Opinion for the court filed by Circuit Judge MICHEL. Opinion concurring in the result filed by Senior Circuit Judge NIES.
MICHEL, Circuit Judge.
Athletic Alternatives, Inc. (“AAI”) appeals from the June 29, 1994 order of the U.S. District Court for the District of Arizona, No. CIV-92-176-PHX-RGS (SLV), granting summary judgment of noninfringement in favor of Prince Manufacturing, Inc. (“Prince”).1 Because AAI cannot lawfully prevail on its infringement allegations against Prince when the asserted patent claim is correctly construed, we affirm.
Background
AAI designs athletic products, including string systems for sports rackets, and Prince manufactures and distributes tennis rackets. In February 1990, after entering into a confidentiality agreement, the parties began a collaborative effort to develop a commercially available tennis racket with splayed strings, ie., with string ends anchored to the racket frame alternately above and below its central plane. AAI had already conceived of and applied for a patent covering its prototype “Redemption Stringing System” in August 1988, and it shared the prototype with Prince under the confidentiality agreement. During the course of their collaboration, Prince and AAI shared other designs and ideas but ultimately failed to reach a mutually satisfactory licensing agreement.
Prince abandoned work on AAI’s prototype in favor of an alternative splayed string system. In February 1991, it placed a racket with this system on the market under the model name “Vortex.” The sides of the Vortex frame include strings splayed from the central plane at two, and only two, distinct offset distances: a distance of 2 millimeters at the upper and lower corners, and a dis-[1575]*1575tanee of 4.5 millimeters along the sides of the frame.
AAI, meanwhile, continued to prosecute its splayed string system patent application. As originally filed, the application contained 19 claims, the first of which broadly claimed “[a] sports racket ... wherein at least some of [the] ends of [the] string segments are successively and alternately secured to [the] frame at locations a distance d¡ in front of and behind [the] central plane.” Claim 5 of the application, depending from Claim 1, recited a racket “wherein d¡ is uniform for all strings.” Claim 7, also drafted in dependent form, recited a racket “wherein [the] distance d¡ ... varies continuously between mínimums as small as about zero for the lateral strings near the tip and the heel portions of said frame, and a maximum of up to about % inch for the ends of the lateral string segments through the center of the side portions of [the] frame.” This language, unlike that of Claim 1, closely tracked the specification’s description of the preferred embodiment.2
In a May 1989 office action, the examiner rejected, inter alia, Claims 1 through 5 and objected to Claim 7. Specifically, the examiner rejected Claims 1 and 5 as anticipated by British Patent No. 223,151 to Lewis (“Lewis”). Like AAI’s Claim 5, Lewis discloses a racket in which all the string ends are offset from the central plane by a uniform distance. Having rejected Claim 1, the independent claim from which it ultimately depended, the examiner objected to Claim 7. At the same time, however, the examiner indicated that Claim 7 contained patentable subject matter and thus “would be allowable if rewritten in independent form.”
In response to this office action, AAI filed a set of amendments to the application in October 1989. First, AAI canceled Claim 1 in favor of a new Claim 20, reciting “[a] sports racket ... where [the offset] distance d¡ varies between a minimum distance for the first and last string ends in [the] sequence [of adjacent string ends] and a maximum distance for a string end between [the] first and last string ends in [the] sequence.” After noting that it “rewrote Claim 1 to more succinctly define the present invention ... [in] Claim 20”, AAI explained the responsiveness of the new claim to the rejection on Lewis as follows:
Lewis is primarily concerned with the problem of splitting of wooden frames where all holes are located in the center plane of the frame. Another objective is to stabilize the frame against twisting. He places the holes somewhat away from and alternately above and below the center plane. Consequently, the stringing surface superficially resembles the present string suspension arrangement. However, we do not believe that the degree of splay or separation is sufficient to achieve the present effect of improved performance and feel of the string surface. Moreover, the Lewis racket is structurally different in that it maintains the same [s]play (d¡) all the way around.
Claim 1 has now been rewritten and resubmitted as Claim 20 to positively recite the feature of the splay (d¡), varying between a minimum and a maximum for a series of adjacent parallel strings, to thus clearly distinguish over Lewis.
In other words, according to this portion of the record, AAI drafted Claim 20 in a manner calculated to retain as much of the scope of Claim 1 as possible while still avoiding the Lewis reference.
Consistent with the examiner’s suggestion, AAI also redrafted Claim 7 in independent form and submitted it as Claim 21. It recited “[a] sports racket ... wherein [the] distance d¡ ... varies continuously between mínimums as small as about zero for the ends of laterial [sic, lateral] strings near the tip and heel portions of [the] frame, and maximum of up to about %-inch for the ends of lateral string segments near the center of [the] side portion of [the] frame.” As with Claim 7, the language of Claim 21 closely tracks that of the specification’s description of the preferred embodiment. Claim 21 thus [1576]*1576differed from Claim 20 in two ways: first, with respect to the limitation directed to the pattern of splay, Claim 20 recited an offset distance that “varies between” minimum and maximum values, whereas Claim 21 recited an offset distance that “varies continuously between” such values (emphasis added); second, Claim 21 recited particular numerical values for the minimum and maximum offset distances, whereas Claim 20 did not. Finally, AAI added a new claim that, like the rejected Claim 1, broadly covered any pattern of splayed string ends.3
The examiner’s Interview Summary Record indicates that the parties discussed proposed Claims 20 and 21 during a personal interview on November 16, 1989. In sharp contrast to the explanation AAI offered with the amendments that it filed, the summary notes that AAI’s “[ajttorney explained how el. 20 is directed to [the] preferred embodiment and how [its] recitation distinguishes over [the] cited references.” This notation calls into serious doubt the premise that AAI drafted Claim 20 with the sole purpose of overcoming the Lewis-based rejection of Claim 1.
After reviewing the submitted amendments, the examiner rejected Claims 20 through 22 in the final office action of January 29, 1990. Specifically, the examiner concluded that all three claims would have been obvious from Lewis in view of U.S. Patent No. 4,664,380 to Kuebler. AAI contested the rejection, arguing that the Kuebler patent, directed to improving the stiffness of racket frames, teaches nothing at all about splayed string ends and reiterating that “Lewis ... exhibits splay, but uniform splay all the way around the racket frame.” According to AAI, its “previous amendment ... introduced the structural limitations of variously specifying the non-uniform splay characteristics which clearly differentiate the present invention from the 1924 wooden tennis racket art represented by the Lewis reference.” The examiner nevertheless adhered to his final rejection of the claims by notice dated April 17,1990.
AAI appealed the final rejection to the Board of Patent Appeals. Consistent with the Interview Summary Record discussed above, and in contrast to the supporting explanation filed with the amendments, AAI implied in its appeal brief that Claims 20 and 21 had the same scope with respect to the claimed pattern of splay. First, early in its appeal brief, AAI noted that “Claims 20 and 21 represent independent claims based on rewritten dependent claim[ ] 7 ... which [was] formerly held to be allowable if rewritten in independent form.” Second, in explaining the nature of the invention, AAI stated that “[i]n the preferred embodiment, best seen in Fig. 4a, the distance d¡ varies between minima for the first and last lateral strings near the tip section 12 and heel section 13, respectively, and a maximum in the central region of the lateral section of the frame.” (Emphasis added). AAI thus described the preferred embodiment as a pattern in which the offset distance “varies between” a minimum and a maximum without using the intervening modifier “continuously.” Finally, AAI affirmed that “Claim 20 requires the distance d, (the degree of splay) to vary between a minimum ... and a maximum,” and that “Claim 21 ... recites the same element of varying splay of the ends of lateral string segments.” (Emphasis added). After further communication between AAI and the examiner, as well as an amendment deleting Claim 22, the examiner withdrew the final rejection and issued a Notice of Allowability in December 1990.
AAI was accordingly granted U.S. Patent No. 5,037,097 (the ’097 patent) on August 6, 1991. It then filed the infringement suit at bar in January 1992. Claim 1 of the ’097 patent, formerly Claim 20 of the application, was the only claim AAI asserted against Prince. It reads as follows:
1. A sports racket, comprising:
a peripheral frame having a tip, heel, and side portions disposed in and about a central plane of symmetry; said tip, heel, and side portions having a width in [1577]*1577a direction perpendicular to said plane of symmetry;
a handle with a grip thereon extending from and secured to the heel of said frame;
a first plurality of string segments extending across said central plane between opposite locations of said frame in a first direction;
a second plurality of string segments extending across said central plane between opposed locations of said frame in a second direction;
said first and second direction being at an angle with respect to each other, and said first and second plurality of string segments being interwoven into a planar webbing defining a ball contact area in said central plane within said frame;
each of said string segments having two ends, a first of said ends extending in a first direction between the periphery of said contact area and a proximate location on said frame, and the second of said ends extending between the opposite side of the periphery of said ball contact area to a proximate location on said frame in a direction opposite to said first direction;
where at least said first ends of at least said first plurality of string segments are secured to said frame at a distance dj, where d¡ is the perpendicular distance between the central plane and the location on said frame to which the ith string end is secured, i designating the order of the ith string end in the sequence of adjacent first string ends of said first plurality of string segments, the distance dj being alternately measured in opposite directions from said central plane, and where said distance d/ varies between minimum distances for the first and last string ends in said sequence and a maximum distance for a string end between said first and last string ends in said sequence.
The language critical to this appeal has been emphasized. Claim 14 of the ’097 patent, formerly Claim 21 of the application, is the patent’s only other independent claim. Again, it covers AAI’s preferred pattern of variation in the string offset distances, according to which distance d¿ “varies continuously between mínimums as small as about zero for the ends of lateral strings near the tip and heel portions of said frame, and a maximum of up to about ié-inch for the ends of lateral string segments near the center of said side portions of said frame.”
On December 7, 1992, Prince moved for summary judgment of noninfringement on the grounds that (a) Claim 1 of the ’097 patent, which, according to its construction, requires that the offset distance d¿ take on at least three values (ie., a minimum, a maximum, and at least one intermediate value), does not read on the Vortex racket, which has only two offset distances; and (b) the prosecution history of the ’097 patent, during which AAI amended Claim 1 to require this minimum of at least three values for d in order to overcome the rejection on Lewis, estops AAI from attempting to hold Prince liable for infringement under the doctrine of equivalents.4
The district court granted Prince’s motion for summary judgment over AAI’s opposition in its June 29, 1994 order, adopting the substance of Prince’s arguments as to both claim construction and prosecution history. Specifically, the court concluded that “according to the plain language of Claim 1, d; must change at a point that is in an intermediate position in relation to the minimum distances for the first and last string ends and the maximum distance for a string end between the first and last string end. Consequently, there must be more than two values for d¡.” The court further concluded that AAI was estopped from alleging that the Vortex racket infringes Claim 1 under the doctrine of equivalents, having already sacrificed cover[1578]*1578age of a splay pattern with less than three offset distances in order to overcome the rejection on Lewis. Providing an alternative ground for its decision of noninfringement under the doctrine of equivalents, the district court also concluded that AAI’s failure to respond to Prince’s motion for summary judgment with citation to “specific facts from which a factfinder could conclude that Defendant’s racket ‘performs substantially the same function in substantially the same way to obtain the same result’ ” entitled Prince to summary judgment in its favor on the issue.
AAI appeals from the district court’s decision, contending that Claim 1, when properly construed, reads on the splay pattern with two offset distances used in the Vortex racket'. In other words, according to AAI, the district court erred in construing Claim 1 to require that the offset distance d, take on at least three values. AAI further contends that the district court repeated its error in claim construction when it undertook its prosecution history estoppel analysis and erred by entering judgment against AAI for failure to point to specific record evidence of the fact of equivalence at a time when AAI was not on notice that such citation to specific evidence was required to resist an adverse summary judgment.
Analysis
The moving party is entitled to summary judgment under Federal Rule of Civil Procedure 56(c) “if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.” Fed.R.Civ.P. 56(e). We review the district court’s grant of summary judgment de novo, Conroy v. Reebok Int'l, Ltd., 14 F.3d 1570, 1575, 29 USPQ2d 1373, 1377 (Fed.Cir.1994). The disposition of this appeal turns largely on whether the trial court properly construed Claim 1 of the ’097 patent, a question of law that wé review de novo. Markman v. Westview Instruments, Inc., 52 F.3d 967, 979, 34 USPQ2d 1321, 1329 (Fed.Cir.) (in banc), cert. granted, — U.S. -, 116 S.Ct. 40, 132 L.Ed.2d 921 (1995).
A Literal Infringement
Patent infringement analysis involves two steps: the threshold construction of the meaning and scope of the asserted claim, followed by the determination whether the accused product infringes the properly construed claim. Id. at 976, 34 USPQ2d at 1326. Where, as here, the parties do not dispute any relevant facts regarding the accused product but disagree over which of two possible meanings of Claim 1 is the proper one, the question of literal infringement collapses to one of claim construction and is thus amenable to summary judgment. See, e.g., Chemical Eng’g Corp. v. Essef Indus., 795 F.2d 1565, 1572-73, 230 USPQ 385, 390 (Fed.Cir.1986).
The task of claim construction requires us to examine all the relevant sources of meaning in the patent record with great care, the better to guarantee that we determine the claim’s true meaning. As we have often noted, these sources include the patent’s claims, specification, and, if in evidence, its prosecution history. Markman, 52 F.3d at 979-80, 34 USPQ2d at 1329-30. In addition, a number of canons, such as the doctrine of claim differentiation, guide our construction of all patent claims. Finally, we have a rich body of case law on which to rely.
1. The Claim Language
We begin with the language of Claim 1 itself. As was noted above, Claim 1 of the ’097 patent recites a pattern of splay “where [the offset] distance d¿ varies between minimum distances for the first and last string ends in [the] sequence and a maximum distance for a string end between [the] first and last string ends in [the] sequence.” Because the specification contains neither a definition of the phrase “varies between” nor a suggestion that AAI sought to assign to claim terms anything but their ordinary and accustomed meanings, those are the meanings we must give them. Intellicall, Inc. v. Phonometrics, Inc., 952 F.2d 1384, 1388, 21 USPQ2d 1383, 1386 (Fed.Cir.1992); Envirotech Corp. v. Al George, Inc., 730 F.2d 753, 759, 221 USPQ 473, 477 (Fed.Cir.1984).
[1579]*1579AAI contends, citing a number of dictionary definitions of “between” in its favor, that “the plain meaning of the words ‘varies between mínimums ... and a maximum’ is that the distance of string splay must change ... [but] nothing in Claim 1 prescribes (or limits) how much or how many times the distance must change.”5 Prince, for its part, cites alternative and equally valid dictionary definitions of “between” in an effort to demonstrate that the trial court - properly construed the claim to require that the offset distance d¿ take on at least three values, i.e., a minimum, a maximum, and at least one intermediate value.6 A comparison of these definitions with those relied upon by AAI suggests that the claim phrase “varies between” is, at best, equivocal.
According to one ordinary sense of “between” — “in the space that separates: BETWIXT <an alleyway" two tall buildings > <a vacuum " two electrodes > : in the midst of : surrounded by <a lion rampant" eight crosses>,” Webster’s Third New International Dictionary of the English Language 209 (1976) (definition 2b of preposition “between”) — only a splay pattern in which the offset distance takes on more than two values has an offset distance that varies between minimum distances and a maximum distance. According, however, to another equally ordinary sense of “between” — “from one to the other of <air service ' the two cities>,” id. (definition 3a of preposition “between”) — a splay pattern in which the offset distance takes on only two values has an offset distance that varies between minimum distances and a maximum distance. In sum, the dis-positive claim language on its face is susceptible to two equally plausible meanings, under one of which the Vortex racket literally infringes Claim 1, and under the other of which it does not. As a result, the scope of Claim 1 cannot be defined by resort to the ordinary and accustomed meanings of its terms alone, and the specification is completely silent with regard to the meaning of “varies between.”
2. The Prosecution History
Because the disputed claim and its specification do not establish the meaning of the operative claim language, we turn to the prosecution history for guidance. Our reading of the prosecution history is informed by the understanding that the documents recording this history are created by persons schooled in the art of patent prosecution.
To the trained eye, the prosecution history of the ’097 patent presents a muddled and self-contradictory story. As originally filed, independent Claim 1 broadly claimed a racket with any pattern of splay, while dependent Claims 5 and 7 each claimed a particular pattern of splay. Specifically, Claim 5 recited the pattern of uniform splay, while Claim 7 recited the pattern of splay described as the preferred embodiment in the patent’s specification. Thus, as originally filed, Claim 1 was certainly broad enough to cover a racket with a pattern of splay containing less than three offset distances. The examiner initially rejected Claims 1 and 5 as anticipated by Lewis, a reference teaching the pattern of uniform splay. Such a rejection was, of course, appropriate as to Claim 1 because the pattern of uniform splay taught by Lewis surely anticipates a claim broadly covering a pattern with “at least some” splay.
Before focusing on AAI’s response to this rejection, we pause to note a practical aspect of patent prosecution. Specifically, when amending a claim so as to avoid a rejection based on a particular reference, the skilled patent prosecutor usually seeks' to draft an amendment that narrows the claim only as much as is thought necessary to overcome the rejection. Prior to looking at AAI’s response to the initial rejection, an experienced patent prosecutor might thus have expected AAI to attempt to overcome Lewis by amending Claim 1 to cover all patterns of «ore-uniform splay, including those with only two offset distances.
AAI’s written response to the first office action does not disappoint this expectation. [1580]*1580As was discussed above, AAI’s response included two distinct amendments. First, AAI rewrote Claim 1 as new Claim 20, expressly stating that it was doing so in a manner calculated to overcome the Lewis rejection. Second, AAI rewrote Claim 7 as new Claim 21. Since the subject matter of this latter claim had already been deemed allowable by the examiner, AAI did not discuss Claim 21 beyond briefly remarking that it was an independent version of Claim 7. This brief remark sharply contrasts with AAI’s detailed explanation of why, in its view, Lewis did not anticipate Claim 20, thereby lending strong support to the conclusion that AAI drafted Claim 20 to be broader in scope than Claim 21. We might well conclude that Claim 20 was directed to any non-uniform splay pattern, including a pattern having only two offset distances, were this the only evidence in the file history. It is not.
The record also contains documents clearly indicating that, with respect to the pattern of splay, AAI understood the scope of Claim 20 to approximate that of Claim 21, which both parties concede recites a pattern of splay containing at least three offset distances. The first such document is the examiner’s Interview Summary Record of November 16, 1989, in which the examiner noted that AAI’s “[ajttorney explained how cl. 20 is directed to [the] preferred embodiment and how [its] recitation distinguishes over [the] cited references.” In other words, according to this summary, the intended scope of Claim 20 was not the broadest scope that would nonetheless overcome the Lewis reference, but was instead the same as that of Claim 21, at least with respect to the basic pattern of splay. On this reading, Claims 20 and 21 differed only in that Claim 21, unlike Claim 20, recited specific numerical values for the minimum and maximum offset distances.
AAI’s appeal brief, like the interview summary, features the same characterization of Claim 20. There, as outlined above, AAI stated that both Claims 20 and 21 were based on the allowable subject matter that had originally appeared in Claim 7, reciting the pattern of splay in the preferred embodiment. Furthermore, AAI used the phrase “varies between” to describe the variable splay of the preferred embodiment and maintained that the “element of varying splay” was the same in Claims 20 and 21. Again, these assertions indicate that the scope of Claims 20 and 21 differed only with respect to the particular numerical embodiment described in Claim 21, but not with respect to the basic pattern of splay.
Two strong and contradictory interpretive strands thus run through the patent’s prosecution history. According to the first, Claim 20 was drafted simply to avoid the Lewis reference and differs significantly in scope from Claim 21. According to the second, Claims 20 and 21 recite the same basic pattern of splay, while Claim 21 adds a numerical value limitation. Each strand, considered alone, leads to a coherent and distinct meaning of the disputed claim. One does not prevail over the other, and together they are irreconcilable. The prosecution in this case is thus unhelpful as an interpretive resource for construing the “varies between” claim limitation.
3. Claim Differentiation
Neither the claim, the specification, nor the prosecution history establishes the meaning of the phrase “varies between” in Claim 1. AAI, resorting to an interpretive guide for secondary support of its position regarding the meaning of “varies between,” contends that the doctrine of claim differentiation compels us to interpret the claim as it does, ie., to require that the offset distance take on two or more, rather than three or more, values. According to AAI, while “[tjhere is no question Claim 14 requires three or more different distances of string splay,” it would constitute legal error to read the “continuously” limitation of Claim 14 into Claim 1.
AAI’s argument overlooks the fact that Prince does not contend that Claim 1, like Claim 14, requires that the offset distance recited therein vary continuously between mínimums and a maximum. Continuous variation, as AAI itself concedes, requires a continuous gradual change in the offset distance of parallel strings; this implies that the offset distance for each string differs from that of both adjacent strings. In contrast, [1581]*1581Prince contends only that Claim 1 requires that the offset distance take on three or more values, a potentially broader limitation. For example, suppose a tennis racket were to employ seventeen horizontal strings but only four offset distances. Under Prince’s construction, this arrangement would infringe Claim 1 because it employs more than three offset distances, but would not infringe Claim 14 because the offset distances do not vary continuously. Therefore, assent to Prince’s contention that Claim 1 requires three or more offset distances would not render the “continuously” limitation in Claim 14 superfluous. In short, the choice between Prince’s and AAI’s construction of Claim 1 simply does not implicate the doctrine of claim differentiation, and we are left with two equally plausible meanings of Claim 1. We must therefore pursue the interpretive process to state which of the two meanings is correct.
4. Section 112
In order to decide which of the two senses of “varies between” to employ in construing the ’097 patent, we refer to the statutory provision that prescribes the would-be paten-tee’s claim drafting burden, 35 U.S.C. § 112 (1988). Specifically, paragraph 2 of section 112 requires that the “specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.” (Emphasis added). As courts have recognized since the requirement that one’s invention be distinctly claimed became part of the patent law in 1870,7 the primary purpose of the requirement is “to guard against unreasonable advantages to the patentee and disadvantages to others arising from uncertainty as to their [respective] rights.” General Electric Co. v. Wabash Appliance Corp., 304 U.S. 364, 369, 58 S.Ct. 899, 902, 82 L.Ed. 1402 (1938). See, e.g., McClain v. Ortmayer, 141 U.S. 419, 424, 12 S.Ct. 76, 77, 35 L.Ed. 800 (1891) (“The object of the patent law in requiring the patentee [to distinctly claim his invention] is not only to secure to him all to which he is entitled, but to apprise the public of what is still open to them.”); Rengo Co. v. Molins Mach. Co., 657 F.2d 535, 551 (3d Cir.) (“Its purpose is to demarcate the boundaries of the purported invention, in order to provide notice to others of the limits beyond which experimentation and invention are undertaken at the risk of infringement.”) (internal quotation omitted), cert. denied, 454 U.S. 1055, 102 S.Ct. 600, 70 L.Ed.2d 591 (1981); Hoganas AB v. Dresser Indus., 9 F.3d 948, 951, 28 USPQ2d 1936, 1939 (Fed.Cir.1993) (function of claims is “putting competitors on notice of the scope of the claimed invention”). Were we to allow AAI successfully to assert the broader of the two senses of “between” against Prince, we would undermine the fair notice function of the requirement that the patentee distinctly claim the subject matter disclosed in the patent from which he can exclude others temporarily. Where there is an equal choice between a broader and a narrower meaning of a claim, and there is an enabling disclosure that indicates that the applicant is at least entitled to a claim having the narrower meaning, we consider the notice function of the claim to be best served by adopting the narrower meaning.
We conclude that Claim 1 of the ’097 patent includes the limitation that the splay-creating string end offset distance take on at least three values, i.e., a minimum, a maximum, and at least one intermediate value. We thus affirm the district court’s conclusion that Claim 1 does not literally read on the Vortex racket.
B. Infringement Under the Doctrine of Equivalents
Under the doctrine of equivalents, an accused product that differs from the claim, and thus does not literally infringe, nonetheless infringes if its difference from that claim is insubstantial from the perspective of one of ordinary skill in the relevant art. Hilton Davis Chem. Co. v. Warner-Jenkinson Co., Inc., 62 F.3d 1512, 1518, 35 USPQ2d 1641, 1645 (Fed.Cir.) (in banc), petition for cert. filed, 64 U.S.L.W. 3349 (U.S. Nov. 14, 1995) (No. 95-728). As we have [1582]*1582often observed, however, the doctrine of equivalents is not a license to ignore or “erase ... structural and functional limitations of the claim,” limitations “on which the public is entitled to rely in avoiding infringement.” Perkin-Elmer Corp. v. Westinghouse Elec. Corp., 822 F.2d 1528, 1532, 3 USPQ2d 1321, 1324 (Fed.Cir.1987). Liability for infringement thus requires, without exception, that an accused product contain each limitation or its equivalent.8 Dolly, Inc. v. Spalding & Evenflo Cos., 16 F.3d 394, 398, 29 USPQ2d 1767, 1769 (Fed.Cir.1994); Pennwalt Corp. v. Durand-Wayland, Inc., 833 F.2d 931, 935, 4 USPQ2d 1737, 1739 (Fed.Cir.1987) (in banc), cert. denied, 485 U.S. 961, 108 S.Ct. 1226, 99 L.Ed.2d 426, 485 U.S. 1009, 108 S.Ct. 1474, 99 L.Ed.2d 703 (1988). Two additional principles further circumscribe the doctrine’s sweep: it cannot be used to protect subject matter in, or obvious in light of, the prior art, Wilson Sporting Goods Co. v. David Geoffrey & Assocs., 904 F.2d 677, 684, 14 USPQ2d 1942, 1948 (Fed.Cir.), cert. denied, 498 U.S. 992, 111 S.Ct. 537, 112 L.Ed.2d 547 (1990), nor can it be used to recapture subject matter that was relinquished during prosecution of the patent, Hoganas AB, 9 F.3d at 951-52, 28 USPQ2d at 1939.
As we noted above, the district court granted Prince’s motion for summary judgment of noninfringement under the doctrine of equivalents on two grounds. First, the district court held that AAI surrendered coverage of a two-offset-distanee splay pattern during prosecution of the ’097 patent and was thereby estopped from asserting that the Vortex racket infringes Claim 1. Second, the district court ruled that AAI failed to shoulder its burden of citing specific record evidence sufficient to raise a triable issue as to the fact of equivalence and thus to withstand Prince’s motion for summary judgment. AAI appeals from both conclusions, raising close questions as to the soundness of the district court’s reasoning.
While it is true that a patentee’s unmistakable assertions to the PTO in support of patentability, whether or not they were truly required to secure allowance of the claim, will estop the patentee from obtaining protection of the subject matter surrendered thereby, Kinzenbaw v. Deere & Co., 741 F.2d 383, 389, 222 USPQ 929, 933 (Fed.Cir.1984), cert. denied, 470 U.S. 1004, 105 S.Ct. 1357, 84 L.Ed.2d 379 (1985), it is difficult to conclude, in light of the truly equivocal nature of the claim phrase “varies between” and the conflicting evidence in the prosecution history, that AAI surrendered protection of the two-offset-distance splay pattern with the clarity that prosecution history estoppel requires. It is also difficult to conclude that the judgment of noninfringement can be affirmed on the alternative ground stated by the district court. Its entry of judgment for Prince due to AAI’s purported failure to cite specific evidence of record regarding the insubstan-tiality vel non of the difference between Claim 1 and the Vortex racket cannot stand unless we conclude that AAI “was on notice that [it] had to come forward with all of [its] evidence” on this precise issue. Celotex Corp. v. Catrett, 477 U.S. 317, 326, 106 S.Ct. 2548, 2554, 91 L.Ed.2d 265 (1986). Could AAI have reasonably thought it was only required to respond to Prince’s assertions regarding prosecution history estoppel in order to withstand summary judgment on the doctrine of equivalents question? Again, this presents a close question. Given the more fundamental flaw in AAI’s case, however, we need not resolve the challenges AAI presses against the two grounds relied upon by the district court.
As a corollary to the “all limitations” rule discussed above, we have held that “the concept of equivalency cannot embrace a structure that is specifically excluded from the scope of the claims.” Dolly, Inc., 16 F.3d at 400, 29 USPQ2d at 1771. Applying this formulation to the undisputed facts of the instant case, we conclude that the intermediate offset distance required by the properly construed claim cannot have an equivalent in a racket with only two offset distances. In other words, the two-distance splayed string system was “specifically excluded from the scope of the claims.” Id. To hold Prince liable for infringement of Claim 1 of the ’097 patent for its production and sale of the [1583]*1583Vortex racket would thus run afoul of our holding in Dolly, Inc. and the “all limitations” rule from which it derives. As a result, AAI is precluded as a matter of law from successfully asserting that the Vortex racket infringes Claim 1 of the ’097 patent under the doctrine of equivalents.
This conclusion, based on undisputed facts of record, establishes the correctness of district court’s summary judgment of nonin-fringement under the doctrine of equivalents, thereby eliminating any need for further review of the district court’s decision and rendering remand unnecessary. See Consolidated Aluminum Corp. v. Foseco Int’l Ltd., 910 F.2d 804, 814, 15 USPQ2d 1481, 1489 (Fed.Cir.1990) (“An appellate court need not close its eyes to the record where, as in this case, there is a way clearly open to affirm the district court’s action.”); Bio-Rad Labs., Inc. v. Nicolet Instrument Corp., 807 F.2d 964, 969, 1 USPQ2d 1191, 1195 (Fed.Cir.1986) (noting the principle that “a court of appeals may affirm the judgment of a district court on any ground, including grounds not relied upon by the district court” itself), cert. denied, 482 U.S. 915, 107 S.Ct. 3187, 96 L.Ed.2d 675 (1987).
CONCLUSION
Because AAI cannot lawfully prevail on its infringement allegations against Prince when the asserted patent claim is correctly construed, summary judgment for Prince was proper. Accordingly, the district court’s order and the judgment of noninfringement are
Affirmed.
Costs
Each party to bear its own costs.