dissenting.
The only issue remaining in this case is the reapplication of the law of equivalency, as adjusted by the Supreme Court in Festo VIII, 535 U.S. 722, 122 S.Ct. 1831, 152 L.Ed.2d 944 (2002), to the technological facts of the accused SMC rodless cylinder. However, this court has confounded the issue by creating a new and incorrect criterion for the measurement of “foreseeability,” the court now holding that an existing structure need not be recognized, or even recognizable, as an equivalent at the time of the patent application or amendment, in order to be “foreseeable” if it is later used as an equivalent. This is a significant departure from precedent, as well as from the Court’s guidance in understanding the estoppels that arise when claims are amended during prosecution. Today’s new rule further erodes the residue of the doctrine of equivalents, for its foreseeable result is to deprive amended claims of access to the doctrine of equivalents.
The panel majority reviews only the magnetizable sleeve aspect of the district court’s opinion, and I limit my remarks to that element. The question is the foreseeability of the SMC sleeve made of aluminum alloy, when the '125 patent specification describes only a sleeve made of magnetizable metal. The claims were all limited to the magnetizable metal sleeve by voluntary amendment, for there was no rejection related to this aspect.1 The court now rules that the equivalency of the magnetizable and nonmagnetizable metal was foreseeable as a matter of law, because a nonmagnetizable sleeve previ*1384ously was known although its technological equivalency was unknown. The panel majority holds that “An equivalent is foreseeable if one skilled in the art would have known that the alternative existed in the field of art as defined by the original claim scope, even if the suitability of the alternative for the particular purposes defined by the amended claim scope were unknown.” That is, even if unforeseeable as a matter of fact, even if technologically unexpected or unlikely, the equivalent must be ruled to be foreseeable if the structure is later found to be a usable equivalent.
Festo’s position throughout this litigation is that SMC’s aluminum alloy sleeve actually provides sufficient shielding for the magnetic field leakage of the SMC device, thereby meeting the function/way/result criteria of equivalency as Graver Tank; and in response to the new issue of this remand, Festo points out that the use of aluminum alloy to this effect was not foreseeable when Festo’s '125 patent application was filed and prosecuted. Thus Festo argues that it is not estopped to reach the SMC aluminum alloy sleeve under the doctrine of equivalents. However, the panel majority holds that this equivalent was foreseeable as a matter of law, and thus barred from access to the doctrine of equivalents.
This is not what the Court held. The Court held that an accused structure that is not within the literal scope of amended claims, but that is the technological equivalent thereof, cannot be captured under the doctrine of equivalents if the claims before amendment did not exclude that structure and the equivalency was “readily” foreseeable. My colleagues now hold that such structures can be foreseeable as a matter of law, although the equivalency was unknown in fact. My colleagues also deem it irrelevant whether the accused structure could have been included in allowable claims.
Although the technological equivalency of an accused structure is determined as of the time of the infringement, foreseeability is determined as of the time of filing and amendment of the patent application. The Court so recognized in Festo VIII, in its remand for finding whether the assertedly equivalent aluminum alloy sleeve used by SMC was “unforeseeable at the time of the application.” 535 U.S. at 740, 122 S.Ct. 1831. We then remanded to the district court with instructions to find “whether an ordinary skilled artisan would have thought an aluminum sleeve to be an unforeseeable equivalent of a magnetizable sleeve in the context of the invention.” Festo X, 344 F.3d at 1371-72.
The district court found that “any shielding capability of non-magnetizable aluminum alloy was unknown in the literature or to one of ordinary skill in 1981,” and that “no one disputes that the use of an aluminum alloy sleeve for purposes of shielding magnetic fields was not known at the time of amendment.” Festo XI, slip op. at 2. The district court nonetheless ruled that the use of aluminum alloy for shielding magnetic fields was foreseeable because the SMC magnetic leakage fields are “very small” and therefore little shielding is needed. The district court erroneously determined foreseeability based on the particular properties of the SMC device at the time of infringement, instead of *1385on the knowledge at the time the '125 patent application was filed and presented.
My colleagues on this panel do not directly endorse the district court’s reasoning, although similar reasoning is subsumed in the panel majority’s holding that it is irrelevant whether the later-observed technologic equivalency was foreseeable at the time of the amendment. Instead, the panel majority rules that the aluminum alloy shield was retrospectively foreseeable at the time of the amendment because it later was used as an equivalent, although it was not known to be equivalent and would not have been deemed equivalent at the time of the amendment. Hindsight is not foreseeability. No error has been shown in the district court’s finding that persons of skill in the field of the invention would not have deemed magnetizable and non-magnetizable sleeves to be equivalent at the time the '125 patent application was filed and prosecuted. It cannot be irrelevant that the then-existing knowledge in the field of the invention would not have deemed an aluminum alloy sleeve equivalent to a magnetizable metal sleeve in shielding magnetic leakage fields.
Evidence of foreseeability must be limited to prior art, not future art. It is not the correct interpretation of foreseeability, to rely on SMC’s later equivalent use of an aluminum alloy sleeve to prove that the nonmagnetizable aluminum alloy was a foreseeable equivalent of a magnetizable metal, “at the time of the application” for the '125 patent.2 Foreseeability is determined as of the time of the application. The unforeseen does not become foreseeable after someone later discovers it. If the prior art does not support a finding of equivalency, the applicant cannot be charged with foreseeability of the equivalent.
My colleagues compound their error, in holding that “foreseeability does not require the applicant to be aware that a particular equivalent would satisfy the insubstantial differences test or the function/way/result test with respect to the claim as amended.” Maj. op. at 1370. Indeed, if the particular technology is not recognized as equivalent at the time of the application — whether recognized by the applicant or by others of skill in the field— that technology cannot be foreseeable. Foreseeability cannot be found as fact if the equivalence of magnetizable metal and aluminum alloy sleeves was not reasonably known at the time of filing and prosecution. In Graver Tank the Court explained: “An important factor is whether persons reasonably skilled in the art would have known of the interchangeability of an ingredient not contained in the patent with one that was.” Graver Tank & Mfg. Co. v. Linde Air Prods. Co., 339 U.S. 605, 609, 70 S.Ct. 854, 94 L.Ed. 1097 (1950). The inquiry surrounding when one becomes “aware” of equivalence must focus on the era of the application. If persons of ordinary skill in this field cannot be charged with foreseeing that an aluminum alloy sleeve would be a technological equivalent to a magnetizable metal sleeve, then the aluminum sleeve does not meet the criterion of foreseeability. The panel majority is incorrect in ruling that “the foreseeability requirement does not require the knowledge that the equivalent would satisfy the function/way/result test or the insubstantial differences test.” Maj. op. at 1380. How can a particular equivalent be foreseeable, if it was not known that the tech*1386nology was equivalent in the context of the invention?
I also take note of the panel majority’s unnecessary obfuscation of the technological criteria of equivalency as founded on Graver Tank, where the Court explained that “What constitutes equivalency must be determined against the context of the patent, the prior art, and the particular circumstances of the case,” 339 U.S. at 609, 70 S.Ct. 854, the Court defining an equivalent as one that “performs substantially the same function in substantially the same way to obtain the same result.’ ” Id. at 608, 70 S.Ct. 854. The function/way/result test was endorsed by the court in Festo VIII, along with the insubstantial-differences test, and is no more susceptible to “endless bickering” as any other fact-dependent issue.
The panel majority states that it recognizes that the applicant can be charged with foreseeability only as to “[equivalents] known before the amendment, not equivalents known after the amendment,” but then contravenes this statement, for my colleagues hold that it suffices to estop access to the doctrine of equivalents if the then-believed to be nonequivalent structure was a known structure in the field of the invention. My colleagues rule that it is irrelevant that the structure was believed not to be equivalent, as long as an original claim before amendment could have generally included a device having that structure. Such a complete bar based on amendment revives the approach rejected by the Court when it called our previous assault on the doctrine of equivalents a “complete bar by another name,” Festo VIII, 535 U.S. at 741, 122 S.Ct. 1831, the Court overturning the Federal Circuit’s holding in Festo VI.
The fact of whether technologic equivalency was known cannot be irrelevant to foreseeability, for the foreseeability bar is directed to subject matter that was fore-seeably equivalent at the time of filing and amendment. This is evident from Festo VIII’s guidance to “readily known equivalents.” 535 U.S. at 740, 122 S.Ct. 1831. If the equivalency of the accused technology would not have been “readily known” to a person of ordinary skill in the field of the invention, under proper application of the Court’s guidance the “foreseeability” bar does not arise. In Festo VIII the Court ratified the value of the doctrine of equivalents to “ensurfe] the appropriate incentives for innovation,” id. at 732, 122 S.Ct. 1831, by rejecting this court’s ruling that had led to “excessive uncertainty and burdens legitimate innovation.” Id. at 737, 122 S.Ct. 1831. The panel’s holding today strays from controlling precedent as well as from logic. I respectfully dissent.
. In Festo VIII the Court explained that “[a] rejection indicates that the patent examiner does not believe the original claim could be patented. While the patentee has the right to appeal, his decision to forgo an appeal and submit an amended claim is taken as a concession that the invention as patented does not reach as far as the original claim.” Festo *1384VIII, 535 U.S. at 734, 122 S.Ct. 1831. Here there was no substantive rejection but only an objection due to claim form.
. In using these words, the Court did not distinguish among the time of filing, amendment, or grant; such distinction need not be resolved on the facts of this case.