Mr. Justice Woods

delivered the opinion of the court.

This suit was brought for an alleged infringement Of the complainant’s reissued letters-patent, No. 5216, dated Jan. 7,1878, for an improvement in corset-springs.

The original letters bear date July 17,1866, and were issued to Samuel H. Barnes. - The reissue was made to the complainant, under her then name, Frances Lee Barnes, executrix of the original patentee.

*334The specification for the reissue declares-: —

This invention consists in forming the springs of corsets of two or more metallic plates, placed one upon another, and so connected as to prevent them from- sliding off each other laterally or edgewise, and at the same time admit of their playing or sliding upon.each other, in the direction of their length or longitudinally, whereby their flexibility and elasticity are greatly increased, while at the same time much strength is obtained.”

The second claini is as follows: —

“ A pair of corset-springs, each member of the pair being composed of two or more metallic plates, placed one on another, and fastened together at their centres, and so connected at or near each end that they can move or play on each other in the direction of their length.”

The bill alleges that Barnes was the original and first inventor of the improvement covered by the reissued letters-patent, and that it had not, at the time of his application for the original letters, been for more than two years in public use or on sale, with his consent or allowance.

. The answer takes issue on this averment and also denies infringement. On a final hearing the court dismissed the bill, and the complainant appealed.

As to the second defence above mentioned, it is sufficient to say that the evidence establishes beyond controversy the infringement by the defendants of the second claim of the reissue.

We have, therefore, to consider whether the defence that the patented invention had, with the consent of the inventor, been publicly used for more than two years prior to his application for the original letters, is sustained by the testimony in the record.

The sixth, seventh, and fifteenth sections of the act of July 4, 1836, c. 357 (5 Stat. 117), as qualified by the seventh section of the act of March 3, 1839, c. 88 (id. 353), were in force at the date of his application. Their effect is to render letters-patent invalid if the invention which they cover was in public use, with the consent and allowance.of the inventor, for more than two years prior to his application. Since the passage of the act of 1839 it has been strenuously contended *335that the public use of an invention for more than two years' before such application, even without his consent and allowance, renders the letters-patent therefor void.

It is unnecessary in this case to decide this question, for the alleged use of the invention covered by the letters-patent to Bai’nes is conceded to have been with his express consent.

The evidence on which the defendants rely to establish a prior public use of the invention consists mainly of the testimony of the complainant.

She testifies that Barnes invented the improvement covered by his patent between January and May, 1855; that between the dates named the witness and her friend Miss Cugier were .complaining of the breaking of their corset-steels. Barnes, who was present, and was an intimate friend of the witness, said he thought he could make her a pair that would not break. At their next interview he presented her with a pair of corset-steels which he himself had made. The witness wore these steels a long time. In 1858 Bai’nes made and presented to her another pair, which she also wore a long time. When the corsets in which these steels were used wore out, the witnes's ripped them open and took out the steels and put them in new corsets. This was done several times.

It is admitted, and, in fact, is asserted, by complainant, that these steels embodied the invention afterwards patented by Barnes and covered by the reissued letters-patent on which this suit is brought.

Joseph H. Sturgis, another witness for complainant, testifies that in 1863 Barnes spoke to him about two inventions made by himself, one of which was a corset-steel, and that he went to the house of Barnes to see them. Before this time, and after the transactions testified to by the complainant, Barnes and she had intermarried. Barnes said his wife had a pair of steels made according to his invention in the corsets which she was then wearing, and if she would take them off he would show them to witness. Mrs. Barnes went out, and returned with a pair of corsets and a pair of scissors, and ripped the corsets open and took out the steels. Barnes then explained to witness how they were made and used.

This is the evidence presented by the record, on which the *336defendants rely to establish the public use of the invention by ■the patentee’s consent and allowance.

The question for our-decision is, whether this testimony shows a public.use within the meaning of the statute.

We observe, in the first place, that to constitute the public use of an invention it' is not necessary that more than one of the patented articles should be publicly used-. The use of a. great number may tend to strengthen the proof, but one well-defined case of such use is just as effectual to annul the patent as-many; McClurg v. Kingsland, 1 How. 202; Consolidated Fruit-Jar Co. v. Wright, 94 U. S. 92; Pitts v. Hall, 2 Blatchf. 229. For instance, if the inventor of a mower, a printing-press, or a- railway-car 'makes and sells only one of the articles invented by him, and allows the vendee to use it for two years, without restriction or limitation, the' use is just as--public as if he had sold and allowed the use of a. great ■number.

■ We remark, secondly, that, whether the use of an invention is public-or private does not necessarily depend'upon the number of persons to whom its use is known. If an inventor, having-made his device; gives or sells it to another, to'be used by t-he donee or vendee, without limitation or restriction, or injunction of secrecy, and it is so used, such irse is public, even though -the' use and knowledge of the use may be confined to one person.

We say', thirdly, that some inventions are by their vei’y character only capable of being used where they cannot be seen or observed by the public eye. An invention may consist of a lever or spring, hidden in the running gear of a watch, or of a rachet, shaft, or cog-wheel covered from view in the recesses of a machine for spinning or weaving. Nevertheless, if its inventor, sells a machine of which Ins invention forms a part, and allows it to be used without restriction of any kind, the use is a public one. So, on the other hand, a use necessarily open to public view, if made in good faith solely to test the qualities of the invention, and for the purpose of experiment, is not a public use within the meaning of the statute; Elizabeth v. Pavement Company, 97 U. S. 126; Shaw v. Cooper, 7 Pet. 292.

*337. Tested .by these principles, we- think the evidence ¡of. the complainant herself sliows that for mpre.than two years.before the ■ application for the original letters there was, Jby the ,consent and allowance of- Barnes, a public use of the invention, covered by them. He made and gave to her two pairs of corset-steels, constructed according to his device, one in 1855 and one in 1858. They were presented to her for use. He imposed no obligation of secrecy, nor any condition or restriction whatever. They -were not presented for the purpose of experiment, ,-norto test their qualities. No such claim is set up in her testimony. The invention was at the time complete, and there is no evidence that it was afterwards changed or improved. The donee of the steels used:-them for years for the purpose and in the manner designed by the inventor. They were not ■ capable of any other use. She .might have exhibited them to any person, or made other steels of the same kind, and used or-sold them without violating any condition or restriction imposed on her by the inventor.

According to the testimony of the complainant, thp invention was completed and put to use in 1855. The inventor slept on. .his rights for eleven years. Letters-patent were not applied for till' March, 1866. In the mean time, the invention had found its way into general, and almost universal, use. A great part of the record is taken up with the testimony of the manufacturers and venders of corset-steels, showing that before he applied for letters the principle of his .device'was almost universally used in the manufacture of corset-steels. It is fair to. presume that having learned from this general use that there was some value in his invention, he -attempted to resume, by his application, -what by his acts he had clearly dedicated to the' public.

“ An abandonment of an invention to the public may be evinced bj” the conduct of the inventor at any time, even -within the two years named in the law. The effect of the law is that no such consequence will necessarily follow from the invention being in public use or on sale, with the inventor’s consent and allowance, at any time within two years before his application; but that, if the invention is in public use or on sale prior to that time, it will be conclusive evidence of *338abandonment, and the patent will be void.” Elizabeth v. Pavement Company, supra.

Wé are of opinion that the defence of two years’ public use, . by the consent and allowance of the inventor, before-he made application for letters-patent, is satisfactorily established by the evidence.

'Decree affirmed.

Mr. Justice Miller

dissenting.

The sixth section of the act of July 4, 1806, c. 357, makes it a condition of the grant- of a patent that the invention for which it was asked should not, at the time of the application for a patent, “ have been in public 'use or on salé with the consent or allowance” of the’inventor or discoverer. Section fifteen of the same act declares that it shall be a good-defence to an action for infringement of the patent, that it had been in public use or on sale with the. consent or’ allowance o.f the patentee before his application. This was afterwards modified by the seventh section of the act of March 3, 1839, c. 88, which 'declares that no patent shall be void on that ground .'unless the prior use has been for more than two years before the application..'

This is the law under which the patent of the complainant is held void by the opinion just delivered. The previous part of the same section requires that the invention must be one “ not known or used by others ” before the discovery or invention made by the applicant. In this limitation, though in the same sentence as the other, the word “ public ” is not used, so that the use by others which would defeat the applicant, if without his consent,.need not be public; but where the use of his invention is by his consent or allowance, it must be public or it will not have that effect.

The reason of this is undoubtedly that, if without his con- / sent others have used the machine, composition, or manufacture, it is strong proof that he was not the discoverer or first inventor. In that case he was not’entitled to a patent. If the use was with his consent or allowance, the fact that such consent tir allowance was first obtained is evidence that he was the inventor, and claimed to be such. In such case, he was not to *339lose his right to a .patent, unless the use which he permitted was such as showed an intention-of' abandoning his invention to the public. It must, in the language óf' the act, be in public use or on sale. If on sale, of course the public who buy can use it, and if used in .public with, his .consent, it may be copied by others. In either event there is an end of his exclusive right of use or.sale.

The word public is, therefore, an important member of the sentenep. A private use with consent, which could lead to no copy or reproduction of the machine, which taught the. nature • of the invention to no one' but the party to whom such consent was given,'which left the.public at large as ignorant of this as it was before- the’ author’s discovery, was no abandonment to the public, and did not defeat his claim for a patent. , If the little steel spring inserted1 in a single. pair of corsets, and used by only one woman, covered by her outer-clothing, and in a position always withheld from public obseiv vation, is a public use of' that piece of steel, I am at a loss to know the line between a private and a public use..

The opinion argues that the use'was public, because, with the consent of the inventor to its use, no limitation-was imposed in regard to its use in public. • It may be well imagined that a prohibition to the party so permitted against exposing her use of the steel spring to public observation would have been supposed to be a piece of irony. An objection quite the opposite' of this suggested by the opinion is, that the invention was incapable of a public use. That is to say, that while the statute says the right to the patent can only be defeated by a use which is public, it is equally fatal to the claim, when it is permitted to be used at all, that the article can never be used in public.

I cannot on such reasoning'as this eliminate from the statute the word public, and disregard its obvious importance in connection with the remainder of the act, for the purpose of defeating a patent otherwise meritorious.

Mr. Chief Justice TANEY

delivered the opinion of the court.

Three objections have been taken to the instructions given by the Circuit Court at the trial, and neither' of them is, perhaps, entirely free from difficulty.

The first question arises upon the assignment of Fitzgerald to Enos Wilder. The assignment was made and recorded in the Patent-Office before the patent issued. It afterwards issued to Fitzgerald. And the plaintiffs in error insist that this *493assignment did not convey to Wilder the legal right to the monopoly subsequently conferred by the patent, and that the plaintiff, who claims under him, cannot therefore maintain this action.

The inventor of a new and useful- improvement certainly has no exclusive right to it, until he obtains a patent. This right is created by the patent, and no suit can be maintained by the inventor against any one for using it before the patent is issued. But the discoverer of a new and useful improvement is vested by law. with an inchoate-right-to its exclusive use, which he may perfect and make absolute by proceeding in the manner which the law requires. Fitzgerald possessed this inchoate right at the time of the assignment. The discovery had been made, and the specification prepared to obtain a patent. And it appears'by the language of the assignment, that it was intended to operate upon the perfect legal title which Fitzgerald then had a lawful right to obtain, as well as upon the imperfect and inchoate interest which he actually possessed. The assignment requests that the patent may issue to the assignee. And there would seem'to be no sound reason for defeating the intention of the parties by restraining the assignment-to the latter interest, and .compelling them to execute another transfer, unless the act of Congress makes it'necessary. The court think it does not. The act of 1836 declares that every patent shall be assignable in • law, and that the assignment must be in writing, and recorded within the time -specified. But the thing to be assigned is not the mere parchment on which the grant is written. It is the monopoly which the grant confers: the right of property which it creates. And when the party has acquired an- inchoate right to .it, and the power to make that right perfect and absolute at his pleasure, the assignment of his whole interest, whether executed before or after the patent issued, is equally within the provisions of the act of Congress.

And we are the less disposed to give it a different construction, because no purpose of justice would be answered by it, and the one we now give was the received construction of the act of 1793, in several of the circuits; and there is no material difference in this respect between the two acts. As long ago as 1825, it was held by Mr. Justice Story, that in-a case of this kind an action could not be maintained in the name of the patentee, but must be brought by the assignee. 4 Mason, 15. We understand the same rule has prevailed in other circuits; and if it were now changed, it might produce much injustice to assignees who have relied on such assignments, and defeat pending suits brought upon the faith *494of long established judicial practice and judicial decision. Fitzgerald sets up no claim against the assignment, and to require another to complete the transfer would be mere form. We do not think the act of Congress requires it; but that, when the patent issued to him, the legal right to the monopoly and property it created was, by operation of the assignment then on record, vested in Enos Wilder.

The next question is upon the agreement between the defendant in error and Herring. Is this' instrument an assignment to Herring for the State or city of New York, upon which he might have sued in his own name?’ If it is, then this action cannot be maintained by the defendant in error.

Now the monopoly granted to the patentee is for one entire thing; it is the exclusive right of making, using, and vending to others to be used, the improvement he has invented, add for which the patent is granted. The monopoly did not exist at common law, and the rights, therefore, which may be exercised' under it cannot be regulated by the rules of the^ommon law. It is created by the act of Congress; and no rights can be acquired in it unless authorized by statute, and in the manner the statute prescribes.

By the eleventh section of the act of 1836, the patentee may assign his whole interest, or an undivided part of it. But if he assigns a part under this section it must be an undivided portion of his entire interest under the patent, placing the assignee upon an • equal footing with himself for th,e part assigned. Upon such an assignment, the patentee and his assignees become joint owners of the whole interest secured by the patent, according to the respective proportions which the. assignment creates.

By the fourteenth section, the patentee may assign his exclusive right within and throughout a specified part of the United States, and upon such an assignment the assignee may sue in his own name for an infringement of his rights. But in order to enable him to sue, the assignment must undoubtedly convey to him the entire and unqualified monopoly which the patentee held in the territory specified,— excluding the patentee himself, as well as others.. And any assignment short of this is a mere license. For it was obviously not the' intention of the legislature to permit several monopolies to be made out of one, and divided among different persons within the same limits. Such a division would inevitably lead to fraudulent impositions upon persons who desired to purchase the use of the improvement, and would subject a party who, under a mistake as to his rights, used the invention without authority, to be harassed *495by a multiplicity of suits instead of one, and to successive recoveries of damages by different persons holding different portions of the patent right in the same place. Unquestionably, a contract for the purchase of any portion of the patent right may be good as between the parties as a license, and enforced as such in the courts of justice. But the legal right in the monopoly remains in the patentee, and he alone can maintain an action against o, third' party who commits an infringement upon it. This is the view taken of the subject in the case of Blanchard v. Eldridge, J. W. Wallace, 337, and we think it the true one.

Applying these principles to the casé before us, the action was properly brought by the plaintiff below, and could not have been maintained by Herring.

The agreement is singularly confused and complicated. . It purports to grant to Herring the exclusive right, to make and vend the Salamander safe in the city, county, and State of New York; and Herring agrees to pay to the defendant in error a cent a pound,for every pound the safes might weigh, to be paid monthly. But at the same time it reserves to Wilder the right to set up a manufactory or works for making these safes in the State of New York, provided it is not within fifty miles of the city, and to sell them in the State of New York, paying to Herring a cent a pound on each safe so sold within the State.

It is evident that this agreement is not an assignment of an undivided interest in the whole patent, nor the assignment of an exclusive right to the entire monopoly in the State or city of New York. It is therefore to be regarded as a license only, and under the act of Congress does not enable Herring to maintain an action for an infringement of the patent right. The defendant in error continues the legal owner of the monopoly created by the patent.

The remaining question is upon the validity of the patent on which the suit was brought.

It appears that James Conner, who carried on the business of a stereotype founder in the city of New York, made a safe for his own use between the years 1829 and 1832, for the protection of his papers against fire; and continued to use it until 183S, when it passed into other hands. It was kept in his counting-room and known to the-persons engaged in the fouridery; and after it passed out of his hands, he used others of a different construction.

It docs not appear what became of this safe afterwards. And there is nothing in the testimony from which it can be' inferred that its mode of construction was known to the person into whose possession it fell, or that any value was attached *496to it as a place of security for papers against fire ; or that it was ever used for that purpose.

Upon these facts the court instructed the jury, “ that if Conner had not made his discovery public, but had used it simply for his own private purpose, and it had been finally forgotten or abandoned, such a discovery and use would be no obstacle to the taking out of a patent by Fitzgerald or those claiming under him, if he be an original, though not the first, inventor or discoverer.”

The instruction assumes that the jury might find from the evidence that Conner’s safe was substantially the same with that of Fitzgerald, and also prior in timé. And if the fact was so, the question then was whether the patentee was “ the original and first inventor or discoverer,”' within the meaning of the act of Congress.

The act of 1836, ch. 357, § 6, authorizes a patent where the party has discovered or invented a new and useful improvement, “ not known or used by others before his discovery or invention.” And the 15th section provides that, if it appears on the trial of an action brought for the infringement of a patent that the patentee “ was not the original and first inventor or discoverer of the thing patented,” the verdict shall be for the defendant.

Upon a literal construction of these particular words, the patentee, in this case certainly was not the original and first „ inventor or discoverer, if the Conner safe was the same with his, and preceded his discovery.

But we do not think that this construction would carry into effect the intention of the legislature. It is "not by detached words and phrases that a statute ought to be expounded. The whole act must be taken together, and a fair interpretation given to it, neither extending nor. restricting it beyond the legitimate import of its language, and its obvious policy and object. And in the 15th section, after making the provision above mentioned, there is a further provision, that, if it shall appear that the patentee at the time of his application for the patent believed himself to be the first inventor, the patent shall riot be void on account of the invention or discovery having been known or used in any foreign country, it not appearing that it had been before patented or described in any printed publication.

In the case thus provided for, the party who invents is not strictly speaking the first and original inventor. The law assumes that the improvement may have been known and used before his discovery. Yet his patent is valid if he discovered it by the efforts of his own genius, and believed himself to be *497the original inventor. The clause in question qualifies the words before used, and shows that by knowledge and use the legislature meant knowledge and use existing in a manner accessible to the public. If the foreign invention had been printed or patented, it was already given to the world and open to thé people of this country, as well as of others, upon reasonable inquiry. They would therefore derive no advantage from the invention here. It would confer no benefit upon the community, and the inventor therefore is not considered to be entitled to the reward. But if the foreign discovery is not patented, nor described in any printed publication, it might be known and used in remote places for ages, and the people of this country be unable to profit by it. The means of obtaining knowledge would not be within their reach; and, as far as their interest is concerned, it would be the same thing as if the improvement had never been discovered. It is the inventor here that brings it to them, and places it in their possession. And as he does this by the effort of his own genius, the law regards him as the first and original inventor, and protects his patent, although the improvement had in fact been invented before, and used by others.

So, too, as to the lost arts. It is well known that centuries ago discoveries were made in certain arts the fruits of which have.come down to us, but the means by which the work was accomplished are at this day unknown. The knowledge has been lost for ages. Yet it would hardly be doubted, if any one now discovered an art thus lost, and it was a useful improvement, that, upon a fair construction of the act of Congress, he would be entitled to a patent. Yet he would not literally be the first and original inventor. But he would be the first to confer on the public the benefit of the invention. He would discover what is unknown,' and communicate knowledge which the public had not the means of obtaining without>his invention.

Upon the same principle and upon the same rule of construction, we think that Fitzgerald must be regardéd as the first and original inventor of the safe in question. The case as to this point admits, that, although Conner’s safe had been kept and used for years, yet no test had been applied to it, and its capacity for resisting heat was not known; there was no evidence to show that any particular value was attached to it after it passed from his possession, or that it was ever afterwards used as a place of security for papers; and it appeared that he himself did not attempt to make another like the one hé is supposed to have invented, but used a different one. And upon this state of the evidence the court put it to the jury to say, whether this safe *498had been finally forgotten or abandoned before Fitzgerald’s invention, and whether he was the original inventor of the safe for which he obtained the patent $ directing them, if they found these two facts, that their verdict must be for the plaintiff. We think there is no error in this instruction. For if the Conner safe had passed away from the memory of Conner himself, and-of those who had seen it, and the safe itself had disappeared, the knowledge of the' improvement was as completely lost as if it had never been discovered. The public could derive no benefit from it until it was discovered by another inventor. And if Fitzgerald made his discovery by his own efforts, without any knowledge of Conner’s, he invented an improvement that was then new, and at that time unknown; and it was not the less new and unknown because Conner’s safe was recalled to his memory by the success of Fitzgerald’s.

We do not understand the Circuit Court to have said that the omission of Conner to try the value of his safe by proper tests would deprive it of its priority ; nor his omission to bring it into public use. He might have omitted both, and also abandoned its use, and been ignorant of the extent of its value ; yet, if it was the same with Fitzgerald’s, the latter would not upon such grounds be entitled to a patent, provided Conner’s safe and its mode of construction were still in the memory of Conner before they were recalled’ by Fitzgerald’s patent.

The circumstances above mentioned,-referred to in the opinion of the Circuit Court, appear to have been introduced as evidence tending to prove that^the Conner safe might have been finally forgotten, and upon which this hypothetical instruction was given. Whether-this evidence was sufficient for that.purpose or not, was a question for the jury, and the court, left'it to them. And, if the jury found the fact to be so, and that Fitzgerald again discovered it, we regard him gs standing upon the same ground with the discoverer of a lost art, or an unpatented and unpublished foreign invention, and like him entitled to a patent. For there was no existing and living knowledge of this improvement, or of its former use, at the time he made thé discovery. And whatever benefit any individual may derive from it in the safety of his papers, he owes entirely to the genius and exertions of Fitzgerald.

Upon the whole, therefore, we think there is no error in the opinion of the Circuit Court, and the judgment is therefore affirmed.

Mr. Justice McLEAN.

I dissent from the opinion of a majority of the judges in this .case. The point of difference, I think, is essential to the maintenance of the rights of the public and also of inventors.

*499It was proved by James Conner, as appears from the bill of exceptions, “ that between 1829 and 1832 he was engaged in business as a stereotype founder, and, knowing that plaster of Paris was a non-conductor of heat, he constructed a safe with a double chest, and filled the space between the inner and outer one with plaster of Paris; the same, substantially, as testified to and claimed by Fitzgerald, except there was no plaster used on the top of the safe. It was made for his own private use in his establishment, and was used by him as a safe from the time it was made till 1838, when it passed into other hands. It was kept in the counting-room while he used it, and was known to the persons working in' the foundery.” This evidence was “confirmed by another witness.

By the sixth section of the patent act of 1836, it is provided, “ that any person or persons having discovered or invented any new or useful art, machine, manufacture, or composition of matter, or any new or useful improvement on any art, machine, manufacture, or composition of matter, not known or used by others before his or their discovery or invention thereof f may apply for a patent, &c. The applicant is required to “ make oath or affirmation that he does verily believe that he is the original and, first inventor,” &c., “ and that he does not know or believe that the same was ever before known or used.”

The seventh section authorizes and requires the Commissioner of Patents “ to make or cause to be made an examination of the alleged new invention or discovery ; and if on such examination it shall not appear to the Commissioner that the same had been invented or discovered by any other person in this country prior to the alleged invention or discovery thereof by the applicant, or that it had been patented or described in any printed publication in this or any foreign country,” &c., the Commissioner may grant a patent.

In the fifteenth section it is provided, that whenever it shall satisfactorily appear that the patentee, at the time of making his application for the patent, believed himself to be the first inventor or discoverer of the thing patented, the same shall ndt be held to be void on account of the invention or discovery, or any part thereof, having been before known or used in any foreign country, it not appearing that the same or any substantial part thereof had before been patented or described in' any printed publication.”

From the above extracts, it is seen that the patentee must be the'inventor of the machine, or the improvement of it, or he can have no right. If the thing was known or used by others, be cannot claim a patent. Or if it was patented in a foreign country, or described in any publication at home or in any *500foreign country, he has no right to a patent. To this there is only the exception in the fifteenth section above cited. But this can have no influence in the present case.

Let these provisions of the statute be compared with the two last paragraphs of the charge of the court, as stated in the third exception : —

“ And said court further instructed the jury, that if they found that the use made by James Conner of plaster of Paris was confined to a single iron chest, made for his own private use after said Fitzgerald’s discovery and experiments, then it was not in the way of Fitzgerald’s patent, a'nd the same was valid; but if the jury found that said James Conner made his said safe, as claimed, and tested it by experiments, before Fitzgerald’s invention and improvement, and before he tested the same, then said Fitzgerald was not the first inventor, as claimed, and was not entitled to said patent.”

This charge stands disconnected with any other facts in the case, except those named, and, in my judgment, it is erroneous. . If Conner’s safe were identical with Fitzgerald’s, and though it was of prior invention, yet if it were not tested by experiments before Fiizgerald’s improvement, and before he tested the same, the jury under the instruction were bound to find for Fitzgerald. And the case was thus made to turn, not on the priority of invention only, but upon that and the fact of its having been tested by experiments. This introduces a new principle into the patent law. The right under the law depends upon the time of the invention. An experimental test may show the value of the thing invented, but it is no part of the invention.

“ The court further charged, that, independently of these considerations, there was another view of the case, as it respected the Conner safe; that it was a question whether the use of it by him had been such as would prevent another inventor from taking out a patent; that if Conner had not made his discovery public, but had used it simply for his own private purpose, and it had been finally forgotten or' abandoned, such a discovery and use would be no obstacle to the taking out of a patent by Fitzgerald, or those claiming under him, if he be an original, though not the first, inventor or discoverer of the improvement.”

If there be any thing clear in the patent' law, it is that the original inventor means the first inventor, subject only to the provision stated in the fifteenth section. This instruction presupposes that the safes are the same in principle. Now, if the invention was patented abroad, or was described in a foreign publication, both of which were unknown to the inventor in *501this country, still his patent is void. So it is void if such invention has been known to any person in this country. The instruction says, if Conner’s invention “ had been forgotten or abandoned,” it was no obstacle to Fitzgerald’s right. Can a thing be forgotten or abandoned that was never known ? If known before Fitzgerald’s invention, it is fatal to it. By whom must it have been forgotten ? By the inventor, or the public, or both ? And how must it have been abandoned ? When an invention is abandoned, it is said to be given up to the public, and this"is the sense in which the term abandonment is used in the patent law. Such an abandonment would be fatal to the right of Fitzgerald.

Conner’s safe, as appears from the bill of exceptions, was used in his counting-house, being accessible to every one, some six or eight years. In 1838 it passed into other hands ; but into whose hands it does not appear. In 1843, Fitzgerald obtained his patent... How long before that he made experiments to test the invention ife not proved. At most, the time must have been less than five years. This is a short period on which to found a presumption of forgetfulness. The law authorizes no such presumption. It can never become the law. It is not founded on probability or reason. The question is, Was Conner’s invention prior to that of Fitzgerald ? That it was of older date by some ten or twelve years is proved. And the instruction, it must be observed, was founded on the supposition that both inventions were similar.

The instruction seems to attach great-importance to the fact that Conner’s safe was used only for his private purpose. This is of no importance. The invention is the question, and not the manner in which the inventor used it. The safe was constructed at the foundery, and must have been known to the hands there employed. How can it be ascertained that Fitzgerald was not informed by some of these hands of the structure of Conner’s safe, or by some one of the many hundreds who had seen it in his counting-house in the city of New York ? It was to guard against this, which is rarely if ever susceptible of' proof, that the act is express, — if the thing patented was known before, the patent is void. If the fact of this knowledge in any one be established, it is immaterial whether the patentee may have known it or not, it avoids his patent.

-The law,. o.n this subject, is not founded upon any supposed notions of equity. A foreign patent for the samé' thing, or a description of the thing in a foreign publication, is as effectual to. avoid the patent as if the patentee had seen the prior invention. Notice to him is not important. The law is adopted op. *502a settled public policy, which, while it is just to inventors, protects the rights of the public. Any other basis would open the door for endless frauds, by pretended inventors, without the probability of detection. And. especially does this new doctrine of forgetfulness, or abandonment, used in any other sense than as recognized in the patent law, leaving such matters to a jury, overturn what I consider to be the settled law on this subject.. Of the same character is the fact, that the invention was used for private purposes. A thing may be used in that way, and at the same time be public, as was the case with the Conner safe, and yet the jury are necessarily misled by such an instruction.

Mr. Justice DANIEL,

dissenting.

Differing from the majority in the decision just pronounced, I proceed to state the grounds on which my dissent from that decision is founded.

' On two essential points in this pause, it seems to me that the learned justice who tried it at the Circuit has erred, and that the decision here should therefore have been for a reversal of his judgment. Those points involve, first, the- right of the plaintiff below to maintain his action upon the title or right of action deduced from Fitzgerald through Enos and Benjamin Wilder; and secondly, a right to or interest in the subject of the suit on the part of the .plaintiff below, admitting that subject to have been originally invented and used by some other person than Fitzgerald; a right founded upon an assumption that this subject had been used in private only, or had, in the language of th‘e learned justice, been “ finally forgotten or abandoned ” by such first inventor. These points are presented by the first and third exceptions of the plaintiffs in error to the rulings at the trial below. The plaintiff in the Circuit Court claimed by assignment from B. G-. Wilder, assignee of Enos Wilder, assignee of Daniel Fitzgerald, alleged to have been the inventor of the Salamander safe. By the paper deduction of title, it appears that, on the 11th day of April, 1839, Fitzgerald, alleging that he had invented an improvement called the Salamander safe, for which he was about to apply for letters patent, for the consideration of five thousand dollars, sold the interest he then had, or might thereafter have, in this invention, to Enos Wilder; that Enos Wilder, on the 1st day of September, 1843, for the consideration of one dollar, assigned and transferred to the plaintiff all the right, title, and interest which he had derived from Fitzgerald, under the agreement of the 11th of April, 1839 ; that no patent issued for this Salamander safe until the year 1843, when a patent was granted to Daniel Filz *503 gerald, as the original inventor; that no patent for this invention has ever been granted either to Enos or B. G. Wilder, either as inventor or assignee of this safe; that the title, whatever it may be, rests upon the agreément between Fitzgerald and Enos Wilder, of the 11th of April, 1839, before the patent to the former.

It must be recollected, that this is an- action at law; and in order to maintain it, the plaintiff was bound to set out and to prove a legal title. Has he done either ? What was the character of the interest or title transferred from Fitzgerald to Enos Wilder ? This could not tran'scend the interest or title possessed by Fitzgerald himself; and what was this ? A title to any specific machine which he may have constructed, and of which no person could rightfully deprive him; and a claim upon the good-will and gratitude of the community, if in truth he should have conferred upon them a benefit by the discovery and construction of his machine. I speak now in reference to rights derivable from the common law; and independently of the Constitution or of statutory provisions. The mere circumstances of inventing and constructing .a machine' could no more inhibit its imitation, than would the structure Or interior arrangement of a house of peculiar ingenuity or convenience prevent the like imitation by any one who could possess himself of its plan. The mere mental process of devising an invention enters not into the nature of property according to the common law; it forms no class or division in any of its enumerations or definitions of estates or property, and is a matter quite too shadowy for the practical character of that sturdy system.

A doctrine contrary to this, though with some discrepancy amongst the judges as to its extent, seems at one time to have obtained in the King’s Bench, as propounded in the case of Millar v. Taylor, in 4 Burrow, 2305, in opposition to the profound and unanswerable reasoning of Mr. Justice Yatés; but upon a review of the same question in-the Lords, in the case of Donaldsons v. Becket and others, the doctrine of the King’s Bench was repudiated, and that of the common law, as asserted by Yates, Justice, vindicated and restored. And,' indeed, if, according to the opinions of some of the judges in the case of Millar v. Taylor, the mere mental process of invention constituted an estate or property at the common law, and property vested in perpetuo, except so far as it should be transferred by the owner, it is difficult to perceive the necessity of a cautious and complicated system for the investment and security of interests already perfect, and surrounded with every guard and' protection- which is inseparable under the *504common law from every right it has created or recognized. But if the mere mental and invisible process of invention, apart from the specific,.sensible, and individual structure, can be classed at all as property at law, it must partake of the character of a chose in action, much more so than an obligation or contract, the terms and conditions of which are defined and assented to by the contracting parties. To choses- in action, it can scarcely be necessary here to remark, assign-ability is imparted by statutory enactment only, or by commercial usage. To hqld that the single circumstance of invention creates an estate or property at law, and an estate and legal title transmissible by assignment, appears to me a doctrine not merely subversive of the common law, but one which contravenes the origin and course of- legislation in England in relation to patent rights, and renders useless and futile both the constitutional provision and all the careful enactments of Congress for. the security and transmissibility of the same rights. For why, as has been already remarked; should that provision and these enactments have been made for the establishment and security of that which was established and safe independently of both ? I hold it, then, to be true, that the circumstance of invention invests no such perfect estate or-right of property as can be claimed and enforced at law. or in equity against the user of the same invention, either by subsequent inventors or imitators, and that any estate or property in the mere mental process of invention must be traced to and deducibl'e from thé Constitution and the acts of Cohgress alone. I cannot but regard .as mischievous and alarming an attempt to introduce a quasi and indefinite, indefinable, and invisible estate, independently of the Constitution and acts of Congress, and unknown to the rules and principles of the common law.

It is the patent alone which creates an estate or .interest in the invention known to the law, and which can be enforced either at law or in equity, either by the inventor .or by the person to whom, by virtue of the statute, he may assign his rights. Down to the act of Congress of 1837, nothing but the estate, interest, or property created or' invested by the patent itself, was made assignable. The language of • the law is, that “ every patent,” “ the exclusive right under any patent,” “ the thing patented,” may be assignable. The fact or existence of a patent is in.every instance inseparable from the right given. It is this’ fact and this only which impresses the quality of. assigna'bility. Of course, under these .provisions there could be no .transfer of the legal title previously to a^patent.

By section sixth of the, act of Congress approved March *5053d, 1837, it is provided-that thereafter any patent to be issued may be made to the assignee of the inventor or discoverer, upon the conditions set forth in that section. Yet still it is presumed that, until the issuing of a patent, so far is it from being-true that a legal estate or title existed in such assignee, it is clear, on the contrary, that no legal title existed before the patent in the inventor himself, for it is the patent which constitute^. his title. • Of. course, then, the assignee can at most hold nothing but an equity under such an assignment, which he may insist upon under this assignment against-the inventor or against the government; but he has no legal title by force merely of such an assignment, and a fortiori he has no legal title, if the patent, notwithstanding such an assignment, is in fact issued to the inventor, but is thereby entirely excluded from all pretension to a legal title. Thus, in the case befóte us, the patent under which the plaintiff claims was, subsequently to the 'agreement between Fitzgerald and Enos Wilder, issued to Fitzgerald, the inventor, and, according to the proofs in the cause, has never been renewed to Enos Wilder, nor to any claimant under him, nor been assigned to any such claimant, but remains still in the alleged inventor, Fitzgerald. It seems to me, then, indisputable, that the legal title indispensable for the maintenance of this suit at law never was in the plaintiff, and that he could not maintain the action.

The second instance in which I hold the learned justice who tried this cause to have erred is that in which he instructed the jury as follows: — “ That if Conner had not made his discovery public, but had used it simply for his own private purpose, and it had been finally forgotten or abandoned, such discovery and use would be. no obstacle to the taking out of a patent by Fitzgerald, or those claiming under him, if he .be an original, though not the first, inventor or- discoverer of the improvement.” In considering this instruction of the learned judge, the first vice with which it appears to be affected is its violation of a rule thought to be universally applicable to instructions to juries in trials at law; and that rule is this, that instructions should always arise out of, and be limited to, the facts or the evidence in the cause to which the questions of law propounded from the bench should be strictly applicable ; and that instructions which are general, abstract, or not springing from, and pertinent to, the facts of the case, are calculated to mislead the jury, and are therefore improper. Tried by this rule, the instruction of the learned judge, so far as it relates to Conner’s not having made his discovery public, or having finally forgotten or abandoned it, is certainly irrelevant to and unsustained by, any evidence in the record. So far is the *506existence of such testimony from being shown, the converse is proved and is justly inferrible throughout; for although it does hot appear that Conner advertised his invention in the public papers, or claimed a patent for it, it is admitted that he used this, safe in an extensive business establishment, to which it is certain from the nature of. his business the public had access; and it is not pretended that he made any effort at concealment of what he had invented, and the record is entirely destitute of evidence of an abandonment of his invention. Ás to the assumption of his having forgotten it, there is neither a fact, an inquiry, nor conjecture in thé testimony pointing to such a conclusion. The instruction appears to me to be wholly gratuitous and irrelevant. But supposing this instruction to. have been founded upon testimony introduced before the jury, let us consider for a momént its correctness as a rule of law applicable to this cause. This charge, it must be recollected, admits that Conner'was, or might have been, the first inventor; and, notwithstanding, asserts that Fitzgerald, though posterior in time, might, upon the conditions and considerations assumed by the judge, become the owner of the right. Are these conditions warranted, either by the rules of public policy, or by the terms aUd language of legislative provisions on such subjects ? It is said that patent privileges are allowed as incitements to inventions and improvements by which the public may be. benefited. This position,- that may be conceded in general, should not be made a means of preventing the great and public purposes its legitimate enforcement is calculated to secure. The admission of this principle leaves entirely open the. inquiries, Whether he is more, the benefactor of the public who makés a Useful improvement which he generously shares with his fellow-citizens, or he who studies some device whidh .he denies to all, and limits by every means in his power to a lucrative monopoly ; and still more, whether the latter shall be permitted to seize upon that which had already (ás is here admitted) been given to the public, thereby to levy contributions, not only on the community' at large, but upon him.,even who had been its generous benefactor. It was doubtless to prevent consequences like those here presented* that the priority and originality of inventions are so uniformly and explicitly insisted upon in all the legislation of Congress, as will presently be shown. The tendency of the learned judge’s charge to mislead the jury, from its want of precision, and its failure to define any certain predicament Upon whieh the action of the jury should be founded, is of itself an insuperable objection-to that charge. Thus it is said, if Conner (t had not made his discovery public.” In what mo'de t it may be asked. What form-of publicity did *507the learned judge intend the jury should'require? It is shown that Conner used his safe publicly; that is, he concealed it from no one ; and if any mode or kind of publication or concealment was requisite, either to establish or conclude the right of Conner, or to conclude common right (a delinquency in the nature of a forfeiture), surely that mode, if found either in any statute, or in the rules of the common law, ought to have been clearly laid down, so as to guard the rights of all. In the next place, it is said by the learned judge, that, if Conner had abandoned this improvement which the charge admits him to have invented, this would justify a patent to another who had not known of the improvement, although a subsequent inventor. I have always understood it to be indisputable law, • that wherever an inventor abandons or surrenders an invention or improvement which he has certainly made, and neither claims an exclusive right in himself nor transfers .it to another, the invention or 'improvement is given to the public; but by the charge in this case, such an abandonment transfers an exclusive right to one who, by the case supposed, is admitted not'to be the first inventor. So, too, with respect to the hypothesis of the learned judge that the invention had, or might have, been forgotten. To this the same objections of vagueness and uncertainty, and the graver objection of injustice to the real inventor, or to the public, are applicable. By whom and for what interval of time must this improvement have been forgotten, in order to transfer it from the originator thereof ? For a term of years ? And if so, for how long a term ? But suppose he forgets it for his lifetime, shall his executor or his posterity, upon the exhibition of indisputable proofs of the invention, yea, the very machine itself, perfect in all its parts and in its operation, be cut off? This surely cannot be; but, at any rate, the jury should have been furnished with some rule or measure of obliviousness, if this was to be made the substantive' cause' of deprivation as to the original inventor, or the foundation of right and of exclusive right in one confessedly not the first inventor. An attempt has been made to compare the doctrine propounded by the court to what it might be thought is the law as applicable to the discovery, or rather recovery, of the processes employed in what have been called the lost arts. This illustration is in itself somewhat equivocal, and by no means satisfactory ; for if that process could certainly be shown to be the same with one claimed by the modern' inventor, his discovery could scarcely have the merit of originality; or be the foundation of exclusive right. But, in truth, the illustration attempted to be drawn from a revival of a lost art is not apposite to the present case. The term lost art is applicable pecu*508liarly. to certain monuments of. antiquity still remaining in the world, the process of whose accomplishment has been lost for centuries, has been irretrievably swept from the earth, with every vestige of the archives or records of the nations with- whom those arts existed, and the origin or even the identity of which process none can certainly establish. And if a means of producing the effect we see and have amongst us be discovered, and none can either by history or tradition refer to a similar or to the identical process, the inventor of that means may so far claim the merit of originality, though the work itself may have been produced possibly by the same means. But not one principle drawn from such a state of things can be applied to a recent proceeding, which counts from its origin scarcely a period of fifteen years. In fine, this ruling of the learned judge is regarded as being at war not less with the policy and objects than it is with the express language of all the legislation by Congress upon the subject of patent rights, which legislation has uniformly constituted priority of invention to be the foundation and the test of all such rights. Thus in the act of April 10th, 1790,-the first patent law, (Stat. at Large, 109,) it is declared by the first section, “ That upon the application of any person or persons,' &c., setting-forth that he, she, or they hath or have invented or discovered any useful art, &c., not before known or used,” &c.; and the second section of the same statute, requiring a specification of any invention or discovery, declares that it shall be so described as to distinguish'it from all other things known or used.”

,. The act of February 21st, 1793, (1 Stat. at Large, 318,) provides, that when any citizen or citizens of the United States shall allege that .he or they have invented any “ new and useful art, &c., not known or used before the. application," &c.

By the act of April 17th, 1800, (2 Stat. at Large, 38,) which extends the privilege of patents to aliens, proof is required that the art, invention, or discovery hath not been known or used in that or any foreign, country. It is true that this requisition has been so far relaxed as to admit of the patenting in this country inventions which had' been invented and used abroad,-but with respect to this country the invention, &c. must still be original.

In the act of July 4th, 1836, (5 Stat. at Large, 117,) reorganizing the Patent-Office, the language of the sixth section is as follows: “ That any person or persons having discovered or invented any new and useful art, &c., not known or used by others before his or their discovery .” &c. ' The language and import of the laws here cited are too plain to require comment, *509and I think that the production of a single instance from tile statute-book may safely be challenged by which the requisites above mentioned have been dispensed with. Every law, on the contrary, has emphatically demanded originality and priority as indispensable prerequisites to patent privileges, and every aspirant to such privileges is. expressly required to swear to .these prerequisites, as well as to establish them. These tests ordained by the laws are not only founded upon the true reason for the privileges conferred, but they are simple and comprehensible ; whereas the innovations permitted by the ruling of the learned judge not only conflict with the true reason and foundation of patent privileges, but tend to an uncertainty and confusion which cannot but invite litigation and mischief. I think that the judgment of the Circuit Court should be reversed^ and the cause remanded for a venire facias de novo.

Mr. Justice GRIER also dissented.

Order.

This cause came on to be heard on the transcript of the record from the Circuit Court of the United States for the Southern District of New York, and was argued by counsel. On consideration whereof, it is now here ordered and adjudged by this court, that the judgment of the said Circuit . Court in this cause be, and the same is hereby, affirmed, with costs and damages at the rate of six per centum per annum.

KATZMANN, Circuit Judge.

This case calls upon us to decide whether the New York Mercantile Exchange, Inc. (“NYMEX”) can enforce a copyright in the settlement prices it produces to value customers’ open positions. We hold that even if these prices are created by NYMEX — a question we do not resolve— there has been no infringement because enforcing the copyright here would effectively accord protection to the idea itself. We also consider the district court’s dismissal of IntercontinentalExchange, Inc.’s (“ICE”) state law claims and conclude that it did not abuse its discretion in declining to exercise supplemental jurisdiction. The judgment of the district court is affirmed.

I.

The New York Mercantile Exchange, Inc. is an exchange for the trading of futures and options contracts for energy commodities. It operates a physical trading floor in New York City where brokers and traders transact. Two of its most successful futures contracts are those for Henry Hub natural gas and West Texas Intermediate crude oil. Intercontinenta-lExchange, Inc. operates an electronic, Internet-based, trading market for trading of physical commodities and over-the-counter derivative contracts.

A futures contract requires the delivery of a commodity at a specified price at a specified future time, though most contracts are liquidated before physical delivery occurs. For each day when the contract remains open (i.e. before delivery or liquidation), NYMEX’s Clearing House1 evaluates the change in value of its cus*111tomers’ open contracts. This process, known as “marking-to-market” the customer’s open position, determines whether a customer must post additional margin or, instead, receives payments on margin. The settlement prices are used to value the open positions.

To that end, on a daily basis NYMEX determines the settlement prices for each futures contract. The Commodity Futures Trading Commission (“CFTC”), pursuant to the Commodities Exchange Act (“CEA”), requires NYMEX, as a Designated Contract Market, to record and disseminate these prices. See 17 C.F.R. 16.01(b). NYMEX defines a settlement price as “the value, at the end of trading each day, of a particular futures contract for a particular commodity for future delivery at a particular time.” For example, today’s settlement price for an October 2007 crude oil contract is the fair market value, today, of a contract obliging the purchase or sale of a specified amount of crude oil in October 2007.

The subcommittees of the NYMEX Settlement Price Committee (the “Committee”) are charged with determining the settlement price of each open futures contract for each commodity. Unlike on a securities exchange, the settlement price may not be the final trade, for two reasons. First, because of the nature of the trading, it is not always clear which trade was the closing trade. Traders handwrite their transactions on cards which are thrown into the center of trading rings, scooped up, time stamped, and sent for processing. Because the cards may be “scooped up” out of order, the card with the latest time stamp may not represent the final trade of the day. Second, on any given day, 32 or 33 months of crude oil futures contracts and 72 months of natural gas futures contracts are being traded. For the “outer” months, those further from the trading date, there is often little or no trading on a particular day.

After the trading floor closes, the Committee determines the appropriate price for the delivery of crude oil for each of the next 32 or 33 months and for delivery of natural gas for each of the next 72. NY-MEX’s rules distinguish between months with sufficient trading and open interest2 and months without (for simplicity, we will call the former “high-volume” months and the latter “low-volume” months). For high-volume months, settlement prices are based on a formula: “a weighted average of all trades done within the closing range.” While only a small number of delivery months are high-volume on any given day, ICE claims that these months “represent a large percentage of the total daily trading volume in these contracts.”

For low-volume months, the extent of the Committee’s creative judgment is disputed. NYMEX asserts that the membership “considers, sifts, weighs and extrapolates from a wealth of data at the close of trading to reach an opinion” as to the appropriate settlement price. ICE contends that there is little judgment involved because the subcommittees only review “objective market data,” and, in practice, “look at settlement prices for the near month contracts — ie., those that are determined by mathematical formula — and then extrapolate to determine the remainder of the settlement prices based on the changes in the month-to-month spread relationships in the various contracts compared to the previous day.”

NYMEX’s rules also provide that the Committee may override the settlement price for a high- or low-volume month. *112The frequency of the use of the override provision is disputed.

After determining the settlement prices, NYMEX uses and disseminates them in several ways. First, its Clearing House uses the settlement prices to place current values on the accounts of all NYMEX Clearing Members whose clients have open positions. The members then use the settlement prices to mark-to-market their customers’ open positions. Second, it publicly discloses those prices by the next business day, as required by the CFTC. See 17 C.F.R. Pt. 38, App. B, Core Principle 8. Third, between the time of creating the prices and the required public disclosure the following day, NYMEX supplies them to market data vendors such as Reuters pursuant to license agreements. These vendors then disclose the prices to their subscribers. Fourth, the prices are disclosed to the public on NYMEX’s website.

ICE is one of the subscribers that receives settlement prices through a licensed market data vendor. While ICE cannot clear trades itself, it contracts with the London Clearing House (“LCH”) to do so. ICE copies NYMEX’s settlement prices and forwards them to LCH which clears ICE’s customers’ trades. One small exception, however, is when there has been trading on an ICE contract and it is not the final day of trading for that particular contract. In that case, the ICE committee adjusts the NYMEX settlement price one “tick” (one cent for crude oil and one tenth of a cent for natural gas) closer to the weighted average price of the ICE trades. This exception is used more often for natural gas contracts, but even there, most of the prices transmitted to LCH are copied from NYMEX.

In March 2002, NYMEX sought a copyright for its database including the settlement prices. After the Copyright Office informed NYMEX that it was unwilling to provide a copyright in settlement prices, NYMEX filed a replacement application and obtained a copyright for its database only.

NYMEX brought this suit in November 2002, in the Southern District of New York, alleging copyright infringement,3 trademark infringement under federal and state law, and a state law claim of tortious interference with contract.

ICE moved for summary judgment on all of NYMEX’s claims; NYMEX cross-moved for partial summary judgment on the eopyrightability of NYMEX’s settlement prices and on its state law tortious interference claim. The district court granted ICE summary judgment on both the copyright and trademark claims; it then declined to exercise supplemental jurisdiction and dismissed the state law claims as well. New York Mercantile Exch., Inc. v. IntercontinentalExchange, Inc., 389 F.Supp.2d 527, 530, 547 (S.D.N.Y.2005).

The district court resolved the copyright question without “determining] the degree of creativity that goes into the setting of individual prices.” Id. at 541 n. 10. Instead, it found two grounds for depriving NYMEX of copyright protection in its settlement prices. First, it found that these prices are “real-world facts,” the “actual price at which a futures contract must be settled,” and are “widely publicized and used as benchmarks by market participants.” Id. at 542. “[B]ecause ... the *113only way to express the idea of a settlement price,” is a single number, the idea “cannot be distinguished from its expression, [and] the merger doctrine applies.” Id. at 542-43. Second, the settlement prices are “non-copyrightable words or short phrases.” Id. at 543. The district court did not address ICE’s additional arguments for denying NYMEX’s copyright claim. Id. at 544 n. 13. NYMEX does not appeal the trademark ruling so that portion of the district court’s decision is not relevant to this discussion.

NYMEX argues that the district court erred because settlement prices are copyrightable. Furthermore, it argues that the district court abused its discretion in declining to exercise supplemental jurisdiction over the state law claims. ICE defends the district court’s decision and argues for several alternate grounds to affirm. The United States has filed an ami-cus cuña brief in support of ICE, arguing that the settlement prices are facts, even if they are not facts the idea of the prices have merged with their expression, and even if merger did not apply, the prices are not copyrightable because they are short phrases.

II.

We review grants of summary judgment de novo. See Island Software & Computer Serv., Inc. v. Microsoft, Corp., 413 F.3d 257, 260 (2d Cir.2005); Dallas Aerospace, Inc. v. CIS Air Corp., 352 F.3d 775, 780 (2d Cir.2003). In doing so, we “construe the facts in the light most favorable to the non-moving party and must resolve all ambiguities and draw all reasonable inferences against the movant.” Dallas Aerospace, 352 F.3d at 780. We review for abuse of discretion a district court’s decision to decline to exercise supplemental jurisdiction. Valencia ex rel. Franco v. Lee, 316 F.3d 299, 305 (2d Cir.2003).

The Constitution provides that “Congress shall have power to ... promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” U.S. Const. Art. I, § 8, cl. 8. Pursuant to this power, Congress enacted the Copyright Act, 17 U.S.C. § 101 et seq.; the latest version of the Act states: “Copyright protection subsists ... in original works of authorship fixed in any tangible medium of expression, now known or later developed, from which they can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.” Id. § 102(a). One of the categories of protectable works listed in the Act is “literary works,” id. § 102(a)(1), defined as “works, other than audiovisual works, expressed in words, numbers, or other verbal or numerical symbols or indicia, regardless of the nature of the material objects, such as books, periodicals, manuscripts, phonorecords, film, tapes, disks, or cards, in which they are embodied,” id. § 101.

“The sine qua non of copyright is originality.” Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., Inc., 499 U.S. 340, 345, 111 S.Ct. 1282, 113 L.Ed.2d 358 (1991). Originality is a constitutional requisite and requires that the “work was independently created by the author (as opposed to copied from other works) and that it possesses at least some minimal degree of creativity.” Id. Originality has been described as “founded in the creative powers of the mind ... the fruits of intellectual labor,” In re Trade-Mark Cases, 100 U.S. 82, 94, 25 L.Ed. 550 (1879) (emphasis omitted), and evidencing “intellectual production, ... thought, and conception,” Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 60, 4 S.Ct. 279, 28 L.Ed. 349 (1884).

*114In contrast, “facts do not owe their origin to an act of authorship. The distinction is one between creation and discovery: The first person to find and report a particular fact has not created the fact; he or she has merely discovered its existence.” Feist, 499 U.S. at 347, 111 S.Ct. 1282. “[A]ll facts — scientific, historical biographical, and news of the day ... may not be copyrighted and are part of the public domain available to every person.” Id. at 348, 111 S.Ct. 1282 (internal quotation marks omitted). As Professor Nim-mer explains in his leading copyright treatise:

The “discoverer” of a scientific fact as to the nature of the physical world, an historical fact, a contemporary news event, or any other “fact,” may not claim to be the “author” of that fact.... The discoverer merely finds and records. He may not claim that the facts are “original” with him, although there may be originality and hence, authorship in the manner of reporting, ie., the “expression,” of the facts. As copyright may only be conferred upon “authors,” it follows that ... discoveries as facts per se may not be the subject of copyright.

1-2 Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 2.03[E] (2006) (“Nimmer”) (internal footnote call numbers omitted). NYMEX concedes, and indeed it is “universally understood,” that if a settlement price is a fact, it cannot be copyrighted. Feist, 499 U.S. at 344, 111 S.Ct. 1282.

In elucidating the distinction between original creation and discovery, the Feist Court provided a helpful example: census takers are not authors of the census data. Census takers merely discover the appropriate population figure; “in a sense, they copy these figures from the world around them.” Id. at 347, 111 S.Ct. 1282.

The question then, is one of characterization: does the Committee create the settlement prices, or is it more accurate to view the Committee’s task as like that of a census taker, copying the market’s valuation of futures contracts? While the line between creation and discovery is often clear-cut, we recognize that it is a difficult line to draw in this case. For reasons we explain below, we believe there is a strong argument that, like the census taker, NY-MEX does not “author” the settlement prices as the term is used in copyright law.4 See Burrow-Giles, 111 U.S. at 57-*11558, 4 S.Ct. 279 (defining “author” as “he to whom anything owes its origin; originator; maker” (internal citation omitted)). Still, because this is a close question, particularly at the summary judgment stage, we do not decide whether settlement prices are unoriginal, and instead affirm based on the merger doctrine, as discussed in Part III.

In considering how to characterize the Committee’s task we start with what a settlement price measures. For high-volume months, settlement prices are determinations of how the market values a particular futures contract. After all, there is no dispute that the Committee members seek to determine the appropriate market valuation of each commodity contract, not how the market should value them or will value them. Under this view, the market is an empirical reality, an economic fact about the world, that Committee members are seeking to discover. Thus, there is a strong argument that, as a matter of law, the Committee is discovering facts, not creating predictions or estimates.5 So characterized, there is one proper settlement price; other seemingly-accurate prices are mistakes which actually overvalue or undervalue the futures contract. After all, as NYMEX’s Executive Vice President and Chief Operating Officer has stated, NYMEX’s settlement prices are heavily relied upon because they “reflect actual trades,” not mere sentiment.6

That said, characterizing the Committee’s task as discovery, even for high-volume months, is disputable. The analogy to the census taker is not perfectly apt *116because the discovery of a settlement price is a more abstract concept than the discovery of a population figure; an economic reality is certainly less tangible than a population reality. For example, with sufficiently precise tools we could perhaps measure the population of a metropolis, whereas it is not clear that we could ever precisely calculate the appropriate valuation of a particular futures contract.

The characterization inquiry is even more difficult for low-volume months, particularly at the summary judgment stage. We recognize that settlement prices are determined for 32 or 33 months of crude oil futures contracts and 72 months of natural gas futures contracts every day. There is generally less trading in contracts for the outer months. It may be the case that there is a material factual dispute whether, for certain outer months, there is so little trading that there is no real market to speak of. Where there is no market the Committee’s work appears closer to creation, to making “predictions ... of expected values.” CCC, 44 F.3d at 63. For such months, then, we are particularly reluctant to hold, as a matter of law, that the Committee is discovering the settlement prices.7 In any event, it is unnecessary to resolve this question and we instead affirm because even if NYMEX creates the settlement prices, NYMEX’s claim fails due to application of the merger doctrine, as explained below.

III.

Assuming that the settlement prices are created and not discovered, we consider whether the merger doctrine should be applied to withhold copyright protection.8 “It has been long accepted that copyright protection does not extend to ideas; it protects only the means of expression employed by the author.” CCC, 44 F.3d at 68; see also 17 U.S.C. § 102(b) (“In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.”). Because “ideas are too important to the advancement of knowledge to permit them to be under private ownership,” and because “open public debate, which is essential to a free democratic society, requires free access to the ideas to be debated,” ideas cannot be copyrighted. CCC, 44 F.3d at 69. Instead, “only the manner of [an idea’s] ‘expression’ ” is copyrightable. Id.

To ensure free access to ideas, courts have applied the merger doctrine such that “even expression is not protected in those instances where there is only one or so few *117ways of expressing an idea that protection of the expression would effectively accord protection to the idea itself.” Kregos v. Associated Press, 937 F.2d 700, 705 (2d Cir.1991); see also 4-13 Nimmer § 13.03[B][3] (explaining that “courts have invoked the merger doctrine” where “rigorously protecting the expression would confer a monopoly over the idea itself, in contravention of the statutory command”). We must exercise “considerable care,” in analyzing merger: “if the merger doctrine is applied too readily, arguably available alternative forms of expression will be precluded; if applied too sparingly, protection will be accorded to ideas.” Kregos, 937 F.2d at 705.

We begin by identifying the “idea” that might be merging with its expression. The district court stated that the idea is “the price of a particular futures contract at the close of trading.” NYMEX, 389 F.Supp.2d at 541. NYMEX disputes this approach and argues that the idea is “that a sound and reasonable opinion of fair market value for each NYMEX contract as of the close of open outcry trading on the NYMEX floor each day may be achieved by assessing trades, bids, and offers and (in various instances) off exchange information, particularly developed late in the trading day.” The United States disagrees, and argues instead that a “settlement price is the expression of a much more specific idea, for example, that a Henry Hub natural gas futures contracts is worth $25 dollars [sic] at the end of a given day.”

Even if we accept NYMEX’s formulation, however, the expression here merges with the idea. In this Circuit, we look at the range of possible expressions and consider whether all possible expressions are so “substantially similar” that granting the copyright would bar others from expressing the underlying idea.9 Hart v. Dan Chase Taxidermy Supply Co., Inc., 86 F.3d 320, 322 (2d Cir.1996). To survive summary judgment, .then, NYMEX must demonstrate that the range of possible settlement prices is broad enough that any possible expression will not “necessarily be ‘substantially similar.’ ” Id.

It cannot meet that burden. NYMEX is not seeking a copyright in the process whereby settlement prices are created.10 Instead, the “fair market value for each NYMEX contract” is expressed as a settlement price, and it is these prices that *118NYMEX seeks to bar ICE from using. It is undisputed that all possible expression takes the same form, a number. The question then becomes the possible range of that number. Because any settlement price for a particular futures contract would be determined based on the-same underlying market facts, any dissension would be exceptionally - narrow. While NYMEX contends that there are “numerous possible variations ... as to what the Settlement Prices should be,” it has not demonstrated a range of possible variations that would preclude application of the merger doctrine. See Colavito v. New York Organ Donor Network, Inc., 486 F.3d 78, 81 (2d Cir.2007) (“A party may not defeat a Rule 56 motion based on conjecture alone.”). Instead, the record merely demonstrates that, at times, Committee members have disagreed on the exact settlement price. To grant NYMEX copyright protection here “would effectively accord protection to the idea itself,” Kregos, 937 F.2d at 705, and bar ICE or other competitors from valuing NYMEX’s contracts. We decline to do so.

Furthermore, policy considerations weigh heavily in determining the appropriate application of the merger doctrine. See CCC, 44 F.3d at 72 (noting that “Kregos, [937 F.2d 700] ... makes a- policy judgment as between two evils,” and withholding application of the merger doctrine where “[u]nbridled application of the merger doctrine would undo the protection the copyright law intends to accord”). “[T]he objectives pf the copyright law ... [are] to promote the advancement of knowledge and learning by giving authors economic incentives (in the form of exclusive rights to their creations) to labor on creative, knowledge-enriching works.” Id. at 65. Without such incentives authors “might direct their energies elsewhere, depriving the public of their creations and impeding the advance of learning.” Id. at 66. Put another way, “copyright is intended to increase and not to impede the harvest of knowledge,” while also “assuring] contributors to the store of knowledge a fair return for their labors.” Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539, 545-46, 105 S.Ct. 2218, 85 L.Ed.2d 588 (1985).

NYMEX needs no such incentives here. In order to establish a functioning commodities market it must have a price at which to settle open positions. Furthermore, NYMEX is required by law to record settlement prices, and it does not challenge the legality of that rule. See 17 C.F.R. 16.01(b)(2) (“Each reporting market shall record the following information ... [t]he settlement price established by each reporting market or its clearing organization.”). Even without copyright in its settlement prices we are confident that NYMEX will not “direct [its] energies elsewhere,” CCC, 44 F.3d at 66, as its exchange requires settlement prices to clear contracts.

We therefore apply the merger doctrine and hold that, in using the settlement prices, ICE “took nothing more than ideas, for which the copyright law affords no protection to the author.” Id. at 68. Because “the expression is essential to the statement of the idea, the expression [is] also ... unprotected.” Id.

As NYMEX cannot prevail due to application pf the merger doctrine, we reach neither the district court’s analysis of the short phrases doctrine nor ICE’s alternative bases on which to affirm.

IV.

Having determined that the district court correctly found that NYMEX’s federal copyright claim fails, we must now consider whether the district court abused its discretion in declining to exercise sup*119plemental jurisdiction over NYMEX’s state law claims. We find that the district court acted well within the bounds of its discretion. “In general, where the federal claims are dismissed before trial, the state claims should be dismissed as well.” Marcus v. AT&T Corp., 138 F.3d 46, 57 (2d Cir.1998). Furthermore, here the district court determined that resolving the state law claim would entail resolving additional issues of fact, NYMEX, 389 F.Supp.2d at 547, making dismissal of those claims after the federal claims had been dismissed particularly appropriate.

V.

For the foregoing reasons, we hereby AFFIRM the judgment of the district court granting summary judgment to the defendants, denying plaintiffs cross-motion for partial summary judgment, and declining to exercise supplemental jurisdiction.

HALL, Circuit Judge,

concurring in part.

I agree with the result reached in the majority opinion; NYMEX settlement prices do not merit copyright protection. I write separately, however, to register my disagreement with the majority’s analysis in Part II of its opinion, where it asserts, unnecessarily in my view, that there exists a “strong argument” NYMEX settlement prices lack the requisite originality to qualify for copyright protection. In so doing, the majority renders dicta that contemplate heightening the standards by which we determine whether a work exhibits “some minimal degree of creativity.” Feist Publ’ns v. Rural Tel. Serv. Co., Inc., 499 U.S. 340, 345, 111 S.Ct. 1282, 113 L.Ed.2d 358 (1991). This is a step contraindicated both by our jurisprudence and by the procedural posture of this case. Because the application of the merger doctrine, as explained in Part III of the majority’s opinion, makes it unnecessary for us to consider this question of originality, I concur in all but Part II of the majority’s opinion.

In this case, NYMEX faced the task of meeting the originality requirement in the context of an appeal from summary judgment. Rooted in the Constitution, the originality requirement of copyright presents an extremely low bar. Feist, 499 U.S. at 345-46, 111 S.Ct. 1282. To prove possession of a valid copyright, a party must show a subject is an “original work of authorship.” 17 U.S.C. § 102(a); Nimmer on Copyright, § 2.03[A] (2004 ed.). The Supreme Court has emphasized that “[ojriginal, as the term is used in copyright, means only that the work was independently created by the author ... and that it possesses at least some minimal degree of creativity.” Feist, 499 U.S. at 345, 111 S.Ct. 1282. “To be sure, the requisite level of creativity is extremely low; even a slight amount will suffice. The vast majority of works make the grade quite easily, as they possess some creative spark, no matter how crude, humble or obvious it might be.” Id. (internal citations omitted); see also Key Publ’ns v. Chinatown Today Publ’g Enters., 945 F.2d 509, 512-13 (2d Cir.1991).

Accordingly, to survive summary judgment on the issue of originality, NYMEX must show, in “view of the facts in a light most favorable” to it, Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986), there exists some “genuine issue as to any material fact,” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 250, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986), of whether the settlement prices contained an “extremely low ... slight amount” of “creative spark, no matter how crude, humble or obvious.” Feist, 499 U.S. at 345, 111 *120S.Ct. 1282. This combined standard is a permissive one indeed, posing a virtually non-existent bar to NYMEX’s efforts to demonstrate the originality of its work.

NYMEX has vouched forcefully for the originality of its settlement prices. NY-MEX asserts it uses “judgment and discretion” to create them; according to one witness, “[t]he whole thing is about opinions.” Indeed, as NYMEX would have it, its Settlement Price Committee is the Shakespeare of the commodities clearing world, crafting its settlement prices like so many numeric sonnets. Naturally, ICE disagrees, and the majority largely credits, albeit in dicta, ICE’s assertions. But, given the permissive originality requirement of our jurisprudence as applied on appeal from summary judgment, I cannot so easily accept that ICE has made such a “strong argument” against the originality of NYMEX’s settlement prices.

The majority attempts to skirt this issue, implicitly acknowledging there may be an element of creativity in producing settlement prices, but asserting that NYMEX “is discovering facts, not creating predictions or estimates.” I respectfully suggest such a rationale is mistaken for at least three reasons.

First, in supposing that NYMEX discovers, rather than creates, the settlement prices, the majority contradicts prior precedent of this Circuit. In CCC Information Services, Inc. v. Maclean Hunter Market Reports, 44 F.3d 61 (2d Cir.1994), we held that an individual car price valuation could be creative; the creativity lay in the integration of “a multitude of data sources, but also on professional judgment and expertise.” Id. at 67. Settlement prices are no different; they arise from the integration of data sources (the prices and circumstances of trades during the day), as well as professional judgment or expertise (the opinion of the Settlement Price Committee). Just as CCC found the creativity inhered in the creation of a valuation, not the discovery of a value, so too should we find that any creativity of NYMEX lies in the creation of the settlement price, not in discovery of a price that may or may not exist at any given point in the trading day. See also CDN Inc. v. Kapes, 197 F.3d 1256 (9th Cir.1999) (O’Scannlain, J.).

Second, it would appear the rationale is circular. The majority instructs that although “it is not clear that we could ever precisely calculate the appropriate valu[e]” of a futures contract, settlement prices nevertheless are unoriginal because they are discoverable facts, not creations of NY-MEX. But why are the settlement prices facts, not creations? Unless I misread the majority opinion, their answer turns on the conclusion that settlement prices lack the spark of originality. In other words, the settlement prices are facts, and therefore unoriginal; they are unoriginal because they are facts. This is a tautology, to which I cannot subscribe.

Third, there is good reason to doubt one predicate of the tautology, that settlement prices are “facts.” A settlement price is an “arbitrary price used as the basis for the settlement of contracts through a clearinghouse.” Webster’s Third New International Dictionary 2079 (2002) (emphasis added). It is “the amount ... treat[ed] as the value, at the end of trading each day, of a particular futures contract for a particular commodity for future delivery at a particular time” (Appellant’s Brief, 6) (emphasis added). The settlement price does not replicate individual trades; nor does it replicate a weighted average of those trades. To the contrary, NYMEX’s settlement price formula includes an override mechanism, such that even if NYMEX did know the precise “fact” of the day’s weighted average, it could adjust that average to conform to *121what it considers, in its judgment or opinion, to be the better “arbitrary price,” or “amount ... treat[ed] as the final value.” Thus, given the posture of this case, one could easily conclude that settlement prices are not preexisting facts about the world; they are evaluative opinions created by NYMEX. True, they derive in part from facts about the world, such as the prices and circumstances of trades during the day (See Majority Opinion at 114, fn 4). But this proves nothing; many copyright-worthy creative works derive in part from facts.

As the majority acknowledges, however, this panel need not — and, indeed, does not — reach the question of originality of settlement prices other than in dicta. Instead, the majority grounds its affirmance of the denial of copyright protection to NYMEX’s settlement prices exclusively by applying the merger doctrine. Because I find the majority’s analysis of the merger doctrine to be persuasive, and because its discussion of originality is unnecessary, I concur in all but Part II of the majority’s opinion.

POSNER, Circuit Judge.

Stephen Bucklew brought this suit in a federal district court in Wisconsin against two affiliated companies that we’ll call “HAB.” The suit charges copyright infringement, fraud, conversion, and violation of RICO. The district judge dismissed all but the copyright claim, which was tried to a jury, resulting in a verdict for Buck-lew of $660,000, which the judge cut down to $395,000. HAB appeals; Bucklew cross-appeals, challenging the dismissal of his noncopyright claims.

Local housing authorities that want grants from the federal department of Housing and Urban Development have to complete forms prescribed by HUD. The forms require not only specific basic data such as salaries and other categories of proposed expenditure by the grant applicant but also simple arithmetical transformations of the data, such as adding the numbers in particular cells in the table of basic data. Bucklew developed and copyrighted software, intended to be used in conjunction with standard spreadsheet ap*926plications such as Lotus 1-2-3 and Excel, for doing these transformations and displaying them in tables. The essential transformation involves entering the basic data in an electronic copy of the HUD-prescribed form and applying to those data an algorithm that picks out the relevant cells and performs the relevant operation (namely addition) on them and displays the results in tabular form. Standard spreadsheet applications make this transformation relatively easy to program because they include functions such as DSUM, a simple command for adding up the numbers in the cells identified by a criterion specified by the programmer and displaying the results of the addition. Bucklew does not claim copyright in either the spreadsheet applications or DSUM. But there is more, though not a great deal more, to his product than these programs. Decisions have to be made regarding choice and size of font, the size of cells and columns, whether and where to use color, the wording of labels and headings (other than those prescribed by the HUD forms), and whether to use boldface or italics for column headings. These decisions were made by Bucklew. The ones we have named all involve the appearance of the forms, but software that is read only by the computer and not by its human user is also copyrightable, and Bucklew does claim copyright in the “construction” as well as display of his forms; but it is unexplained what he means by this or whether it has been infringed.

The trial focused on one of the four forms (as the parties refer to the conversion of a prescribed HUD form into an electronic form that computes and displays the arithmetic manipulations that HUD requires) copyrighted by Bucklew that he claimed were copied by HAB. This form is a transformation of HUD form 52566, which requires salary data; the other three forms require data for other categories of expenditure. HAB grudgingly concedes that the evidence compels an inference that it copied Bucklew’s form 52566 to create its own form 52566. Some similarities between a copyrighted work and a work alleged to infringe it are consistent with an inference of independent creation, and in that case evidence that the alleged infringer had access to (that is, saw or at least could have seen) the copyrighted work is indispensable. Ty, Inc. v. GMA Accessories, Inc., 132 F.3d 1167, 1169-70 (7th Cir.1997). But when the similarities concern details of such an arbitrary character that the probability that the infringer had duplicated them independently is remote, an inference of copying may be drawn without any additional evidence. Id.; Selle v. Gibb, 741 F.2d 896, 901, 903-OS (7th Cir.1984); Gaste v. Kaiserman, 863 F.2d 1061, 1067-68 (2d Cir.1988); Bouchat v. Baltimore Ravens, Inc., 241 F.3d 350, 355-56 (4th Cir.2001). These cases say that access can be inferred from a sufficiently striking similarity between the two works, and that is true; but as the Fifth Circuit suggested in Ferguson v. National Broadcasting Co., 584 F.2d 111, 113 (5th Cir.1978), it is more straightforward to say that in some cases proof of access isn’t required.

It is in order to avoid having to prove access that mapmakers will sometimes include a fictitious geographical feature in their maps; if that feature (what is called in the trade a “copyright trap”) is duplicated in someone else’s map, the inference of copying is compelling. Rockford Map Publishers, Inc. v. Directory Service Co. of Colorado, Inc., 768 F.2d 145, 147 (7th Cir. 1985); Melville B. Nimmer & David Nimmer, Nimmer on Copyright, vol. 4, § 13.03[C], pp. 13-75 to 13-77 (2002); see also General Drafting Co. v. Andrews, 37 F.2d 54, 56 (2d Cir.1930) (errors as distinct from deliberately set traps). And that is *927the case with regard to form 52566. Buck-lew’s version contains an arbitrary pattern of boldfacing of cells; HAB’s duplicates it exactly. Bucklew’s form also contains an “output range,” an intermediate table that he carried over from a previous spreadsheet program but that has no function in his current program — or in HAB’s. Had HAB written its program from scratch, it would have had no reason to include an output range — yet it did. And in another intermediate table, one that contains the identical data as the form 52566 itself (the starting point for the transformation), HAB’s version uses the identical headings as Bucklew’s corresponding intermediate table, rather than the headings in the form 52566 with which HAB had begun. (For unexplained reasons, that form had different headings from the form from which Bucklew had begun.) And HAB’s program like Bucklew’s used a font (Swiss font) that was not available in the version of the Windows operating system that HAB used to develop its software. This is further evidence that rather than creating its own form using the software available to it on Windows, HAB had copied Buck-lew’s form.

The evidence of copying with regard to the other three forms is scanty. Indeed, they were barely mentioned at trial. HAB complains that they were smuggled into the trial and that the jury should not have been permitted to consider them. Before trial the district judge, in conformity with his established practice (though we cannot find it stated in writing in the rules of his court or anywhere else), explained to the parties that any objections to exhibits, other than objections based on relevance, had to be made before trial and any exhibit not objected to before trial would become a part of the trial record. The exception for relevance was presumably intended for the situation in which, because of the course the trial might take, a document that had seemed relevant to the issues in the case before trial no longer was. HAB did not object before trial to the introduction into evidence of any of the four forms of Buck-lew’s that were alleged to be infringed. But it argues that it assumed that the forms would be discussed at trial rather than just being handed to the jury, along with the other exhibits that had not been ruled inadmissible, when the jury retired to deliberate.

There was no basis for the assumption. Admissible documents are evidence, just like testimony. They do not require sponsorship. The procedure followed by the district judge is common and sensible. It saves trial time. The Federal Rules of Civil Procedure encourage judges to resolve evidentiary disputes so far as possible before trial in order to shorten trials and minimize interruptions, Fed.R.Civ.P. 16, 26(a)(3) — an especially important consideration in jury trials. Evidently Bucklew was content, having established that HAB had copied his form 52566 to make its form 52566, to let the jury make up its own mind about the other forms. He testified that he believed his other three forms had been copied as well, and his expert’s reports, which though hearsay were admitted into evidence without objection, said the same thing though without elaboration. Had HAB wanted to challenge the inference of copying with evidence of its own, there was nothing to stop it from doing so. Maybe it held back because it thought that the jury would be unlikely to find infringement of the other three forms when the plaintiff had presented so little evidence about them and there was no evidence that HAB had had access to Bucklew’s versions of those forms, though this is irrelevant because Bucklew’s claim is that HAB’s versions of those forms were copied from Bucklew’s form 52566, to which HAB did have access.

*928In this court, HAB’s principal argument about the admissibility of the three forms is that they confused the jury. Just as a document that seemed admissible before trial might be objected to at trial as lacking relevance, so it might be objected to as confusing because it had not been the subject of testimony and without testimony could not be comprehended by a lay person. But there was testimony about the forms — elicited indeed by HAB itself, on cross-examination — enough to justify the judge in overruling the objection.

HAB remains free to argue that there was a failure of proof with respect to the three forms. Bucklew, however, testified without contradiction to certain similarities that indicated copying (though, to repeat, from Bucklew’s form 52566, not from Bucklew’s versions of the three other forms), notably a missing vertical fine where a line would otherwise be expected, explicable only as a trace of the deletion of the unneeded output range that HAB had copied from Bucklew’s form. (A cell to the left of the deleted area had a box around it; apparently when an HAB programmer deleted material to the right of that cell, he stripped the box around the cell of its right-hand side vertical border.) What is true is that there are more differences between HAB’s version of the other three forms and form 52566, and this may be relevant to infringement, as we are about to see; but it does not rebut an inference of copying. To return to the map example, the presence of a fictitious geographical feature in the alleged infringer’s map might be conclusive proof of copying even though the copier had changed the colors or other features of the map, so that at a glance the two maps appeared quite different.

With respect to all four forms HAB argues that even if there was copying there was no infringement. It appeals primarily to the copyright doctrines of merger and scenes á faire. The first of these confusing labels refers to the situation in which there is only one feasible way of expressing an idea, so that if the expression were copyrightable it would mean that the idea was copyrightable, and ideas are not copyrightable. E.g., Ets-Hokin v. Skyy Spirits, Inc., 225 F.3d 1068, 1082 (9th Cir.2000); Kregos v. Associated Press, 937 F.2d 700, 705 (2d Cir.1991); Kern River Gas Transmission Co. v. Coastal Corp., 899 F.2d 1458, 1459, 1463-64 (5th Cir.1990). The standard citation remains Baker v. Selden, 101 U.S. 99, 25 L.Ed. 841 (1879). Selden had published (and copyrighted) a book describing a bookkeeping system that he had invented, and he illustrated the book with blank bookkeeping forms. Baker copied the forms, rearranging columns and using different headings, and sold them to people who wanted to use Selden’s system. This was held not to be copyright infringement, even though Baker had copied part of a copyrighted work, since otherwise Selden would have had a monopoly over his bookkeeping system (which was an idea, and hence not copyrightable) that he could have exploited by insisting that anyone wanting to use the system buy the forms necessary for using it from him. If Bucklew were claiming copyright in the tabular presentation of the summary data required by HUD, this case would be governed by Baker v. Sel-den. But he is not. He is claiming copyright in tables configured in an optional way, tables that are the product of format choices that are not unavoidable, for which indeed there were an immense number of alternative combinations any one of which HAB was free to use in lieu of Bucklew’s. See, e.g., Mason v. Montgomery Data, Inc., 967 F.2d 135, 138-41 (5th Cir.1992); Kregos v. Associated Press, supra, 937 F.2d at 706-07.

*929The fact that Bucklew’s formatting choices do not reflect a high degree of originality is irrelevant. When as in this case a work in which copyright is claimed is based on work in the public domain, the only “originality” required for the new work to be copyrightable (the very term is a misnomer) is enough expressive variation from public-domain or other existing works to enable the new work to be readily distinguished from its predecessors. Alfred Bell & Co. v. Catalda, Fine Arts, 191 F.2d 99, 102-03 (2d Cir.1951). “Originality in this context means little more than a prohibition of actual copying.” Id. at 103 (internal quotation marks deleted); Three Boys Music Corp. v. Bolton, 212 F.3d 477, 489 (9th Cir.2000). That undemanding requirement is satisfied in this case; any more demanding requirement would be burdensome to enforce and would involve judges in making aesthetic judgments, which few judges are competent to make.

The doctrine of scenes á faire (another confusing label, literally “scenes for action,” which the Oxford English Dictionary tells us is a theatrical term meaning “the most important scene in a play or opera, made inevitable by the action which leads up to it” — which is not the legal doctrine at all) teaches, sensibly enough, that a copyright owner can’t prove infringement by pointing to features of his work that are found in the defendant’s work as well but that are so rudimentary, commonplace, standard, or unavoidable that they do not serve to distinguish one work within a class of works from another. Reed-Union Corp. v. Turtle Wax, Inc., 77 F.3d 909, 913-14 (7th Cir.1996); Atari, Inc. v. North American Philips Consumer Electronics Corp., 672 F.2d 607, 616-17 (7th Cir.1982); Williams v. Crichton, 84 F.3d 581, 587-89 (2d Cir.1996). Every expressive work can be decomposed into elements not themselves copyrightable— the cars in a car chase, the kiss in a love scene, the dive bombers in a movie about Pearl Harbor, or for that matter the letters of the alphabet in any written work. The presence of such elements obviously does not forfeit copyright protection of the work as a whole, but infringement cannot be found on the basis of such elements alone; it is the combination of elements, or particular novel twists given to them, that supply the minimal originality required for copyright protection.

A more fundamental limitation on the scope of copyright infringement is, as we have already noted, that copyright protects only expression and not what is expressed, the “idea,” which in' the case of Action is more likely to be a plot or situation than, as in Baker v. Selden, an algorithm or other abstraction. Suppose one wanted to prove that the Tom Hanks-Meg Ryan movie You’ve Got Mail infringed the Jimmy Stewart-Margaret Sullavan movie The Shop Around the Comer. You could point to the fact that in both movies the hero falls in love with the heroine on the basis of correspondence (letters in the earlier movie, email in the later) without having met her and without realizing that it is someone he knows — and dislikes (and the feeling is reciprocated) — though everything comes right in the end. But it would not be evidence of infringement that both movies were love stories, or that they were comedies, or that the lead actors were a man and a woman. These are generic characteristics, obtained by abstracting the commonest features of the specific work; they do not indicate uniqueness. They are at the opposite extreme from the arbitrary details, such as the pattern of boldfacing in Bucklew’s form 52566, that contribute such originality as a new expressive work may have. The format choices that Bucklew made were not generic. It is not as if everyone who writes programs of this sort uses Swiss font or displays an output *930range or uses a particular pattern of bold-facing.

Nor is it important that after copying Bucklew’s form 52566 and using it as the template for the other three forms, HAB introduced a number of variations in its version of form 52566. Variants that result from tinkering with a copied form are derivative works from that form, and it is a copyright infringement to make or sell a derivative work without a license from the owner of the copyright on the work from which the derivative work is derived. Pickett v. Prince, 207 F.3d 402, 405-407 (7th Cir.2000); Nimmer & Nimmer, supra, vol. 1, § 3.06, pp. 3-34.24 to 3-34.26; Mark A. Lemley, “The Economics of Improvement in Intellectual Property Law,” 75 Tex. L.Rev. 989, 992 (1997). In some cases a work, however, though derivative in a literal sense, is so utterly transformed as to bear no traces of the original; and then there is no infringement. Castle Rock Entertainment, Inc. v. Carol Publishing Group, Inc., 150 F.3d 132, 143 n. 9 (2d Cir.1998). Such cases are in fact common, though they are rarely litigated. Suppose one copied a long passage from a copyrighted work and then edited it to produce a paraphrase so loose that it would not be similar enough to the original to constitute an infringement. The fact that the paraphrase had been “derived” in a genetic sense from a copyrighted original would not make it infringing. Pickett v. Prince, supra, 207 F.3d at 407 (“works only loosely connected with some ancestral work claimed to be their original”). Similarly, if the original expression added by the unauthorized preparer of a derivative work is clearly detachable from the original work itself, so that no confusion, or disruption of the copyright owner’s plans for the exploitation of his work, would be created by allowing the unauthorized preparer to copyright his original expression, the unauthorized preparer might be allowed to do so, Williams v. Broadus, 60 U.S.P.Q.2d 1051, 1055 (S.D.N.Y.2001); Paul Goldstein, Copyright, vol. 1, § 2.16, p. 2:209 and n. 11 (2d ed. 2003), though this principle may be limited to compilations, where “the infringing portion would be easily severable and the scope of the compilation author’s own work... would be easily ascertainable.” Anderson v. Stallone, 11 U.S.P.Q.2d 1161, 1167-69 (C.D.Cal.1989). But HAB does not argue that its alterations of the copied form went so far as to bring it within the scope of either of these qualifications of the general rule concerning derivative works.

The analysis to this point sufficiently establishes that HAB’s version of form 52566 infringed Bucklew’s version. But that leaves the other three forms, where there are fewer traces of borrowing from Bucklew’s form. The principal traces are the captions “input range” and “criteria range,” the use of “crit1,” “crit2,” etc. to designate specific criteria, and the designation of each category of data as a “type.” These are good examples of scenes á faire — standard expressions, like language itself, without which the would-be author of an expressive work would be speechless. The only other residue of a borrowed-from-Bucklew feature in the three forms was that telltale missing vertical line — the tipoff that HAB had indeed copied Bucklew’s form 52566 in creating its own version of the other three forms and had then deleted the superfluous output range. But a nonexistent line has no expressive content, except possibly in the most advanced modern art. (Cf. Georges Perec, La Disparation (1969) — a 300-page novel written without the letter e.) The only element of an expressive work that is infringed by copying is an expressive element; what remain from the copying in the case of the three forms under consideration either *931were scenes á faire or weren’t expressive elements at all.

So the jury verdict on infringement can stand only with respect to the first of the four allegedly infringing forms. This has obvious implications for the damages award, but before getting to those let us consider HAB’s challenges to that award that are independent of any doubts about liability. After the judge reduced the jury’s award by granting a remittitur that Bucklew accepted, the award consisted of four items: $100,000 for Bucklew’s lost profits; $125,000 for HAB’s profits from the infringing forms; $70,000 for a savings in time that HAB obtained as a result of the infringement; and $100,000 for profits obtained by HAB on separate products by virtue of its being able to offer its customers “one-stop shopping,” i.e., a complete line of HUD financial software, including the copied forms.

A copyright owner can sue for his losses or for the infringer’s profits, but not for the sum of the two amounts. 17 U.S.C. § 504(b); Deltak, Inc. v. Advanced Systems, Inc., 767 F.2d 357, 363 (7th Cir. 1985); Taylor v. Meirick, 712 F.2d 1112, 1120 (7th Cir.1983); Hamil America, Inc. v. GFI, 193 F.3d 92, 108 n. 7 (2d Cir.1999); Nimmer & Nimmer, supra, vol. 4, § 14.01[A], pp. 14-5 to 14-6. That would be double counting. The profits that HAB obtained from selling forms copied from Bucklew came at his expense; his loss was HAB’s gain. If the infringer is a more efficient producer, his gain may exceed the copyright owner’s loss. If he is not more efficient, his gain is likely to be less than the copyright owner’s loss because competition will tend to force price down to cost, minimizing the infringer’s profits but depriving the copyright owner of the supra-competitive return that he enjoyed before he faced competition from the seller of an exact substitute for his product. The copyright owner is allowed to waive damages (lost profits) and sue for the infringer’s gain. Copyright infringement unlike patent infringement is an intentional tort, and by forcing the infringer to disgorge his profit should it exceed the copyright owner’s loss the law discourages infringement and encourages the would-be infringer to transact with the copyright owner rather than “steal” the copyrighted work. Taylor v. Meirick, supra, 712 F.2d at 1120. But there is no basis in the law for requiring the infringer to give up more than his gain when it exceeds the copyright owners’ loss. Such a requirement would add a punitive as distinct from a restitutionary element to copyright damages, and while the copyright statute does authorize statutory damages unrelated to losses or gains, see 17 U.S.C. § 504(c), these were not sought here and — a point to which we shall return — the statute contains no provision for punitive damages. But, astonishingly, HAB has failed to complain about double counting, and so it has booted away a winning issue.

HAB argues that Bucklew’s lost profits were due not to the infringement but simply to the fact that HAB was competing with him. Remember that Bucklew had no copyright in the database programs that he used (Lotus and Excel) or in the spreadsheet formula (DSUM) that transformed the raw data into the data required by HUD. HAB could have designed its own, noninfringing program that would have duplicated the essential — and nonco-pyrighted — features of Bucklew’s. Since it would have been producing a close substitute, it would have taken sales from Buck-lew, reducing Bucklew’s profits.

The essential distinction at which this argument gestures is between infringer’s profits that are due to the infringer’s being a more efficient producer than the plaintiff, which the copyright owner can *932sue for in order to discourage infringement, and infringer’s profits that are due to features of his work that do not infringe; those profits belong to him and not the copyright owner. But there is a difference between an infringer’s adding expressive features to what he has copied, and a copier who adds no expressive features but argues that since the copied work contains mainly public-domain material, damages should be limited to the perhaps trivial commercial value attributable to the expressive features that the author of that work added to the public-domain materials with which he began. The argued distinction is artificial. The value of an expressive work is the result of the operation of the author’s creativity on public-domain materials, ideas, and so forth; partitioning the two sources of the value, the public domain and the author’s creative transformation of it, would not be practical. The apportionment that the law permits is not between the creative and the derived features of the plaintiffs work, but between the infringing and the noninfringing features of the defendant’s work.

Once the plaintiff proves his losses, or the defendant’s profits, from the defendant’s sale of an infringing work, the burden shifts to the defendant to apportion the profits or losses between the infringing and noninfringing features of the defendant’s work. 17 U.S.C. § 504(b); Ty, Inc. v. Publications Int’l Ltd., 292 F.3d 512, 523-24 (7th Cir.2002), and cases cited there. HAB failed to do that.

The item of damages for time savings is not duplicative of the other damages that were awarded, though it was improper on another ground. The savings are in the time that HAB takes to service its customers, the local housing authorities, and do not show up in the profits that HAB makes from the sale of the infringing products and so the loss that Bucklew suffered from the infringement. But HAB argues compellingly that any savings in time could not have reflected the copyrighted features of Bucklew’s forms, because those features— the formatting choices — do not reduce the amount of time that is required to process data, using the spreadsheets with their built-in DSUM function. There is an analogy to the principle of “antitrust injury.” Brunswick, Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 486-88, 97 S.Ct. 690, 50 L.Ed.2d 701 (1977); Atlantic Richfield Co. v. USA Petroleum Co., 495 U.S. 328, 337, 339, 110 S.Ct. 1884, 109 L.Ed.2d 333 (1990); Dial A Car, Inc. v. Transportation, Inc., 82 F.3d 484, 486-87 (D.C.Cir. 1996). If two competing firms merge in violation of antitrust law, but the merger’s effect on the market price from eliminating competition between the merging firms is dominated by its effect in reducing the costs of the firm resulting from the merger below those of the two predecessor firms, and in consequence the firm reduces its price below the old competitive price, the remaining firms in the market will have to lower their price in order to remain competitive and so they are injured by the merger. But the injury is not actionable; it is not “antitrust injury” (that is, injury for which antitrust law provides a remedy) because it is not the sort of injury that antitrust law is concerned with preventing or remedying. Likewise here, HAB may have obtained cost savings by copying Bucklew’s form, but insofar as those cost savings reflected not the appropriation of the expressive features of Bucklew’s form but the appropriation merely of the idea of using spreadsheet programs with the DSUM function to expedite the preparation of grant applications to HUD, they were not the fruit of conduct for which copyright law is intended to provide a remedy.

What is true, and is the basis of Buck-lew’s claim of conversion, of which more *933shortly, is that HAB did not buy Bucklew’s form; it borrowed and, without authorization, copied it. The unauthorized copying was an infringement, but the only damages to which it can give rise are the price of the form, not the value of the ideas in the form because the ideas are not protected by copyright law.

There was also a failure of proof with regard to the “one-stop shopping” damages. HAB argues that damages can never be obtained that are based on the profits that the infringer made from the sale of a product that did not infringe the copyright. That is incorrect. Business Trends Analysts, Inc. v. Freedonia Group, Inc., 887 F.2d 399, 404, 407 (2d Cir.1989); Frank Music Corp. v. Metro-Goldwyn-Mayer, Inc., 772 F.2d 505, 516-17 (9th Cir.1985). Remember that the purpose of allowing suit for the infringer’s lost profits is to make infringement worthless to the infringer. This will sometimes require tracing those profits into another product, as where it is bundled with the infringing product. Suppose a defendant had copied a copyrighted book verbatim and he then offered to sell copies of the book for nothing to anyone who would pay him $25 for a bookmark that had cost him 10 cents and had a market value of 50 cents. To hold in such a case that the defendant’s profits from infringement were zero would approve a formula for evading copyright law. Cf. Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2, 13 n. 19, 104 S.Ct. 1551, 80 L.Ed.2d 2 (1984); United States v. Microsoft Corp., 253 F.3d 34, 86 (D.C.Cir.2001) (en banc) (per curiam); Howard P. Marvel, “The Resale Price Maintenance Controversy: Beyond The Conventional Wisdom,”63 Antitrust L.J. 59, 63-64 (1994).

But Bucklew’s evidence that HAB shifted profits from the infringing forms to a separate, noninfringing financial software package (“HMS for Windows”) was too speculative to sustain an award of damages. His theory is that the prospect of one-stop shopping attracted customers who might otherwise have gone to a competitor, and so increased HAB’s revenue and presumably its profits. One of HAB’s employees testified that the infringing forms would indeed help with the sale of HMS for Windows, but no evidence was presented that would have enabled the market value of this “help” to be gauged. The testimony by an expert witness for the plaintiff that 10 percent of the profits on HAB’s sales of HMS for Windows were due to the buyers’ being able to buy the infringing forms from HAB had no factual basis whatsoever, and we repeat previous reminders to the bench and bar of this circuit that proof of damages requires— proof. CSC Holdings, Inc. v. Redisi, 309 F.3d 988, 995 (7th Cir.2002); MindGames, Inc. v. Western Publishing Co., 218 F.3d 652, 658 (7th Cir.2000); Harbor House Condominium Ass’n v. Massachusetts Bay Ins. Co., 915 F.2d 316, 318-19 (7th Cir.1990); see also Hillside Enterprises v. Carlisle Corp., 69 F.3d 1410, 1414 (8th Cir.1995).

Besides all these problems with the computation of damages, it will be necessary on remand to apportion damages between the infringing form and what we have determined to be the three noninfringing forms.

The cross-appeal need not detain us for long. Bucklew seeks punitive damages for fraud and conversion under Wisconsin law. He argues that HAB obtained a computer disk containing his copyrighted program by falsely representing that it wanted merely to evaluate the program with a view toward possibly licensing it from Bucklew, and having thus obtained the program by false pretenses then copied it, with the consequences de*934scribed earlier. We may assume that this is a good claim under Wisconsin law. But Bucklew is not asking to have the disk returned, or for damages equal to the value of the disk plus punitive damages proportioned to that value. The compensatory damages that he seeks for the fraud and conversion are identical to the damages that it seeks for copyright infringement, so that its request for punitive damages is in fact a request for punitive damages for copyright infringement. The copyright statute does not authorize such damages, as we have noted, and the statute’s preemption clause forbids states to add sanctions for a wrongful act that is identical to a violation of the statute. See 17 U.S.C. § 301(a); Baltimore Orioles, Inc. v. Major League Baseball Players Ass’n, 805 F.2d 663, 674, 676-77 (7th Cir. 1986), and with specific reference to punitive damages Budget Cinema, Inc. v. Watertower Associates, 81 F.3d 729, 733 (7th Cir.1996); Hays v. Sony Corp. of America, 847 F.2d 412, 415 (7th Cir.1988); On Davis v. The Gap, Inc., 246 F.3d 152, 172 (2d Cir.2001); Nimmer & Nimmer, supra, vol. 4, § 14.02[B], p. 14-28.

The RICO claim is another loser. RICO provides a remedy against conducting the activities of an enterprise through a pattern of racketeering activity, and the definition of racketeering includes criminal copyright infringement. But apart from whether Bucklew could show that HAB has engaged in a pattern of such crimes, the RICO claim fails because the enterprise alleged to have been conducted through a pattern of racketeering activity (defendant HAB, Inc.) is a wholly owned subsidiary of the alleged racketeer (the other defendant, Hawkins, Ash, Baptie & Co.). The claim is that the parent stole the software and gave it to the subsidiary to market. A parent and its wholly owned subsidiaries no more have sufficient distinctness to trigger RICO liability than to trigger liability for conspiring in violation of the Sherman Act, see Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 777, 104 S.Ct. 2731, 81 L.Ed.2d 628 (1984), unless the enterprise’s decision to operate through subsidiaries rather than divisions somehow facilitated its unlawful activity, which has not been shown here. Bachman v. Bear, Steams & Co., 178 F.3d 930, 932 (7th Cir.1999); Emery v. American General Finance, Inc., 134 F.3d 1321, 1324 (7th Cir.1998); Fitzgerald v. Chrysler Corp., 116 F.3d 225, 226-28 (7th Cir.1997); Bessette v. Avco Financial Services, Inc., 230 F.3d 439, 448-49 (1st Cir.2000); Brannon v. Boatmen’s First National Bank of Oklahoma, 153 F.3d 1144, 1147-49 (10th Cir.1998).

To summarize, the judgment on liability is affirmed in part and reversed in part, and the case is remanded for a redetermi-nation of the damages to which the plaintiff is entitled.

Affirmed in Part, Reversed IN Part, and Remanded.

MESKILL, Chief Judge:

This is an appeal from a preliminary injunction entered in the United States District Court for the Eastern District of New York, Spatt, J., in an action brought under section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a), and various state unfair competition laws. The preliminary injunction prevents. the appellant, McNeil-P.P.C., Inc. (McNeil) from marketing its product, “Tylenol PM,” in the trade dress that it had been using prior to the injunction. Also before us is a cross-appeal by Bristol-Myers Squibb Company (Bristol) from the denial of preliminary injunctive relief preventing McNeil from using the term “PM” in connection with a combination analgesic/sleep aid.

As' in any case under section 43(a) of the Lanham Act, the ultimate question here is whether, because of the challenged markings, consumers are likely to be confused as to the origin or sponsorship of goods. Because we believe that the district court erred in finding that the “Tylenol PM” trade dress was likely to cause confusion among purchasers due to its similarity to Bristol’s “Excedrin PM” trade dress, we reverse the preliminary injunction. Because we agree with the district court’s denial of Bristol’s request preliminarily to enjoin McNeil’s use of the term “PM,” we affirm as to the cross-appeal.

Affirmed in part and reversed in part.

BACKGROUND

Bristol is a major pharmaceutical company that produces and markets “Excedrin,” *1037a nationally known over-the-counter analgesic pain reliever. Since 1968 Bristol has manufactured and distributed “Exeedrin PM,” a product that combines an analgesic with a sleep aid.

Although there have been some alterations over the years, the packaging ■ of “Exeedrin PM” has remained fairly constant. That packaging consists of an outer carton with a solid deep blue background, white lettering for the name “Exeedrin PM,” which appears on a single line, with “Exeedrin” in lower case lettering except for the initial “E” and “PM,” which are in capital letters. In the' lower right portion of the face of the box is a depiction of two light blue tablets, each marked “PM.” In 1988, Bristol introduced a caplet form of “Exeedrin PM.” The trade dress for this caplet form is similar to that of the tablet form except that the background is a solid dark green and there is a depiction of two caplets each displaying the full name “Exeedrin PM.” On both packages the word “Exeedrin” occupies approximately one-third of the total face of the package.

McNeil is also a major pharmaceutical company. Among McNeil’s products is “Tylenol,” which, like “Exeedrin,” is a nationally famous over-the-counter analgesic. In 1991 McNeil introduced “Tylenol PM,” which, like “Exeedrin PM,” is a combination analgesic/sleep aid. Although there are slight differences in the composition of the two products, it is undisputed that those differences are not material and that the two products are functionally interchangeable. Like “Exeedrin PM,” “Tylenol PM” comes in two forms: tablet and caplet.

The packaging for the tablet form of “Tylenol PM” consists of a box with a green background that shifts from dark green at the top of the box to light green at the bottom and the name “Tylenol” in white capital letters with “PM” in yellow capital letters on the same line. In addition, the package depicts two tablets on the lower right face of the package, one of which is imprinted with the word “Tylenol” and the other of which is marked “PM.” The trade dress for the caplet form of “Tylenol PM” resembles that of the tablet form except that the background fades from dark to light blue rather than green and two caplets, each of which reads “Tylenol PM,” are depicted instead of the tablets. On both forms of the “Tylenol PM” trade dress, the word' “Tylenol” occupies approximately one-third of the total area of the face of the package.

Shortly after “Tylenol PM” was introduced, Bristol filed this action before Judge Spatt in the United States District Court for the Eastern District of New York, seeking relief ünder section 43(a) of the Lan-ham Act, 15 U.S.C. § 1125(a), and on various state law grounds. Bristol moved for a preliminary injunction to prevent McNeil from using the term “PM” in connection with a combination analgesic/sleep aid and to prevent McNeil from marketing an analgesic/sleep aid in the “Tylenol PM” packaging described above.

Judge Spatt referred the matter to Magistrate Judge Orenstein pursuant to 28 U.S.C. § 636(b)(1)(B). The magistrate judge heard evidence and examined exhibits and the results of discovery. The magistrate judge filed a report in the district court in which he made certain findings of fact and ultimately recommended that the preliminary injunction be denied. Both parties made timely objections to various portions of the magistrate judge’s report and the district court addressed the matter de novo pursuant to 28 U.S.C. § 636(b)(1).

The district court denied Bristol’s request for an injunction preventing McNeil from using the term “PM” in connection with an analgesic/sleep aid, finding that “PM” in this context did not qualify for protection under section 43(a) of the Lan-ham Act. The district court granted, however, Bristol’s request for an injunction with regard to the “Tylenol PM” trade dress because it found that Bristol, had demonstrated that the similarities between the two packages were likely to cause consumer confusion 786 F.Supp. 182. Both parties appealed and a panel of this Court stayed the preliminary injunction pending appeal.

*1038DISCUSSION

I. PRELIMINARY INJUNCTION

In order to obtain a preliminary injunction, the moving party must demonstrate both (1) irreparable harm in the absence of the requested relief, and (2) either (a) a likelihood that it will succeed on the merits of the action, or (b) a sufficiently serious question going to the merits combined with a balance of hardships tipping decidedly in favor of the moving party. See Jackson Dairy v. H.P. Hood & Sons, 596 F.2d 70, 72 (2d Cir.1979) (per curiam). In the context of a Lanham Act claim both the likelihood of success on the merits and the potential for irreparable harm in the absence of preliminary relief may be demonstrated by a showing that a significant number of consumers are likely to be misled or confused as to the source of the products in question. Western Publishing Co. v. Rose Art Industries, 910 F.2d 57, 59 (2d Cir. 1990).

We have jurisdiction to review the grant or denial of a preliminary injunction pursuant to 28 U.S.C. § 1292(a)(1). We review a district court’s decision whether to issue a preliminary injunction only for an abuse of discretion. Doran v. Salem Inn, 422 U.S. 922, 931-32, 95 S.Ct. 2561, 2567-68, 45 L.Ed.2d 648 (1975). “Applying legal standards incorrectly or relying upon clearly erroneous findings of fact may constitute an abuse of discretion.” Haitian Centers Council v. McNary, 969 F.2d 1326, 1338 (2d Cir.1992) (citation omitted). We are not, however, limited to “reversing only when the lower court’s action exceeds any reasonable bounds and to rubber-stamping with the imprimatur of an affirmance when it does not.” Coca-Cola Co. v. Tropicana Products, 690 F.2d 312, 315 (2d Cir.1982) (citation omitted). Although our scope of .review is limited, if we are left with “ ‘the definite and firm conviction that a mistake has been committed’ ” we will reverse the grant or denial of a preliminary injunction. See Standard & Poor’s Corp. v. Commodity Exchange, 683 F.2d 704, 708 (2d Cir.1982) (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948)). With these principles in mind, we turn to the parties’ specific claims.

II. THE LANHAM ACT CLAIMS

Neither the “PM” designator nor the “Excedrin PM” trade dress is a registered trademark. Bristol, therefore, relies on that part of the Lanham Act that addresses unregistered marks. Section 43(a) of the Lanham Act prohibits any person from using

in connection with any goods ... or any container for goods, ... any word, term, name, symbol, or device, or any combination thereof ... which ... is likely to cause confusion, or to cause mistake, or to deceive ... as to the origin, sponsorship, or approval of his or her goods ... by another person.

15 U.S.C. § 1125(a).1 “[I]t is common ground that § 43(a) protects qualifying unregistered trademarks and that the general principles qualifying a mark for registration under § 2 of the Lanham Act are for the most part applicable in determining whether an unregistered mark is entitled to protection under § 43(a).” Two Pesos, Inc. v. Taco Cabana, Inc., — U.S.-,-, 112 S.Ct. 2753, 2757, 120 L.Ed.2d 615 (1992) (citations omitted). As the words of the statute indicate, the central inquiry is whether the use of a marking is likely to cause consumer confusion.

Bristol makes two related claims under section 43(a). First, Bristol claims that the use by McNeil of the term “PM” in connection with its combination analgesic/sleep aid is likely to cause confusion as to the source, sponsorship or approval of “Tylenol *1039PM.” Second, Bristol claims that the overall look of the “Tylenol PM” packaging, its trade dress, is so similar to the trade dress of Bristol’s “Excedrin PM” that consumers will confuse the two.

Before we will address whether an allegedly infringing mark is likely to cause consumer confusion, the plaintiff must show that its trademark or trade dress is of the type that deserves protection under section 43(a). See Thompson Medical Co. v. Pfizer, Inc., 753 F.2d 208, 213 (2d Cir.1985). The plaintiff then must demonstrate that the use by the defendant of a mark or dress is likely to cause consumer confusion because of the similarity between the defendant’s mark or dress and the mark or dress of the plaintiff. See id.

A. Trademark Infringement: The “PM” Designator

To facilitate our analysis, we will deal first with the cross-appeal. Bristol appeals from the district court’s refusal to grant preliminary injunctive relief preventing McNeil from using the term “PM” appended to its “Tylenol” trade name to designate its combination analgesic/sleep aid. Under the terms of the preliminary injunction, McNeil remains free to market “Tylenol PM” in a trade dress that does not resemble the “Excedrin PM” trade dress. The district court agreed with the magistrate judge that the term “PM” as used by Bristol did not warrant protection under section 43(a).

In making the preliminary inquiry into whether a particular mark is eligible for protection under section 43(a), we have established several categories into which we classify various marks. “Arrayed in an ascending order which roughly reflects their eligibility to trademark status and the degree of protection accorded, these classes are (1) generic, (2) descriptive, (3) suggestive, and (4) arbitrary or fanciful.” Abercrombie & Fitch Co. v. Hunting World, 537 F.2d 4, 9 (2d Cir.1976). Suggestive, arbitrary and fanciful marks, “because their intrinsic nature serves to identify a particular source of a product, are deemed inherently distinctive and are entitled to protection.” Two Pesos, — U.S. -, 112 S.Ct. at 2757. Generic marks are never entitled to protection, while a descriptive mark is eligible for protection “if it ‘has become distinctive of the [producer’s] goods in commerce.’ This acquired distinctiveness is generally called ‘secondary meaning.’ ” Id. — U.S. at-, 112 S.Ct. at 2757 (citations omitted).

Although we have not expressly articulated the standard of review appropriately applied to a district court’s classification of a mark, other courts of appeals have found that classification to be a factual question, review of which is limited to whether the district court was clearly erroneous in its determination. See, e.g., Ford Motor Co. v. Summit Motor Products, 930 F.2d 277, 292 n. 18 (3d Cir.) (“The characterization of a mark is a factual issue.”) (citations omitted), cert. denied, — U.S. -, 112 S.Ct. 373, 116 L.Ed.2d 324 (1991); G. Heileman Brewing Co. v. AnheuserBusch, 873 F.2d 985, 992 (7th Cir.1989) (“[T]he district court’s classification of a term on the trademark spectrum is a factual determination subject to the ‘clearly erroneous’ standard of review.”); Wiley v. American Greetings Corp., 762 F.2d 139, 141 (1st Cir.1985) (“Whether a design is ‘inherently distinctive,’ i.e., whether it is arbitrary or merely descriptive, is ordinarily a question of fact.”) (citation omitted); Anheuser-Busch v. Stroh Brewery Co., 750 F.2d 631, 635 (8th Cir.1984) (“[T]he categorization of a term for which trademark protection is claimed is considered to be a factual issue and thus is to be reviewed under the clearly erroneous standard of Fed.R.Civ.P. 52(a).”) (citations omitted); Soweco, Inc. v. Shell Oil Co., 617 F.2d 1178, 1183 n. 12 (5th Cir.1980) (“Any given term’s correct categorization is a factual issue.”) (citations omitted), cert. denied, 450 U.S. 981 (1981); cf. In re Northland Aluminum Products, 777 F.2d 1556, 1559 (Fed.Cir.1985) (“Whether a term is a common descriptive name is a question of fact” in context of trademark registration.) (citation omitted). We agree with these courts that the initial classification of a mark to determine its eligibility for protec*1040tion is a question of fact left to the determination of the district court. We will substitute our own judgment on the matter for that of the district court only if the district court’s determination is clearly erroneous. Fed.R.Civ.P. 52(a).

The district court classified the “PM” mark as descriptive. Bristol argues that the mark is suggestive. Neither party claims that the mark is generic, arbitrary or fanciful. The question in this case, therefore, is whether “PM,” when affixed to an analgesic trade name, is descriptive or suggestive of the product, a combination analgesic/sleep aid.

A descriptive mark is one that “ ‘forthwith conveys an immediate idea of the ingredients, qualities or characteristics of the goods.’ ” Abercrombie & Fitch, 537 F.2d at 11 (citation omitted). “[A] term can be descriptive in two ways. It can literally describe the product, or it can describe the purpose or utility of the product.” 20th Century Wear v. Sanmark-Stardust, 747 F.2d 81, 88 (2d Cir.1981), cert. denied, 470 U.S. 1052, 105 S.Ct. 1755, 84 L.Ed.2d 818 (1985). A term is suggestive “ ‘if it requires imagination, thought and perception to reach a conclusion as to the nature of goods.’ ” Abercrombie & Fitch, 537 F.2d at 11 (citation omitted).

Although the line between descriptive and suggestive may be difficult to discern, see id. at 10-11, the consequence of the classification is important. A descriptive term is subject to protection under section 43(a) only if the proponent of protection demonstrates that, in addition to the ordinary common meaning of the word or words, the term has acquired a secondary meaning in its particular market — that the consuming public primarily associates the term with a particular source. See Centaur Communications, Ltd. v. A/S/M Communications, 830 F.2d 1217, 1221 (2d Cir.1987); Thompson Medical, 753 F.2d at 212-13. The public presumably will not be confused by a descriptive term, but if the proponent of protection can show that the descriptive term is primarily associated with a single producer, a sufficient question is raised to justify further inquiry into the likelihood of confusion. In contrast, if its mark is suggestive a plaintiff need not prove such secondary meaning in order to qualify for trademark protection. Id. at 213.

In this case, the “PM” acts as a modifier to the analgesic brand name. Both “Tylenol” and “Excedrin” are well known brand names for analgesics; the “PM” modifies each to show that they are a particular type of analgesic. “PM,” usually abbreviated “p.m.” or “P.M.” (for “post meri-diem”), is a common term that refers to the period of time between noon and midnight. It is often associated with the time when most people go to sleep.

As used here, “PM” does not literally describe the presence of a sleep aid in the product. The “PM” refers to the purpose or utility of the product — it is an analgesic that should be used at night. The issue, therefore, is whether the connection between “PM” and a nighttime sleep aid is direct enough that the term may be categorized as descriptive or whether the connection is more indirect, requiring categorization as suggestive.

Bristol argues that “PM” is suggestive because a consumer must engage in a mul-ti-step analysis before coming to the conclusion that it denotes the presence of a sleep aid in the analgesic. Bristol asserts that the consumer first must eliminate possible alternate meanings for “PM” — e.g., “Pre-Menstrual” or “Pain Medication”— before arriving at “Post Meridiem.” Next, Bristol argues, the consumer must make a leap from all the post meridiem hours to those at night, and a further intellectual leap from nighttime to sleeping.

The magistrate judge found that several other over-the-counter products are designated as nighttime products by the use of some close variant of “PM.” The magistrate judge held in that context that “[t]he direct connotation of ‘PM’ is nighttime. There is no ‘multi-stage reasoning process’ that a consumer must indulge in to associate the term ‘PM’ with a nighttime product.”

*1041We cannot say that the district court’s adoption of this finding was clearly erroneous. The focus in categorizing a mark is on how the words are used in context rather than their meaning in the abstract. See Abercrombie & Fitch, 537 F.2d at 12. One of the leading commentators has offered the following example to demonstrate the context-dependent nature of the classification: “[T]he word ‘apple’ would be arbitrary when used on personal computers, suggestive when used in ‘Apple-A-Day’ on vitamin tablets, descriptive when used in ‘Tomapple’ for combination tomato-apple juice and generic when used on apples.” 1 J.T. McCarthy, Trademarks and Unfair Competition § 11:22, at 498-99 (2d ed. 1984). There was sufficient evidence before the magistrate judge to support the finding that several nighttime products were sold using some variant of “PM.” Given that context, the conclusion that “PM” describes rather than suggests a nighttime product was not clearly erroneous. Once the consumer arrives at an awareness that the product is useful at nighttime, the “purpose or utility” of the product has been conveyed, even though the consumer is not aware of why the product is useful at night. See Thompson Medical, 753 F.2d at 216 & n. 15 (“Sportscreme” descriptive of cream useful in sports even though inadequately descriptive as to significance of product). Therefore, the consumer need not conclude that the analgesic contains a sleep aid.

Given the deferential standard of review, and the fact that this matter is before us at the preliminary injunction stage, we will not disturb the district court’s finding that “PM” is descriptive.

Because “PM” is a descriptive term it is not entitled to trademark protection unless Bristol demonstrates secondary meaning. The district court found that Bristol had not established that the “PM” designator had acquired secondary meaning. We agree with the district court.

“To establish secondary meaning, a manufacturer must show that, in the minds of the public, the primary significance of a product feature or term is to identify the source of the product rather than the product itself.” Inwood Laboratories v. Ives Laboratories, 456 U.S. 844, 851 n. 11, 102 S.Ct. 2182, 2187 n. 11, 72 L.Ed.2d 606 (1982) (citation omitted). Secondary meaning is an essentially factual determination, proof of which “ ‘entails vigorous evidentiary requirements.’ ” Thompson Medical, 753 F.2d at 217 (citation omitted); see Coach Leatherware v. AnnTaylor, Inc., 933 F.2d 162, 169 (2d Cir.1991). We will reverse the district court’s determination that a term has not acquired secondary meaning only if that determination is clearly erroneous. See Murphy v. Provident Mut. Life Ins., 923 F.2d 923, 928 (2d Cir.1990), cert. denied, — U.S. -, 112 S.Ct. 65, 116 L.Ed.2d 40 (1991); 815 Tonawanda Street Corp. v. Fay’s Drug Co., 842 F.2d 643, 647 (2d Cir.1988).

Among the factors that we have found relevant to this inquiry in the past are advertising expenditures, consumer studies, sales success, unsolicited media coverage, attempts to plagiarize and length and exclusivity of use. Thompson Medical, 753 F.2d at 217. There are undoubtedly other types of evidence that would also be relevant to a claim of secondary meaning. The fundamental question, however, is whether “ ‘the primary significance of the term in the minds of the consuming public is not the product but the producer.’ ” Centaur Communications, 830 F.2d at 1221 (citations omitted).

“The existence of secondary meaning is a question of fact with the burden of proof on the party claiming exclusive rights in the designation.” PaperCutter, Inc. v. Fay’s Drug Co., 900 F.2d 558, 564 (2d Cir.1990) (citation omitted). The district court found that Bristol had not presented sufficient evidence that consumers recognized the source of “Excedrin PM” by the “PM” designator and that there was no evidence that Bristol had ever marketed the product as “PM.”

Bristol suggests two grounds in support of its contention that the district court’s finding that the “PM” mark had not acquired secondary meaning was clearly *1042erroneous. First, Bristol argues that McNeil’s intentional copying of its mark is indicative of secondary meaning. Although imitative intent can help support a finding of secondary meaning, see Centaur Communications, 830 F.2d at 1224, it does not necessarily mandate one, see American Footwear Corp. v. General Footwear Co., 609 F.2d 655, 663 (2d Cir.1979) (“[N]o single factor is determinative.”), cert. denied, 445 U.S. 951, 100 S.Ct. 1601, 63 L.Ed.2d 787 (1980). Therefore, we believe that, even assuming McNeil’s imitative intent, the district court was not bound to find that the “PM” designator had acquired secondary meaning.

Second, Bristol argues that the fact that each “Excedrin PM” tablet is imprinted with “PM” shows that the consuming public must associate “PM” with Bristol. However, “[although the mark owner strives to create a secondary meaning for its product, it is the consuming public which, in effect, determines whether that effort has succeeded.” Centaur Communications, 830 F.2d at 1221 (citations omitted). The mere fact that Bristol places the name “PM” on its tablets does not persuade us that the consuming public associates that term with Bristol. Moreover, on the larger caplet version of the product, Bristol placed the entire “Excedrin PM” mark.

The district court was not clearly erroneous in classifying “PM” as a descriptive term in these circumstances. Because “PM” had not become primarily associated with a single producer in these circumstances, “PM” as used by Bristol is not by itself subject to protection under the Lan-ham Act. The district court therefore properly declined to enjoin McNeil from using the “PM” designator in connection with its combination analgesic/sleep aid.

B. Trade Dress Infringement

McNeil appeals from the entry of a preliminary injunction preventing it from marketing “Tylenol PM” in the packaging that the district court found to have infringed on Bristol's trade dress. The “trade dress” of a product “ ‘involves the total image of a product and may include features such as size, shape, color or color combinations, texture, [or] graphics.’ ” LeSportsac, Inc. v. K mart Corp., 754 F.2d 71, 75 (2d Cir.1985) (citation omitted). In examining trade dress the focus is on the entire look of the product or packaging. Individual aspects of a trade dress may be eligible for trademark protection in their own right, but in an action for trade dress infringement each aspect should be viewed in relation to the entire trade dress. The inquiry is whether the trade dress of a “junior user” — one who has adopted its mark or dress later than the “senior user” — is likely to cause consumer confusion as to the source of the product. As in the trademark arena, we must first determine whether a particular trade dress is eligible for protection under the Lanham Act.

The district court proceeded on the assumption, formerly the law in this Circuit, that, unlike trademarks, a trade dress was always ineligible for protection under section 43(a) in the absence of proof of secondary meaning. See, e.g., LeSportsac, 754 F.2d at 75. The Supreme Court, however, recently rejected this distinction between trademarks and trade dress. Two Pesos, — U.S. at-, 112 S.Ct. at 2760 (“We see no basis for requiring secondary meaning for inherently distinctive trade dress protection under § 43(a) but not for other distinctive words, symbols, or devices capable of identifying a producer’s product.”).

The district court made no finding as to whether the “Excedrin PM” trade dress was “inherently distinctive.” However, if the district court was correct in determining that the trade dress had acquired secondary meaning, that dress is eligible for protection under section 43(a).

1, Secondary Meaning

The secondary meaning inquiry is the same for trade dress as it is for a trademark. Bristol must show that the primary significance of its “Excedrin PM” trade dress is to identify the source of the product rather than the product itself. See Inwood Laboratories, 456 U.S. at 850-51 n. *104311, 102 S.Ct. at 2187 n. 11. The question whether a mark or dress has acquired secondary meaning is a factual one that we will not reverse unless it is clearly erroneous. See Murphy, 923 F.2d at 928; 815 Tonawanda Street Corp., 842 F.2d at 647. The district court properly examined the Thompson Medical factors and concluded that Bristol’s “Excedrin PM” trade dress had acquired secondary meaning. See Thompson Medical, 753 F.2d at 217.

The district court found that Bristol’s advertising expenditures in connection with “Excedrin PM” and its sales success weighed in favor of a finding of secondary meaning for the trade dress. It further held that consumer survey evidence introduced by Bristol, “though problematic in terms of some of its methodology,” supported Bristol’s claim of secondary meaning. The district court finally found that McNeil intentionally copied the “Excedrin PM” trade dress.

Given the deferential standard of review, we cannot say that the district court clearly erred in determining that the overall trade dress of “Excedrin PM” had acquired secondary meaning. This conclusion is strengthened by the well known trade name “Excedrin” prominently displayed on the “Excedrin PM” box as an integral part of the whole trade dress. That name strongly supports the district court’s determination that consumers associate the trade dress, taken as a whole, with a particular producer.

2. Likelihood of Confusion

Having determined that the district court did not err in finding that the “Excedrin PM” trade dress is eligible for protection under section 43(a), we turn to the central question in this case: Whether McNeil’s use of the “Tylenol PM” trade dress is likely to cause confusion as to the “origin, sponsorship, or approval” of “Tylenol PM” with “Excedrin PM.” See 15 U.S.C. § 1125(a)(1).

Whether the public is likely to confuse two products is a question that is not easily answered. Our cases, however, have established an analytical framework designed to add structure to this nebulous inquiry and to guide judicial decisions toward some degree of predictability and uniformity.

In Polaroid Corp. v. Polarad Electronics Corp., 287 F.2d 492 (2d Cir.), cert. denied, 368 U.S. 820, 7 L.Ed.2d 25 (1961), we examined similar marks used for dissimilar products. We set forth the following nonexclusive list of factors relevant to success on a Lanham Act claim:

the strength of [the prior owner’s] mark, the degree of similarity between the two marks, the proximity of the products, the likelihood that1 the prior owner will bridge the gap [between the two products], actual confusion, and the reciprocal of defendant’s good faith in adopting its own mark, the quality of defendant’s product, and the sophistication of the buyers.

Id. at 495. We have subsequently held that these factors are appropriately considered when examining the likelihood of confusion between two competing products. See, e.g., Thompson Medical, 753 F.2d at 213-14.

The standard of appellate review of a district court’s conclusion regarding the likelihood of confusion between two products has split the courts of appeals. See Euroquilt, Inc. v. Scandia Down Corp., 475 U.S. 1147, 106 S.Ct. 1801, 90 L.Ed.2d 346 (1986) (White, ,/., dissenting from the denial of certiorari to resolve whether the determination of likelihood of confusion under section 43(a) is subject to de novo review as a conclusion of law or “clearly erroneous” review as a question of fact). In this Circuit, a district court’s determination of the individual Polaroid factors are subject to review as findings of fact, subject to reversal only if clearly erroneous, while the ultimate balancing of all the Polaroid factors to determine the likelihood of confusion in any given case is done de novo by this Court. See Murphy, 923 F.2d at 928; Western Publishing, 910 F.2d at 61.

Some earlier cases in this Circuit have held that to the extent that the Polaroid factors rest on a comparison of the marks *1044themselves we are in as good a position to evaluate them as the district court and thus will evaluate those factors de novo. See, e.g., McGregor-Doniger v. Drizzle, 599 F.2d 1126, 1133 (2d Cir.1979). However, Rule 52(a) of the Federal Rules of Civil Procedure was amended in 1985 to require that “[findings of fact, whether based on oral or documentary evidence, ... not be set aside unless clearly erroneous.” Fed. R.Civ.P. 52(a). The Notes of the Advisory Committee on Rules accompanying that amendment emphasize that even where the appellate court is in as good a position as the trial court to make a given factual determination, the factfinding function should be left to the district courts because of “the public interest in ... stability and judicial economy.” Notes of Advisory Committee on Rules, 1985 Amendment. Thus, each of the Polaroid factors should be reviewed under Rule 52(a). Cf. Johnson & Johnson v. GAC International, 862 F.2d 975, 979 (2d Cir.1988) (finding of falsity in advertisement, formerly reviewed de novo, reviewed under amended Rule 52).

The only case that applies a de novo review to any Polaroid factor since the amendment to Rule 52(a) is Centaur Communications, 830 F.2d at 1226 (“Unlike the other factors, this finding [with regard to similarity of the marks] is reviewed de novo because an appellate court is in as good a position to evaluate this factor as is a trial court.”) (citations omitted). In that case, the Court did not examine the amendment to Rule 52, and its statement that we review similarity of the marks de novo was not necessary to its holding. Therefore, we do not find that statement controlling.

Despite the existence and importance of the Polaroid factors, we must remember that they are merely tools “designed to help grapple with the ‘vexing’ problem of resolving the likelihood of confusion issue,” and that “the ultimate conclusion as to whether a likelihood of confusion exists' is not to be determined in accordance with some rigid formula.” Lois Sportswear v. Levi Strauss & Co., 799 F.2d 867, 872 (2d Cir.1986) (citation omitted).

We now turn to an examination of the Polaroid factors. In the case of competing products, one of the Polaroid factors, proximity of the products in the marketplace, is necessarily answered in favor of the senior user, and another factor, the likelihood of bridging the gap, is not a relevant inquiry. Where, as here, the competitive products are of nearly identical quality, a comparison of the quality of the products is not particularly helpful in determining the likelihood of confusion. The remaining factors, however, require additional discussion.

a. Strength of the “Excedrin PM” Trade Dress

The strength of a particular mark or dress is measured by its distinctiveness or the degree to which it indicates the source or origin of the product. McGregor-Doniger, 599 F.2d at 1131; Banff, Ltd. v. Federated Dep’t Stores, 841 F.2d 486, 491 (2d Cir.1988). The strength of a mark should be examined in its commercial context. Centaur Communications, 830 F.2d at 1226.

We agree with the district court that the “Excedrin PM” trade dress, taken as a whole, strongly indicates the source of the product. The presence of the “Excedrin” trade name prominently displayed as an integral part of the trade dress demonstrates beyond question that the entire trade dress indicates the origin of the product.

b. McNeil’s Good Faith

Evidence of intentional copying by a junior user may be indicative of an intent to create a confusing similarity between the products. Charles of the Ritz Group v. Quality King Distrib., 832 F.2d 1317, 1322 (2d Cir.1987). “Although this factor alone is not dispositive, it bolsters a finding of consumer confusion.” Id.

The district court here found that McNeil intentionally copied elements of Bristol’s trade dress. There is evidence in the record to support that finding. McNeil concededly was aware of Bristol’s trade dress. There was evidence that the target *1045market for “Tylenol PM” was the “Excedrin PM” user. Evidence also existed tending to show that the similarities in the trade dresses were raised by McNeil employees during the design process. There were many proposed trade dresses that did not incorporate substantial aspects of the “Excedrin PM” trade dress yet, as the district court noted, “every choice [McNeil] made brought it closer to the EXCEDRIN PM trade dress.” An internal memorandum regarding the text to be placed on the package stated that it should appear “as in Excedrin PM.” Based on this and other evidence in the record, the district court was not clearly erroneous in finding that McNeil had not exercised good faith in adopting its trade dress.

Although there was sufficient evidence in the record to support the district court’s finding with regard to McNeil’s good faith, McNeil argues that the district court erred in rejecting, without receiving additional evidence, the findings of the magistrate judge, who had found that McNeil acted in good faith. District courts may designate a magistrate judge to conduct hearings and to submit to the district court proposed findings of fact and dispositions. 28 U.S.C. § 636(b)(1)(B). If a party files objections to those findings, as both parties did in this case, the district court is to

make a de novo determination of those portions of the report or specified proposed findings or recommendations to which objection is made. [The district court] may accept, reject, or modify, in whole or in part, the findings or recommendations made by the magistrate. The judge may also receive further evidence.

Id. at 636(b)(1). McNeil argues that before a district court may reject credibility findings of a magistrate judge it must recall witnesses and hear and observe them itself. We disagree.

McNeil’s reliance on dicta from United States v. Raddatz, 447 U.S. 667, 681 n. 7, 100 S.Ct. 2406, 2415 n. 7, 65 L.Ed.2d 424 (1980), is unavailing. In that case, a criminal prosecution, the Court expressed skep-tieism as to whether a district court could reject dispositive findings of credibility without seeing and hearing the witnesses. Arguably, the rejection of the magistrate judge’s conclusion by the district court was not based on witness credibility in this case. Moreover, we have indicated that the Raddatz dicta may be inapplicable outside the criminal context, see Moore v. Ross, 687 F.2d 604, 609 n. 5 (2d Cir.1982), cert. denied, 459 U.S. 1115, 103 S.Ct. 750, 74 L.Ed.2d 969 (1983), and where, as here, the civil party has raised no objection on those grounds before the trial court we will not reverse the district court’s findings on that ground.

c. Evidence of Actual Confusion

Before the magistrate judge, Bristol presented evidence of actual confusion that included two different types of consumer surveys, the results of McNeil’s own television advertisement survey and questions from McNeil’s employees regarding the similarity in the two trade dresses. The magistrate judge reviewed all of the evidence presented and found that Bristol had not demonstrated any actual confusion between the two products.

The district court addressed only the consumer surveys and found that they did not demonstrate any meaningful actual confusion between the products. The district court had before it evidence critical of the manner in which the surveys were conducted. Bristol has not demonstrated to us on appeal that the district court was clearly erroneous in placing little credence in those surveys. Nor has Bristol given us any reason to supplant the finding of the magistrate judge with regard to the other evidence of actual confusion. Therefore, Bristol has not demonstrated any meaningful instances of actual confusion resulting from the “Tylenol PM” trade dress.

d. The Similarity of the Trade Dresses

The district court found that the two trade dresses were strongly similar. After examining the two trade dresses, we conclude that, although they share many similar elements, the prominence of the trade *1046names on the two packages weighs heavily against a finding of consumer confusion resulting from the overall look of the packaging. Therefore, we find that the district court’s finding on this issue was clearly erroneous.

Undeniably, the “Tylenol PM” trade dress shares many attributes with the “Excedrin PM” trade dress. The color scheme used by the two products is similar. Both products have the “PM” designator on the same line as the trade name. The text on the packaging is also similar. However, in a trade dress infringement case the question is not how many points of similarity exist between the two packages but rather whether the two trade dresses “create the same general overall impression.” RJR Foods v. White Rock Corp., 603 F.2d 1058, 1060 (2d Cir.1979). The issue is whether the similarity between the two trade dresses will contribute to consumer confusion as to the origin of the product. See Western Publishing, 910 F.2d at 61 (“ ‘Ultimately, the crucial question is whether the similarity is likely to create confusion.’ ”) (citations omitted); McGregor-Doniger, 599 F.2d at 1133 (“ ‘Similarity in and of itself is not the acid test. Whether the similarity is likely to provoke confusion is the crucial question.’ ”) (quoting 3 R. Callman, The Law of Unfair Competition, Trademarks and Monopolies § 82.1(a), at 601-02 (3d ed.1969)).

The presence and prominence of markings tending to dispel confusion as to the origin, sponsorship or approval of the goods in question is highly relevant to an inquiry concerning the similarity of the two dresses. When prominently displayed it can go far towards eliminating any possible confusion, see, e.g., American Rolex Watch Corp. v. Ricoh Time Corp., 491 F.2d 877, 879 (2d Cir.1974) (“Rolex’s reliance upon [section 43(a)] appears to be misplaced in view of the prominent display of ‘Ricoh’ on the challenged watch.”); Bose Corp. v. Linear Design Labs, 467 F.2d 304, 310 (2d Cir.1972) (“[Tjhere is hardly likelihood of confusion or palming off when the name of the manufacturer is clearly displayed.”) (footnote omitted), even though the appearance of the junior user’s name does not in all cases eliminate the possibility of consumer confusion, see Banff, 841 F.2d at 492 (“Bloomingdale’s attachment of its company name below its standard typestyle ‘B Wear’ mark does not offset the marks’ similarity because the name is in very small letters and may actually increase the misappropriation by linking defendant’s name to plaintiff’s goodwill.”) (citation omitted).

In this case, by far the most prominent feature of the “Excedrin PM” trade dress is the trade name “Excedrin.” At least as prominent on the “Tylenol PM” packaging is the trade name “Tylenol.” These trade names are the major features of otherwise ordinary boxes. The “Tylenol” trade name is displayed in the same typeface used on other “Tylenol” products. In fact, except for the color of the box and the presence of “PM,” the “Tylenol PM” trade dress is extremely similar to the trade dress of the other “Tylenol” analgesic products. The differences between the two trade dresses are therefore significant.

Despite the similarity between certain aspects of the two trade dresses, when taken as a whole, including the prominently displayed names, they are not similar in any manner that is likely to cause consumer confusion; in fact, the prominent presence of well known trade names goes far toward countering any suggestion of consumer confusion arising from any of the other Polaroid factors.

e. Sophistication of the Purchasers

Generally, the more sophisticated and careful the average consumer of a product is, the less likely it is that similarities in trade dress or trade marks will result in confusion concerning the source or sponsorship of the product. See Centaur Communications, 830 F.2d at 1228. The district court found that consumers of non-prescription analgesic/sleep aids were not typically careful or sophisticated in making their purchases. Given the expert testimony available to the district court that over-the-counter analgesics were “low involvement” purchases, and given the fact *1047that these products are relatively inexpensive, we cannot say that the district court was clearly erroneous in this regard. Thus, consumer sophistication in making a purchase of these products militates toward a likelihood of confusion.

f Ultimate Likelihood of Confusion

Having reviewed the district court’s findings with regard to the Polaroid factors, we now engage in a de novo balancing of those factors in order to determine whether McNeil’s use of the “Tylenol PM” trade dress is likely to cause confusion among consumers because of its similarity to the “Excedrin PM” trade dress. See Murphy, 923 F.2d at 928.

Although the overall trade dress of “Excedrin PM” strongly identifies the product with a particular source, it does so largely because the word “Excedrin” is emblazoned over almost half the package. The “Tylenol PM” trade dress differs significantly from the “Excedrin PM” trade dress in that it does not include the word “Excedrin.” The absence of the name “Excedrin” from McNeil’s product is a significant difference. Additionally, that absence is replaced by a different prominent name. The origin-indicating name “Tylenol” appears at least as prominently on the “Tylenol PM” trade dress as does “Excedrin” on the “Excedrin PM” trade dress. McNeil did not appropriate the aspect of the “Excedrin PM” trade dress that most strongly indicates the origin of the product and McNeil supplied the “Tylenol PM” trade dress with a marking that strongly indicated that the product emanated from a different source.

Even McNeil’s bad faith in adopting many elements of the “Excedrin PM” trade dress does not alter that conclusion. Absence of good faith, like the other Polaroid factors, is not by itself dispositive in any given ease. See Centaur Communications, 830 F.2d at 1228 (“ ‘[I]f comparison of the [marks] reveals no fair ... issue concerning the likelihood of confusion, then intent to copy, even if found from the proffered evidence, would not establish a Lan-ham Act violation.’ ”) (quoting Warner Brothers v. American Broadcasting Co., 720 F.2d 231, 247 (2d Cir.1983)). The prominent placement of the “Tylenol” name goes far toward countering any suggestion that McNeil intended to confuse its customers as to the source of its product. We believe that even “low involvement” consumers will have a hard time missing the famous names that form the most prominent feature of the trade dress in each case. It is thus not surprising that Bristol was unable to present any meaningful evidence of actual confusion between the two products.

We do not mean to intimate that the distinctive elements of any trade dress may be freely appropriated as long as the junior user clearly identifies the source of the goods. In many cases, the distinctive elements of a trade dress may themselves be eligible for trademark protection. In other cases the trade name may be a less dominant feature of the entire trade dress and thus have less force in countering other similarities between two trade dresses. Also, the junior user’s trade name may less strongly identify a particular source than the “Tylenol” name at issue here.

In this case, however, the trade name is the most prominent and distinctive feature in each of the two trade dresses. The fact that they each are nationally prominent names and are different on the respective packages outweighs the evidence that tends to demonstrate a likelihood of confusion.

Because there is no likelihood of confusion between these two products resulting from any similarities in their overall trade dress, we are left with a “ ‘definite and firm conviction,’ ” see Standard & Poor’s, 683 F.2d at 708, that the district court erred in preliminarily enjoining McNeil from marketing “Tylenol PM” in the proposed trade dress. Thus we reverse the decision to grant preliminary relief on the Lanham Act claim.

*1048III. STATE UNFAIR COMPETITION CLAIMS

The district court also enjoined McNeil’s use of its trade dress on the basis of New York and Florida common law of unfair competition. In order to prevail under New York law, a plaintiff must demonstrate a likelihood of confusion between the two products. See Coach Leatherware, 933 F.2d at 169. Similarly, likelihood of confusion is a necessary element of an unfair competition claim under Florida law. See American Bank of Merritt Island v. First American Bank and Trust, 455 So.2d 443, 445-46 (Fla. Dist. Ct.App.), review denied, 461 So.2d 114 (Fla.1984). As we have determined with regard to the Lanham Act claim, there is no likelihood of consumer confusion arising from McNeil’s use of the “Tylenol PM” trade dress. Therefore, Bristol is not entitled to relief under its common law unfair competition theory.

Bristol also claims on appeal that it is entitled to a preliminary injunction, notwithstanding absence of confusion, under New York state common law under a theory of “misappropriation.” Bristol seizes on our language in Flexitized, Inc. v. National Flexitized Corporation, 335 F.2d 774, 781-82 (2d Cir.1964), cert. denied, 380 U.S. 913, 85 S.Ct. 899, 13 L.Ed.2d 799 (1965), to the effect that New York law will prevent a party from misappropriating “ ‘the results of the skill, expenditures and labors of a competitor.’ ” Id. at 781 (citations omitted). Bristol claims that the district court’s finding of bad faith, without more, entitles it to relief under this theory.

We disagree. In Flexitized, we were faced with one party that had appropriated the name of another. Relief was unavailable under the Lanham Act because the name “Flexitized” was deemed descriptive and had not acquired secondary meaning. Because the appropriating party had exhibited bad faith and because the name had acquired some familiarity among the consuming public, we held that use of the name constituted unfair competition. We later clarified that this “misappropriation” theory still requires some confusion, if not as to source at least as to sponsorship or permission. See American Footwear Corp., 609 F.2d at 662. That type of confusion has not been demonstrated here.

More troubling is the application of New York law to Bristol’s request to enjoin use of the “PM” mark. New York law, unlike federal law, does not conclusively presume that there is no likelihood of confusion if a descriptive mark has not acquired secondary meaning. See Coach Leatherware, 933 F.2d at 169. Although it may be very difficult to show a likelihood of confusion where the senior mark is not associated in the public mind with a particular producer, under New York law it is not impossible. The district court did not analyze the Polaroid factors with regard to the use of the “PM” designator and thus we can make no determination to that effect. We note only that, under New York law, an absence of secondary meaning in a descriptive mark does not necessarily result in defeat for the owner. The district court should consider this issue in the first instance.

IV. THE NEW YORK ANTI-DILUTION STATUTE CLAIM

Bristol also appeals from the refusal of the district court to enter an injunction against McNeil on the basis of New York’s “Anti-Dilution” statute, New York Gen.Bus.Law § 368-d. That statute states:

Likelihood of injury to business reputation or of dilution of the distinctive quality of a mark or trade name shall be a ground for injunctive relief in cases of infringement of a mark registered or not registered or in cases of unfair competition, notwithstanding the absence of competition between the parties or the absence of confusion as to the source of the goods or services.

New York Gen.Bus.Law § 368-d. As the plain language of the statute indicates, relief is available under section 368-d even when there is no likelihood of confusion as *1049to the source of the product. See Sally Gee, Inc. v. Myra Hogan, Inc., 699 F.2d 621, 624 (2d Cir.1983); Allied Maintenance v. Allied Mechanical Trades, 42 N.Y.2d 538, 545, 399 N.Y.S.2d 628, 632, 369 N.E.2d 1162, 1166 (1977).

The district court denied relief under section 368-d because Bristol and McNeil are direct competitors and the district court determined that section 368-d applied only to non-competing goods. Whether section 368-d applies to products in direct competition is a question that has split the district courts in this Circuit. See E.P. Lehmann Co. v. Polk’s Modelcraft Hobbies, 770 F.Supp. 202, 206 (S.D.N.Y.1991) (collecting and discussing cases). New York courts have not directly addressed this issue.

We need not resolve this conflict over New York law because, even assuming ar-guendo that the statute applies to competitors, under the appropriate legal standard Bristol cannot prevail on its claim under section 368-d. “[0]nly those trade names which are truly of distinctive quality or which have acquired a secondary meaning in the mind of the public should be entitled to protection under the anti-dilution statute.” Allied Maintenance, 42 N.Y.2d at 546, 399 N.Y.S.2d at 633, 369 N.E.2d at 1166. In interpreting this New York statute, we have held that it “protects only extremely strong marks.” Sally Gee, Inc., 699 F.2d at 625. The “PM” designator appended to the analgesic trade name does not qualify as an “extremely strong mark.” Nor do the aspects of the “Excedrin PM” trade dress that are common to both products constitute “extremely strong marks.” Thus, Bristol is not entitled to protection under section 368-d.

V. DISMISSAL OF THE ACTION

Although this is an interlocutory appeal from a preliminary injunction, McNeil argues that we should dismiss Bristol’s action in its entirety. The Supreme Court has cautioned that “it is generally inappropriate for a federal court at the preliminary-injunction stage to give a final judgment on the merits.” University of Texas v. Camenisch, 451 U.S. 390, 395, 101 S.Ct. 1830, 1834, 68 L.Ed.2d 175 (1981) (citations omitted). Although in the past we have dismissed Lanham Act cases on appeal from the grant or denial of a preliminary injunction, see, e.g., American Cyanamid v. Connaught Laboratories, 800 F.2d 306, 307 (2d Cir.1986), we do not believe that this case appropriately lends itself to such extraordinary summary disposition.

On this record, the balancing of the Polaroid factors does not convince us that Bristol is likely to succeed in its claim. However, findings of fact at the preliminary injunction stage are not binding at a trial on the merits. See Camenisch, 451 U.S. at 395, 101 S.Ct. at 1834. There remains the possibility, albeit slight, that Bristol will present evidence at trial — numerous instances of actual confusion between the two products by actual customers for example — that might lead to a change in the ultimate conclusion regarding the likelihood of consumer confusion. We therefore decline to dismiss Bristol’s action in its entirety.

CONCLUSION

For the reasons stated above, the district court erred in preliminarily enjoining McNeil from marketing its combination analgesic/sleep aid in the “Tylenol PM” trade dress. The injunction is therefore vacated and the matter is remanded to the district court.

OPINION

OWEN, District Judge.

Plaintiff Parrot Jungle, Inc. (“Parrot-Miami”) moves this court for an order pursuant to Fed.R.Civ.P. 65 preliminarily enjoining defendants Parrot Jungle, Inc. (“Parrot-New York”) and its president, Steven Kates, from using the name “Parrot Jungle.” Plaintiffs allege that such use constitutes a violation of § 43(a) of the Lanham Act, 15 U.S.C. § 1125(a), which prohibits, inter alia, the use of a false designation of origin of goods or services in commerce.

In two days of testimony and argument at the preliminary injunction hearing, the following emerged: Parrot-Miami owns a tourist attraction in Miami, Florida called Parrot Jungle, which was created in 1936 by one Franz Scherr, the father of plaintiff’s current president, Jerome Scherr. It is situated on twelve acres of lush-foliaged land and features parrots and other birds, uncaged and in a natural habitat, much like a jungle. Parrot Jungle in Miami is a nationally-known enterprise, which each year attracts some 400,000 visitors who pay five dollars per adult admission to view the birds and grounds and to watch a show of trained parrots. Plaintiff’s gross revenues from its operations amounted to $3.2 million in 1980.

Defendant (“Parrot-New York”), using the identical name, Parrot Jungle, Inc., commenced operations in 1977 in a store in Brooklyn. Its president and owner is defendant Steven Kates. Parrot-New York, a New York corporation, offers expensive parrots for sale at retail outlets. Kates, who has a considerable background in birds,1 testified that he was “obviously aware” of Parrot-Miami when he selected a name for his operation since he had visited *268the Miami attraction in 1972 and again in 1974.2 Kates insisted, however, that he chose the name “Parrot Jungle” for his New York store because customers at a friend’s exotic plant store at which he earlier had displayed his uncaged birds frequently commented that the store looked and sounded like a jungle. Kates said that when he opened his own store, he merely combined those comments with the product he was selling to arrive at the name. Defendants’ stores today do display the birds perched freely on simulated tree branches against a background of jungle-like wallpaper.

In 1978 Kates opened additional retail parrot stores, one in Paramus, New Jersey, one in Manhattan, and two, which failed, on Long Island. Plaintiff’s president, Jerome Scherr, concedes that by very early 1979 he was aware that perhaps three stores in the New York area were using plaintiff’s name, but he took no action because he regarded the stores as nothing “more than pet shops” which “come and go.” In August 1980, however, Scherr heard that Parrot-New York was embarking upon further expansion, and in October he learned from an interview show aired on national television and an article in a magazine that Parrot-New York had begun a national franchising program. This discovery immediately prompted Scherr to contact counsel and retain a trademark specialist. By that time, however, defendants had franchised their first three stores, one in Westbury (Long Island), one in Scarsdale, New York, and one in Stamford, Connecticut. •

In assessing plaintiff’s claim under § 43(a) of the Lanham Act, I turn first to an analysis of plaintiff’s trademark to determine the scope of protection to which that mark is entitled. The cases interpret the Lanham Act as identifying four categories of trademarks, each with its own degree of protection against infringement. In ascending order of strength, the four categories are (1) generic, (2) descriptive, (3) suggestive, and (4) arbitrary or fanciful. McGregor-Doniger Inc. v. Drizzle, Inc., 599 F.2d 1126, 1131 (2d Cir. 1979). It should be noted that to some extent, the categories and the legal principles which they embody have “become lost in a welter of adjectives.” Abercrombie & Fitch Company v. Hunting World, Incorporated, 537 F.2d 4, 9 (2d Cir. 1976). Applying the teaching of McGregor, supra, I conclude on the record before me that the name “Parrot Jungle” is in the area of descriptive or suggestive. It is not generic or arbitrary or fanciful. See generally 3 R. Callman, The Law of Unfair Competition, Trademarks, and Monopolies § 71, at 115-196 (3d ed. 1969).

Further guidance in the suggestive-descriptive area may be found in Stix Products, Inc. v. United Merchants and Manufacturers Inc., 295 F.Supp. 479 (S.D.N.Y.1968), where the court stated at page 488:

A term is suggestive if it requires imagination, thought and perception to reach a conclusion as to the nature of goods. A term is descriptive if it forthwith conveys an immediate idea of the ingredients, qualities or characteristics of the goods.

Keeping these guidelines in mind, I conclude, that plaintiff’s m'ark belongs properly in the descriptive category.3 As a descriptive mark, it is entitled to protection against infringement by an identical name if its senior user can demonstrate (1) a likelihood that an appreciable number of ordinarily prudent consumers are likely to be confused as to the source of goods and services defendants offer for sale, McGregor-Doniger, Inc. v. Drizzle, Inc., supra, 599 F.2d at 1132, and (2) that its mark has obtained “secondary meaning” in its market, i. e., that the consuming public has come to associate the name “Parrot Jungle” with plaintiff’s attraction in Miami. W. E. Basset Company v. Revlon, Inc., 435 F.2d 656, 661 (2d Cir. 1970).

*269In reaching a determination on the secondary meaning issue, the court will consider, among other factors, (1) the duration of plaintiff’s exclusive use of its mark, (2) the nature and extent of plaintiff’s advertising and promotion of its mark, (3) the extent to which plaintiff’s business has, other than by its own efforts, come to the attention of the consuming public, (4) plaintiff’s patronage record, and (5) defendants’ conscious copying of plaintiff’s mark. E. g., Harlequin Enterprises Limited v. Gulf & Western Corporation, 503 F.Supp. 647, 650 (S.D.N.Y.1980), aff’d, 644 F.2d 946 (2d Cir. 1981); RJR Foods, Inc. v. White Rock Corp., 603 F.2d 1058, 1060 (2d Cir. 1979); see generally 3 Callman, supra, § 77.1.

Examining these factors, I find that plaintiff has used its mark continuously and exclusively for more than forty years. Plaintiff advertises modestly but extensively, primarily by providing brochures to travel agencies and tour groups worldwide which have booked tours to Parrot-Miami. Its advertising expenditures in 1980 neared $60,000 and, since 1957, have totalled almost $600,000. Plaintiff widely promotes its Florida attraction, most notably by permitting various companies, such as Eastman Kodak and Union Carbide, to use photographs of plaintiff’s establishment in their advertising and also by travel promotions and advertising sponsored by several airlines, including Eastern, National, and Pan Am. This latter advertising has appeared in such major national publications as Time, Newsweek, U.S. News and World Report, and Sports Illustrated. In addition, plaintiff, through the Florida Department of Commerce, distributes a twenty-eight minute film about The Parrot Jungle in Miami which has been screened for a number of years throughout the country on television, in school classrooms, and on airplanes. Apparently more than forty million persons have seen this film. Parrot-Miami has also been the subject of scores of articles in publications around the country and the world, among them, Life, National Geographic, The Chicago Tribune, The Saturday Evening Post, the Detroit News, the Miami Herald, This Week (the weekly magazine of the now-defunct New York Herald Tribune), and the Toronto Star. The promotion and advertising apparently has been successful; the number of visitors to the attraction has grown steadily since 1936 to approximately 400,000 in -1980.

I also find on the record of this hearing that, when he named his business, defendant Steven Kates deliberately copied plaintiff’s mark to take advantage of plaintiff’s good will and reputation. Kates admitted that, prior to starting to use the name “Parrot Jungle,” he had twice been to the Miami attraction and therefore knew its name.4

Kates’s state of mind is further evidenced by the fact that he must have told his lawyer during preparation for the hearing that plaintiff had granted him permission to use the name, for prior to the hearing his lawyer told the court the following:

Mr. Kates visited Parrot Jungle, Inc. in Florida in 1977, and thereafter told Mr. Scherr and a Mr. Springer, who at the time was a trainer at the amusement park, that he had opened a store in New York to sell birds and that he was going to use the name Parrot Jungle, Inc. in connection with his retail activities.

Mr. Scherr said nothing. As a matter of fact, he wished him good luck and made no objection to his use of the name.

Kates later testified before me, however, to having had only two conversations with Mr. Scherr, neither of which related to this crucial subject. The fact that this alleged conversation, first recited to the lawyer, was not later supported under oath strongly suggests that Kates approached the hearing aware that he had made a mistake in copying plaintiff’s name and was groping for an explanation. There is no question that even in 1977 Kates was aware of the existence of a problem, for he acknowledges conferring with a lawyer about the propriety of using plaintiff’s name before he adopted it for his own business. And final*270ly, the acknowledgements in Kates’s book, “Encyclopedia of Cockatoos”, are expressed in such a way that a reader might be led to conclude that Parrot-Miami assisted Kates in preparing the book.5

The foregoing constitutes abundant evidence that defendants intentionally imitated plaintiff’s mark. Deliberate copying, of course, is persuasive evidence that plaintiff has succeeded in creating consumer recognition and good will for its service business. RJR Foods, Inc. v. White Rock Corp., supra, 603 F.2d at 1060; Socony-Vacuum Oil Co., Inc. v. Rosen, 108 F.2d 632, 636 (6th Cir. 1940). Thus, from this evidence of intentional copying, combined with the other factors discussed above, I conclude for purposes of a preliminary injunction that plaintiff’s mark has achieved secondary meaning.

Next to be determined is whether there is a likelihood that the relevant public is likely to confuse Parrot-New York with Parrot-Miami. The principal factors to consider on this issue include intentional copying and secondary meaning, already determined, as well as the “proximity” of plaintiff’s and defendants’ businesses. McGregor-Doniger, Inc. v. Drizzle, Inc., supra, 599 F.2d at 1130. First, defendants’ deliberate use of plaintiff’s exact name unquestionably gives rise to a legal presumption that the similarity will cause consumer confusion. RJR Foods, Inc. v. White Rock Corp., supra, 603 F.2d at 1060; Harold F. Ritchie, Inc. v. Chesebrough-Ponds, Inc., 281 F.2d 755, 760 (2d Cir. 1960) quoting My-T-Fine Corp. v. Samuels, 69 F.2d 76, 77 (2d Cir. 1934). In addition, the fact that plaintiff’s mark has secondary meaning enhances its strength and thus the probability of confusion. And finally, the proximity of the two businesses is undisputed; each party displays tropical birds, the plaintiff for viewing and the defendants for sale. Thus, for purposes of this preliminary injunction, plaintiff has sufficiently demonstrated its claim of trademark infringement under § 43(a) of the Lanham Act.

Defendants contend, however, that in any event, relief to the plaintiff is barred by laches, because after plaintiff first learned of defendants’ use of its name it delayed unreasonably in bringing this action. In my view, however, there is a substantial difference between plaintiff’s awareness of a pet store or stores in New York and its awareness of a national franchising effort. It was the latter which triggered plaintiff’s prompt assertion of its rights. A modest encroachment is one thing, a sudden proposed national exploitation of plaintiff’s name is quite another, and plaintiff’s failure to challenge the former will not entirely disable plaintiffs from preventing the latter. See Miss Universe, Inc. v. Patricelli, 271 F.Supp. 104, 110 (D.C.Conn.), aff’d, 386 F.2d 997 (2d Cir. 1967).

Finally, I note that among the legitimate interests the senior user of a mark is entitled to protect are (1) its right to expand into an obviously related field in which an infringer is operating and (2) its right to run and develop its business free from damage to its reputation by the potentially improper practices of an infringer not under its control.6 Avon Shoe Co. v. David *271 Crystal, Inc., 279 F.2d 607, 613 (2d Cir.), cert. denied, 364 U.S. 909, 81 S.Ct. 271, 5 L.Ed.2d 224 (1960). Both of these rights are entitled to protection here.

Given the record before me I conclude that plaintiff has met its burden of demonstrating (1) that defendants’ planned expansion of their use of the name “Parrot Jungle” will, in all reasonable probability, cause plaintiff to suffer irreparable harm and (2) that plaintiff is likely to succeed on the merits of its claim. Jack Kahn Music Co., Inc. v. Baldwin Piano & Organ Co., 604 F.2d 755, 758 (2d Cir. 1979).

Plaintiff’s motion for a preliminary injunction is therefore granted to the following extent: defendants Parrot Jungle, Inc. and Steven Kates and anyone acting in concert or association with them are hereby preliminarily enjoined from establishing any new stores, whether by franchise or some other method, using the name “Parrot Jungle”. All other requested relief is, at this time denied.

Settle Order on reasonable notice.