In converting the factual question of experimental purpose into a matter of law, our court has cut another notch in the removal of patent issues from the trier of fact." The appellate/trial relationship is distorted in the panel’s treatment of'the factual question of whether this use of Mr. Lough’s engine seal assembly was primarily experimental and therefore not a “public use” in terms of § 102(b). The panel’s treatment is squarely in conflict with precedent. Thus I must dissent, respectfully, from the court’s decision not to consider this issue en banc.
The Appellate/Trial Relationship
As discussed by my colleagues in dissent, the issue of experimental use is a question of fact, not a ruling of law. The trial process is the mechanism, in our system of justice, for finding facts, determining credibility, weighing the evidence, balancing the circumstances, and ascertaining the truth. The trier of fact, applying the rules of statute and the experience of precedent to the facts of the particular case, must exercise judgment in light of the probative value of the evidence, informed by the complex of societal and policy considerations that are the foundation of our system of law. The trial court is not a mere compiler of data; it is a tribunal of justice.
The role of the appellate court is to correct errors at trial that prejudiced the outcome, not to make de novo fact-based rulings as if there had been no decision at trial. When the law has been correctly stated and the evidence has been weighed to plausible effect, such that the conclusion is not clearly erroneous (upon bench trial) or is not without substantial supporting evidence (upon jury trial), the appellate court must give appropri*1520ate deference to the judgment of the trier of fact. The Federal Rules and judicial tradition require this appellate role. See Anderson v. City of Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985) (“If the district court’s account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently.”); Fed.R.Civ.P. 52(a).
The trial courts are not divested of their primary role in resolving disputes, whenever appellate review is sought. The issue raised by this en banc petition is not whether we, as the appellate court, have authority to review the law, the facts, and the application of the law to the facts of this case. Of course we do. The issue is whether we have authority to reweigh and reevaluate and rebalance the evidence, with no deference to the procedures and judgment at trial. That is an improper appelláte role. In the resolution of disputes the trial is the “main event,” not a “tryout on the road,” in the words of Wainwright v. Sykes, 433 U.S. 72, 90, 97 S.Ct. 2497, 2508, 53 L.Ed.2d 594 (1977).
Whether the installation of Mr. Lough’s six handmade seal assemblies was primarily for experimental purpose, upon the entirety of the circumstances, was found by the trier of fact. Review of that finding is not de novo, as the panel majority holds. This court’s arrogation of authority to decide this question de novo diminishes the trial process and elevates the appellate role. As a commentator has explained:
The classification of ultimate facts as questions of law amounts to a manipulation of the law-fact doctrine to take questions from the jury or to subject the trial level’s resolution of questions to free appellate review.
Martin B. Louis, Allocating Adjudicative Decision Making Authority Between the Trial and Appellate Levels: A Unified View of the Scope of Review, the Judge/Jury Question, and Procedural Discretion, 64 N.C. L.Rev. 993, 1028 (1986). Whether experimental use is viewed as an historical fact or an ultimate fact, it is not a matter of law. Its removal from the jury was an improper “manipulation of the law-fact, doctrine.”
Public Use and. Experimental Use
The patent statute provides that an inventor is barred from the patent system if the invention has been in public use for more than a year before applying for the patent:
35 U.S.C. § 102 Conditions for patentability; novelty and loss of right to patent
A person shall be entitled to a patent unless—
(b) the invention was ... 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, or____
This provision has been in the patent statute since 1839, varying only in whether the period during which the bar accrued was one or two years'.
The statute states the law. The boundary between law and fact was explained in simple terms by Professor Morris: the law is the general principle, stated in advance, while the fact is a case-specific inquiry based on evidence. Clarence Morris, Law and Fact, 55 Harv. L.Rev. 1303, 1304 (1942). This usage is illustrated in City of Elizabeth v. Pavement Co., 97 U.S. 126, 24 L.Ed. 1000 (1877), wherein the Supreme Court discussed the relationship between experimentation and public use, stated certain additional principles of law, and discussed their application in that case. Thus, the Court stated 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, has never been regarded as such a [public] use.
97 U.S. at 134. This is a principle of law, of general applicability. The Court also stated a generally applicable principle of procedural law, as follows: *152197 U.S. at 133. The “law” is these statements of principles. Applying these principles to the facts in City of Elizabeth the Court found, on examining the circumstances, that the inventor Nicholson’s purpose was primarily experimental. That finding was not a matter of law, but a case-specific finding upon weighing and balancing and evaluating all of the circumstances.
*1520To determine this question, it is necessary to examine the circumstances under which this pavement was put down, and the object and purpose that Nicholson had in view.
*1521Both before and after City of Elizabeth numerous cases have involved questions of experimental and public use. The relevant criteria have been summarized in several opinions of this court. E.g., Baker Oil Tools, Inc. v. Geo Vann, Inc., 828 F.2d 1558, 1564, 4 USPQ2d 1210, 1214 (Fed.Cir.1987); TP Laboratories, Inc. v. Professional Positioners, Inc., 724 F.2d 965, 971, 220 USPQ 577, 582 (Fed.Cir.), cert. denied, 469 U.S. 826, 105 S.Ct. 108, 83 L.Ed.2d 51 (1984). These criteria include the necessity for the testing or evaluation, the extent of the use and by whom, the degree of openness to the public during evaluation, whether the inventor monitored the use and kept a laboratory notebook or other experimental records, whether there was payment for the use, and whether there was a confidentiality agreement. The ultimate question is whether the primary purpose of the activity was experimental. See Allied Colloids, Inc. v. American Cyanamid Co., 64 F.3d 1570, 1576, 1577, 35 USPQ2d 1840, 1844 (Fed.Cir.1995) (“The dis-positive consideration is whether the inventor was in fact testing the invention.”);. TP Laboratories, 724 F.2d at 972, 220 USPQ at 582 (“[I]f a use is experimental, even though not secret, ‘public use’ is negated.”)
The trier of fact must weigh and balance all of the evidence in order to find the primary purpose of the activity. Although scholars of trial/appellate relationships have drawn elegant distinctions in defining fact and law, this determination of experimental purpose is surely not a matter of law. In Armour & Co. v. Wilson & Co., 274 F.2d 143, 155, 124 USPQ 115, 124-25 (7th Cir.1960) the court stated:
We have come to speak of questions of ‘facts,’ ‘primary facts,’ ‘subsidiary facts,’ ‘evidentiary facts,’ ‘ultimate facts,’ ‘physical facts,’ ‘documentary facts,’ ‘oral evidence,’ ‘inferences,’ ‘reasonable inferences,’ ‘findings of fact,’ ‘conclusions,’ ‘conclusions of law,’ ‘questions of fact,’ ‘questions of law,’ ‘mixed questions of law and fact,’ ‘correct criteria of law,’ and so on ad infinitum. The simple answer is that we are all too frequently dealing in semantics, and our choice of words does not always reflect the magic we would prefer to ascribe to them.
Indeed, let us not be misled by semantics. The “magic” of de novo appellate determination of experimental use serves not to reflect a nuaneed definition of law and fact, but to affect trial/appellate authority and, in the case at bar, the role of the jury.
The Case of Lough v. Brunswick Corporation
Mr. Lough, a workman at a Florida marina, observed that a commercial marine engine had a deficient upper seal assembly. He designed an improved seal assembly, worked it on his grandfather’s lathe, and installed it in his own boat. He made five additional prototypes and persuaded five other persons, all personal acquaintances in the boating community, to install these handmade assemblies. Mr. Lough testified that he needed to know whether his seal assembly would hold up during a practical period of marine use, and that he sought more extensive and varied use than in his own boat:
Well, Jack Wherry’s boat constantly sat in the water. My boat, I take — it was on a trailer____
And I explained to Jack what I want to do, I wanted to put a seal in his boat for testing purposes to see if it would last, because his boat sat in the water all the time. And when it comes back to the marina, then I could look at it and I could test — I could, you know, make sure everything in it was just fine.
App. at A000230.
Mr. Lough testified that he wanted to be sure that his seal would last for at least a year, the typical warranty period for boats:
I figured if they lasted a year and got them through the warranty period — the Mercruiser seal wouldn’t even get them hardly through the warranty period on a lot of boats.
*1522So, I figured if this thing here that I made lasted them at least a year, a boat dealership could put these in there and at least tell the man he’s going to fix his boat, you know, and it’s going to last for at least a year.
App. at A000229.
Mr. Lough testified about the free provision of the handmade seal after he persuaded his friends to try it out, App. at A000228-29 (“Q: Did you receive any money or anything for any of these five seals? A:' No, I did not.”), and his eventual conclusion that his seal worked sufficiently better, with respect to the life of the then-used marine engine seals, as to have commercial potential. He testified about when he reached this conclusion, and about his decision to enter into the patent system:
Q: All right. What prompted you to patent this seal?
A: I guess all . the friends at the marina; everybody told me that, you know, you got a good thing there, and why don’t you take it to a patent attorney and have him patent the thing for you so ...
Q: All right. Did you take it to a patent attorney?
A: I did.
App. at A000229-30.
Witnesses at the trial included some of the persons who offered their boats for the evaluation. Thomas Nickla testified about his understanding that this seal assembly was a new creation of Mr. Lough, App. at A000543 (“Q: Did there come a time when Mr. Lough told you that he had a new — he had made a new seal to correct this problem? A: Yeah, there was.”), and that he agreed to its installation in his boat in view of the known deficiency in existing seals:
Q: Did you or he, or the two of you, put a seal in your boat?
A: Right. Both of us did one day.
Q: And you’re talking about now the 23 foot Wellcraft?
A: Right.
Q: All right, sir. What did you understand the purpose of putting that seal in your boat to be? Why did you all do it?
A: The reason I first understood was to fix the problem that MerCruiser had with that seal was — the aluminum was so corroded that the seal wouldn’t stay in the housing.
Q: And why did Steve give you a seal to put in yourboat?
A: Just to try it out.
App. at A000543. Other witnesses testified similarly that there was no payment for the seal, and that they knew it was a handmade prototype whose performance was being evaluated. Mr. Lough testified that after his patent application was filed he started a business to make the seals.
All of this evidence occupied many days of trial, with examination and cross-examination of the witnesses. The jury was correctly instructed to consider all of the circumstances and to determine whether the primary purpose of the use of the seal in this period was experimental, in which event, the jury was instructed, the use was not a public use under the law. The jury found, by special verdict, that none of “the seal assemblies distributed by Mr. Lough were in public use more than one year before the filing of his patent application.”
The panel majority, redetermining the issue de novo, explains the factors that it found persuasive in deciding that the use was not primarily experimental: that one of the boats was sold without retrieving the seal, that Mr. Lough did not keep written records, that he did not inspect the seal assemblies, and that he did not “maintain any supervision” over the séals during the testing. I am not in a position to evaluate and weigh these findin'gs, some of which appear to be directly contrary to unchallenged testimony. Those who were present at the trial were surely better positioned to sort, out the truth, weigh the evidence, and balance all of the circumstances.
The panel majority, in reaching a different balance than did the trier of fact, must have disbelieved the testimony that the purpose was to “try out” the seal assembly, and must have disbelieved the testimony describing the disassembly and inspection of the seals, for the majority explicitly finds that Mr. Lough “fail[ed] to monitor the use of his prototypes,” 86 F.3d at 1122, 39 USPQ2d at 1106, and that he received no comments concern*1523ing their operability. Id. at 1116, 86 F.3d 1113, 39 USPQ2d at 1102. Mr. Lough testified that he took pictures of the seal assembly in the boat of a customer of the marina where he worked, to show that the assembly was working properly:
I was trying to show that there was no— they still had the red grease on the bell crank after about probably six to eight months of testing. And when we pulled the drive off, there still wasn’t any water in it____
James Yow, owner of the marina where Mr. Lough worked, testified about observations of the performance of Mr. Lough’s seal in a tow boat at the marina:
Q: How long was the seal in this particular boat?
A: Approximately four years.
Q: Did you ever have any problems with the seal?
A: No, sir. We pulled the boat out of the water about every three to six months, pulled the drive, greased the U-joints and stuff like that, and checked the seal and there was no water in the cavity.
It is hard to square this evidence with the panel majority’s statements that Mr. Lough did not monitor the performance. The proper appellate analysis is not how a panel of this court would have believed the witnesses and weighed the evidence, but whether the evidence and credibility determinations that the trier of fact deemed probative of experimental purpose were supported by substantial evidence and within the scope of a reasonable jury verdict.
The need to exercise wisdom or judgment in considering the totality of the circumstances does not convert fact into law and remove the issue from the trier of fact. To the contrary, the exercise of wisdom and judgment when weighing and balancing the evidence is the role of the trier of fact. See Best v. District of Columbia, 291 U.S. 411, 415, 54 S.Ct. 487, 489, 78 L.Ed. 882 (1934) (“Where uncertainty arises either from a conflict of testimony or because, the facts being undisputed, fair-minded men may honestly draw different conclusions from them, the question is not one of law but of fact to be settled by the jury.”) While the appellate court has full authority to correct jury verdicts that are not supported by substantial evidence, such correction does not permit ignoring the trial process. When the jury verdict is sustainable on the evidence, the appellate role is fulfilled.
Why Does It Matter?
If the Federal Circuit wants to make de novo determinations of factual matters, isn’t this a good thing? Won’t it add consistency to patent decisions by removing harried trial judges and sympathetic juries from another area of patent litigation? Is it not a force for stability if the Federal Circuit will decide all fact-driven questions de novo?
Indeed, when there are fewer decision-makers, consistency is more likely. However, the appellate court is not structured to find facts without recourse to the procedures of trial. On issues that require weighing and balancing several confluent fact-laden circumstances, how can we presume to reach a more just result than those who were present for days or weeks of testimony, with demonstrations, exhibits, cross-examination, and argument? The procedures of trial and the deference owed to the judgment grounded in these procedures are the fabric of our system of justice and the trial/appellate relationship.
Although appellate courts have on occasion departed from deferential review of findings of fact, such departures are rare and invoke important policy issues, as in questions of constitutional fact, e.g., Bose Corp. v. Consumers Union of United States, Inc., 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984) (the issue of product disparagement required interpretation of the First Amendment) and most recently in the policy considerations surrounding patent claim construction, established in Markman v. Westview Instruments, Inc., — U.S. -,-, 116 S.Ct. 1384, 1396, 134 L.Ed.2d 577, 38 USPQ2d 1461, 1470 (1996) (“[W]e see the importance of uniformity in the treatment of a given patent as an independent reason to allocate all issues of construction to the court.”)
However, I doubt that the finding of experimental purpose in particular factual circumstances raises sufficiently grand portents *1524to warrant further disruption of the trial/appellate relationship. Even if the removal of patent issues from the mainstream processes in which tradition (and wisdom) have placed them is the direction of the future, this piecemeal attrition of district court authority is pernicious. As Professor Louis remarked, swpra at 1018, appellate action to treat fact as law “amounts to a direct judicial assault on the prerogatives of fact finders.” If the majority of this court believes that the special niche of de novo review is nonetheless the correct place for the factual question of experimental use, then the court should speak with one voice so that trial courts and parties will know how to try and appeal the issue. Thus I must, respectfully, dissent from the court’s refusal to take this issue en banc in order to establish a position on experimental use by which we will all be bound.
PLAGER, Circuit Judge, with whom RADER, Circuit Judge joins, dissenting from the Order declining the suggestion for rehearing in banc.In my dissent to the original panel opinion, I stated my view as to why the case was wrongly decided. I write now to state my reasons why, under the standards governing in banc review by this court, see Federal Circuit Rule 35(a), I would take this case in banc as appellee has suggested and clarify the legal standard by which we review cases such as this. In the course of doing so, we would also correct a gross miscarriage of justice.
An inventor is free to dedicate his invention to the public, in which ease the Patent Act denies him the right subsequently to obtain a patent. There are several things an inventor may do that will have that effect; these are spelled out in section 102 of the Act. Included is to have “the invention ... in public use ... in this country, more than one year prior to the date of application for patent.” 35 U.S.C. § 102(b) (1994).
The statute does not describe the circumstances under which an invention is to be considered “in public use.” Presumably the distinction is between a public and a private use. Courts recognize that private use, strictly defined, is not always possible because a certain amount of testing might be needed in order to perfect the invention, and such testing might have to be done in a setting which, in a strict sense, was public, that is, open to the public, or involving persons other than the inventor. The classic case is City of Elizabeth v. American Nicholson Pavement Co., 97 U.S. 126, 24 L.Ed. 1000 (1877), in which the Supreme Court found a new type of pavement material used on a public road for six years not to be “in public use” because the inventor’s purpose was to test the material under actual use conditions.
City of Elizabeth and cases like it establish that when an accused infringer defends by claiming patent invalidity based on a § 102(b) public use bar, the patentee may counter with what has become known as the “experimental use” counter-defense. If the use of the invention is experimental, then the fact that the use resulted in public exposure does not invalidate the patent even if the exposure occurred more than one year before application for patent was made.
Whether an inventor intended to dedicate his invention to .the public, or whether he was simply experimenting in order to perfect the invention before seeking his patent, is a question the answer to which is entirely fact driven. It involves intent, purpose, behavior, relationships actual and perceived, timing, the nature of the invention and its stage of development, the .environment in which the testing was done, the experience and training of the inventor, and perhaps even the extent and sophistication of advice readily available to the inventor as. the invention matures.
When presented after-the-fact to a fact finder, the reconstruction of these historical events is the typical fodder of a trial, whether it be a jury trial or a bench trial. The subjective nature of the reconstruction process gives rise to notions like the virtual unreviewability of the fact finder’s credibility determinations regarding the witnesses who testify, and the deference given to the factual findings and conclusions reached by the fact finder. If the fact finder is a jury, on appeal we ask whether a reasonable jury, on the evidence before it, could have arrived at the conclusion it did. If the matter is tried to a *1525judge in a bench trial, we ask not whether as judges we all agree, but whether the trial judge’s findings and conclusions were clearly erroneous.
It is sometimes suggested that there is a legally-significant difference between what are called “underlying facts,” on the one hand, and other facts, including factual inferences and fact-based conclusions that presumably follow from the “underlying facts.” If a fact falls in the first category, it is said, the fact findings are entitled to deference on appeal; if not, the fact findings get no such deference. This artificial bifurcation when dealing with what is a singular, fact-intensive, issue is I believe misplaced and misunderstood. The question of whether an inventor’s activities placed the invention “in public use” cannot be separated from the factual underpinnings on which the answer stands. The factual conclusion — was he experimenting or was he intending to dedicate his invention to the public — does not derive from some independent basis of judgment or knowledge, but simply from the total experience presented to the fact finder by the evidence in any given case.
Thus the concept of “totality of the circumstances” makes sense as a description of what fact finders wrestle with, and out of which emerges the “truth” of what happened and why. However, when invoked as an appellate standard of review, “totality of the circumstances” becomes simply a thin veil behind which appellate judges substitute their view of the facts for that of the fact finder. As such it has the propensity to be arbitrary, capricious, and an invitation to standardless decision making, the sort of thing that appellate courts are supposed to correct, not employ.
I fully support the rationale and result in Markman v. Westview Instruments, Inc., 52 F.3d 967, 34 USPQ2d 1321 (Fed.Cir.1995) (in bane), aff'd, — U.S. -, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996), in which, in the interest of an efficient and fair judicial system, we made a careful distinction between the role of judge and jury in the decisional process. However, our decision that the process of claim construction does not raise any factual issues appropriate for jury participation does not necessarily mean that other issues in patent law should be similarly treated. It depends on the nature of the issue. What is needed is functional analysis rather than a priori labeling , as “legal issue” or “factual issue.”
When the issue is reconstruction of historical events, the jury has long been deemed an appropriate fact finder. That the “public use” standard is found in statute is of no moment; it is not the source of the standard but its content that matters. Whether a use is experimental, and thus not a public use under the statute, is unlike the question of the meaning of a patent claim. In claim construction, certain “facts” may be relevant — what might one skilled in the art have thought the claim language meant, and what precisely was the prior art against which this invention was played — but the ultimate judgment of the meaning of the claim derives from the skill and specialized knowledge of persons trained to make those judgments. The announced conclusion reached incorporates learning and, not infrequently, technological grasp and policy sensitivity, which are aided by distance from the immediacy of the trial stage.
In the issue before us, whether Mr. Lough’s exposing of his invention to others was for the purpose of experimenting to perfect it or not, and thus in public use or not, the jury under proper instructions decided the .fact-intensive question in his favor; the trial judge who was present during the proceedings approved what the jury did; and on appeal the record contains more than enough evidence by which we can conclude that a reasonable jury could have arrived at the conclusion this one did. For me, that is the end of the case.
Mr. Lough was not' engaged in market testing or other commercial activity that, though perhaps falling short of the related on-sale bar, might be cause for concern. Nor did the panel offer a new, specific test for experimental use, such as requiring written confidentiality agreements, detailed reports contained in lab notes, or some other such requirement that Mr. Lough is deemed *1526to have violated.1 Absent that, the decision by the panel majority to deny the jury any deference and thus arrogate to itself the fact finding has not only caused a gross miscarriage of justice to Mr. Lough, but profoundly contributes to the confusion and misunderstanding that had already begun to emerge around this question of what is, and who decides, experimental use.
As attested to by the several accompanying opinions also dissenting from the denial of in banc action, this court bears some responsibility for the state of affairs reflected by the decision in this case. I would take the case in banc for the purpose of clearly and unequivocally stating the rule that accords the fact finder proper deference when the issue is whether the experimental use counter-defense has been successfully established as a fact in face of an invalidity attack based on an alleged public use under section 102. Even under such a deferential standard, this court can effectively police the behavior of fact finders to ensure that the appropriate criteria, such as they are or may become, are properly applied. The outcomes in these cases may vary, but that ordinarily will be a natural consequence of the different facts in different cases, and should not be a matter of concern. Even if it were, other than in the exceptional case it is not a concern that an appellate court is equipped to address effectively. Our job is to provide uniformity of rule and process, not to impose our view of the facts on the constituted fact finder under the mistaken notion that an appellate court routinely can prdvide on a cold record some better understanding of what actually occurred.
MICHEL, Circuit Judge,dissenting from the Order declining the suggestion for rehearing in banc.
I would grant the suggestion for, rehearing in banc for three reasons not fully explicated in any of the separate opinions of my colleagues: (1) our ease law regarding the application of the public use bar is too vague, and outcomes are too unpredictable, (2) dispute accordingly arises regarding whether public use and the reciprocal issue of experimental use are questions of fact or of law, and (3) dispute likewise arises regarding the proper standard of review by both the trial judge after a jury verdict and by this court. Unlike most of the other dissenting judges, however, I do not agree that public use should be reclassified as an issue of fact. Therefore, I write separately to explain that I would take the ease in banc to reaffirm our classification of public use as ultimately an issue of law and also to identify certain indicia of “control,” e.g., record keeping, secrecy agreements, testing protocols, supervision of testing, reports to the inventor or restricted access to others, proof of which should be required before any potentially barring use in public can, on grounds of experimentation, avoid the public use bar. Our declaring such a requirement in banc would make the potential for invalidation of particular patents more predictable.
I.
Certain of our cases describe the public use bar as turning on the “totality of the circumstances.” E.g., Manville Sales Corp. v. Paramount Sys., Inc., 917 F.2d 544, 549, 16 USPQ2d 1587, 1591 (Fed.Cir.1990). So labeling the test may suggest that all relevant circumstances are merely factors to be weighed and that no fact can definitively determine whether a patent is invalid because of public use of the invention. Thus, lawyers and business people can . only guess about the invalidity of a particular patent until a panel of our court has ruled. However, to say that many factors may be relevant is not to say that none can ever be controlling. For example, I believe existing case law, when properly read, suggests that control is such a requirement, i.e., a use cannot be experimental if the inventor failed to maintain sufficient control over the invention and its testing. Indeed, that is what the panel seems to hold. See Lough v. Brunswick Corp., 86 F.3d 1113, 1121 (Fed.Cir.1996) (“Thus, Lough did not maintain any supervision and control over the seals during the alleged testing.”). Requiring indicia of control guarantees both that the inventor was actually experimenting in an attempt to ac*1527quire information about whether the invention would work for its intended purpose and that the invention has been reasonably protected from disclosure to the relevant public so that competitors could not reasonably conclude it has been dedicated to the public. Absent such control, there can be no realistic conclusion that the inventor was truly experimenting or that the competitors and the public were not mislead. Unfortunately, control as a prerequisite to experimentation may not seem absolutely required by the panel opinion and, in any event, the panel opinion lacks the full weight of an in banc court behind it. Therefore, I would rehear the appeal in banc.
While requiring control would improve predictability, redesignating public use as an issue of fact would not. My dissenting colleagues who favor so reclassifying the issue would increase unpredictability by, granting greater deference to jury verdicts involving a conclusion of public use. Different juries could render different verdicts on the public use bar in the face of similar facts, and what a given jury would decide on particular facts could not be foretold. Consequently, reclassification would also encourage more litigation. Patent invalidity on public use grounds could not be determined on examination of known facts by mere lawyers, no matter how skillful, but would require the full time and expense of a jury trial, followed by post-trial motions and appeal, whenever the patentee seeks to assert his right to exclude others, such as through warning letters or licensing negotiations. Although de novo judicial re-determination of invalidity may be a burdensome and exacting task, I think it is better for the trial bench and our court to perform this task than to leave undisturbed an array of unpredictable and inconsistent jury verdicts.
Moreover, reclassifying the public use issue as one of fact might create confusion throughout our invalidity jurisprudence. If we so reclassify public use, would we not, in order to be jurisprudentially consistent, likewise reclassify other legal invalidity grounds that we have said comprise both legal and factual issues, such as obviousness?
II.
In any event, I consider reclassifying the public use bar as an issue of fact to have been foreclosed by the Supreme Court’s decision long ago’ in City of Elizabeth v. American Nicholson Pavement Co., 97 U.S. 126, 24 L.Ed. 1000 (1877). There, the Court distinguished between “use in public,” judicially reviewable as fact, id. at 134 (“That the use of the pavement in question was public in one sense cannot be disputed.”), and “public use,” the statutory standard, reviewable as law, id. at 136 (“The public had the incidental use of the pavement, it is true; but was the invention in public use, within the meaning of the statute?”) (emphasis added). Our court since its inception has drawn the same distinction. While, if we conclude the issue is not foreclosed by City of Elizabeth, we aré free to reverse our view, in a stare decisis system a court in banc should only upset long settled law in the face of a compelling change of circumstances. None is presented here.
III.
If our classification of public use as a question of law with underlying factual issues is to operate more successfully, as I believe it can, the court in banc must articulate how we and trial judges are to review jury verdicts on this and, by implication, other legal grounds of invalidity. To say only that such grounds of invalidity are legal conclusions based on underlying facts is not to enlighten. Nor is it sufficient to say of the “underlying facts” that those that are historical warrant deferential review. Another important category of facts might be termed “legal facts.” An example of such a legal fact is the question of control when considering asserted experimental use. I believe that on legal issues of invalidity, these “legal facts,” too, deserve deferential review by judges. Indeed, in my view, only the ultimate legal conclusion of invalidity ever warrants free review by the trial judge following jury verdicts and by us on appeal. Accordingly, I believe judges at both levels, as well as patent lawyers and business people who could then make better predictions about patent validity, would all benefit if we were clearer about how judges are to decide these issues.
*1528I believe the most difficult challenge is to distinguish between those cases in which the facts regarding public use, as duly found, could only support one result, and those in which the facts as duly found could support either result. In the latter class of cases, judges must redecide the issue based on the underlying facts ás properly found by the jury, while in the former, I believe judges must leave undisturbed the jury’s resolution of the ultimate issue as well as the underlying facts, so long as the underlying facts are supported by substantial evidence. Such a standard of review, although variable and hence complex, would not risk encroachment on the fact-finding, evidence-weighing function- of the jury. Rather, it would guarantee deference to the jury’s verdict in those eases where such deference is due, i.e., where the facts dictate the result, while still allowing appropriate judicial redetermination in circumstances that merit it, i.e., where the facts as found do not dictate the result. Although evaluating the jury’s fact-finding is often difficult in the face of a general verdict, the jury’s proper role can be further protected by encouraging the use of special verdict forms or written interrogatories, see Fed. R.Civ.P. 49, thus leaving little doubt as to the jury’s findings.
I had hoped that in this appeal our court in banc might have settled these questions regarding judicial review of jury verdicts on the public use bar, if not legal invalidity issues generieally. That occasion is not' to arise in the present appeal. I hope our court will soon be presented with such occasion in another appéal. Everyone affected by the patent system stands to gain.
RADER, Circuit Judge,dissenting from Order declining the suggestion for rehearing in bane.
In Lough v. Brunswick Corp., 86 F.3d 1113, 39 USPQ2d 1100 (Fed.Cir.1996), a split panel of this court pronounced public use under 35 U.S.C. § 102(b) a question of law. Id. 86 F.3d at 1120. Under this standard of review — a departure from Supreme Court and Federal Circuit law — the Lough court gave no deference to the trial jury’s findings. Because I believe the Lough court lacks authority for this departure and complicates a quagmire of already confusing law on statutory bars, I must respectfully dissent from this court’s decision to decline in banc rehearing.
No subject matter within the jurisdiction of this court of appeals prompts more outright scholarly criticism than this court’s review of section 102(b) bars.2 Moreover, if this court presumes to extend plenary appellate review to the vague totality of circumstances test, this single sentence in title 35 can only produce more conflicting judgments. The unrestricted review of a wide-ranging “totality of circumstances” creates ample latitude for inconsistency. By inexplicably designating both public use and experimental use as questions of law, Lough threatens to multiply the chaos. Until the Federal Circuit restores the emphasis on factual inquiry set forth in the Supreme Court’s standard of review, certainty and predictability — not to mention unwary inventors — will be the victims.
I.
Steven G. Lough is the most recent victim of the uncertainties of public use law. He won a unanimous jury verdict after an exhaustive seventeen-day trial. After a lengthy consideration of evidence, the jury decided that Lough did not publicly use his invention. Moreover, the trial judge who presided over the trial sustained the jury’s verdict.
Lough invented an upper seal assembly to solve a corrosion problem in inboard/outboard motor boats — a solution that had escaped Brunswick for many years. After some trial and error with his grandfather’s metal lathe, Lough made six usable proto*1529types in the spring of 1986. He installed one prototype in his own boat. He placed a second prototype with a friend who installed it in his boat. He also installed prototypes in the tow boat at the marina where he worked and in the boat of a marina customer — both of which Lough regularly maintained. He gave the remaining prototypes to a friend who was employed at another marina in Sarasota, Florida. At no point did Lough charge for these prototypes or try to profit from his invention. Instead, Lough testified of his intent to perfect his invention through experimentation. He relied on his close relationship and association with the owners of the experimental boats to maintain the confidentiality of his invention and to monitor the performance of his prototypes.
On June 6, 1988, Lough filed a patent application entitled “Liquid Seal for Marine Stern Drive Gear Shift Shafts,” which issued as U.S. Patent No. 4,848,775 on July 18, 1989. Lough sued Brunswick on June 12, 1993, for infringement of the ’775 patent. The jury found no public use before the critical date of June 6, 1987.
On appeal to this court, Brunswick argued that Lough’s prototypes were not experimental, but public uses. Lough relied on his evidence at trial that he had experimented with the prototypes to complete his invention. He stressed his refusal to seek any commercial gain from the prototypes. In fact, he did not place the seal assemblies on sale until after he filed his patent application. He also referred to his selection of a few friends and personal acquaintances to check the performance of the invention under different conditions.
This court, however, reversed the decision of the jury and the trial judge. Over a vigorous dissent, this court reweighed the record evidence on its own and, for the first time on appeal, found that Lough’s six prototypes were a public use.
II.
The outcome of Lough turns on the standard of review applied by this court. With any level of proper deference to the fact-finding of those who heard seventeen days of evidence, this court must have affirmed. Both logic and Supreme Court rules call for that deference. Logic calls for deference because this court used the intensely factual “totality of circumstances” analysis. A factual analysis logically calls for the deferential review accorded questions of fact. Respect for the Supreme Court also calls for deference because our nation’s highest court has repeatedly instructed that public use is predominantly a question of fact. See, e.g., Pennock v. Dialogue, 27 U.S. (2 Pet.) 1, 16, 7 L.Ed. 327 (1829) and cases cited infra pp. 5-6.
Notwithstanding logic and binding precedent, the Lough court ventured to designate public use a question of law and extended that error by announcing that experimentation is also subject to unfettered review.3 Logie sometimes escapes even this distinguished court, but disregard of binding precedent, including the prior jurisprudence of this court and its predecessors, invites irreconcilable conflict and confusion.
Turning first to the standard of review for the public use and on sale bars of section 102(b), the Supreme Court has already settled this1 question. The Supreme Court’s uniform treatment of statutory bars as predominantly questions of fact began when Justice Story articulated the justification for statutory bars in Pennock v. Dialogue. Justice Story, a consummate trial judge as well as Justice, took care to state from the outset the duties of judge and jury on statutory bar issues:
The single question then is, whether the charge of the court was correct in point of law. It has not been, and indeed cannot be denied, that an inventor may abandon his invention, and surrender or dedicate it to the public____ The question which generally arises at trials, is a question of fact, rather than of law; whether the acts or *1530acquiescence of the party furnish, in the given case, satisfactory proof of an abandonment, or dedication of the invention to the public. But when all the facts are given, there does not seem any reason why the court may not state the legal conclusion deducible from them. In this view of the matter, the only question would be, whether, upon general principles, the facts stated by the court would justify the. conclusion.
Id. 27 U.S. (2 Pet.) at 16 (emphasis added). This rule could not be clearer. The public use question which generally arises at trial is a question of fact, not of law.
Any provision of law invariably involves some interpretation of the legal meaning of the statutory terms. When a statute uses straightforward words, such as those in 35 U.S.C. section 102(b), however, that legal interpretation is minimal. Indeed with section 102(b), the definition of the statutory language occurred long ago at. the Supreme Court. See, e.g., Egbert v. Lippmann, 104 U.S. 333, 26 L.Ed. 755 (1881) (a use need not be open and visible to be “public”). Throughout its lengthy experience with the public use and on sale bars, the Supreme Court has emphasized that these questions are overridingly factual. E.g., Gayler v. Wilder, 51 U.S. (10 How.) 477, 484, 13 L.Ed. 504 (1850) (“unless the plaintiff ... had abandoned said improvement to the public, and suffered the same to go into public use before the application for said patent, of which facts the jurors were the judges ”) (emphasis added); Kendall v. Winsor, 62 U.S. (21 How.) 322, 331, 16 L.Ed. 165 (1858) (“The real interest of an inventor with respect to an assertion or surrender of his rights ... is aw inquiry or conclusion of fact, and peculiarly within the province of the jury____”) (emphasis added); Consolidated Fruit-Jar Co. v. Wright, 94 U.S. 92, 94, 24 L.Ed. 68 (1876) (“The result must always depend upon the purpose and incidents accompanying the act or acts relied upon [as public uses].”); City of Elizabeth v. Pavement Co., 97 U.S. 126, 133, 24 L.Ed. 1000 (1877); Manning v. Cape Ann Isinglass & Glue Co., 108 U.S. 462, 465, 2 S.Ct. 860, 863, 27 L.Ed. 793 (1883) (“The decided weight of the evidence shows that there was also a public use of the invention____ These conclusions of fact are fatal to the complainant’s case.”); Cantrell v. Wallick, 117 U.S. 689, 696, 6 S.Ct. 970, 974, 29 L.Ed. 1017 (1886) (“The proof of prior use in this case depends on the testimony.”); Smith & Griggs Mfg. v. Sprague, 123 U.S. 249, 8 S.Ct. 122, 31 L.Ed. 141 (1887); International Tooth-Crown Co. v. Gaylord, 140 U.S. 55, 64, 11 S.Ct. 716, 720, 35 L.Ed. 347 (1891) (“[i]n the light of this testimony we are compelled to hold”); General Talking Pictures Corp. v. Western Elec. Co., 304 U.S. 175, 182-83, 58 S.Ct. 849, 853, 82 L.Ed. 1273 (1938) (“The district court and circuit court of appeals found that there was no public use____ These findings were made upon adequate evidence and petitioner’s contentions as to them will not be considered here.”). In resolving statutory bar questions, the Supreme Court unquestionably stressed time and again that appellate courts must defer to the factual findings of trial courts.
On at least ten occasions, the Supreme Court clarified that factual inquiry drives section 102(b) issues. Moreover, the Supreme Court repeatedly stressed its deference to the fundamental questions of fact in reviewing statutory bars. At no point does the Supreme Court refer to statutory bars solely as legal issues. The Supreme Court’s clear and unwavering treatment of the statutory bars calls into question this court’s departure.
This court’s predecessor courts both recognized and followed the Supreme Court’s rule. The United States Court of Customs and Patent Appeals stated expressly that section 102(b) is. an issue of fact in In re Reese:
The board’s finding that the tablets defined in the claims of the instant application were “in public use for twenty years prior to the filing date of the application,” and were therefore not patentable under 35 U.S.C. § 102(b), raises an issue of fact.
48 CCPA 1015, 290 F.2d 839, 841, 129 USPQ 402, 403 (1961) (emphasis added). In re Reese, an opinion binding on this court, see South Corp. v. United States, 690 F.2d 1368, 1370, 215 USPQ 657, 657-58 (Fed.Cir.1982) (in banc), also followed the Supreme Court’s rule.
*1531Even if the Supreme Court and the Court of Customs and Patent Appeals had not spoken, this court would have been bound by the United States Court of Claims precedent. The Court of Claims followed the Supreme Court rule in Smith v. United States: “The main issue is validity of the patent in suit over prior domestic and foreign patents, and over a prior public use. This issue is factual.” 136 Ct.Cl. 487, 145 F.Supp. 396, 396, 111 USPQ 135, 135 (1956) (emphasis added). Based on the principle of respect for binding precedent, this court should recognize that it lacks authority to fashion a plenary standard of review for section 102(b) bar issues.4
The clear rule of the Supreme Court and this court’s predecessors raises the obvious question: How could this court have departed from binding precedent? The first departure occurred in Barmag Barmer Maschinenfabrik v. Murata Machinery, Ltd., 731 F.2d 831, 221 USPQ 561, 565-66 (Fed.Cir. 1984) (the on sale bar is a question of law). Notably Barmag did not cite to Supreme Court opinions which uniformly treated the question on factual terms. Rather, Barmag stated:
[A section 102(b) bar is not] itself a fact nor an inference of fact drawn from established facts. In re Corcoran, 640 F.2d at 1333, 208 USPQ at 869 (one-year time bar applied under “principles of law”). What Barmag is actually contesting is the court’s conclusion on ... [this] legal issue[ ].
731 F.2d at 837 (citation omitted). With these few words, Barmag fashioned an unrestricted standard of review for statutory bars.
Even a passing analysis calls the reasoning of the Barmag court into question. Barmag’s reliance on Corcoran is bizarrely out of context. See In re Corcoran, 640 F.2d 1331, 208 USPQ 867 (CCPA 1981). Corcoran says nothing about' section 102(b) issues as questions of law. Corcoran instead refers to resolving a section 102(b)/103 rejection under “the principles of law enunciated in In re Foster.” 640 F.2d at 1333 (emphasis added). Surprisingly, In re Foster does not even hint that section 102(b) is a question of law. See In re Foster, 52 CCPA 1808, 343 F.2d 980, 145 USPQ 166 (1965). Statutory bars were not an issue in Foster. Neither Corcoran nor the cases upon which it relied addressed statutory bars. Corcoran simply did not and could not supply authority to depart from binding Supreme Court precedent.
In sum, Barmag, the first opinion of this court to stray from the Supreme Court’s standard of review, was wrong in at least four ways. First, it erred by citing Corcoran for a principle that that case did not even address. Second, it erred by taking the Corcoran phrase “principles of law” drastically out of context when that phrase had no reference at all to standards of review. Third, it erred by ignoring the Supreme Court rule. Fourth, it erred by ignoring binding precedent from both the Court of Customs and Patent Appeals and the Court of Claims.
Barmag dealt with on sale bars under section 102(b). Until Lough, this court had remained more faithful to the Supreme Court’s rule for review of public uses under section 102(b). In Hycor Corp. v. Schlueter Co., 740 F.2d 1529, 1535, 222 USPQ 553, 559 *1532(Fed.Cir.1984), this court stated, that facts underlying public use are subject to clearly erroneous review. Hycor had the virtue of keeping the focus properly on factual inquiry. Moleculon Research Corp. v. CBS, Inc., 793 F.2d 1261, 1265, 229 USPQ 805, 808 (Fed.Cir. 1986), next stated that public use is a question of law, with underlying facts subject to clearly erroneous review. Still the focus remained on factual inquiry. This appellate court did not presume to reassess the facts without deference to the fact-finders.
The court in Lough, however, purported to abandon factual inquiry: “Whether an invention was in public use prior to the critical date within the meaning of section 102(b) is a question of law.” 86 F.3d at 1120. For that proposition, the Lough court cites Manville Sales Corp. v, Paramount Systems, Inc., 917 F.2d 544, 549, 16 USPQ2d 1587, 1591 (Fed. Cir.1990). Manville, however, does not stand for the proposition for which Lough cites it:
Whether or not an invention was on sale or in public use within the meaning of section 102(b) is a question of law that this court reviews de novo; however, factual findings underlying the trial court’s conclusion are subject to the clearly erroneous standard of review. See, Envirotech, 904 F.2d at 1574, 15 USPQ2d at 1232; Moleculon, 793 F.2d at 1266, 229 USPQ at 808.
917 F.2d at 549 (emphasis added). The Lough court follows the first part of the above sentence in Manville almost verbatim. The Lough court, without explanation, refuses to quote or follow the rest of the sentence in Manville. In its entirety, the sentence from Manville does not make public use a question of law, but instead honors the predominant factual inquiry standard.5 Thus, Lough departs from the deference due trial court findings without any legal basis.
The Lough court’s error is painfully reminiscent of the Barmag error. First, it used an incomplete Manville citation to arrive at an inaccurate characterization of the standard of review governing public uses. Second, it erred by declining to honor Supreme Court and Court of Custom and Patent Appeals precedent. Third, it erred by declining to honor Federal Circuit precedent that had uniformly recognized the clear error standard within the public use context. According to normal rules of precedence, Lough lacked authority for its bold change of policy.
Turning next to the standard of review for experimental use,6 Lough again replicates the Barmag error. Once again, logic and binding precedent make experimentation to negate public use a fact-driven inquiry, subject to the deferential review which the Lough court refused to accord this jury verdict.
Again logic dictates that a “totality of circumstances” analysis depends on factual distinctions. With factual considerations dictating the outcome of the underlying analysis, logic relegates the standard of review to the factual realm as well. Thus, the following statement of the Lough court creates a strange illogical juxtaposition: “To determine whether a use is ‘experimental,’ a question of law, the totality of circumstances must be considered____”7 Lough, 86 F.3d at 1120. Frankly the proposition that experimental use — a judge-made doctrine without any express basis in the Patent Act — is a question *1533of law is absurd on its face. It is hard to imagine how case law decisions made by judges to resolve unique factual cases on a fact-driven issue could create questions of law.
To make matters worse, Lough’s assertion that experimental use is a question of law hides as a brief subordinate clause in a sentence of no less than eighty-five words. Id. The citations at the end of this loaded sentence do not support at all the assertion that experimental use is a question of law. In fact, each of the three eases cited by the Lough court at the end of its paragraph-sentence stand for the proposal that experimental use is a factual inquiry. Baker Oil Tools, Inc. v. Geo Vann, Inc., 828 F.2d 1558, 1565, 4 USPQ2d 1210, 1215 (Fed.Cir.1987) (“The disputed factual issue of experimental purpose was material to the decision.”) (emphasis added); Sinskey v. Pharmacia Ophthalmics, Inc., 982 F.2d 494, 498, 25 USPQ2d 1290, 1294 (Fed.Cir.1992) (insufficient evidence was submitted to create a genuine issue of material fact as to party’s experimental purpose); TP Lab., 724 F.2d at 972.
The Supreme Court first acknowledged, experimental use, a concept without an analog in the patent system of any other country, in City of Elizabeth, 97 U.S. at 135. Setting aside the question of whether the law today would allow an inventor to experiment in full public view for over six years with no alterations to the claimed invention over that period,8 the Supreme Court acknowledged that an inventor with a “bona fide intent of testing” would not encounter a public use bar. Id. In excluding Nicholson’s experiment from public use, the Supreme Court “examine[d] the circumstances under which this pavement was put down, and the object and purpose that Nicholson had in view.” Id. at 133. In other words, the Supreme Court set forth a clear factual test. Questions depending on the “bona fide intent” of a party hardly could be otherwise than factual. Indeed, each time the Supreme Court encountered an experimental use, it deferred to. factual findings. See, e.g., Pennock, 27 U.S: (2 Pet.) at 16; Root v. Third Ave. R.R. Co,, 146 U.S. 210, 225, 13 S.Ct. 100, 104, 36 L.Ed. 946 (1892); Electric Storage Battery Co. v. Shimadzu, 307 U.S. 5, 20, 59 S.Ct. 675, 683-84, 83 L.Ed. 1071 (1939). Until Lough, this circuit had consistently followed this Supreme Court standard. See, e.g., Baker Oil Tools, 828 F.2d at 1565; Sinskey, 982 F.2d at 498.
The Lough court disregarded the findings of the jury and trial court and forced Lough to re-prove (with all the strictures of a brief appellate record) his experimental use. Thus, the Lough court defied logic and binding precedent to make experimental use a question of law. The casualties of this error, in addition to Mr. Lough, are the certainty and predictability of statutory bar law.
III.
A significant part of the Lough court’s error is abstracting experimentation out of the public use inquiry and designating it a separate “question of law.” Lough, 86 F.3d at 1120. This incorrect methodology flies in the face of the teaching in TP Lab.: “[I]t is incorrect to impose on the patent owner, as the trial court in this case did, the burden of proving that a ‘public use’ was ‘experimental.’ These are not two separable issues. It is incorrect to ask: ‘Was it a public use?’ and then, Was it experimental?’ ' Rather, the court is faced with a single issue: Was it a public use under section 102(b)?” 724 F.2d at 971. This court in TP Lab. taught that experimental use is not an exception to public use, but a negation. In other words, experimental use is not a separate exception with its own standard of review or its own multi-factored test. Instead an experimental *1534use by definition does not abandon the invention to the public.
Disregarding TP Lab., the Lough court addressed itself to a separate doctrine of experimental use, for which it created a new standard of review separate from public use. The jury and the trial court had correctly answered the correct question: whether Lough gave his invention into public use. The Lough court transformed the question: “The question framed on appeal is whether Lough’s alleged experiments lacked enough of these required indicia so that his efforts cannot, as a matter of law, be recognized as experimental.” 86 F.3d at 1121. To reach its incorrect result, the Lough court had to separate experimental use from its proper context of public use, then create a separate list of required elements for a factual inquiry, and finally make the separate list a question of law to avoid the deference due to the jury verdict. Each error sent the Lough court fishing in deeper waters without a license.
The Lough court made the same mistake as the trial court in TP Lab.: “[I]t is incorrect to impose on the patent owner, as the trial court in this case did, the burden of proving that a ‘public use’ was ‘experimental.’ ” 724 F.2d at 971. In fact, in TP Lab., the inventor had used the invention nearly four years before filing, id. at 967, had kept only records that were “scanty at best,” id. at 969, and had shown control over the invention only “inherently by the dentist-patient relationship,” id. at 972. Yet the Federal Circuit, in a case strikingly similar to Lough, discerned no public use. Id. at 973. More important, the Federal Circuit in TP Lab. warned against divorcing experimentation from its public use context — the Lough court’s initial false turn on its unauthorized fishing expedition.
IV.
The mischief of Lough’s plenary review standard is even more egregious because public use rests on an evaluation of the totality of the circumstances. 35 U.S.C. § 102(b) (1994); see Manville Sales, 917 F.2d at 549. As noted earlier, the Supreme Court set forth a predominantly factual inquiry for application of the statutory bars. See cases cited supra at p. 1530. The Supreme Court, for good reason,9 assigned the public use factual inquiry to trial, not appellate, courts. See, e.g., Kendall, 62 U.S. (21 How.) at 331 (“The real interest of an inventor with respect to an assertion or surrender of his rights ...' is an inquiry or conclusion of fact, and peculiarly within the province of the jury____”). Without the discipline of the limited standard of review under the Supreme Court’s rule, an appellate court may pick, choose, and reassess various circumstantial factors. No wonder this undisciplined exercise produces inconsistent and irreconcilable results.10 No wonder commentators excoriate this court’s random application of the “totality of circumstances” test.11
*1535Indeed Lough illustrates the arbitrariness of reapplying a factual “totality of circumstances” test on an appellate record without any deference to the fact-finding of the trial court. Lough lists several indicia of experimentation, including: 1) the number of prototypes and the duration of testing; 2) the attention to records or reports during testing; 3) the existence of a confidentiality arrangement; 4) the receipt of any commercial advantage by the patentee; and 5) the inventor’s control over the testing. Although a brief review of prior Federal Circuit cases calls into question the sufficiency of this listing,12 at this point it is more important to show that the Lough court’s application of its own indicia conflicts with binding Federal Circuit cases.
Beginning with the first indicator, Lough gave no analysis to its own first factor at all. The Lough court might have noted Lough’s few prototypes or the limited duration of his testing. In fact, in view of the six years of open air testing in City of Elizabeth, 97 U.S. at 133, or more than a year of open display in Moleculon, 793 F.2d at 1263, the Lough court might well have followed prior cases to assess this indicator strongly in favor of experimentation. After all, Lough conducted a far more limited test than in City of Elizabeth, limiting his testing to six invisible prototypes over the course of one year to ensure that his invention would work in the harsh marine environment.
On the second indicator, the Lough court weighed this harshly against Lough to undercut his experimentation claim. Lough, 86 F.3d at 1121. The Lough court declined to note that, although not keeping formal written records, Lough had taken photographs to monitor the installed seal assemblies during the experimentation period. In view of several eases that detected an experimental use despite a lack of formal record keeping on the testing, TP Lab., 724 F.2d at 969 (finding experimental use with “scanty” records “at best”); Manville Sales, 917 F.2d at 551 (finding experimental use with one memorandum of intent to experiment, but no records of testing), the Lough court might well have followed prior cases in seeing this factor as no obstacle to Lough’s claim of experimentation.
The Lough court did not analyze its own third indicator. Lough, however, testified that he only shared his experimental' prototypes with a few individuals — with whom he shared a close personal or professional relationship — who he knew would maintain his secrecy. In view of Federal Circuit eases that inferred confidentiality on the basis of the close association of the patentee and the *1536testers, see, e.g., Moleeulon, 793 F.2d at 1266, TP Lab., 724 F.2d at 972, the Lough court might well have weighed this indicator in favor of, rather than against, experimentation.
On the fourth indicator, the Lough court luke-warmly acknowledged Lough’s steadfast refusal to commercially exploit his invention before filing. Lough, 86 F.3d at 1121. In view of Federal Circuit cases that attached great, perhaps dispositive weight, to this factor, see, e.g., Ferag AG v. Quipp, Inc., 45 F.3d 1562, 1566, 33 USPQ2d 1512, 1515 (Fed. Cir.1995) (“Foremost among these is the policy of preventing inventors from exploiting the commercial value of their inventions while deferring the beginning of the statutory term.”) and Moleculon, 793 F.2d at 1266-67, the Lough court might well have given this factor considerable weight in favor of a finding of experimentation.
On the fifth indicator, the Lough court weighed this strongly against Lough. Once again, however, the record shows that Lough selected friends and professional associates over whom he had control by virtue of their relationship. In view of prior Federal Circuit cases permitting close relationship and notification of experimental purpose to substitute for full control, see, e.g., Moleculon, 793 F.2d at 1266-67, and LaBounty Mfg. v. United States International Trade Commission, 958 F.2d 1066, 1072, 22 USPQ2d 1025, 1029 (Fed.Cir.1992) (suggesting that notification of experimental purpose may substitute for control), the Lough court might well have balanced this factor in favor of Lough.
Further, to buttress its own factual finding on appeal that Lough lacked control over the experiment, the Lough court used reasoning in conflict with other Federal Circuit cases. For instance, the Lough court noted that the experimental use involved Lough’s associates rather than Lough himself. Lough, 86 F.3d at 1121-22. Other binding Federal Circuit eases have held that participation by others in the experiment does not show a lack of control. See TP Lab., 724 F.2d at 972 (“routine checking of [success of invention] by one of the [inventor’s' associates] does not indicate the inventor’s lack of control”); Grain Processing Corp. v. American Maize-Prods. Co., 840 F.2d 902, 906, 5 USPQ2d 1788, 1792 (Fed.Cir.1988) (submitting samples of invention to food manufacturers for determination of the product’s utility did not negate experimentation). Additionally, the Lough court criticized the fact that one of the boats containing the experimental seal was sold at some point after the conclusion of the experiment. Lough, 86 F.3d at 1121. This indicator has also been held not to constitute a loss of control. See Watson v. Allen, 254 F.2d 342, 347, 117 USPQ 68, 71 (D.C.Cir.1958) (“We believe the protective umbrella of the experimental use doctrine should include reasonable disposal of models and prototypes of the invention once their usefulness to the inventor has ended — reasonable in view of the nature of the device and the probability of discovery and appropriation of the invention by strangers.”); Goodwin v. Borg-Warner Corp., 157 F.2d 267, 273, 70 USPQ 387, 393 (6th Cir.1946) (finding experimentation although automobile containing seal was later sold).
Moreover, several binding Federal Circuit eases analyze indicators the Lough court chose to ignore. These unacknowledged indicators further buttress Lough’s case against a public use. For instance, another indicator appearing in binding precedents is the inventor’s careful tailoring of the tests to the nature of the invention. See City of Elizabeth, 97 U.S. at 136; Manville Sales, 917 F.2d at 550-51; Allied Colloids, 64 F.3d at 1575; see also Caterpillar Tractor Co. v. Berco, S.P.A., 215 USPQ 948, 958 (D.Wyo. 1982) (finding experimental purpose where “tests for durability and performance are conducted by those who will eventually be using the product under conditions which are likely to be difficult to duplicate in the lab”), ajfd on other grounds, 714 F.2d 1110 (Fed. Cir.1983). The record shows Lough designed his experiments to test the invention in various kinds of boats in various marine environments. Indeed a careful assessment of this indicator might also recognize that the necessity of testing under varying circumstances militates against maintenance of full control at all times over the testing.
Still another indicator appearing in binding precedents is the hidden nature of an experi*1537ment.13 In determining whether an inventor intended to experiment, an invisible use may add plausibility to the inventor’s explanation that he perceived little need to take further precautions to protect confidentiality. See, e.g., Manville Sales, 917 F.2d at 550 (“[T]he invention was mounted atop a 150-foot tall pole in a rest area still closed to the public, making it very unlikely that the public would even see the new design.”); Allied Colloids, 64 F.3d at 1574 (public access to and knowledge of the public use is one of the factors to consider). This factor, too, if considered by the Lough court, would have helped Lough. In picking and choosing amongst its “totality of circumstances,” the Lough court sadly did not consider the totality of the circumstances.
In sum, this brief comparison of the Lough court’s assessment of the totality of the circumstances with the same assessment by prior Federal Circuit cases shows the plastic malleability of a totality of circumstances test when divorced from the discipline of a deferential standard of review. Ah appellate panel can pick and choose amongst malleable circumstances to reach results irreconcilable with other Federal Circuit decisions.
This comparison also shows the plausibility of the jury’s unanimous verdict in Lough. For each and every indicator .under the Lough court’s abbreviated totality of circumstances analysis, this record provides factual support for the jury’s conclusion. Under Federal Circuit law, the Lough jury had a very solid foundation for its result. Frankly, this comparison makes the Lough court’s conclusion (“the jury had no legal basis to conclude that the uses of Lough’s prototypes were experimental,” Lough, 86 F.3d at 1122 (emphasis added)) seem very strained.
V.
As noted earlier, no area of Federal Circuit jurisprudence has produced more outright criticism than its review of statutory bars. Plenary standards of review contribute to unpredictability and inconsistency by multiplying at the appellate level the uncertainties of a “totality of circumstances” analysis. Nonetheless the “totality of circumstances” format itself may well be chiefly responsible for the confusion.
This observation may seem to lay the blame for chaos at the steps of the Supreme Court because this nation’s highest court repeatedly called for factual inquiry to resolve statutory bar cases. To the contrary, factual inquiry differs from a standardless “totality of circumstances.” The “totality” test recognizes no hierarchy or principle to guide factual inquiry. The experience of the common law teaches that generally some facts are more important than others. These “more important” facts serve as touchstones lending predictability and consistency to the law. The “totality” test, however, starts each inquiry from scratch and abandons the experience that Holmes denoted the “life of the law.” Oliver Wendell Holmes, The Common Law 1 (1881).
The chief mischief of the “totality” test lies in beginning each factual inquiry anew without reference to the experience of the common law. Thus, for instance, in ascertaining whether an inventor possessed a bona fide intent to experiment, the absence or presence of premature commercial exploitation might disclose more about the inventor’s intent than the duration of the testing or some other factor. Thus the twin touchstones for statutory bar analysis — premature economic exploitation (usually) in the marketplace14 (particularly applicable to on sale bars) and creation of a widespread expectation that the public is free to use the invention (particular*1538ly applicable to public use bars) — ought to guide and restrain factual inquiry at the trial court level. Indeed Federal Circuit cases have so held. Ferag AG, 45 F.3d at 1566.
In this case, the jury determined that Lough did not sell or otherwise exploit his seal, nor did he create the public expectations by his uses. Instead he conducted limited tests to ensure his invention seal worked. Moreover the jury determined that Lough maintained adequate monitoring and control over his testing to show that he was really experimenting, not exploiting or surrendering his invention. The Lough court, however, reweighed these findings as “questions of law” and chose to find in its own peculiar “totality of circumstances” a bar. Without a more rule-based standard for assessing the facts and with appellate courts reassessing the facts without deference, predictability and certainty will remain barred from section 102(b) jurisprudence.
. I do not intend by this to suggest that any single test can or should be imposed.
. See, e.g., Thomas K. Landry, Certainty and Discretion in Patent Law: The On Sale Bar, The Doctrine of Equivalents, and Judicial Power in the Federal Circuit, 67 S. Cal. L.Rev. 1151 (1994); William C. Rooklidge & Stephen C. Jensen, Common Sense, Simplicity and Experimental Use Negation of the Public Use and On Sale Bars to Patentability, 29 J. Marshall L.Rev. 3 (1995); James A. Jorgensen, Comment, Environmentally Dependent Inventions and the "On Sale" and “Public Use” Bars of 102(b): A Proffered Solution to a Statutory Dichotomy, 49 U. Miami L.Rev. 185 (1994).
. Despite the Lough court's presentation, experimental use is not a doctrine separate from public use with its own required indicia and standard of review. A fuller discussion of this Lough court error appears in part III infra. For now, this opinion examines Lough in its own erroneous context.
. In addition, pre-Federal Circuit appellate decisions followed the Supreme Court rule. The uniformity of their conclusion that section 102(b) is predominantly a question of fact both buttresses the Supreme Court’s rule and strikingly illustrates the aberration of the Federal Circuit's departure. See, e.g., Watson v. Allen, 254 F.2d 342, 346-47 (D.C.Cir.1958); Atlas v. Eastern Air Lines, Inc., 311 F.2d 156, 159-60 (1st Cir.1962); McGraw-Edison Co. v. Central Transformer Corp., 308 F.2d 70, 73 (8th Cir.1962); FMC Corp. v. F.E. Myers & Bro. Co., 384 F.2d 4, 10 (6th Cir. 1967); Lear Siegler, Inc. v. Ark-Ell Springs, Inc., 569 F.2d 286, 291 (5th Cir.1978); DeLong Corp. v. Raymond Int'l, Inc., 622 F.2d 1135, 1141 (3d Cir. 1980); Del Mar Eng'g Lab. v. Physio-Tronics, Inc., 642 F.2d 1167, 1169 (9th Cir.1981).
Further, early treatises uniformly acknowledgéd that public use and on sale doctrines are questions of fact, e.g.: "What constitutes public use is ordinarily a question of fact.” 2 Deller’s Walker on Patents § 144 (1964). The 1985 edition of Lipscomb’s Walker on Patents also states that public use is a question of fact, but then notes: "The decision of the Court of Appeals for the Federal Circuit in Moleculon Research Corp. v. CBS, Inc., regarding whether "public use” under § 102(b) is a question of fact or a question of law appears to be at odds with the prior cases.” 2 Lipscomb’s Walker on Patents § 7:28.
. Subsequent cases have followed in this distortion of binding precedent. See, e.g., Petrolite Corp. v. Baker Hughes Inc., 96 F.3d 1423, 1425, 40 USPQ.2d 1201, 1203 (Fed.Cir. 1996) (citing Barmag, 731 F.2d at 836-37); Baxter Int’l, Inc. v. COBE Lab., Inc., 88 F.3d 1054, 1058, 39 USPQ2d 1437, 1441 (Fed.Cir. 1996).
. As noted earlier, the law would not separate experimental use from its context within a section 102(b) bar question and grant it a distinct standard of review. See TP Lab., 724 F.2d at 971. Nonetheless, the Lough court adopted this methodology. This dissent examines Lough within its own format.
. Perhaps the only analog for plenary review of a "totality of circumstances” test within the sweep of law governed by our sister circuits is review of so-called constitutional facts. See, e.g., Swint v. City of Wadley, 51 F.3d 988, 996 (11th Cir.1995); United States v. Edenfield, 995 F.2d 197, 200 (11th Cir.1993); United States v. Bloom, 975 F.2d 1447, 1456 (10th Cir.1992). With life and liberty at stake, a context hardly similar to the statutory bars, appellate courts review probable cause findings without deference. See Ker v. California, 374 U.S. 23, 33-34, 83 S.Ct. 1623, 1629-30, 10 L.Ed.2d 726 (1963); see also Ornelas v. United States, - U.S. -, 116 S.Ct. 1657, 134 L.Ed.2d 911 (1996).
. The law generally recognizes that an experimental use ends with an actual reduction to practice. RCA Corp. v. Data Gen. Corp., 887 F.2d 1056, 1061, 12 USPQ2d 1449, 1453 (Fed.Cir. 1989). On the other hand, the Supreme Court specifically noted: “If durability is one of the qualities to be attained, a long period, perhaps years, may be necessary to enable the inventor to discover whether his purpose is accomplished. And though, during all that period, he may not find that any changes are necessary, yet he may be justly said to be using his machine only by way of experiment; and no one would say that such a use, pursued with a bona fide intent of testing the qualities of the machine, would be a public use, within the meaning of the statute.” City of Elizabeth v. Pavement Co., 97 U.S. 126, 135, 24 L.Ed. 1000 (1877).
.Those who doubt the difference between a cold, abridged, written appellate record and the living, detailed, and emotion-charged courtroom need to read Ludwig Wittgenstein’s philosophical observations (communication lacks meaning divorced from its entire context) and contemplate: What does a witness mean by the answer, “I believe so.” Is this assent to the proposition? Is this an expression of doubt in the proposition? Is this a description of the witness's state of mind (tentative belief) or a description of an actual state of affairs (it is so!)? In other words, how is an appellate court to assess this factual record: ”Q: Mr. Inventor, did you intend to experiment? A: I believe so.”?
This ambiguity and complexity multiples with the number of witnesses — each with their own peculiar personality, perception, and perspective. An appellate tribunal is simply not equipped in time or tools to unlock the mysteries of the single testimonial snowflake (e.g., "I believe so.”), let alone the swirling blizzard of trial evidence. Our trial system — multiple jurors with an experienced trial judge as a check — may have flaws, but remains a “palladium of free government," •and “an excellent method of determining questions of property.” The Federalist No. 83, at 499, 501 (Alexander Hamilton) (Clinton Rossiter ed., 1961). For wise purpose, the Supreme Court commended factual inquiry to the trial system.
. The dissent in Atlantic Thermoplastics Co. v. Faytex Corp., 5 F.3d 1477, 1482-88, 28 USPQ2d 1343, 1347-52 (Fed.Cir.1993), notes that the Federal Circuit has set forth at least three different tests for application of section 102(b).
. E.g., William K. West, Jr. & Nancy J. Linck, The Law of “Public Use" and “On Sale": Past Present and Future, 72 J. Pat. & Trademark Off. Soc’y 114, 115 (1990) ("[T]he court’s insistence *1535on a 'totality of the circumstances’ test may tend to make predictability of result ... even more difficult than in former times.”); Jorgensen, supra note 1, at 213 ("[T]he Federal Circuit has lacked consistency in its application of the on sale and public use bars, of " section 102(b)____ Instead, confusion remains the central theme....”); Stephen R. Schaefer, Comment, Envirotech Corp. v. Westech Engineering, Inc.: The On-Sale Bar to Patentability and Executory Sales Offers, 75 Minn. L.Rev. 1505, 1508 (1991) ("[T]he application of the on-sale bar provision lacks predictability, fails to inform the inventor ... and, thus, fosters litigation.”). This confusion has also been critically noted by commentators:
[T]he Federal Circuit needs to resolve inconsistencies in its precedent regarding the policy underlying experimental use itself, the factual/legal nature of experimental use.... Only by returning to the law as it stood over 100 years ago and by keeping an eye on the policy underlying experimental use could the Federal Circuit return the law in this area to a state of common sense and reasonable, if not sublime, simplicity.
Rooklidge & Jensen, supra note 1, at 49-50.
. Other panels of the Federal Circuit, as might be expected of a "totality of circumstances” analysis, list different "indicia.” See, e.g., TP Lab., 724 F.2d at 971 (indicia include length of test, whether payment was made, conditions of secrecy, whether records were kept, who conducted the tests, and the number of tests conducted); Manville Sales, 917 F.2d at 550 (indicia for totality of the circumstances are the policies underlying the bar, not separate experimental use indicia); Harrington Mfg. v. Powell Mfg., 815 F.2d 1478, 1480, 2 USPQ2d 1364, 1366 (Fed.Cir. 1986) (adding indicia of public necessity); Baker Oil Tools, 828 F.2d at 1564 (adding indicia of public testing in relation to nature of experiment); Allied Colloids Inc. v. American Cyanamid Co., 64 F.3d 1570, 1574, 35 USPQ2d 1840, 1842 (adding indicia of public access to and knowledge of the public use). These differing lists disclose again the appellate disposition, in the absence of a disciplined standard of review, to pick and choose amongst factors to reach inconsistent results.
. Case law distinguishes the notoriety or openness of an experiment from the notoriety or openness of a public use in general. As the Supreme Court held in Egbert v. Lippmann, 104 U.S. 333, 336, 26 L.Ed. 755 (1881), a barring use — in this case an innovative undergarment— need not be open and notorious to be “public.” Because Egbert himself put the invention into use eleven years before filing for patent protection, the corset worn by Egbert’s intimate friend, though not in view, was abandoned to the public. Id. 104 U.S. at 337-38. Such evidence might have been viewed differently, however, if experimentation had been at issue in Egbert. See id. at 337 ("They were not presented for the purpose of experiment, nor to test their qualities. No such claim is set up in her testimony.”).
. Egbert, the famous corset case, is the classic example of exploiting an invention without seeking commercial benefit. 104 U.S. at 336.