dissenting.
Two inventors at J.T. Eaton & Company, Inc. (Eaton) built a better mousetrap and, as promised in the old adage,* the world beat a path to their door. In an effort to reap its rightful reward for advancing the state of the art, Eaton surmounted extensive legal hurdles. Eaton persevered through eight years of patent prosecution in the Patent and Trademark Office (PTO). After a competitor protest, two reexamination proceedings, and two appeals, the Board of Patent Appeals and Interferences (Board) twice confirmed the patentability of Eaton’s invention. Thereafter, Eaton spent twelve years in district court proceedings, ultimately winning a judgment of infringement against its competitor, Atlantic Paste & Glue Company, Inc. (Atlantic). All this effort, however, came to naught when this court concocted its own novel interpretation of the lengthy prosecution history of Eaton’s patent. Based on its reading of the administrative record, this court arrives at a claim interpretation that no party advocated throughout the protracted history of this patent and is likely to cover neither Eaton’s nor Atlantic’s products.
To reach its idiosyncratic claim interpretation, this court relies on a few isolated excerpts from over 1400 pages of prosecution history. In particular, this court abstracts two paragraphs from an eleven-page declaration of John M. Questel, filed in the second reexamination, to convert a stringent vertical flow test into a claim requirement. In fact, neither Eaton, nor the Board, nor even Mr. Questel himself, relied on his 120°F vertical flow test to distinguish prior art. The administrative record, when considered as a whole, does not support this court’s adoption of this test as the measure of Eaton’s claims.
This court need not have strained to interpret the claims. The patent record—the claims, specification, and prosecution history—provides ample support for the claim interpretation adopted by the United States District Court for the Eastern District of New York. In fact, the examiner, the Board, and at least two federal judges have already accepted the same interpretation—a reading *1573compelled by the prosecution history as a whole.
I.
The phrase “plastic flow temperature” has no fixed meaning in the adhesive art. However, an inventor may freely define unfamiliar claim terms in the specification or in the prosecution history. See Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582, 39 USPQ2d 1573, 1576 (Fed.Cir.1996); Lear Siegler, Inc. v. Aeroquip Corp., 733 F.2d 881, 889-90, 221 USPQ 1025, 1031 (Fed.Cir.1984). Eaton took that course to define its claim term.
Eaton’s Patent No. 4,438,584 (the ’584 patent) distinguished its ready-to-use glue trap for mice and rats from the prior art with an adhesive having a “plastic flow temperature above 120°F.” The inventors, Stanley and Benjamin Baker, recognized that the adhesives used in prior art traps would melt and flow at temperatures normally encountered during shipping and storage. The Bakers learned from the United States Department of Transportation that the maximum ambient temperature encountered during shipping was 120°F. To solve the problem with the prior art, the Bakers set out to stabilize their adhesive at temperatures up to 120°F. Thus, the requirements of secure shipping became the benchmark for “plastic flow temperature.”
Throughout the lengthy administrative proceedings, Eaton and the PTO clarified this meaning of plastic flow temperature. For example, in its Amendment filed July 24, 1981, Eaton explained:
[Bjecause the plastic flow temperature is above 120°F, the body of the adhesive will not flow plastically at normally encountered ambient temperatures while being shipped or stored. The traps can be disposed horizontally, upside down or vertically at normally encountered temperatures with no fear of touching of adjacent surfaces or the adhesive of one trap to the other.
Thus Eaton explained that its claimed adhesive would not flow upside-down or “vertically at normally encountered temperatures.”
To underscore this message, Eaton submitted the declaration of Kenneth A. Nelson, the chemist who developed the claimed adhesive. Mr. Nelson’s declaration disclosed a two-prong test for the claim requirement. That test requires,' first, placing a 1/16 inch layer of adhesive upside-down in an oven preheated to 120°F for sixteen hours. Second, the test requires that the adhesive be hung vertically in an oven at 77°F for sixty-three hours. An adhesive meets Eaton’s requirement if the glue does not flow from the tray during either of these tests.
The Board understood and accepted this specific definition of plastic flow temperature. In its November 30,1983, opinion, the Board recognized “plastic flow temperature” as the “characteristic of the adhesive ... required to prevent the adhesive from flowing or sagging at temperatures at which the trap is shipped, stored, or used.” The Board noted that prior art publications did not anticipate Eaton’s claims because they lacked adhesives with these plastic flow properties:
Prior to the instant invention, glue traps were prepared by the user rather than the manufacturer in that ready made traps could not be readily shipped and stored. Workers in this art were working to find various means to solve these problems. Appellants recognized this problem and conceived of the idea embodied in the instant claims of utilizing an adhesive having special plastic flow properties.
Thus, the full administrative record gives an explicit and reasonable meaning of “plastic flow temperature” that focuses on shipping and storage of the commercial trap product without adhesive spillage. Mr. Nelson’s two-prong test makes Eaton’s definition unmistakable' to one of skill in the art. For this reason, the Eastern District of New York adopted the Nelson tests as the definition of “plastic flow temperature.” Indeed, another district court interpreted “plastic flow temperature” in the same ’584 patent precisely in accordance with Mr. Nelson’s declaration. See J.T. Eaton & Co. v. Bell Lab., Inc., No. 95-C-0441-S (W.D.Wis. March 29, 1996) (oral jury instructions). Thus, the examiner, the Board (twice), and *1574two district courts read the plastic flow limitation in accordance with the Nelson tests.
Trial testimony from one skilled in the art confirmed the logic of this uniform claim interpretation. According to this testimony, Mr. Nelson’s sixteen hour inverted horizontal test at 120°F properly tested the product under typical shipping conditions. After all, elevated temperatures would not last longer than normal daylight hours. Testing the adhesives beyond sixteen hours made little sense in the context of shipping these products in trucks and boxcars. Likewise, according to the same testimony, the lengthy vertical test at 77°F properly reflected the prolonged periods of shelf storage at normal room temperatures. Given this logical explanation, there is little wonder every other agency and court that reviewed Eaton’s claims accepted the same interpretation.
II.
Despite the overall content of the prosecution history, this court stretched the vertical flow test concept out of proportion, exaggerated the stringency of that test beyond the realities of storage and shipping, and made it the sole talisman of the Eaton claim. In its brief, Atlantic urged this court to find a stringent vertical flow test in a single sentence of the voluminous prosecution history:
Plastic flow temperature, as used herein, is that temperature below which the adhesive will not flow or sag even if stored with the surface of a thick (1/16 to 1/8 inch) layer of adhesive vertical for unlimited periods of time. (Applicant’s Brief to Board of Patent Appeals and Interferences, December 27,1982)
Based on this statement, Atlantic argued that an accused trap product could not infringe unless it could dangle in a vertical position at 120°F for an infinite period of time without any glue flow. Like the district court before it, however, this court recognized that one skilled in the art could not reasonably rely on such obvious hyperbole as if it affected the scope of the claim. See Intervet Am., Inc. v. Kee-Vet Lab., Inc., 887 F.2d 1050, 1054, 12 USPQ2d 1474, 1477 (Fed.Cir.1989) (holding that claim language controls over an attorney’s “erroneous remarks” made during the course of prosecution). Accordingly, the panel rejected this rationale. In so doing, it rejected the only claim interpretation proffered by Atlantic for “plastic flow temperature.”
Without any basis in argument from either party or testimony from the trial record, the panel’s ultimate claim construction rests on its own collection of orphaned passages from the file history. According to the panel, these passages show that Eaton did not provide a test for measuring vertical flow. To the contrary, Eaton provided Mr. Nelson’s 77°F vertical test—-a test designed for temperatures where the products are likely to hang on vertical displays. Nonetheless, this court created the myth of a “missing vertical test” to justify its reliance on two paragraphs in Mr. Questel’s declaration (paragraphs 22 and 23)—a declaration filed two years after the ’584 patent originally issued.
This court’s reading of the record goes astray in several ways. First, nowhere does the ’584 patent prosecution suggest the necessity of a 120°F vertical flow test. The “missing” test was not missing at all, but was simply not required by the claims. Before the PTO, Eaton consistently stressed that the Nelson tests defined “plastic flow temperature” under temperature conditions likely to be encountered during shipping and storage. Second, the brief extract from Mr. Questel’s declaration simply lacks the meaning and the importance attributed to it by the panel.
A.
The premise underlying this court’s opinion is that Mr. Questel supplied a missing test for vertical flow two years after the patent issued. The panel divines this premise from a few passages in which Eaton generally describes the advantages of the claimed adhesive. For example, the panel opinion relies on Eaton’s statements to the Board on December 27,1982:
Also importantly, the adhesive must have temperature-flow characteristics such that even if the surface of the adhesive is vertical, the adhesive will not flow or sag at ordinarily encountered ambient storage *1575temperatures (less than 120°F). That is to say, the plastic flow temperature of the adhesive must be above 120°F.
According to the panel, this quote demonstrates that “plastic flow temperature” had a vertical component that Mr. Nelson did not measure. To the contrary, Eaton did not shirk its duty to define the claim terms. In the very same 1982 brief, when actually distinguishing the prior art cited by the examiner, Eaton specifically cited to the Nelson tests as the definition of “plastic flow temperature.”
The record is replete with other instances in which Eaton mentioned the word “vertical” in contexts fully consistent with using the Nelson tests as the measure of plastic flow temperature. For example, in its Amendment filed July 24, 1981, Eaton argued:
A trap made [in the manner described in the application] has been evaluated by Kenneth A. Nelson and found to have acceptable plastic flow characteristics. The trap can therefore be shipped or stored in vertical or inverted horizontal positions without fear of plastic flow and thus contamination of adjacent surfaces or spilling.
Eaton simply did not argue for its patent on the basis of some undefined vertical flow capabilities. Eaton had no need to test vertical flow above 77°F because its traps would be displayed vertically only at room temperature.
More importantly, the PTO understood this simple principle. The PTO did not understand Eaton to rely on a stringent vertical test for flow characteristics. Both the examiner and the Board expressly disclaimed any reliance on vertical or horizontal testing to determine patentability of the claimed invention. For example, the examiner stated on May 27,1986:
[Wjhether the body tests were done in vertical, inverted horizontal or flat horizontal position has no relevance. The claims do not specify what position the adhesive is in regarding vertical flow.
The Board expressed a similar view in its January 16, 1987, decision on the second reexamination:
The claims on appeal do not include any limitations restricting the orientation of the commercial trap product defined therein to any particular horizontal, vertical, or inverted disposition, nor are there any limitations requiring flow of the adhesive from the container to be prevented for any particular length of time.
These passages make absolutely clear that the PTO did not decide to issue the ’584 patent based on any of Eaton’s statements about vertical flow or any of Mr. Questel’s tests. Instead, it is this court that first imports into the claims a supposed limitation from the prosecution history—a limitation on which neither the applicant nor the PTO relied during acquisition of the patent.
In fact, the PTO never objected to the phrase “plastic flow temperature” as indefinite for lack of a vertical flow test. If vertical flow was the pivotal claim parameter perceived by this court, then the PTO should have noted its absence. Instead, the lengthy prosecution culminated in the issuance of the ’584 patent without so much as a mention of a mythical “missing” vertical test.
Finally, Eaton did not, as the panel suggests, switch its focus from the Nelson tests to the Questel tests during the final chapter of the prosecution. Rather, at the conclusion of the second reexamination, as at the beginning, Eaton relied on its consistent definition of plastic flow temperature:
The terminology “plastic flow temperature” in appellants’ claims on appeal has that meaning which is made clear elsewhere in the patent or in the file wrapper Appellants defined their terminology “plastic flow temperature” throughout the prosecution of their application which matured as the subject patent as being a characteristic of the adhesive which would prevent the adhesive from sagging or flowing from its support at temperatures at which their product is shipped, stored, or used.
This passage shows that Eaton did not attempt to switch arguments on the PTO in the closing chapter of the prosecution. .
Perhaps aware of the danger of stringing together isolated passages out of context, *1576this court tries to correlate these passages by reference to private communications between Eaton’s lawyers and scientists. Specifically, the panel relies on Eaton’s attorney’s instructions to Mr. Nelson about vertical flow testing. Far from supporting the missing test myth, this tactic betrays this court’s flawed method of seeking claim meaning. These privileged exchanges are not part of the public record. Even if Eaton’s attorneys believed patentability turned on vertical flow at 120°F, such private correspondence, not a part of the public record, cannot possibly supply evidence of claim meaning. See Markman v. Westview Instruments, Inc., 52 F.3d 967, 979, 34 USPQ2d 1321, 1329 (Fed.Cir.1995) (patent claims are construed in light of the public record), aff'd, — U.S. —, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996). Thus, this court relies not only on out-of-context extracts, but on extracts without any relevance to claim interpretation. Indeed, few of the isolated passages on which this court relies had anything to do with the meaning of the plastic flow characteristic. Neither Eaton nor the Board relied on the passages to which this court attaches disposi-tive significance.
For all these reasons, I must conclude that this court has created a requirement for a stringent vertical flow test that the claims, read in light of the prosecution history, do not require. Had Eaton truly acquired its patent by requiring such a test, the record would surely contain far more emphatic evidence than the few isolated passages cited in the panel opinion. See York Prods., Inc. v. Central Tractor Farm & Family Center, 99 F.3d 1568, 1575, 40 USPQ2d 1619, 1624 (Fed.Cir.1996) (“Unless altering claim language to escape an examiner rejection, a patent applicant only limits claims during prosecution by clearly disavowing claim coverage.”).
B.
Without doubt, the precedent of this court indicates that a patentee’s remarks during patent prosecution can illuminate the meaning of the claims. See Markman, 52 F.3d at 980; Vitronics, 90 F.3d at 1582-83. However, our precedent makes equally clear that this court construes an administrative record in its full context, not on the basis of snippets lifted out of context. See Unique Concepts, Inc. v. Brown, 939 F.2d 1558, 1565, 19 USPQ2d 1500, 1506 (Fed.Cir.1991) (“We construe claims in the light of the language of the claim itself, the specification on which it is based, and the whole prosecution history.”) (Rich, J., dissenting); see also Mark I Mktg. Corp. v. R.R. Donnelley & Sons Co., 66 F.3d 285, 292, 36 USPQ2d 1095, 1100 (Fed.Cir.1995), cert. denied, — U.S. —, 116 S.Ct. 917, 133 L.Ed.2d 847 (1996). Moreover, the reviewing court must examine the full context of prosecution history to discern not only what the applicant said, but why he said it. Cf. Hebert v. Lisle Corp., 99 F.3d 1109, 1118, 40 USPQ2d 1611, 1617 (Fed.Cir.1996) (“It is of course necessary to consider not only the amendments to the claims but the reason why they were made_”).
These basic legal principles disclose the infirmity of the court’s reliance on two paragraphs from a single declaration at the end of the second reexamination. A fair reading of the declaration indicates that Mr. Questel distinguished the prior art Formula No. 31 adhesive on two grounds: (1) it exhibited unacceptable levels of cold flow (i.e., 72°F) in a vertical orientation; and (2) it exhibited unacceptable levels of hot flow when tested at 100°F in a vertical orientation. Thus, although Mr. Questel conducted twenty-four hour vertical tests at 120°F, neither he nor Eaton ever relied on such tests to distinguish any prior art. Nor did Mr. Questel in any way suggest that he considered a twenty-four hour vertical test at 120°F to be the measure of “plastic flow temperature.” This record simply does not support the court’s conclusion that Eaton’s claims are measured by the 120°F vertical flow test. See Ethicon Endo-Surgery, Inc. v. United States Surgical Corp., 93 F.3d 1572, 1581, 40 USPQ2d 1019, 1025-26 (Fed.Cir.1996) (looking to “the prosecution history as a whole” to determine how one skilled in the art would understand a claim term) (Clevenger, J.).
Sadly, this court’s fixation on the 120°F vertical flow test foreclosed the one claim construction that was used repeatedly by all parties throughout the public record. The district court properly used that claim inter*1577pretation. This court errs in departing from the full context of the administrative record, and importing a false limitation from the prosecution history.
III.
Finally, even if I agreed with the panel’s claim construction, I could not join a reversal of the district court’s finding of infringement. This court decides this issue against Eaton on the basis of a failure of proof—that is, Eaton’s failure to prove infringement under a claim meaning no one had ever imagined before this court’s pronouncement. The panel reasons that Eaton should have known this court’s interpretation of the claims, and therefore had the burden of presenting adequate proof of infringement. Contrary to this assumption, no patentee in Eaton’s position could have predicted this court’s peculiar claim construction. Even Atlantic, a company with a strong incentive to adopt any argument to avoid infringement, did not proffer this court’s late-coming interpretation.
In light of the time and resources invested in this patent and this infringement battle, this court might at least have afforded Eaton an opportunity to present evidence of infringement under this panel’s novel claim construction. Accordingly, even if I could accept the court’s claim meaning, I would remand the case for further proceedings.
IV.
This decision stands as a monument at the troubled intersection between legal and factual analyses in this court’s post-Markman jurisprudence. That claims must be construed by the court does not divorce the interpretive process from a host of inherently factual subsidiary matters, such as how one skilled in the art would understand claim terms and prosecution history statements. Markman v. Westview Instruments, Inc., — U.S. — , — , 116 S.Ct. 1384, 1395-96, 134 L.Ed.2d 577 (1996) (“We accordingly think there is sufficient reason to treat construction of terms of art like many other responsibilities that we cede to a judge in the normal course of trial, notwithstanding its evidentiary underpinnings.”). This court’s role in reviewing claim meanings discerned by the district courts calls for modesty and restraint—bom not of timidity, but of recognition of the limits inherent in appellate review. When an appellate court arrives at a novel claim interpretation after nearly twenty years of prosecution and litigation, it is inadequately equipped to test its new and unprecedented reading against the views of those skilled in the art. I believe the court today has overstepped the boundaries of effective appellate review.
Perhaps this court was influenced by the district court’s finding that, but for the commercial success of Eaton’s trap products, the ’584 patent would be obvious. Unfortunately, this court does not address the issue of obviousness directly, see, e.g., Newell Cos., Inc. v. Kenney Manufacturing Co., 864 F.2d 757, 768-69, 9 USPQ2d 1417, 1426-27 (Fed.Cir.1988) (concluding that the patent-in-suit was obvious despite strong evidence of commercial success), but instead distorts claim interpretation techniques to defeat infringement. Although sensitive to questions about the validity of the ’584 patent, I believe that the court should not give voice to those questions by means of a disingenuous reading of the prosecution history. Accordingly, I must respectfully dissent.
The exact development of the adage remains a mystery. American essayist and poet Ralph Waldo Emerson (1803-1882) wrote in an 1855 essay in Journals:
I trust a good deal to common fame, as we all must. If a man has good com, or wood, or boards, or pigs, to sell, or can make better chairs or knives, crucibles, or church organs, than any body else, you will find a broad, hard-beaten road to his house, though it be in the woods.
The mousetrap, however, did not appear until seven years after Emerson's death. Mrs. Sarah S.B. Yule, in her book Borrowings (1889), insisted that at a lecture Emerson had stated a catchier variation: "If a man write a better book, preach a better sermon, or make a better mousetrap than his neighbor, though he build his house in the woods, the world will make a beaten path to his door.” More recently, in a twentieth century adaptation, Newman Levy wrote: "If a man builds a better mousetrap than his neighbor, the world will not only beat a path to his door, it will make newsreels of him and his wife in beach pajamas, it will discuss his diet and his health, it will publish heart-throb stories of his love life.”