MACK, Circuit Judge.
Injunction, surrender of plates, accounting of profits, and damages for infringement of copyrights were sought by and denied to plaintiff, appellant. Defendant's one-volume work, Throckmorton's Code of Ohio, published in 1921, with appendix thereto of 1923, is alleged to be an infringement of plaintiff's copyrights of Page & Adams' seven-volume Annotated General Code (1912) and four-volume supplement (1916), of the Topical Index to the official General Code (1910), of the unannotated first three of the four-volume Page & Adams Compact Code (1920), and of volumes 9 and 10 of Page's Ohio Digest (1915).
Infringement is charged in three respects: (1) The copying or paraphrasing of annotations and indices; (2) the use of plaintiff's analyses of cases and grouping of material; and (3) the taking of citations and/or duplicate citations from plaintiff's books, with or without checking them up, either as a substitute for or as a supplement to an independent survey of the sources. As to the first two methods of infringement, while the facts are denied, the conclusion of infringement as a matter of law is conceded, if the facts be established as alleged. In respect to the third, there is disagreement as to the legal conclusion to be drawn from the facts, but a partial agreement as to the facts themselves.
After three weeks of trial in court, at the suggestion of the District Judge and with the consent of the parties, the matter was referred to a special master, to ascertain and report his findings of fact and conclusions of law. Under this reference, he properly heard no witnesses, and based his report on the testimony taken in court and on his examination of the voluminous exhibits prepared by each party, and themselves largely based upon the volumes alleged to be infringed upon and infringing. Plaintiff listed over 150 examples of alleged similarities or identities in expression, content, omissions, and mistakes occurring in the works of
The master reported that there had been a use of citations in plaintiff's annotations to the Code, but merely to check up the results of defendant's own researches; this, in his judgment, did not constitute infringement. As to the "Topical Index," he reported that plaintiff had no copyright therein. He further found no infringement in the taking of duplicate citations and case histories from Page's Ohio Digest. Identities between the defendant's index and the index of the 1916 Supplement, the copyright in which is conceded, he ascribed to legitimate copying from state pamphlets which had reproduced this index material. While he found instances of direct copying from plaintiff's annotations, they were not sufficient in number or importance, relatively to the entire work, to justify, in his judgment, either an injunction or damages. He recommended, therefore, that the bill be dismissed. The District Judge overruled plaintiff's exceptions and confirmed the master's report.
Plaintiff relies upon both direct and circumstantial proof of infringement. Of the former there are two kinds: The testimony of operatives at the printing plant where "Throckmorton's Code" was originally published, and the cross-examination of Baldwin, defendant's president, Throckmorton, the editor of its volume, and the latter's legal assistants. The circumstantial evidence consists of the examples of similarity and identity.
In 1923, defendant herein sued the Breitenbach Linotyping Company, its printer, in the Southern District of Ohio for infringement of the alleged copyright of "Throckmorton's Code." The defense was invalidity, in that the Code itself infringed the present plaintiff's copyrights. After testimony on both sides, a settlement was made at the suggestion of the trial judge; his informal views against the present defendant were, therefore, not embodied in a formal opinion.
At the present trial, four employees of the Breitenbach Company repeated their former testimony that about 25 per cent. of the copy for the Throckmorton Code, upon which they had worked, consisted of clippings set up in the distinctive type of the Page & Adams volumes. Nine of the defendant's editors, practically the whole staff, categorically denied that any clippings from Page & Adams were ever pasted upon the cards sent to the printer. The only clippings, they said, were from the 1890 Smith & Benedict Annotated Statutes, the copyright of which, owned by plaintiff, was in force when defendant began its preparation for, but expired before it had printed, the alleged infringing Code. Liberal use of Smith & Benedict is not only conceded, but asserted by defendant.
Several lines of attack upon the Breitenbach witnesses are pursued. Two of these witnesses, sisters of Breitenbach, are charged with bias; four, who declared that sometimes the card came to the printer with more than one clipped annotation to the card, are contradicted by Throckmorton and Baldwin, who testify that it was the invariable policy to have a separate card for each annotation. Baldwin, however, admitted a departure from this invariable policy in one of his other publications. The cards were not produced. Plaintiff's witnesses testified that they had been returned to defendant; this defendant denied. A great mass of evidence as to the dealings between the parties and the custom of the printing trade was introduced on this point. It is not conclusive either way. The special master found a slight preponderance in favor of the defendant.
On the entire evidence in relation to the Breitenbach controversy, as it affects the present case, we are of the opinion that it cannot be deemed of any considerable weight, much less decisive, for either side.
On behalf of defendant, Baldwin testified that he had spent a good part of his time between 1917 and 1920 in preparing a "plant." He described his method as follows: First, he took a 4×6 card for every subsection of the Ohio Code as amended to date, and pasted upon each card the appropriate clipping from the official statutes. Then he began to clip from Smith & Benedict (before Page & Adams, the standard Annotated Statutes) all annotations and cross-references which were still timely, pasting them on cards to be filed with the proper Code section card. He admitted the use under oral license of a copyrighted volume of Ohio Annotated Statutes published by Bobbs-Merrill Company. Finally he testified that he went through the Ohio Supreme Court reports which postdated the publication of Smith & Benedict, inspecting the "tables of
In 1920, witness Throckmorton, a professor of law at a prominent Ohio university law school, and a stockholder of defendant, was engaged to complete the work. He associated with himself one faculty colleague and a number of students and recent graduates of high academic standing; later he received the assistance of the dean of the school. There is much in the record attesting to the high reputation of the members of this editorial staff. All the editors, it is testified, were warned of the dangers of copyright infringement through improper use of Page & Adams, and were thoroughly familiar with Throckmorton's views of copyright law.
According to Baldwin, his part of the annotations was very meager. In the Breitenbach Case, as he conceded, he had put it at possibly 10 per cent. It is significant that in the testimony in the Breitenbach Case as to the preparation of defendant's volume, there is no mention of the use of the Bobbs-Merrill volume or of the Ohio state reports.
Prof. Throckmorton described his method as follows: Any annotations in the "plant" as it came to him, which seemed satisfactory, were left undisturbed. This, however, was not frequent. If a note was unsatisfactory in form or content, the editors were instructed to refer to the original case and syllabus, and then to rewrite it. Only after this had been done for all the material turned over to them, were the editors to examine Page & Adams, both the original Code and Supplement, to see what cases were there cited that were not in the "plant." With a list of such cases, they were again to examine the reports and write up original notes. Most of the editors testified that they followed the prescribed method. But one admitted that at times he read through the annotations in Page & Adams, not merely to get the list of cases, but to see what Page said of any of them which he had failed to understand. The work of this annotator was found by the special master to have been one of the three portions of defendant's publication in which most of the inexplicably close parallelisms with Page & Adams appear. The two annotators of the other portions alluded to did not testify.
It was further testified that, after about one-third of the work had gone on in this way, a change was made; but no satisfactory reason therefor is offered. The staff was ordered to stop using Page & Adams' citations. Prof. Throckmorton testified that thereafter he checked up the admittedly incomplete list of citations in the "plant" with the Bobbs-Merrill volume, originally a publication of the American Publishing Company. But the staff was still permitted to keep Page & Adams open before them, for the purpose, it is said, of consulting, not the annotations, but solely the text of the statutes. This text included such changes in the Code of 1910 as were made by the Session Laws, volumes 101 and 102. Some of the annotations well toward the end of defendant's volume were admitted, however, upon cross-examination, to have been prepared under the old system. According to the testimony of Page and his assistants, on the one hand, and Throckmorton and his assistants, on the other, the time required by the former for plaintiff's work was at least four times that needed by the latter for 90 per cent. of defendant's work brought down to a later date. Defendant explains this great difference to be due to greater efficiency of defendant's staff and the proper use by them of the earlier works.
Comparison of the Page & Adams publication with that of Throckmorton is made difficult by their marked differences in style, size, and scope. Plaintiff's edition, original and supplemental, of the Code, consists of 11 volumes with lines running the full width of the page; defendant's is a one-volume work, containing some 3,000 double-column pages, with a supplement thereto. The former aims at completeness of statement; the latter at extreme brevity. This justifies giving more than ordinary emphasis to such similarities as do appear. On the other hand, a certain similarity of outline and subject-matter is inevitable, in view of the character of the books and of their common basis in the work of earlier annotators.
A careful comparison of the two works establishes, in our judgment, a substantial and improper appropriation by defendant of the labor which went into the plaintiff's volumes. We find clear proof of methods which go further than defendant's witnesses have admitted; their own testimony, however, suffices to make out a case for the plaintiff.
As to the weight which should be accorded to the special master's findings, it is to be noted that he did not hear the witnesses. Our examination of the evidence upon which he acted has led us to the conclusion that it fails to sustain his findings. We have the less hesitation in reaching this conclusion, inasmuch as the special master evidently considered
One of the most significant evidences of infringement, the Supreme Court has said, is the identity of errors. Callaghan v. Myers, 128 U.S. 617, 662, 9 S.Ct. 177, 32 L. Ed. 547 (1888). In List Pub. Co. v. Keller (C. C.) 30 F. 772 (1887), an injunction was granted on the basis of 39 common misprints and mistakes in a list of some 3,000 names, despite depositions of the defendant attempting innocent explanations of the identity. See, too, Jeweler's Circular Publishing Co. v. Keystone Publishing Co., 281 F. 83, 94, 26 A. L. R. 571 (C. C. A. 2d, 1922).
Plaintiff produced before the master 26 examples of alleged common errors. Of these, 7 either could not be accounted for by the master in any other reasonable way than as copies, or were admitted by him to be highly suggestive thereof. In our judgment at least 8 more of these 26, viewed without the master's preconceptions, also clearly indicate copying. On this appeal, plaintiff has pointed out in the record 79 similar examples, many of which leave with us the same impression.
Plaintiff lists 54 asserted instances of paraphrasing or identities of expression. Fifteen of them are considered by the master clearly to show or to suggest the use of Page & Adams' corresponding notes, in some instances amounting to verbatim copying of long annotations. Quite a few other instances among the 54 seem to us highly suggestive; for example, a 50-page decision, with 2½ pages of syllabi, is abstracted by Page in a 5-line note. Throckmorton devotes 2½ lines to the case, practically restating the first half of the earlier note. True, the last page of the opinion contains language from which both notes could have been independently derived. In another instance, both Page and Throckmorton print a long note, identical with the syllabus and the Smith & Benedict annotation, save that, in both plaintiff's and defendant's annotations, the same two words are changed from the original. This change, while proper enough, was by no means a necessary, or the only possible, one. Numerous other examples of alleged paraphrasing have been pointed out by plaintiff in the argument and briefs on this appeal; a number of them are not explained away to our satisfaction. In view of the great improbability of two workers finding the same needle in a wordy haystack, as well as other equally striking coincidences and some unmistakable improprieties, we conclude that defendant's annotator at the very least derived considerable assistance from the mental labors of his rival.
Some comparative tables have been introduced by the plaintiff to show that, in very long annotations, the analysis of the subject and the subheadings have been taken over from Page & Adams. This evidence is only fairly convincing, especially since the examples are not sufficiently numerous to support a generalization. So, too, it does not seem very important that the two books number chapters and sections of the Ohio Code in the same way, since most of the numbers are official, and the rest part of a conventional scheme. But there is one instance of an identical chapter heading thoroughly out of line with the defendant's usual scheme, which clearly shows copying.
An avoidance of independent work, such as justifies an inference that material which could easily be copied was in fact taken from Page & Adams is shown in Throckmorton's treatment of statutory sections like the one on Banks and Banking. This was codified in 1919 and so after the Page & Adams Supplement. It is unannotated in Throckmorton; the available material in Page & Adams under the old sections on this subject would have had to be completely redistributed, or the original sources would have had to be studied, if these revised statutory provisions were to be annotated properly. Evidently this was too laborious an undertaking, unless Throckmorton's explanation, borne out in some, but not in all, cases, be accepted, namely, that it was not his editorial policy to annotate new sections with cases decided under the old laws.
Finally, there are some very significant repetitions in Throckmorton of Page & Adams peculiarities of citation. For example,
To summarize our conclusions in the light of the testimony of the defendant's witnesses, the admissions of defendant's counsel in the course of attempts to explain away the exhibits, the conflicts in the evidence, and a comparison of the publications — we are satisfied: First, that there was clearly some outright copying and paraphrasing of the plaintiff's annotations and other original material; enough to make it probable that very many similarities in the volume are due to this cause, even though a number of them are capable of an innocent explanation. Second, that Throckmorton's original system and the temptations inherent in his later one, prompted defendant's annotators to peruse Page's annotations and appropriate ideas therefrom, both in the way of literary expression and of designation of the point of the case. Third, that there was a great deal of copying from the old Page & Adams index, the validity of the copyright to which is challenged by defendant, and from the copyrighted index to the Page & Adams supplement, or at least from government pamphlets which had reprinted from these indices without express permission of plaintiff. Fourth, that cumulative references and case histories were freely taken from the alphabetical table of Ohio decisions in volumes 9 and 10 of Page's Ohio Digest. Fifth, that the defendant's staff obtained from Page & Adams a substantial part of their cases by way of "checking up" on glaringly incomplete lists, and that, even if these cases were subsequently read and independently abstracted, the labor of finding all Ohio cases construing statutes and deciding under what section of the Code to deal with them was thus saved to defendant.
The first two propositions are established, in our judgment, by the evidence as hereinabove recited; the question of substantial infringement is thereby concluded in favor of plaintiff.
We come to the third proposition. The evidence is that defendant, while not copying directly from plaintiff's indices to the Page & Adams Code of 1912 and to the Page & Adams Supplement of 1916, did make liberal use of state pamphlets which had reprinted substantial portions thereof without plaintiff's authority. Defendant does not dispute that one who copies from an infringer is himself guilty of infringement; but it denies both that plaintiff had a valid copyright to the 1912 index and that plaintiff is in a position to protest against the republication of either index by the state or against the use of the state pamphlets.
In April, 1906, the General Assembly passed an act authorizing commissioners to revise the statutes and to provide an index thereto. In 1910 the revision, without index, was adopted by the Legislature under the name of "The General Code." It was provided that, each time an edition of the Code was published, the Attorney General should procure a copyright thereon. A separate, later, statute directed the Attorney General to prepare an index for each edition of the Code. Shortly before the enactment of these statutes and the Code, a committee of the Legislature made a contract with plaintiff, which was then engaged in the preparation of its Annotated Code, pursuant to which plaintiff prepared an index suitable for both its own and the state publications. The consideration was $1,500, the bare cost of clerical work and out of pocket expenses, with nothing for overhead or profit. After the passage of the statutes, the Attorney General orally agreed that the copyright of the index contracted for should remain in plaintiff.
On these facts, the copyright granted to plaintiff on the index is valid. An author is not necessarily precluded from copyrighting a work produced under contract with another person; the intent of the parties as to which of them shall have the right to copyright is decisive. Where a contract of employment is silent, there may be an implication in favor of the employer. But in the present case plaintiff was an independent contractor, rather than an employee; moreover, it may properly be inferred that the parties did not intend plaintiff to surrender a copyright in consideration of a sum less than the bare cost of the work. But, in any event, it would seem that by the above-mentioned statutes the Attorney General was impliedly authorized to modify contracts for indexing. There is nothing in the statutes which reserves the copyright on the index to the state; the statute concerning a copyright on the Code makes no reference to the then nonexistent index, while the subsequent statute
Defendant further contends that plaintiff cannot in any event complain of infringement as to either of the two indices, because the publication by the state was known to its vice president and was not objected to. With this we cannot agree. The officer admits knowledge in the case of only a few of the dozen pamphlets in question. He testifies that, before publishing them, the state promised to acknowledge plaintiff as the source of the material, and that, when the pamphlets appeared without such notice, he thought that there were legal obstacles and practical objections to a suit against the state. There is no evidence of an intent to abandon the copyright. A delay in asserting rights, which might be fatal in a suit against an innocent party who had been deceived thereby, will not avail an otherwise infringing defendant.
As to the fourth proposition: The special master found as an uncontroverted fact that defendant had obtained cumulative citations and histories of the affirmances and reversals of cases, without checking back to the original reports, from the tables of cases digested and of cases judicially noted, affirmed, and reversed, in volumes 9 and 10 of Page's Ohio Digest. It is unimportant that a copyright notice was accidentally omitted in the printing of the copy of volume 10 which defendant used, since there is no doubt that the work was in fact copyrighted, and defendant's pleadings so admit. However, we agree with the special master that, in view of the dissimilarity and lack of competition between such tables in a Digest and an Annotated Code, the use of this portion of the Digest for annotation purposes did not constitute infringement. It gave defendant only source material; the labor of abstracting the cases and distributing them to the proper Code sections had to be performed. In our judgment, such a use of source material is not to be denied for noncompetitive work.
The fifth proposition raises important questions of law. Plaintiff claims protection against the appropriation of the result, both of his mechanical work in paging through the volumes of Ohio reports to find the relevant cases and of his mental labor in deciding which cases to select and under what section of the Code to use them. On the other hand, there is the public interest in having subsequent workers in a field allowed to make use of the accumulated knowledge of the past, especially knowledge of the sources. The problem of reconciling these opposing claims has been before the courts many times, often, especially in the second circuit, in cases involving law books.
In considering the authorities, it is necessary to bear certain distinctions in mind. A text-book differs fundamentally from a compilation, such as a directory or list. The latter simply sets forth names, symbols, or the like; its only usefulness lies in its presentation of facts which any one might have learned by repeating the compiler's industry, and which, having learned, he could present to the public practically in only the one way. The former is the product of intellectual labor and literary skill, results of which are unique each time the subject is handled independently, even though the same materials are dealt with. Moreover, in a text-book there is ordinarily a critical discussion of earlier writers and a re-examination of their sources. Again, the directory or list is intended to be used as a guide to something further and different; it is not itself a contribution to knowledge, which is to be availed of by co-workers in the same field. For all these reasons, greater latitude is to be expected in the case of authors consulting other text-books or using directories and lists than in the case of one compiler attempting to make use of another similar compilation. Law treatises and legal encyclopedias are in the class of text-books. A list of cases or case histories is in the directory class. Midway between the two groups stands the law digest, with its list of abstracted cases.
Another important distinction concerns the use to which the rival's citations are put. When the earlier volume is consulted at the start, before the defendant has done any independent gathering of authorities, even if the citations are subsequently looked up, the defendant cannot be said to have done more than verified the plaintiff's work; he has done nothing afresh. On the other hand, a different situation is presented if the competing work is consulted only after the defendant has himself examined all the original sources, and for the sole purpose of discovering possible discrepancies or omissions, which, when found, will be investigated and corrected by further independent work. A very different practice from this last is the "checking up" of obviously incomplete original work, so incomplete that the defendant must know the process of comparison will result in his taking over a large amount of matter from his rival.
We turn now to the cases. In England the problem was opened up by Kelly v. Morris,
Banks v. McDivitt, Fed. Cas. No. 961 (1875), is the earliest American decision. The parties, as in the instant case, had each published an annotated statute book. Verification by defendant of copied citations afforded no defense. An injunction issued.
West Publishing Co. v. Lawyers' Co-op. Publishing Co., 79 F. 756, 35 L. R. A. 400 (2d C. C. A. 1897) involved two digests. The editors of the defendant were found, not only to have taken citations, but also to have read the plaintiff's syllabi to aid their examination of the cases. Exhibits of common errors and of the speed with which the editing was completed were held to outweigh the explicit sworn denials of the members of the editorial staff. The court, in upholding the injunction, said that there may be infringement through labor-saving use of a rival's winnowing of the material, even though there is no close identity in the finished products. The defendant, having put temptation in its editors' way by supplying them with plaintiff's publications of reports containing the copyrighted syllabi subsequently incorporated in the digest, so that they might use the uncopyrighted opinions, was held liable for the foreseeable further improper use of the syllabi themselves.
In Edward Thompson Co. v. American Law Book Co., 122 F. 922 (2d C. C. A. 1903), one encyclopedia was alleged to be an infringement of another. The defendant's workers never saw the earlier book, but, when starting on their task, each man had been supplied with a list of cases, some of which had been taken from the plaintiff's encyclopedia. An injunction was denied, upon considerations of public policy and the nature of treatises, to which we have already adverted, and also on the ground of the plaintiff's unclean hands. The court's whole discussion was in terms of authors and books; it specifically distinguished the case of a compilation of quotations.
George T. Bisel Co. v. Welsh (C. C.) 131 F. 564 (1904), resembles the Banks Case. Here the copying of plaintiff's collection of statutes, as proven by comparison of errors and peculiarities despite defendant's editors' sworn denials, was the basis for an injunction.
A strong case is Sampson & Murdock Co. v. Seaver-Radford Co., 140 F. 539 (1st C. C. A. 1905). The defendant, in compiling his directory, had first made an expensive independent canvass. Next he checked up with the plaintiff's directory for omitted names, which his canvassers then verified. However, what the plaintiff said about the names was copied upon slips, and these slips were used in the verification. The process was responsible for 12 per cent. of the defendant's names. As in the instant case, the two publications were for substantially, but not identically, the same purpose. An injunction was granted; the distinctions between lists and texts above stated was the basis of this decision.
Hartford Printing Co. v. Hartford Directory & Publishing Co. (C. C.) 146 F. 332 (1906), contains a strong dictum by Platt, District Judge, that, if the compiler of a list starts with a really thorough original search, he may later resort to a rival list for further names to be investigated. We express no opinion on the soundness of this distinction. The facts before us show an obviously incomplete, even colorable, preliminary search.
In West Publishing Co. v. Edward Thompson Co., 176 F. 833 (2d C. C. A. 1910), an encyclopedia copying syllabi in a digest was held an infringement. But if the defendant had merely taken lists of cases from the plaintiff for a starting point, said the court, a different result would have been reached. However, the point of this dictum seems to be merely that the use of the cases in a digest is proper in compiling a textbook. On the other hand, as is recognized, "obviously it would not be fair for any publisher of reports of the same cases or of digests of them to copy lists of cases * * * whether by so doing he merely saved mechanical labor or literary work."
Shepard v. Taylor, 193 F. 991 (2d C. C. A. 1912), is a case of proved copying, apparently without checking up. All that was copied, though, was the names of cases; the rival works being citators. Clearly one cannot go to a rival work for a mere list of cases, even if these are then checked up.
Finally, is the significant case of Jewelers' Circular Publishing Co. v. Keystone Publishing
We conclude that, even if defendant had not gone beyond what its president and editor admitted to have been its system of operations for at least a third of the time, namely, to depend for much of its material upon the taking of citations of cases from plaintiff's volumes, to be verified and then used in its annotations, but without copying plaintiff's text, it has infringed plaintiff's copyright.
The decree must be reversed for further proceedings in accordance with this opinion.
As to the injunction, if, in the light of the views herein expressed, defendant desires that some parts of its work be exempted therefrom, opportunity may be given clearly to prove noninfringement in respect to such parts.