Rexam Industries Corporation (“Rex-am”) appeals the judgment of the United States District Court for the Western District of North Carolina, awarding priority in an interference proceeding to the Eastman Kodak Company and Avery Dennison Corporation (“Kodak”). Because we find no error in the district court’s conclusion that Rexam failed to prove prior conception by a preponderance of the evidence, we affirm the judgment of the district court.
I
This case arises from an interference tried before a district court under 35 U.S.C. § 146. Rexam seeks to establish priority over Kodak’s first-filed application by proving an earlier conception and either an earlier actual reduction to practice or diligence up to its reduction to practice. The single count of the interference, identical to claim 1 of Rexam’s U.S. Patent No. 4,931,324 (“the ’324 patent”) reads:
A flexible decorative sheet material suitable for use in surfacing panels, such as automobile body panels, said sheet material being characterized by having the appearance of a glossy, base coat/clear coat automotive paint finish, and comprising a flexible outer layer having a smooth, glossy outer surface with a distinctiveness of image value of 65 percent or greater and formed of a flexible, optically clear weatherable polymer, and a pigmented coating adhered to the opposite surface of said outer layer and visible through the clear outer layer.
The underscored limitation is at issue in this appeal.
Kodak, the senior party, relies upon a constructive reduction to practice from its filing date of November 3, 1987. Rexam, the junior party, has a filing date of September 8, 1988, for the application which matured into the ’324 patent. This application was a continuation-in-part of an earlier application, filed October 28, 1986, which matured into U.S. Patent No. 4,810,540 (“the ’540 patent”). Rexam was denied priority from this earlier filing because the earlier application did not disclose either the distinctiveness of image (“DOI”) limitation of 65 percent or greater or any method of making a sheet with a DOI greater than 65 percent. The application for which Rexam is now trying to claim priority discloses that carrier films with a surface roughness of less than 0.018 mum yield decorative sheets with low haze and DOI of greater than 65 percent. See ’324 patent, col. 6. This disclo*985sure and claim is the only difference between Rexam’s earlier application and the application involved in this interference.
In the original interference proceeding before the Board, Rexam failed to carry its burden of proving priority. Rexam appeared to rely on statements from General Motors (“GM”) engineers who had viewed some Rexam samples in 1986 and declared them a “good match” to GM’s copper me-tale paint standard. The Board discounted this evidence, however, because there was no record of the DOI of GM’s copper metallic paint standard. Moreover, even assuming GM’s paint standards had a DOI of at least 65 percent, the mere fact that GM engineers had proclaimed the Rexam films a “commercially acceptable match” to the GM paint standard would not communicate that the films had a DOI greater than 65 percent. Likewise, evidence that Rexam’s color laboratory described samples as having a “superior” or “outstanding” DOI did not indicate that the inventors recognized that their samples had a DOI greater than 65 percent. Accordingly, the Board held that Rexam’s evidence could establish neither conception nor actual reduction to practice, because both conception and actual reduction to practice would require that the inventors recognized or appreciated that their samples had achieved the minimum DOI required by the interference count.
Rexam commenced a civil action under 35 U.S.C. § 146 in the district court, which conducted a trial including live testimony before a magistrate judge. Discounting the inventors’ testimony in the absence of rehable corroboration, the district court found (for much the same reasons the Board did) that the inventors had not recognized or appreciated that they had achieved a film with a DOI of more than 65 percent prior to Kodak’s filing date. Rex-am therefore could not establish either prior conception or prior reduction to practice. The district court further found that, even if Rexam proved prior conception, Rexam had not shown that it exercised diligence in reducing the invention to practice from just before Kodak’s filing date to Rexam’s own reduction to practice. Because the court did not find that Rexam could establish priority of invention, the district court did not reach the question of whether Rexam had abandoned, suppressed or concealed the invention of the count. Rexam appeals the judgment of the district court, an appeal which is committed to our exclusive jurisdiction by 28 U.S.C. § 1295(a)(4)(C).
II
The junior party in an interference has the burden of proving conception or reduction to practice by a preponderance of the evidence. Priority, conception, and reduction to practice are questions of law based on subsidiary factual findings. Cooper v. Goldfarb, 154 F.3d 1321, 1327, 47 USPQ2d 1896, 1901 (Fed.Cir.1998). A district court conducts a de novo trial, including de novo factual findings, when live testimony is admitted in a civil action under 35 U.S.C. § 146, although the record before the Board is treated as if it were produced before the district court. Winner Int’l Royalty Corp. v. Wang, 202 F.3d 1340, 1347, 53 USPQ2d 1580, 1585-86 (Fed.Cir.2000). We review the district court’s factual findings for clear error and its conclusions of law de novo. Id. at 1344-45, 202 F.3d 1340, 53 USPQ2d at 1583.
A
To establish prior conception in an interference proceeding, the junior party must do more than show he or she imagined or desired the limitations defining the subject matter of the count. Although this principle was referred to only in passing by the district court, it is an ancient one. *986Over a century ago, it was said of prior conception:
the party claiming “must have been first to conceive the thing in controversy; not merely to conceive it possible to construct a device which would produce the result sought.... The conception must not be the result to be obtained, but the means (which is the patentable thing) to produce that result. ”
Mergenthaler v. Scudder, 11 App. D.C. 264 (D.C.Cir.1897) (quoting Voelker v. Gray, 30 O.G. 1091). And more recently, we have said that conception “is complete only when the idea is so clearly defined in the inventor’s mind that only ordinary skill would be necessary to reduce the invention to practice, without extensive research or experimentation.... Because it is a mental act, courts require corroborating evidence of a contemporaneous disclosure that would enable one skilled in the art to make the invention.” Burroughs Wellcome Co. v. Barr Labs., Inc., 40 F.3d 1223, 1228, 32 USPQ2d 1915, 1919 (Fed.Cir.1994). Thus, although reduction to practice is not an element of conception, a junior party claiming prior conception cannot meet its burden of proof merely by showing that the subject matter of the count was desired or spoken of by its inventors.
The count at issue in this case requires that the outer layer of the sheet material possess a smooth, glossy outer surface with a distinctiveness of image value of 65 percent or greater. Regardless of whether the language of the count alone suggests a correlation between the smoothness of the surface and its DOI value, the evidence shows that it is necessary to cast the outer layer on a smooth carrier surface in order to achieve the recited DOI value. See ’324 patent, col. 6, II. 10-61. Contrary to Rexam’s protest that this issue arose for the first time in Kodak’s closing argument, the connection between a smooth casting surface and the high gloss and DOI of the resulting film has been a foundation of these proceedings. For in denying Rexam the benefit of the filing date for its ’540 patent, the Board specifically pointed out that unlike the ’540 patent, the application of the ’324 patent discloses the high gloss of the outer layer and discloses that casting the outer layer on certain grades of PET film is necessary to achieve the DOI level of the count. Were the requirement for a smooth casting surface not an important aspect of the invention, Rexam might have been awarded the benefit of its ’540 filing date .and this interference would have been concluded years ago.
Rexam further protests that such a connection is not taught by the Kodak or Avery applications, and Kodak’s counsel seemed to admit as much at oral argument. The parties have not provided us with Kodak’s senior application, although we note that Avery’s application does disclose the need for a high-gloss carrier sheet, and in particular teaches the Hoechst 3000 grade polyester film disclosed as suitable in the ’324 patent. However, because Kodak’s right to make the count is not at issue in this appeal, we could not in any event attach weight to the presence or absence of this disclosure from Kodak’s senior application.
In light of the requirement to show complete conception of a means for practicing the invention, we think it clear, the district court did not err in concluding that Rexam had failed to carry its burden of proof. The evidence supports a finding that Rex-am’s inventors did not understand the required properties of the casting surface until after Kodak’s constructive reduction to practice. In particular, we note the testimony of Dr. Ellison, apparently elicited to explain that Rexam had not previously concealed the requirement for a smooth casting surface, or unduly delayed in disclosing it. Dr. Ellison testified that al*987though Rexam employees had been casting films on high-grade polyester sheets since 1985, he had not understood the need to use certain smooth grades of polyester in order to achieve a high DOI in a base coat/clear coat (“BC/CC”) film until the spring of 1988-well after Kodak’s filing date. Although Rexam tries to argue that Dr. Ellison’s testimony refers to the date of patenting activity, not the date when the connection between casting surface smoothness and DOI was realized, our own reading of the testimony in context supports the district court’s conclusion that Rexam did not appreciate this relationship before Kodak’s filing date. And given that the work Dr. Ellison referred to constitutes the only significant difference between the disclosures of the ’540 and ’324 patents, we seriously doubt that Rexam would have waited until September of 1988 to file its application if Dr, Ellison had merely been writing up his results for patenting in the spring of 1988.
Likewise, we find that little of Rexam’s corroborating evidence can support a finding of complete conception, much less demand it. Rexam cites the inclusion of U.S. Patent No. 4,215,170 in its Information Disclosure Statement for the earlier-filed ’540 patent as evidence that Rexam’s inventors recognized the correlation between smoothness and DOI. The fact that Rexam’s patent attorney disclosed this document to the Patent Office says little about what the inventors appreciated at the time. More importantly, the cited patent merely states that the carrier film for a metallic coating must have a glossy, polished surface, and discloses that various polymeric substances are suitable for carrier films. It discloses no connection between a certain degree of casting surface smoothness and a resultant DOI.
Rexam also relies on the 1982 GM paint specification, which required DOIs of greater than 65 percent for most of its colors, and the 1985 Hovis meeting notes, which state that Rexam would probably need to maintain an 80 percent DOI for its BC/CC film. These documents corroborate only Rexam’s desire for a film meeting the count limitations, not “a contemporaneous disclosure that would enable one skilled in the art to make the invention.” Burroughs Wellcome, 40 F.3d at 1228, 32 USPQ2d at 1919. As was the case in Hitzeman v. Rutter, 243 F.3d 1345, 58 USPQ2d 1161 (Fed.Cir.2001), Rexam’s problem is not merely its inability to produce a document declaring that the BC/CC project’s aim was to develop a smooth, glossy film with a DOI of greater than 65 percent. Rexam’s problem is that such documents would corroborate only the “hope” of achieving such a film, and could not establish a complete conception without evidence that the inventors understood enough about how to make said films such that only ordinary skill would be required to reduce the invention to practice. Because Rexam failed to introduce such evidence, we cannot say the district court erred in finding that Rexam had not achieved complete conception prior to Kodak’s reduction to practice.1
B
Although we find Rexam’s failure to prove a complete conception sufficient grounds to affirm the judgment of the district court, we would uphold its conclusion on conception even if we required Rexam merely to show appreciation of the count’s 65 percent DOI limitation. We must reject Rexam’s various allegations of fundamental error. The record shows that *988the district court clearly understood the concept of DOI and how it is measured, and clearly understood that the subject matter of the count was a genus defined by a DOI of 65 percent or greater-not by separate limitations of 65 percent and greater than 65 percent. Likewise, we must reject Rexam’s arguments that the DOI limitation flows naturally from the requirement that the sheet have the appearance of a glossy BC/CC finish, or from some other (unspecified) “structural” characteristic of the film, because such arguments merely echo Rexam’s long-rejected position that the DOI limitation should not appear in this interference. Even if we were inclined to allow Rexam to rewrite the interference count at this stage in the proceedings, Rexam has directed us to no evidence that the DOI limitation is immaterial to the patentability of the count, or that those of ordinary skill in the art would recognize the DOI limitation to be necessarily present in the remaining limitations of the count.
Bereft of such legal support, Rexam must bear the burden of showing that the district court’s factual conclusions were clearly erroneous. For although conception is ultimately a conclusion of law, whether or not Rexam’s inventors appreciated the greater than 65 percent DOI limitation is based largely in fact. Our review of such findings is limited:
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. Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.... This is so even when the district court’s findings do not rest on credibility determinations, but are based instead on physical or documentary evidence or inferences from other facts.
Am. Original Corp. v. Jenkins Food Corp., 774 F.2d 459, 462, 227 USPQ 299, 300 (Fed.Cir.1985) (quoting Anderson v. City of Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985)).
Rexam is correct that one need not demonstrate appreciation of the requisite DOI in the exact words of the count, for the invention is not the words of the count but the subject matter so defined. See Silvestri v. Grant, 496 F.2d 593, 599, 181 USPQ 706, 710 (CCPA 1974). But because DOI is not an inherent property of the count’s other limitations, Rexam must still show that the inventors’ conception included-an appreciation of a genus of films with the requisite DOI of the count, even if the inventors did not attach a specific number to this property. That is, even if Rexam could not document conception in terms of a numerical DOI, Rexam would still be required to show that its inventors conceived of the same genus in terms of some property or properties that are interchangeable with numerical DOI measurements.
Viewing the evidence in light of our deferential standard of review, we cannot say that the district court erred by concluding that Rexam had failed to prove appreciation of the count’s subject matter in terms of some appropriate surrogate for a numerical DOI measurement. As we have explained, the GM specifications and Hovis meeting notes can corroborate little but the inventors’ desire to achieve films within the count limitations. And although GM engineers deemed some of Rexam’s samples a “commercially acceptable match,” the district court specifically found that these statements were directed to the overall appearance of Rexam’s films, and cannot corroborate an appreciation of any particular DOI of the films. The fact that *989a DOI of 80 percent was a general goal, or that DOI was one of the characteristics required for an acceptable film, or that DOI numbers were from time to time discussed at Rexam, does not mandate a conclusion that the inventors specifically appreciated the count limitation in the legal sense.
Rexam’s most persuasive evidence is the testimony of Mr. Griswold, who testified that by means of the “hair test” he judged, under controlled conditions, the DOI of certain BC/CC samples to be over 90 percent and so informed Mr. Winton. However, the district court heard Mr. Griswold’s testimony live, and we must be highly deferential of the weight the district court placed upon his testimony, especially in light of the absence of contemporaneous documentation. The district court made a specific finding that the hair test was not a reliable method for determining DOI, and drew from the evidence the inference that Rexam employees used the hair test primarily to gauge overall visual appearance of the sample, and not to classify it as having a relatively high DOI. Thus, even if we do not hold Rexam to a precise numerical definition of the interference count, we cannot say that the district court clearly erred by finding this and similar evidence insufficient to corroborate conception.
We agree with Rexam that one could certainly draw the inference of appreciation from its evidence, and we agree that no evidence rules out such appreciation. But this is not a motion for summary judgment. This court may not substitute its version of the facts for that of the district court-even if we would draw a different conclusion from the evidence in record-and Rexam has identified no genuine error in the district court’s theoretical understanding of DOI, interpretation of the count, evidentiary standards, or application of the law of conception. In the absence of such a showing, we cannot hold that the district court erred in its conclusion that Rexam failed to prove appreciation of the greater than 65 percent DOI limitation.
CONCLUSION
For the reasons set forth above, we conclude that Rexam did not carry its burden of demonstrating an earlier conception date by a preponderance of evidence. Because Rexam has not established an earlier conception date, we need not address whether Rexam exercised diligence in reducing the invention to practice. We therefore affirm the district court’s award of priority to Kodak.
. The fact that Rexam may have actually made some samples meeting the claim limitations does not preclude this conclusion. Without understanding how the claimed subject matter may be made, conception is not complete.