentences by diagraming them. Each lesson in the textbook concluded with a list of sentences to be diagramed by the reader or student. Defendant published a "Teacher's Manual" to the textbook in which all the sentences appearing in the textbook were diagramed. The court found that the sentences were a material and substantial part of plaintiffs' work, and held that plaintiffs' copyright had been infringed. The court went on in dicta, which is quoted in plaintiffs' brief, to equate the protection of a copyright to that of a patent. Whatever the stage of the development of the fair use doctrine in 1884, it is clear today that the existence of that doctrine creates the principal difference between the protection afforded by a copyright and that given by a patent. The existence of the fair use doctrine means that copyrights, unlike patents, do not confer the exclusive right to the use of the protected work. See, e. g., Loew's Inc. v. Columbia Broadcasting System, Inc., 131 F. Supp. 165, 174 (S.D.Cal.1955), aff'd sub nom. Benny v. Loew's Inc., supra.
Plaintiffs state that they are considering publication of a personal name index. Evidence in the record indicates that one of the individual defendants, during his employment by plaintiffs, proposed creation of a personal name index in 1971. In 1973, the same defendant proposed that his corporation, Roxbury Data Interface, Inc., and plaintiffs enter into an agreement for publication of a personal name index. Both proposals were rejected on the grounds that the idea was intrinsically poor and of dubious profitability. The author of these two rejections testified by deposition that he had no information which would cause him to change his prior estimates as to the small market for a personal name index to the Times Index. There is no compelling evidence that plaintiffs will publish a personal name index. It thus appears likely that prohibition of defendants' publication not only will extend plaintiffs' copyrights expansively, but also will impede the public interest in the dissemination of information. In addition, plaintiffs, should they decide to publish a personal name index, are free to do what defendants probably cannot do compile a personal name index based on the Times Index which can refer users directly to The New York Times, although business *226 judgment may direct them not to publish such an index.
Plaintiffs finally argue that appearance of the words "The New York Times" in the title of defendants' index will lead to confusion as to the publisher of the personal name index and will damage the reputation of The New York Times Company. Defendants, however, clearly have a right to label their product correctly. See Societe Comptoir v. Alexander's Dep't Stores, 299 F.2d 33, 36 (2d Cir. 1962). Furthermore, defendants have attempted to dispel any confusion as to the publisher of the personal name index. In labeling their index, defendants have refrained from using the black letter type utilized in the masthead of The New York Times and in the titles of all the New York Times annual and cumulative indexes. In addition, all but the first volume of defendants' index will carry the following notice on their covers: "AN INDEPENDENT WORK NOT PUBLISHED OR APPROVED BY THE NEW YORK TIMES."
Copyrights on indexes have generated little litigation. Neither plaintiffs nor defendants have cited any cases involving indexes and the court has found only one such case. In Kipling v. G. P. Putnam's Sons, 120 F. 631, 57 C.C.A. 295 (2d Cir. 1903), defendants purchased from plaintiff's publishers approximately two hundred copyrighted prose works and several poems. Defendants bound these works in fourteen volumes of a fifteen volume set. The fifteenth volume consisted of a critical commentary on the plaintiff's works and an index, compiled by defendants, to the copyrighted works. The Second Circuit held that defendants' index to the copyrighted works of plaintiff did not infringe plaintiff's copyright "even though the index, as it necessarily must, contains words and phrases found in the text." Id. at 635. See Loew's Inc. v. Columbia Broadcasting System, Inc., supra, at 175.
In Universal Athletic Sales Co. v. Salkeld, 511 F.2d 904 (3d Cir.), cert. denied, 423 U.S. 863, 96 S. Ct. 122, 46 L. Ed. 2d 92 (1975), the court stated:
To establish a copyright infringement, the holder must first prove that the defendant has copied the protected work and, second, that there is a substantial similarity between the two works. The criterion for the latter requirement is whether an ordinary lay observer would detect a substantial similarity between the works. . . . Phrased in an alternative fashion, it must be shown that copying went so far as to constitute improper appropriation, the test being the response of the ordinary lay person.
511 F.2d at 907 (citations omitted). While defendants admittedly have copied the personal names from plaintiffs' indexes, the personal name index differs substantially from the Times Index, in form, arrangement, and function.
In fine, application of the four factors discussed above reveals:
(1) that one purpose of the personal name index is to obtain financial gain for its publishers;
(2) that the personal name index will serve the public interest in the dissemination of information;
(3) that the Times Index is a work more of diligence and fact than of creativity and imagination;
(4) that virtually all of the information appearing in the personal name index was taken from the Times Index;
(5) that defendants expended substantial time, effort and money in compiling their index to the Times Index;
(6) that defendants did not copy the correlation of the personal name index with citations to The New York Times in any case;
(7) that the very nature of the personal name index required that defendants copy the names which they were indexing;
(8) that the effect of the personal name index on the potential market for the Times Index appears slight or nonexistent.
Placing these considerations into the variable equation of fair use indicates that plaintiffs have failed to demonstrate a clear *227 likelihood of success on the merits which would warrant the granting of their motion for a preliminary injunction. This conclusion obviates discussion both of the necessity of plaintiffs showing irreparable harm and of defendants' challenge to the validity of some or all of plaintiffs' copyrights. Plaintiffs' motion for a preliminary injunction is denied. Counsel for defendants will submit an appropriate order with consent as to form within 10 days.NOTES
 Plaintiffs subsequently have compiled and published Times Indexes for those years during which no contemporaneous Times Indexes had been compiled.
 In a few of the cases relied upon by plaintiffs, i.e., those involving social directories, the selection of the name itself may have assumed greater importance than it does in compilations of fact. For example, the compilation of a social directory involves the compiler's subjective judgment of who within a community possesses a social status sufficient to require listing in his directory. Each name in a social directory is a product of the compiler's judgment rather than the publication of a fact. In contrast, the compilation of a telephone directory requires the compiler only to ascertain the fact of telephone subscribership; the compiler is not required to make any significant subjective judgment. Similarly, the compilation of a personal name index does not require the compiler to make the same type of subjective judgments as to what names to include as does the compilation of a social directory.
 In some of the competitive telephone directory cases, the copied lists were rearranged from alphabetical to numerical order. E.g., Leon v. Pacific Tel. & Tel. Co., supra. In that case, however, plaintiff's entire telephone directory was copied and rearranged, offering precisely the same correlation of names and telephone numbers, except that defendants had inverted that correlation and deleted addresses, as appeared in plaintiff's directory.
 Plaintiffs also deny defendants' right to compile an index to The New York Times, presumably because defendants would be copying words from a copyrighted work and unfairly exploiting The New York Times' copyright.
 Only if libraries generally allotted a specified portion of their annual book-buying funds to purchase of volumes of the Times Index and decided to use a portion of that Times Index fund for the procurement of the personal name index might plaintiffs' argument have merit. Even in this situation, however, plaintiffs would have to show that funds allotted to purchase of the Times Index never have been or are likely to be in the future diverted to the purchase of other works.