KENNEDY, Circuit Judge:
The appeal comes from a patent infringement suit where the trial court, within the confines of a summary judgment motion, determined the patent invalid. The court ruled that 35 U.S.C. § 102(b) (1976) operates to invalidate the patent because the invention had been on sale for longer than one year before the date of the patent application. All concede that certain rights to the invention were transferred by an agreement executed more than a year before the patent application. The inventor argues that the transfer was for an experimental purpose and so not within the bar of the cited statute, while the defendant in the infringement suit argues the contract was a sale not limited to that purpose. We agree that summary judgment must be entered for the defendant below based on the cited statute, although on reasoning somewhat different from that of the court below, and we affirm.
Bruce A. Kock, the inventor in this case, was the plaintiff below and is appellant here. Kock undertook a project for Merry Manufacturing Company (Merry). Kock was to be paid one thousand dollars “for the development and building of one working prototype toy watch movement, exclusively for Merry Manufacturing Company.” The toy watch had to be accurate, tick when running, run for one hour per winding, be easily assembled, and be marketable at one dollar retail, allowing for a thirty percent profit margin.
Kock came up with a prototype he thought met these specifications. On April 28,1967, Kock and the Company signed the agreement here in question.1 One clause [652]*652states, “The INVENTOR hereby sells, assigns and transfers all rights to his novelty toy watch movement and any invention embodied therein to the [Merry] COMPANY.” The agreement supplies the basis for the contention of the alleged patent infringers that the patent was void under section 102(b). The district court granted summary judgment for the defendants, reasoning, first, that under Robbins Co. v. Lawrence Mfg. Co., 482 F.2d 426 (9th Cir. 1973), the court may not look beyond the four corners of the written contract in search for an experimental purpose, and second, that the experiments undertaken by Merry were commercial and so not the sort within the experimental purpose exception to the on sale bar of section 102(b).
We begin by exploring the history and rationale of the on sale bar and the experimental purpose exception to it.
35 U.S.C. § 102(b) (1976) states:
.A person shall be entitled to a patent unless . . . (b) the invention was patented or described in a printed publication in this or a foreign country or in public use or on sale in this country, more than one year prior to the date of the application for patent in the United States. (Emphasis added.)
The public use or on sale bar has been part of patent law in some form since 1836. The last amendment to section' 102(b) occurred in 1939, when Congress shortened the grace period from two years to one year. Act of August 5, 1939, ch. 450, § 1, 53 Stat. 1212. See generally 2 D. Chisum, Patents: A Treatise on the Law of Patenta-bility, Validity and Infringement § 6.02 (1979) at 6-5 to 6-16.
The one year grace period gives inventors an incentive to pursue diligently the patent process after their invention is ready for commercial application. This serves to limit the monopoly to the statutory period, Pickering v. Holman, 459 F.2d 403, 406 (9th Cir. 1972), and also encourages inventors to place their handiwork in the public domain as soon as possible so all may benefit from it. Pennock v. Dialogue, 27 U.S. (2 Pet.) 1, 19, 7 L.Ed. 327 (1829); Gould Inc. v. United States, 579 F.2d 571, 580 (Ct.Cl.1978); Atlas v. Eastern Air Lines, Inc., 311 F.2d 156, 159 (1st Cir. 1962), cert. denied, 373 U.S. 904, 83 S.Ct. 1290, 10 L.Ed.2d 199 (1963). See also Note, New Guidelines for Applying the On Sale Bar to Patentability, 24 Stan.L.Rev. 730, 732-736 (1972).
Not all transfers from an inventor to a third party trigger the bar of section 102(b). If a transfer is made for the dominant purpose of experimentation, that is, to perfect the invention, and only incidentally for the profit of the inventor, then there is no public use or sale within the meaning of the statute.
The classic case for the experimental use exception is City of Elizabeth v. American Nicholson Pavement Co., 97 U.S. 126, 24 L.Ed. 1000 (1878):
It is sometimes said that an inventor acquires an undue advantage over the public by delaying to take out a patent, inasmuch as he thereby preserves the monopoly to himself for a longer period than is allowed by the policy of the law; but this cannot be said with justice when a delay is occasioned by a bona fide effort to bring his invention to perfection, or to ascertain whether it will answer the purpose intended . .. [I]t is the interest of the public, as well as himself, that the invention should be perfect and properly tested, before a patent is granted for it. [653]*653Any attempt to use it for a profit, and not by way of experiment, for a longer period than two years [the grace period at that time], would deprive the inventor of his right to a patent.
Id. at 137. Thus, City of Elizabeth held that “the use of an invention by the inventor himself, or of any other person under his direction, by way of experiment, and in order to bring the invention to perfection” was not a public use within section 102(b). Id. at 134.
The experimental purpose exception applies to the sale of an invention under § 102(b), just as it does to its use. Although the Supreme Court has never had occasion to so hold, we think the law is clear on the point. Robbins Co. v. Lawrence Mfg. Co., 482 F.2d 426, 430 (9th Cir. 1973); Red Cross Mfg. Corp. v. Toro Sales Co., 525 F.2d 1135, 1144 (7th Cir. 1975); In re Yarn Processing Patent Validity Litigation, 498 F.2d 271, 277 (5th Cir. 1974), cert. denied, 419 U.S. 1057, 95 S.Ct. 640, 42 L.Ed.2d 654 (1974); see generally Chisum, supra, section 6.02[6], [7].
To insure that abuses section 102(b) is designed to prevent do not arise, the burden of proof that sale is for an experimental purpose is on the inventor. The Supreme Court so indicated in Smith & Griggs Mfg. Co. v. Sprague, 123 U.S. 249, 8 S.Ct. 122, 31 L.Ed. 141 (1887):
In considering the evidence as to the alleged prior use for more than two years of an invention which, if established, will have the effect of invalidating the patent, and where the defense is met only by the allegation that the use was not a public use in the sense of the statute, because it was for the purpose of perfecting an incomplete invention by tests and experiments, the proof, on the part of the patentee, the period covered by the use having been clearly established, should be full, unequivocal, and convincing.
Id. at 264, 8 S.Ct. at 129; see also Omark Industries, Inc. v. Carlton Co., 652 F.2d 783, 787 (9th Cir. 1980); Aerovox Corp. v. Po-lymet Mfg. Corp., 67 F.2d 860, 861 (2d Cir. 1933).
The appellees argue, and the district court agreed, that the testing which occurred in this case, even if it was the purpose for the sale, was “that of a trader, not of an inventor.” They insist that the experimentation required to stop the running of the time period is limited to determining the operability of the patented invention itself. We think that is too narrow a formulation.
In determining whether the transfer fell within the statutory exception, or the closely related question whether the pleadings presented a triable issue of fact on the point, we must examine the distinction between tests and research for experimental purposes and those for commercial purposes. The dichotomy may not be easy to apply since experimentation, for most inventors, is only a means to later commercial exploitation. The precedents allow such a distinction to be stated, however. Where the purpose of a sale is to investigate or stimulate the demand for the product, the object of the transfer is deemed commercial, not experimental, and the exception does not apply. Such a sale is within the bar of section 102(b). Omark Industries, Inc. v. Carlton, 652 F.2d 783, 787-88 (9th Cir. 1980); Kalvar Corp. v. Xidex Corp., 556 F.2d 966, 967-68 (9th Cir. 1977); Cataphote Corp. v. DeSoto Chemical Coatings, Inc., 356 F.2d 24, 27 (9th Cir. 1966), aff’g, 235 F.Supp. 931 (N.D.Cal.1964), cert. denied, 385 U.S. 832, 87 S.Ct. 71, 17 L.Ed.2d 67 (1966); Paeco, Inc. v. Applied Moldings, Inc., 562 F.2d 870, 874-75 (3d Cir. 1977); Dart Industries, Inc. v. E. I. Dupont DeNemours & Co., 489 F.2d 1359, 1366 (7th Cir. 1973).
On the other hand, where the purpose of the sale, as revealed by objective circumstances, is to determine whether the invention can be improved, or reduced to operable, manufacturable, and useful form, a valid experimental purpose exists.
When an invention has been reduced to practice, further public testing and demonstration may well support an inference that the inventor’s intent is to exploit the invention, but “there may be an experimental use even following reduction to practice where the experiments are an attempt to further refine the device.”
[654]*654Del Mar Engineering Laboratories v. Physio-Tronics, 642 F.2d 1167, 1169 (9th Cir. 1981), quoting Cataphote Corp. v. DeSoto Chemical Coatings, Inc., 235 F.Supp. 931, 934 (N.D.Cal.1964). “ ‘[A]n inventor may ... wait until he learns whether his invention is of enough value to justify an application for a patent.’ ” Yarn Processing, 498 F.2d at 282, quoting Aerovox Corp. v. Po-lymet Manufacturing Corp., 67 F.2d 860, 862 (2d Cir. 1933). As the Fifth Circuit held in Yarn Processing, “a public use or sale will be excused if the invention is still undergoing experimentation that is reasonably necessary to test its utility or to determine whether further refinement is needed.” Id. at 285. The Second Circuit is in accord. Cali v. Eastern Airlines, Inc., 442 F.2d 65, 70-71 (2d Cir. 1971). See generally Chisum, Patents Treatise § 6.02[7] at 6-52 to 6-56.
Appellees suggest that the prototype delivered to Merry was identical with the movement later identified in the patent application. The issue, however, is not whether experimentation found that the prototype was sufficiently workable to be patentable without further change, but whether such experimentation took place at all. There are some cases intimating, although none that we have found holding, that the limits of permissible experimentation are defined by the limits of the eventual patent claims. E.g., Application of Theis, 610 F.2d 786 (Cust. & Pat.App.1979); Minnesota Mining and Mfg. Co. v. Kent Industries, 409 F.2d 99 (6th Cir. 1969). In both of those cases, there was commercial purpose so a correct result was achieved, despite a possibly over narrow view of “experimental purpose.” In Minnesota Mining and Mfg. Co., the inventor distributed samples to prospective purchasers and “conducted a substantial sales program.” 409 F.2d at 100. In Theis, there was little, if any, objective evidence of experimental purpose, and the court found “the context in which all of appellant’s activities in the period immediately preceding the critical date exist is clearly that of an attempt at market penetration.” 610 F.2d at 793.
The dictum in these cases is thus at odds with our own view, and with the broad scope of experimentation the Supreme Court allowed in the City of Elizabeth case, supra. To the extent there is a conflict, we adopt what Chisum calls the “better and prevailing view.” Patents Treatise at 6-53. Experimentation may occur even if no change in the patent proves to be necessary. See City of Elizabeth, 97 U.S. at 134-35; Yarn Processing, 498 F.2d at 277. Even if tests are not directly relevant to the claims eventually put forward in the patent, such tests are within the “experimental purpose” exception if their purpose is to determine whether the invention is sufficiently useful to justify a patent or could be improved.
Absent a transfer or use, an inventor may determine when to apply for a patent. An inventor with adequate financing may perfect and refine his invention for a lifetime, if that is his wish. The inventor of more limited means should also be able to test and refine his work before filing for a patent, even if he must sell an interest in the invention to develop a product with which he is fully satisfied, provided always the dominant purpose of the sale is for experiment, and that the transferee may not exercise the right to exploit or use the invention.
As we said in Pickering v. Holman, 459 F.2d 403, 406 (9th Cir. 1972), “the so-called ‘experimental purpose exception,’ [vjiewed in light of the policy underlying the public use provision ... is no exception at all. A bona fide experimental use involves no commercial exploitation.” The same could be said of a sale made to improve or perfect the invention. In such a case, where the profit motive is only incidental to the experimental, the goals of section 102(b) would not be served by applying the on sale bar.2
[655]*655In this case, the affidavits of the inventor in opposition to summary judgment stated that:
When I sold the prototype to Merry, it was for research and study. We did not know at the time that the prototype would be in its final form. It was only after considerable study and work on my part, and research on Merry’s part, that the prototype was adopted as the final form of the toy watch that would be offered for sale to the public.
This would support the view that Kock’s purpose in transferring the invention to Merry was to see if the invention could be improved, or if it was worth patenting in its present form.
An inventor must do more than allege his experimental goals to avoid the statutory bar, however. As we said in Robbins, “an inventor’s testimony of his subjective intent has no probative force against overwhelming evidence to the contrary.” 482 F.2d at 431. Thus, where a prima facie sale has been proven, to come within the exception it is the inventor’s burden to produce clear and objective evidence of experimental purpose as the law defines it. Robbins, id.; Omark Industries, 652 F.2d at 787.
Whether a sale is made for an experimental purpose may turn on many factors. Relevant considerations include the necessity for the inventor to transfer the design to an entity with resources or knowledge superior to the inventor’s, or similar factors tending to show the inventor transferred to obtain development assistance; the need to protect the public safety by limited and cautious experiments, e.g., Del Mar Engineering Laboratories; the stage of completion of the prototype sold; whether reports of the results of the tests or experiments are to be made to the inventor; and whether confidentiality by the buyer is enforced or expected.
It is not, however, enough that an inventor demonstrates by objective evidence that further experiment was necessary and was in fact performed by the transferee. To avoid the on sale bar, the inventor must further show that the transferee lacked authority to use the invention or exploit its commercial value. Where an inventor makes a simple outright sale, with the buyer to manufacture, refine, or sell the invention as the buyer chooses, and with no enforceable obligation on the buyer to hold the invention or design for experimental purpose only, the sale is for profit and, if made before the critical date, will invalidate the patent.
If the invention is out of the inventor’s hands, it is irrelevant what experiments the buyer does or does not do with the invention. Tool Research and Engineering Corp. v. Honcor Corp., 367 F.2d 449, 453 (9th Cir. 1966), cert. denied, 387 U.S. 919, 87 S.Ct. 2032, 18 L.Ed.2d 972 (1967); Dart Industries, Inc. v. E. I. DuPont DeNemours & Co., 489 F.2d 1359, 1366-67 (7th Cir. 1973), cert. denied, 417 U.S. 933, 94 S.Ct. 2645, 41 L.Ed.2d 236 (1974); Cali v. Eastern Airlines, Inc., 442 F.2d 65, 69 (2d Cir. 1971).3 The Supreme Court in the City of Elizabeth case recognized this by requiring that tests made by one other than the inventor be “under his direction” if there is to be an experimental purpose. 97 U.S. at 134. To quote the Court in Egbert v. Lippmann, 104 U.S. 333, 336, 26 L.Ed. 755 (1881), if an inventor “gives or sells [his invention] to another, to be used by the [656]*656donee without limitation or restriction, or injunction of secrecy,” there is public use (or public sale) within the statute.4 The inventor in such a ease has no control over whether commercial exploitation can begin by the buyer immediately.
If the buyer has the authority to use an invention commercially or sell it to others without regard to any duty to experiment further, there is a sale within section 102(b), and the exception for experiment does not apply. It is not enough that the parties do not think it is likely that commercial exploitation will occur. While the limits of permissible experimental purpose are the same in “public use” and “on sale” cases, the relevant question is somewhat different. In “public use” cases, such as Del Mar Engineering Laboratories, the inventor has maintained control, and so the question is whether the use actually made is within the realm of permissible experimentation. In the case of an actual sale, the issue is what use the buyer might make of the invention. If a sale of an invention made before the critical date allows the buyer the possibility of non-experimental use, any patent obtained will be invalid under section 102(b).
For reasons that will be stated shortly, we think the inventor has failed to show all of the elements necessary to invoke the exception here, but our analysis differs from that followed by the court below. While the record is not entirely clear on the point, it appears that the court held that in determining whether experimental purpose was shown, the court is confined to the written agreement of transfer. We agree that the written contract is the beginning point of the analysis, and that considered alone it fails to support the existence of experimental purpose as defined by the standards shown above and indeed points strongly against such a conclusion..
The agreement is an archetype of unfortunate draftsmanship, at least from Merry’s standpoint. It is entitled “Manufacturing Agreement,” and the opening covenant recites a simple sale. The thrust of the entire agreement, arguably, is somewhat broader in that it contemplates that Merry, within ninety days, will determine “whether or not commercial production of a toy watch in accordance with the prototype delivered as specified above is desirable.” If so, then Merry is charged to act with diligence to obtain a patent in the inventor’s name. There is no express recital of any experimental purpose in the contract.
There was no indication on the face of the agreement that Kock would direct the experiments, or even that Merry would make reports of the tests to Kock. The contract’s mention of experiments by Merry thus does not state or clearly imply a permissible experimental purpose.
The inventor here appears to argue that the contract was only an offer to sell for ninety days. If the purpose of the offer is not shown to be experimental by other evidence, an offer is also sufficient to put an invention “on sale.” Robbins, 482 F.2d at 431. See also Johns-Mansville Corp. v. Certain-Teed Corp., 196 USPQ 152 (C.D.Cal.1977) and Yarn Processing, 498 F.2d at 277.
In this case, even if the contract were interpreted so that the buyer had the right to return the invention and rescind the sale if not satisfied with its tests, nevertheless, a transfer for experiment would not necessarily be established. Application of Theis, 610 F.2d 786, 789 (CCPA 1979); Philco Corp. v. Admiral Corp., 199 F.Supp. 797, 816-17 (D.Del.1961).
The contract contains provisions which encourage the inference that the invention was patentable and complete at the time of sale, and hence there is no experimental purpose explaining the transfer. The nine[657]*657ty day limit for deciding whether commercial feasibility exists does not suggest that the operative prototype itself was still subject to experimentation or testing. The contract did not foreclose the possibility of such experimentation, but there is certainly no clear implication that experiment was the dominant purpose for the transfer.
The contract did state that “all improvements and changes in design shall be submitted to the inventor by the company.” Without more, this might have supported an inference of experimentation, but the surrounding contractual language explodes any such argument. “Such submissions to inventor are for his information only, and it is to be understood that the company may make any change in design which in its judgment is desirable.” Clause 10. We are unable, within the boundaries of the written contract, to find any clear implication ■of an experimental purpose, and, if the inquiry were to end here, the inventor would be foreclosed without further explanation.
The district court erred, however, in holding that evidence of experimental purpose must appear on the face of the contract. It is true that in Robbins Co. v. Lawrence Mfg. Co., 482 F.2d 426 (9th Cir. 1973), we held:
[A] sale or an offering for sale precludes any inquiry into the experimental nature of the sale unless the contract of sale or the offering for sale contains an express or clearly implied condition that the sale or offering is made primarily for experimental use.
Id. at 433. The rule is, then, that in an on sale case the purpose to experiment must be express or clearly implied in the original agreement of the parties. The contract which courts are charged to interpret in this respect is not limited to the written agreement, except in the polar case of a document conceded to be completely integrated.5 Robbins did not create some new sort of parole evidence rule which bars proving a transfer for experiment by evidence of conduct after the transfer has occurred.
As with other questions of contractual interpretation, the conduct of the parties and other evidence of events occurring after the transaction may be relevant to show an intent to transfer for an experimental purpose. It would not be sensible to read Robbins as hostile to proof of experimental purpose by objective circumstances relevant to the parties’ intent. If between the parties it is clear that the sale is intended as a step on the way to perfecting the invention, and the transferee’s rights of use and control are appropriately limited, there may be no particular reason to recite such purpose in a written agreement, if indeed a written agreement exists. The court may not ignore the reality of an experimental purpose where it is shown to exist by objective evidence properly admitted to interpret the contract.6
[658]*658In Robbins, moreover, we rested in part on the fact that reports of the experimentation alleged in that case were never in fact made from the buyer to the inventor. Id. at 430 n.2. This court in Cataphote Corp. v. DeSoto Chemical Coatings, Inc., 356 F.2d 24 (9th Cir.), cert. denied, 385 U.S. 832, 87 S.Ct. 71, 17 L.Ed.2d 67 (1966), described the resolution of the question of experimental purpose in an on sale case as depending “principally upon a careful examination of each item of evidence and an evaluation thereof to judge the nature and purpose of the course of conduct of the purported patent holder.” Id. at 26.7 See also Del Mar Engineering Laboratories v. Physio-Tronics, Inc., 642 F.2d 1167, 1169 (9th Cir. 1981).
The various aspects of experimental purpose we have outlined can be shown, therefore, by events after the transfer, but in all cases the inquiry is directed to the purpose and intent of the parties at the time the transfer agreement was made. See, e.g., U.C.C. §§ 1-205, 2-205; Restatement (Second) of Contracts §§ 223, 202(4).
It may be that the inventor here presented sufficient evidence to create a triable issue of fact as to part of the required showing of experimental purpose based on the parties’ subsequent conduct. There was evidence from which the trier of fact might have found the inventor continued to be involved in design after the date of sale. The invention, while concededly a working prototype at the time of sale, might have undergone further refinements before reduction to useful and manufacturable form. The pleadings at least admit the possibility the inventor was unable to determine whether the invention was worth patenting, or could be improved, without use of Merry’s superior testing facilities.
The defect in the appellant’s case, however, is that there is no basis at all for finding that Merry was required to perform further experiments. The agreement must be interpreted to permit Merry to exercise complete control over commercial use and exploitation of the invention, without any obligation to experiment. A limit on the buyer’s full dominion over the invention need not be spelled out in terms in the agreement; it might, for example, be implied where the parties clearly contemplated that the inventor, by participating in permissible experiments, would dictate their scope. The agreement at issue here, in particular clauses 3, 10, and 12, makes clear that the inventor waived all control over the future course of the invention. Nothing that would lead us to a contrary opinion was put forward by the inventor. Mere participation as a volunteer, or as a paid consultant, in experiments decided upon by the buyer and under the direction of the buyer is insufficient to show the existence of permissible experimental purpose. Here, Merry was not limited to experimentation and had the absolute right of immediate commercial exploitation. The defendants were, therefore, entitled to prevail.8 The summary judgment granted them is, accordingly, AFFIRMED.