Mr. Justice Harlan
delivered the opinion of the Court.
In January of 1952, John Adkins, an inventor and mechanical engineer, was hired by Lear, Incorporated, for the purpose of solving a vexing problem the company had encountered in its efforts to develop a gyroscope which would meet the increasingly demanding requirements of the aviation industry. The gyroscope is an essential component of the navigational system in all aircraft, enabling the pilot to learn the direction and attitude of his airplane. With the development of the faster airplanes of the 1950’s, more accurate gyroscopes were needed, and the gyro industry consequently was casting about for new techniques which would satisfy this need in an economical fashion. Shortly after Adkins was hired, he developed a method of construction at the company’s California facilities which improved gyroscope accuracy at a low cost. Lear almost immediately incorporated Adkins’ improvements into its production process to its substantial advantage.
The question that remains unsettled in this case, after eight years of litigation in the California courts, is whether Adkins will receive compensation for Lear’s use of those improvements which the inventor has subsequently patented. At every stage of this lawsuit, Lear has sought to prove that, despite the grant of a patent [656]*656by the Patent Office, none of Adkins’ improvements were sufficiently novel to warrant the award of a monopoly under the standards delineated in the governing federal statutes. Moreover, the company has sought to prove that Adkins obtained his patent by means of a fraud on the Patent Office. In response, the inventor has argued that since Lear had entered into a licensing agreement with Adkins, it was obliged to pay the agreed royalties regardless of the validity of the underlying patent.
The Supreme Court of California unanimously vindicated the inventor’s position. While the court recognized that generally a manufacturer is free to challenge the validity of an inventor’s patent, it held that “one of the oldest doctrines in the field of patent law establishes that so long as a licensee is operating under a license agreement he is estopped to deny the validity of his licensor’s patent in a suit for royalties under the agreement. The theory underlying this doctrine is that a licensee should not be permitted to enjoy the benefit afforded by the agreement while simultaneously urging that the patent which forms the basis of the agreement is void.” 67 Cal. 2d 882, 891, 435 P. 2d 321, 325-326 (1967).
Almost 20 years ago, in its last consideration of the doctrine, this Court also invoked an estoppel to deny a licensee the right to prove that his licensor was demanding royalties for the use of an idea which was in reality a part of the public domain. Automatic Radio Manufacturing Co. v. Hazeltine Research, Inc., 339 U. S. 827, 836 (1950). We granted certiorari in the present case, 391 U. S. 912, to reconsider the validity of the Hazeltine rule in the light of our recent decisions emphasizing the strong federal policy favoring free competition in ideas which do not merit patent protection. Sears, Roebuck v. Stiffel Co., 376 U. S. 225 (1964); Compco Corp. v. Day-Brite Lighting, Inc., 376 U. S. 234 (1964).
[657]*657I.
At the very beginning of the parties’ relationship, Lear and Adkins entered into a rudimentary one-page agreement which provided that although “[a] 11 new ideas, discoveries, inventions, etc., related to . . . vertical gyros become the property of Mr. John S. Adkins,” the inventor promised to grant Lear a license as to all ideas he might develop “on a mutually satisfactory royalty basis.” 1 As soon as Adkins’ labors yielded tangible results, it quickly became apparent to the inventor that further steps should be taken to place his rights to his ideas on a firmer basis. On February 4, 1954, Adkins filed an application with the Patent Office in an effort to gain federal protection for his improvements. At about the same time, he entered into a lengthy period of negotiations with Lear in an effort to conclude a licensing agreement which would clearly establish the amount of royalties that would be paid.
These negotiations finally bore fruit on September 15, 1955, when the parties approved a complex 17-page contract which carefully delineated the conditions upon which Lear promised to pay royalties for Adkins’ improvements. The parties agreed that if “the U. S. Patent Office refuses to issue a patent on the substantial claims [contained in Adkins’ original patent application] or if such a patent so issued is subsequently held invalid, then in any of such events Lear at its option shall have the right forthwith to terminate the specific license so affected or to terminate this entire Agreement . . . .” § 6. (2 App. 138.)
[658]*658As the contractual language indicates, Adkins had not obtained a final Patent Office decision as to the patent-ability of his invention at the time the licensing agreement was concluded. Indeed, he was not to receive a patent until January 5, 1960. This long delay has its source in the special character of Patent Office procedures. The regulations do not require the Office to make a final judgment on an invention’s patentability on the basis of the inventor’s original application.2 While it sometimes happens that a patent is granted at this early stage, it is far more common for the Office to find that although certain of the applicant’s claims may be patentable, certain others have been fully anticipated by the earlier developments in the art. In such a situation, the Patent Office does not attempt to separate the wheat from the chaff on its own initiative. Instead, it rejects the application, giving the inventor the right to make an amendment which narrows his claim to cover only those aspects of the invention which are truly novel.3 It often happens, however, that even after an application is amended, the Patent Office finds that some of the remaining claims are unpatentable. When this occurs, the agency again issues a rejection which is subject to further amendment.4 And so the process of rejection and amendment continues until the Patent Office Examiner either grants a patent or concludes that none of the inventor’s claims could possibly be patentable, at which time a final rejection is entered on the Office’s records.5 Thus, when Adkins made his original application in 1954, it took the average inventor more than three years before he obtained a final administrative decision on the patentability of his ideas, with the Patent [659]*659Office acting on the average application from two to four times.6
The progress of Adkins’ effort to obtain a patent followed the typical pattern. In his initial application, the inventor made the ambitious claim that his entire method of constructing gyroscopes was sufficiently novel to merit protection. The Patent Office, however, rejected this initial claim, as well as two subsequent amendments, which progressively narrowed the scope of the invention sought to be protected. Finally, Adkins narrowed his claim drastically to assert only that the design of the apparatus used to achieve gyroscope accuracy was novel.7 In response, the Office issued its 1960 patent, granting a 17-year monopoly on this more modest claim.
During the long period in which Adkins was attempting to convince the Patent Office of the novelty of his ideas, however, Lear had become convinced that Adkins would never receive a patent on his invention and that it should not continue to pay substantial royalties on ideas which had not contributed substantially to the development of the art of gyroscopy. In 1957, after Adkins’ patent application had been rejected twice, Lear announced that it had searched the Patent Office’s files and had found a patent which it believed had fully anticipated Adkins’ discovery. As a result, the company stated that it would no longer pay royalties on the large number of gyroscopes it was producing at its plant in Grand Rapids, Michigan (the Michigan gyros). Payments were continued on the smaller number of gyros produced at the company’s [660]*660California plant (the California gyros) for two more years until they too were terminated on April 8, 1959.
As soon as Adkins obtained his patent in 1960, he brought this lawsuit in the California Superior Court. He argued to a jury that both the Michigan and the California gyros incorporated his patented apparatus and that Lear’s failure to pay royalties on these gyros was a breach both of the 1955 contract and of Lear’s quasi-contractual obligations. Although Lear sought to raise patent invalidity as a defense, the trial judge directed a verdict of $16,351.93 for Adkins on the California gyros, holding that Lear was estopped by its licensing agreement from questioning the inventor’s patent. The trial judge took a different approach when it came to considering the Michigan gyros. Noting that the company claimed that it had developed its Michigan designs independently of Adkins’ ideas, the court instructed the jury to award the inventor recovery only if it was satisfied that Adkins’ invention was novel, within the meaning of the federal patent laws. When the jury returned a verdict for Adkins of $888,122.56 on the Michigan gyros,8 the trial judge granted Lear’s motion for judgment notwithstanding the verdict, finding that Adkins’ invention had been completely anticipated by the prior art.9
[661]*661Neither side was satisfied with this split decision, and both appealed to the California District Court of Appeal, which adopted a quite different approach. The court held that Lear was within its contractual rights in terminating its royalty obligations entirely in 1959, and that if Adkins desired to recover damages after that date he was “relegated to an action for infringement” in the federal courts. 52 Cal. Rptr. 795, 806. So far as pre-1959 royalties were concerned, the court held that the contract required the company to pay royalties on both the California and Michigan gyros regardless of the validity of the inventor’s patent. 52 Cal. Rptr., at 809.
Once again both sides appealed, this time to the California Supreme Court, which took yet another approach to the problem presented. The court rejected the District Court of Appeal’s conclusion that the 1955 license gave Lear the right to terminate its royalty obligations in 1959. Since the 1955 agreement was still in effect, the court concluded, relying on the language we have already quoted, that the doctrine of estoppel barred Lear from questioning the propriety of the Patent Office’s grant. 67 Cal. 2d, at 907, 435 P. 2d, at 336. The court’s adherence to estoppel, however, was not without qualification. After noting Lear’s claim that it had developed its Michigan gyros independently, the court tested this contention by considering “whether what is being built by Lear [in Michigan] springs entirely” (emphasis supplied) from the prior art. 67 Cal. 2d, at 913, 435 P. 2d, at 340. Applying this test, it found that Lear had in fact “utilized the apparatus patented by Adkins throughout the period in question,” 67 Cal. 2d, at 915, 435 P. 2d, at 341, and reinstated the jury’s $888,000 verdict on this branch of the case.
II.
Since the California Supreme Court’s construction of the 1955 licensing agreement is solely a matter of state [662]*662law, the only issue open to us is raised by the court’s reliance upon the doctrine of estoppel to bar Lear from proving that Adkins’ ideas were dedicated to the common welfare by federal law.10 In considering the propriety of the State Court’s decision, we are well aware that we are not writing upon a clean slate. The doctrine of estoppel has been considered by this Court in a line of cases reaching back into the middle of the 19th century. Before deciding what the role of estoppel [663]*663should be in the present case and in the future, it is, then, desirable to consider the role it has played in the past.
A.
While the roots of the doctrine have often been celebrated in tradition, we have found only one 19th century case in this Court that invoked estoppel in a considered manner. And that case was decided before the Sherman Act made it clear that the grant of monopoly power to a patent owner constituted a limited exception to the general federal policy favoring free competition. Kinsman v. Parkhurst, 18 How. 289 (1856).11 Curiously, a second decision often cited as supporting the estoppel doctrine points clearly in the opposite direction. St. Paul Plow Works v. Starling, 140 U. S. 184 (1891), did not even question the right of the lower courts to admit the licensee’s evidence showing that the patented device was not novel. A unanimous Court merely held that, where there was conflicting evidence as to an invention’s novelty, it would not reverse the decision of the lower court upholding the patent’s validity.
In the very next year, this Court found the doctrine of patent estoppel so inequitable that it refused to grant an injunction to enforce a licensee’s promise never to contest the validity of the underlying patent. “It is as [664]*664important to the public that competition should not be repressed by worthless patents, as that the patentee of a really valuable invention should be protected in his monopoly . . . .” Pope Manufacturing Co. v. Gormully, 144 U. S. 224, 234 (1892).
Although this Court invoked an estoppel in 1905 without citing or considering Pope’s, powerful argument, United States v. Harvey Steel Co., 196 U. S. 310, the doctrine was not to be applied again in this Court until it was revived in Automatic Radio Manufacturing Co. v. Hazeltine Research, Inc., supra, which declared, without prolonged analysis, that licensee estoppel was “the general rule.” 339 U. S., at 836. In so holding, the majority ignored the teachings of a series of decisions this Court had rendered during the 45 years since Harvey had been decided. During this period, each time a patentee sought to rely upon his estoppel privilege before this Court, the majority created a new exception to permit judicial scrutiny into the validity of the Patent Office’s grant. Long before Hazeltine was decided, the estoppel doctrine had been so eroded that it could no longer be considered the “general rule,” but was only to be invoked in an ever-narrowing set of circumstances.
B.
The estoppel rule was first stringently limited in a situation in which the patentee’s equities were far more compelling than those presented in the typical licensing arrangement. Westinghouse Electric & Manufacturing Co. v. Formica Insulation Co., 266 U. S. 342 (1924), framed a rule to govern the recurring problem which arises when the original patent owner, after assigning his patent to another for a substantial sum, claims that the patent is worthless because it contains no new ideas. The courts of appeals had traditionally refused to permit such a defense to an infringement action on the ground [665]*665that it was improper both to “sell and keep the same thing,” Faulks v. Kamp, 3 F. 898, 902 (1880). Nevertheless, Formica imposed a limitation upon estoppel which was radically inconsistent with the premises upon which the “general rule” is based. The Court held that while an assignor may not directly attack the validity of a patent by reference to the prior state of the art, he could introduce such evidence to narrow the claims made in the patent. “The distinction may be a nice one but seems to be workable.” 266 U. S., at 351. Workable or not, the result proved to be an anomaly: if a patent had some novelty Formica permitted the old owner to defend an infringement action by showing that the invention’s novel aspects did not extend to the inclusion of the old owner’s products; on the other hand, if a patent had no novelty at all, the old owner could not defend successfully since he would be obliged to launch the direct attack on the patent that Formica seemed to forbid. The incongruity of this position compelled at least one court of appeals to carry the reasoning of the Formica exception to its logical conclusion. In 1940 the Seventh Circuit held that a licensee could introduce evidence of the prior art to show that the licensor’s claims were not novel at all and thus successfully defend an action for royalties. Casco Products Corp. v. Sinko Tool & Manufacturing Co., 116 F. 2d 119.
In Scott Paper Co. v. Marcalus Manufacturing Co., 326 U. S. 249 (1945), this Court adopted a position similar to the Seventh Circuit’s, undermining the basis of patent estoppel even more than Formica had done. In Scott, the original patent owner had attempted to defend an infringement suit brought by his assignee by proving that his product was a copy of an expired patent. The Court refused to permit the assignee to invoke an estop-pel, finding that the policy of the patent laws would be frustrated if a manufacturer was required to pay for the use of information which, under the patent statutes, was [666]*666the property of all. Chief Justice Stone, for the Court, did not go beyond the precise question presented by a manufacturer who asserted that he was simply copying an expired patent. Nevertheless it was impossible to limit the Scott doctrine to such a narrow compass. If patent policy forbids estoppel when the old owner attempts to show that he did no more than copy an expired patent, why should not the old owner also be permitted to show that the invention lacked novelty because it could be found in a technical journal or because it was obvious to one knowledgeable in the art? As Justice Frankfurter’s dissent indicated, id., at 258-264, there were no satisfactory answers to these questions. The Scott exception had undermined the very basis of the “general rule.”
C.
At about the time Scott was decided, this Court developed yet another doctrine which was profoundly antithetic to the principles underlying estoppel. In Sola Electric Co. v. Jefferson Electric Co., 317 U. S. 173 (1942), the majority refused to permit a licensor to enforce the license’s price-fixing provisions without permitting the licensee to contest the validity of the underlying patent. Since the price-fixing clause was per se illegal but for the existence of a valid patent, this narrow exception could be countenanced without compromising the general estoppel principle. But the Sola Court went further: it held that since the patentee had sought to enforce the price-fixing clause, the licensee could also avoid paying royalties if he could show that the patent was invalid. Five years later, the “anti-trust exception” was given an even more extensive scope in the Katzinger and MacGregor cases.12 Here, licensors [667]*667were not permitted to invoke an estoppel despite the fact that they sought only to collect their royalties. The mere existence of a price-fixing clause in the license was held to be enough to bring the validity of the patent into question. Thus in the large number of cases in which licensing agreements contained restrictions that were arguably illegal under the antitrust laws, the doctrine of estoppel was a dead letter. Justice Frankfurter, in dissent, went even further, concluding that Katzinger and MacGregor had done all but repudiate the estoppel rule: “If a doctrine that was vital law for more than ninety years will be found to have now been deprived of life, we ought at least to give it decent public burial.” 329 U. S., at 416.
D.
The lower courts, both state and federal, have also hedged the impact of estoppel by creating exceptions which have indicated a recognition of the broader policies pointing to a contrary approach. It is generally the rule that licensees may avoid further royalty payments, regardless of the provisions of their contract, once a third party proves that the patent is invalid. See, e. g., Drackett Chemical Co. v. Chamberlain Co., 63 F. 2d 853 (1933). Some courts have gone further to hold that a licensee may notify the patent owner that he is repudiating his agreement, regardless of its terms, and may subsequently defend any action for royalties by proving patent invalidity. Note, The Doctrine of Licensee Repudiation in Patent Law, 63 Yale L. J. 125 (1953); R. Ellis, Patent Licenses § 328 (3d ed., A. Deller 1958). And even in the 19th century, state courts had held that if the licensee had not actually sold products incorporating the patent’s ideas, he could challenge the validity of the patent. See Forkosch, Licensee [668]*668Estoppel in Patent Law, 20 Temp. L. Q. 515, 529, n. 45 (1947).13
III.
The uncertain status of licensee estoppel in the case law is a product of judicial efforts to accommodate the competing demands of the common law of contracts and the federal law of patents. On the one hand, the law of contracts forbids a purchaser to repudiate his promises simply because he later becomes dissatisfied with the bargain he has made.14 On the other hand, federal law requires that all ideas in general circulation be dedicated to the common good unless they are protected by a valid patent. Sears, Roebuck v. Stiffel Co., supra; Compco Corp. v. Day-Brite Lighting, Inc., supra. When faced with this basic conflict in policy, both this Court and courts throughout the land have naturally sought to develop an intermediate position which somehow would remain responsive to the radically different concerns of the two different worlds of contract and patent. The result has been a failure. Rather than creative compromise, there has been a chaos of conflicting case law, proceeding on inconsistent premises. Before renewing the search for an acceptable middle ground, we must reconsider on their own merits the arguments which may properly be advanced on both sides of the estoppel question.
[669]*669A.
It will simplify matters greatly if we first consider the most typical situation in which patent licenses are negotiated. In contrast to the present case, most manufacturers obtain a license after a patent has issued. Since the Patent Office makes an inventor’s ideas public when it issues its grant of a limited monopoly,15 a potential licensee has access to the inventor’s ideas even if he does not enter into an agreement with the patent owner. Consequently, a manufacturer gains only two benefits if he chooses to enter a licensing agreement after the patent has issued. First, by accepting a license and paying royalties for a time, the licensee may have avoided the necessity of defending an expensive infringement action during the period when he may be least able to afford one. Second, the existence of an unchallenged patent may deter others from attempting to compete with the licensee.16
Under ordinary contract principles the mere fact that some benefit is received is enough to require the enforcement of the contract, regardless of the validity of the underlying patent. Nevertheless, if one tests this result by the standard of good-faith commercial dealing, it seems far from satisfactory. For the simple contract approach entirely ignores the position of the licensor who is seeking to invoke the court’s assistance on his behalf. Consider, for example, the equities of the licensor who has obtained his patent through a fraud on the Patent Office. It is difficult to perceive why good [670]*670faith requires that courts should permit him to recover royalties despite his licensee’s attempts to show that the patent is invalid. Compare Walker Process Equipment, Inc. v. Food Machinery & Chemical Corp., 382 U. S. 172 (1965).
Even in the more typical cases, not involving conscious wrongdoing, the licensor’s equities are far from compelling. A patent, in the last analysis, simply represents a legal conclusion reached by the Patent Office. Moreover, the legal conclusion is predicated on factors as to which reasonable men can differ widely. Yet the Patent Office is often obliged to reach its decision in an ex parte proceeding, without the aid of the arguments which could be advanced by parties interested in proving patent invalidity. Consequently, it does not seem to us to be unfair to require a patentee to defend the Patent Office’s judgment when his licensee places the question in issue, especially since the licensor’s case is buttressed by the presumption of validity which attaches to his patent. Thus, although licensee estoppel may be consistent with the letter of contractual doctrine, we cannot say that it is compelled by the spirit of contract law, which seeks to balance the claims of promisor and promisee in accord with the requirements of good faith.
Surely the equities of the licensor do not weigh very heavily when they are balanced against the important public interest in permitting full and free competition in the use of ideas which are in reality a part of the public domain. Licensees may often be the only individuals with enough economic incentive to challenge the patentability of an inventor’s discovery. If they are muzzled, the public may continually be required to pay tribute to would-be monopolists without need or justification. We think it plain that the technical requirements of contract doctrine must give way before the demands of the public interest in the typical situation [?]*?involving the negotiation of a license after a patent has issued.
We are satisfied that Automatic Radio Manufacturing Co. v. Hazeltine Research, Inc., supra, itself the product of a clouded history, should no longer be regarded as sound law with respect to its “estoppel” holding, and that holding is now overruled.
The case before us, however, presents a far more complicated estoppel problem than the one which arises in the most common licensing context. The problem arises out of the fact that Lear obtained its license in 1955, more than four years before Adkins received his 1960 patent. Indeed, from the very outset of the relationship, Lear obtained special access to Adkins’ ideas in return for its promise to pay satisfactory compensation.
Thus, during the lengthy period in which Adkins was attempting to obtain a patent, Lear gained an important benefit not generally obtained by the typical licensee. For until a patent issues, a potential licensee may not learn his licensor’s ideas simply by requesting the information from the Patent Office. During the time the inventor is seeking patent protection, the governing federal statute requires the Patent Office to hold an inventor’s patent application in confidence.17 If a poten[672]*672tial licensee hopes to use the ideas contained in a secret patent application, he must deal with the inventor himself, unless the inventor chooses to publicize his ideas to the world at large. By promising to pay Adkins royalties from the very outset of their relationship, Lear gained immediate access to ideas which it may well not have learned until the Patent Office published the details of Adkins' invention in 1960. At the core of this case, then, is the difficult question whether federal patent policy bars a State from enforcing a contract regulating access to an unpatented secret idea.18
Adkins takes an extreme position on this question. The inventor does not merely argue that since Lear obtained privileged access to his ideas before 1960, the company should be required to pay royalties accruing before 1960 regardless of the validity of the patent which ultimately issued. He also argues that since Lear obtained special benefits before 1960, it should also pay royalties during the entire patent period (1960-1977), without regard to the validity of the Patent Office’s grant. We cannot accept so broad an argument.
Adkins’ position would permit inventors to negotiate all important licenses during the lengthy period while their applications were still pending at the Patent Office, thereby disabling entirely all those who have the strongest incentive to show that a patent is worthless. While the equities supporting Adkins’ position are somewhat more appealing than those supporting the typical [673]*673licensor, we cannot say that there is enough of a difference to justify such a substantial impairment of overriding federal policy.
Nor can we accept a second argument which may be advanced to support Adkins’ claim to at least a portion of his post-patent royalties, regardless of the validity of the Patent Office grant. The terms of the 1955 agreement provide that royalties are to be paid until such time as the “patent ... is held invalid,” § 6, and the fact remains that the question of patent validity has not been finally determined in this case. Thus, it may be suggested that although Lear must be allowed to raise the question of patent validity in the present lawsuit, it must also be required to comply with its contract and continue to pay royalties until its claim is finally vindicated in the courts.
The parties’ contract, however, is no more controlling on this issue than is the State’s doctrine of estoppel, which is also rooted in contract principles. The decisive question is whether overriding federal policies would be significantly frustrated if licensees could be required to continue to pay royalties during the time they are challenging patent validity in the courts.
It seems to us that such a requirement would be inconsistent with the aims of federal patent policy. Enforcing this contractual provision would give the licensor an additional economic incentive to devise every conceivable dilatory tactic in an effort to postpone the day of final judicial reckoning. We can perceive no reason to encourage dilatory court tactics in this way. Moreover, the cost of prosecuting slow-moving trial proceedings and defending an inevitable appeal might well deter many licensees from attempting to prove patent invalidity in the courts. The deterrent effect would be particularly severe in the many scientific fields in which invention is proceeding at a rapid rate. In these areas, a patent may well become obsolete long before its [674]*67417-year term has expired. If a licensee has reason to believe that he will replace a patented idea with a new one in the near future, he will have little incentive to initiate lengthy court proceedings, unless he is freed from liability at least from the time he refuses to pay the contractual royalties. Lastly, enforcing this contractual provision would undermine the strong federal policy favoring the full and free, use of ideas in the public domain. For all these reasons, we hold that Lear must be permitted to avoid the payment of all royalties accruing after Adkins’ 1960 patent issued if Lear can prove patent invalidity.19
Adkins’ claim to contractual royalties accruing before the 1960 patent issued is, however, a much more difficult one, since it squarely raises the question whether, and to what extent, the States may protect the owners of unpatented inventions who are willing to disclose their ideas to manufacturers only upon payment of royalties. The California Supreme Court did not address itself to this issue with precision, for it believed that the venerable doctrine of estoppel provided a sufficient answer to all of Lear’s claims based upon federal patent law. Thus, we do not know whether the Supreme Court would have awarded Adkins recovery even on his pre-patent royalties if it had recognized that previously established estoppel doctrine could no longer be properly invoked [675]*675with regard to royalties accruing during the 17-year patent period. Our decision today will, of course, require the state courts to reconsider the theoretical basis of their decisions enforcing the contractual rights of inventors and it is impossible to predict the extent to which this re-evaluation may revolutionize the law of any particular State in this regard. Consequently, we have concluded, after much consideration, that even though an important question of federal law underlies this phase of the controversy, we should not now attempt to define in even a limited way the extent, if any, to which the States may properly act to enforce the contractual rights of inventors of unpatented secret ideas. Given the difficulty and importance of this task, it should be undertaken only after the state courts have, after fully focused inquiry, determined the extent to which they will respect the contractual rights of such inventors in the future. Indeed, on remand, the California courts may well reconcile the competing demands of patent and contract law in a way which would not warrant further review in this Court.
IV.
We also find it inappropriate to pass at this time upon Lear’s contention that Adkins’ patent is invalid.
Not only did Lear fail to raise this issue in its petition for certiorari, but the California Supreme Court has yet to pass on the question of patent validity in that clear and unequivocal manner which is so necessary for proper adjudication in this Court. As we have indicated, the California Supreme Court considered the novelty of Adkins’ ideas relevant to its decision at only one stage of its extensive analysis. Since Lear claimed that it had developed its Michigan gyros completely independently of Adkins’ efforts, the Supreme Court believed itself obliged to consider whether Adkins’ ideas were not “entirely” anticipated by the prior art. 67 Cal. 2d, at 913, 435 P. 2d, at 340. Apply[676]*676ing this test, the court upheld the jury’s verdict of $888,000 on the Michigan gyros, finding that “Lear utilized the apparatus patented by Adkins throughout the period in question.” 67 Cal. 2d, at 915, 435 P. 2d, at 341. In reaching this conclusion, however, the court did express its belief that Adkins’ invention made a “significant step forward” in the art of gyroscopy. 67 Cal. 2d, at 915, 435 P. 2d, at 341.
It is far from clear that the court, in making this last statement, intended to hold that Adkins’ ideas satisfied the demanding standard of invention explicated in our decision in Graham v. John Deere Co., 383 U. S. 1 (1966). Surely, such a holding was not required by the court’s analysis, which was concerned only with the question whether Lear had benefited from Adkins’ ideas in any degree. In this context, we believe that Lear must be required to address its arguments attacking the validity of the underlying patent to the California courts in the first instance.
The judgment of the Supreme Court of California is vacated and the case is remanded to that court for further proceedings not inconsistent with this opinion.
It is so ordered.