dissenting.
The majority reverses the district court’s summary judgment because it believes the court committed legal error by not requiring “intentional concealment” of a best mode and by ignoring the disclosure of a particular embodiment of the invention. The majority’s decision, however, is itself based on an error of law. It is not in accord with our precedent regarding intent in a best mode analysis and misconstrues as legal error the district court’s reason*1360able conclusion that the best mode the inventors possessed^an issue of fact—was not a disclosed embodiment. Because I do not see any error of law in the district court’s opinion and agree with its conclusions on the relevant factual issues, I respectfully dissent.
A. Intent
The majority holds that a “[violation [of the best mode requirement] requires intentional concealment.” Majority Op. at 1357 (emphasis added). It therefore concludes that the district court “erred in law” by finding intent an unnecessary element of a best mode violation. Id. at 1358. To reach that conclusion, it relies on Brooktree Corp. v. Advanced Micro Devices as the earliest case that holds that concealment must be “intentional.” 977 F.2d 1555, 1575 (Fed.Cir.1992). However, the statement in Brooktree concerning intentional concealment was dicta and clearly conflicted with our precedent at the time— which repeatedly emphasized that “concealment” need not be intentional.1 Indeed, even the case cited in Brooktree for support, Engel Industries v. Lockformer (also cited by the majority), does not contain a single statement or implication that concealment must be “intentional.” See 946 F.2d 1528,1531 (Fed.Cir.1991). Moreover, as the Appellee Broetje submits, this court has repeatedly acknowledged the error in the Brooktree dicta and recognized that we are bound to follow our earlier precedent that intent is not an element of a best mode violation.2 Accordingly, in my view, there is no basis to reverse the district court for properly concluding that a best mode violation does not require an intent to “conceal.”3
*1361B. DISCLOSURE OF A PREFERRED Embodiment
The majority also concludes that the district court legally erred because the best mode of the invention was a tube with three grooves and the inventors disclosed a three-grooved preferred embodiment in the patents. See Majority Op. at 1358-59. Assuming arguendo that the majority is correct that a three-grooved tube was the best mode and was disclosed as an embodiment,4 the inventors’ disclosure in this case does not satisfy the best mode requirement.
The patents show at least four embodiments of the invention. None of those embodiments are specifically identified as the best mode of the invention and, except for falling under a general heading for preferred embodiments, none is specifically identified as preferred. Moreover, each of them appears to have different types and numbers of grooves: one has what appears to be fifty-five grooves, one has three grooves, and two have two grooves.
Thus, the inventors disclosed four embodiments of their invention, only one of which was a possible best mode. That fact alone probably would not lead me to find a best mode violation. But this case, and the inventors’ disclosure, is not so simple. The inventors did not just disclose their best mode amongst many preferred embodiments; they buried it amongst many embodiments that they knew did not work. The inventors discovered that constructing their claimed tubing with a certain number of grooves did not just work better than another number of grooves but was “essential” to implementing their invention in order to prevent rivet rotation and jamming. J.A. 25-27; 101-104. As the district court discussed, preventing rivet rotation and jamming was allegedly a “patentably distinct” feature of the invention that was unsolved by the prior art. 'J.A. 27. And, as the inventors knew at the time of filing, a tube with two grooves would cause short rivets to jam because ' “the stem of the rivet might go inside one of the grooves.” Aurio'l Dep. 44, July 29, 2011. So the inventors did not simply reveal their best mode and disclose others; they buried the best mode among other embodiments, also falling under the preferred embodiment heading, that did not include the “essential” feature that the inventors discovered and knew was necessary to successfully implement their invention. Such disclosure is not the “quid pro quo” demanded by the *1362best mode requirement. See Eli Lilly & Co. v. Barr Labs., Inc., 251 F.3d 955, 967 (Fed.Cir.2001).
The majority arrives at the opposite conclusion by looking to our divided decision in Randomex, Inc. v. Scopus Corp., 849 F.2d 585 (Fed.Cir.1988). That case, however, supports the district court’s conclusion that there was a best mode violation here.
In Randomex, we recognized that, under the second prong of the best mode test, a disclosure can be “so poor as to effectively result in concealment.” Id. at 589 (quoting Spectra-Physics, 827 F.2d at 1535-36). We concluded, however, that the disclosure in that case was adequate to satisfy the best mode requirement based on the peculiar facts in the record. Specifically, we considered it important that the disclosure of the best mode cleaning fluid was mentioned along with just one other cleaning fluid, which was known in the prior art and understood by those of ordinary skill in the art to be “the ‘worst mode,’ not the ‘best,’ ” because it caused explosions. Id. In short, the court found that the disclosure in Randomex did not effectively result in concealment because it was still an adequate guide for one of ordinary skill in the art to determine which of the two disclosed embodiments was the best mode.
The disclosure here, however, is not an adequate guide that one of ordinary skill in the art could follow to determine the best mode for the invention. Unlike in Randomex, one of ordinary skill in the art would have to build and test three other systems—not just one—to tell which was the best mode. And, in contrast to the well-known alternative embodiment disclosed in Randomex, the alternative embodiments disclosed here are not well-known in the prior art. Those of ordinary skill in the art had no way of knowing that the embodiments without three grooves were clearly inferior and should not be used because they might suffer from rivet rotation and jamming. The inventors knew that fact because they apparently were the first to discover it. Yet the inventors chose to not mention that “essential” quality of their invention in their disclosure and buried it amongst several embodiments. In contrast to Randomex, the inventors here not only left those of ordinary skill in the art without any guide as to the identity of best mode, the inventors led them down the wrong path by disclosing multiple other embodiments that only the inventors knew would not work well.
C. Conclusion
The inventors knew that a certain number of grooves was the best mode for implementing their invention. Whether that number was “odd” as the district court found or “three” as the majority concludes, the inventors did not adequately disclose it. Accordingly, the district court was correct to hold that the patents violated the best mode requirement.5
I respectfully dissent.
.See, e.g., Spectra-Physics, Inc. v. Coherent, Inc., 827 F.2d 1524, 1535-36 (Fed.Cir.1987) (recognizing that "concealment” can be accidental or intentional and that an even a deliberate attempt to conform with the best mode requirement can fail if the "quality of the disclosure [is] ... so poor as to effectively result in concealment”); Dana Corp. v. IPC Ltd. P’ship, 860 F.2d 415, 418 (Fed.Cir.1988) (recognizing, that "concealment” can be accidental or intentional and also finding best mode violation without any reference to "intent”); DeGeorge v. Bernier, 768 F.2d 1318, 1324 (Fed.Cir.1985) (recognizing that "concealment” can be' accidental or intentional); Matter of Application of Sherwood, 613 F.2d 809, 816 (CCPA 1980) (holding that "evidence of concealment (accidental or intentional) is to be considered”).
Oddly, the majority quotes yet wholly ignores the statement in Sherwood that concealment can be "accidental or intentional.” Sherwood, 613 F.2d at 816 (emphasis added). Instead, the majority focuses on what concealment might mean only in light of In re Gay, 50 CCPA 725, 309 F.2d 769, 772 (1962). However, Sherwood, not In re Gay, is the precedent we must follow because it is the latest opinion of the Court of Customs and Patent Appeals (“CCPA”). In re Gosteli, 872 F.2d 1008, 1011 (Fed.Cir.1989) (holding that "[t]he CCPA’s later decisions control because that court always sat en banc”).
. U.S. Gypsum Co. v. Nat'l Gypsum Co., 74 F.3d 1209, 1215-16 & n. 7 (Fed.Cir.1996) (holding that intent is not required for a best mode violation and noting that the discussion of intent in Brooktree was dicta); see also, e.g., In re Cyclobenzaprine Hydrochloride Extended-Release Capsule Patent Litig., 676 F.3d 1063, 1085 (Fed.Cir.2012) (holding that "the proper inquiry [in a best mode analysis] focuses on the adequacy of the disclosure rather than motivation for any nondisclosure”), cert. denied, - U.S. -, 133 S.Ct. 933, 184 L.Ed.2d 725 (2013); Star Scientific, Inc. v. R.J. Reynolds Tobacco Co., 655 F.3d 1364, 1373 (Fed.Cir.2011); Graco, Inc. v. Binks Mfg. Co., 60 F.3d 785, 789-90 (Fed.Cir.1995) (holding that "specific intent to deceive is not a required element of a best mode defense”). Indeed, the majority acknowledges that "earlier panel decision[s] control! ] unless overruled en banc,” but they apparently believe Brooktree to be first in time. Majority Op. at 1358 n. 3 (citing Newell Cos. v. Kenny Mfg. Co., 864 F.2d 757, 765 (Fed.Cir.1988)).
. To support its holding that concealment must be intentional, the majority relies on three of our more recent opinions, Wellman, *1361Inc. v. Eastman Chemical Co., 642 F.3d 1355, 1365 (Fed.Cir.2011), High Concrete Structures, Inc. v. New Enterprise Stone & Lime Co., 377 F.3d 1379, 1384 (Fed.Cir.2004), and Cardiac Pacemakers, Inc. v. St. Jude Medical, Inc., 381 F.3d 1371, 1378 (Fed.Cir.2004). But that line of precedent adopts the intentional concealment requirement from Brooktree that we already expressly held in U.S. Gypsum to be nonbinding (and erroneous) dicta. In my view, these cases clearly conflict with earlier precedent, which we are bound to follow, see Newell, 864 F.2d at 765, and have already been dispensed with in dozens of our cases that recognize that the concealment inquiry is objective, see supra note 2.
. The majority declares that "[i]t is not disputed that [a] three groove passageway was the best mode.” Majority Op. at 1358. The majority also claims that “it is not disputed that the preferred embodiment when the patent application was filed was the three-groove embodiment that the specification describes as preferred.” Id. The majority cites no support for those conclusions. I do not believe there is any. The identity of'the best mode and the preference for the embodiments described in'the patents are hotly contested questions in this appeal. And the answer to them formed the basis for the district court’s decision that the best mode requirement was violated because the inventors’ best mode was an "odd number” of grooves and that preference for that mode was not adequately disclosed in the patents.
. Even if the majority disagrees with my analysis of the facts in this case, I see no basis to reverse the decision of the district court. Compliance with the best mode requirement is a question of fact. To reverse the district court would require holding that no reasonable jury could find the facts as both the district court and I see them to be plainly laid out in the record. I do not believe that conclusion can be fairly reached on the record in this case.