BRYSON, Circuit Judge.
American Seating Corporation makes products for use in mass transit vehicles, including a wheelchair restraint system and a vandal-resistant bus seat. American Seating has patents covering both products, one that covers the wheelchair restraint system and two that cover the van[671]*671dal-resistant seating insert. The USSC Group also manufactures wheelchair restraint systems and vandal-resistant seats. Believing that the USSC Group was infringing its patents, American Seating filed suit in the United States District Court for the Western District of Michigan.
American Seating alleged that its patent on the wheelchair restraint system was infringed by two of the USSC Group’s products: the VPRo I and the VPRo II. American Seating also alleged that its patents on the vandal-resistant seating insert were infringed by the “T2C” product sold by the USSC Group. The district court granted summary judgment of noninfringement as to all three patents. American Seating Co. v. USSC Group, Inc., No. 1:01-CV578 (W.D.Mich. Jan. 30, 2003). We affirm in part, reverse in part, and remand.
I
The first patent, U.S. Patent No. 5,888,-038 (“the ’038 patent”), covers a system for securing wheelchairs in mass transit vehicles. The means for securing the wheelchair in the vehicle consists of two parts: (1) a means for engaging a portion of the wheelchair under tension; and (2) a movable securing element (a movable arm) with a housing mounted on it that receives one end of a flexible and adjustable strap. To secure the wheelchair in place, an operator deploys straps that are fastened to the vehicle (the “means for engaging a portion of [the] wheelchair”) and attaches the straps to the body of the wheelchair. The operator then swings the movable arm from its storage position into an operating position and locks the movable arm into place. Once the arm is in the operating position, the operator attaches the straps that are housed on the movable arm to the wheelchair under tension so that the two sets of straps cooperate to hold the wheelchair in place. Claim 1 of the ’038 patent claims the system as follows:
1. In combination,
(a) a vehicle having an area for receiving a wheelchair, and
(b) means for securing a wheelchair in said area to said vehicle, wherein said means for securing is permanently attached to said vehicle and comprises:
(i) means for engaging a portion of said wheelchair under tension, said means for engaging being locked to said vehicle at a predetermined location, and
(ii) a movable securing element having mounted thereon a housing and flexible strap, said flexible strap having one end adjust-ably received in said housing and an opposite end adapted to engage a portion of said wheelchair in said area under tension, said movable securing element being movable with respect to said vehicle between an operating position wherein said housing is locked to said vehicle at a further location for cooperation with said means for engaging to secure said wheelchair and a storage position wherein said housing is remote from said further location.
The district court construed subsection (b)(i) of the claim to require that the “means for engaging a portion of said wheelchair” be locked directly to the vehicle, rather than to a structure within the vehicle such as a sidewall or bulkhead. The court construed subsection (b)(ii) of the claim to require that the movable securing element containing the housing be locked directly to the floor of the vehicle when in the operating position. The court [672]*672construed the term “locked” to mean “made immobile with respect to the vehicle by use of a mechanical means, the means engaging the vehicle in some manner, which means can be disengaged by mechanical force applied by an operator or user.”
Based on its claim construction, the district court granted summary judgment of noninfringement as to both the VPRo I and the VPRo II. The court held that the VPRo I did not infringe because the locking mechanism on the VPRo I “does not directly interface with the vehicle” but instead “interfaces with a sidewall or bulkhead.” The court held that the VPRo II did not infringe for two reasons: (1) because the VPRo II employed, “at least in part, a friction lock through the use of the wheelchair restraint belts,” rather than a mechanical lock; and (2) because the locking mechanism for the movable securing element did not interact directly “with the vehicle.”
A
With respect to the VPRo I, American Seating argues that the trial court improperly required that the “means for engaging a portion of said wheelchair” set forth in subsection (b)(i) of claim 1 directly interface with the vehicle, and not with a bulkhead or sidewall of the vehicle. We agree with American Seating that the court’s interpretation of the claim was incorrect. The claim requires only that the means for securing the wheelchair be “permanently attached to said vehicle at a predetermined location” and that the means for engaging be locked “to said vehicle.” Nothing in the claim or the specification suggests that the securing and engaging means must be attached to the floor or chassis of the vehicle. To the contrary, the reference to the “vehicle” would seem to include any structure that is a permanently part or permanently affixed to the vehicle, such as a bracket, sidewall, or bulkhead. In fact, in the preferred embodiment the adjustable straps that attach to the rear of the wheelchair are attached to what the patent refers to as the “fixed securing element,” which is depicted in the patent figures as a bulkhead rising from the floor of the vehicle. Thus, even in the preferred embodiment, the straps and hooks that constitute the “means for engaging a portion of said wheelchair” are attached to a structure other than the vehicle floor. We therefore reject the district court’s construction of the claim as requiring that the “means for engaging” the wheelchair be locked to a portion of the vehicle other than a sidewall or bulkhead. Because there was evidence in the summary judgment proceeding that the “means for engaging” the wheelchair in the VPRo I is attached to structures that are permanently attached to the vehicle, the court’s summary judgment of noninfringement cannot be sustained on this ground.
In support of its argument that the VPRo I does not infringe claim 1 of the ’038 patent, the USSC Group argues that the claim language requires “that there be a disengageable type lock between the movable securing element and the vehicle without any intermediate parts.” That argument differs from the ground on which the district court held that the VPRo I did not infringe. The district court rested its ruling with respect to the VPRo I on the “means for engaging” limitation (limitation (b)(i) of claim 1) rather than the “movable securing element” limitation (limitation (b)(ii) of claim 1), to which much of the USSC Group’s argument is directed. In any event, however, the USSC Group’s argument with respect to the VPRo I is unavailing.
[673]*673The USSC Group argues that the claim language providing that the housing must be “locked to said vehicle at a further location” requires that the movable arm and housing must “engage the vehicle at a location separate from where the arm is attached to the vehicle at its pivot point.” We find no support for that argument in either the claim language or the specification. The claim language merely requires that the movable securing element (the movable arm) is movable between an operating position, where the housing is locked to the vehicle, and a storage position, where the housing is stored at some distance from the location it occupies when the arm is in the operating position. Nothing in the claim language requires that the locking mechanism that holds the movable arm in its operating position be located somewhere other than at the pivot point of the arm. The claim is silent as to the location of the locking mechanism as long as the locking mechanism secures the movable arm (and the attached housing) in place with respect to the vehicle when the movable arm is in the operating position.
Moreover, one of the embodiments described in the specification features a locking mechanism that is located near the pivot axis of the movable arm. The specification explains that “[o]ther means” may be used for locking the movable element in place when it is in the operating position. As an example, the specification describes a structure consisting of a disk with apertures, in which
the latch for retaining the movable element in the storage position [is] located near the pivot axis [of the movable arm]. Thus, the end of the movable element near the wall could be provided with an element, such as a disk with apertures for cooperating with a removable pin for holding the disk and the movable securing element in any of several predetermined positions, including the operational and storage positions.
’038 patent, col. 2, II. 44-51. Contrary to the USSC Group’s argument, the patent itself contemplates a structure in which the locking mechanism is located at or near the pivot axis of the movable arm and thus the movable arm and the housing lock to the vehicle at or near the pivot point of the arm.
We also reject the USSC Group’s argument that the “movable securing element” limitation must be construed to require that the locking mechanism connect either the housing or the movable arm directly to the floor of the vehicle with no intervening structures. As in the case of the “means for engaging” limitation, the claim language encompasses devices in which the locking mechanism is attached to a structure, such as a sidewall, bulkhead, or bracket, that in turn is permanently attached to the vehicle.
To the extent that the USSC Group suggests that the VPRo I lacks a “lock” to secure the movable arm, we note that the district court did not so hold, and in any event, we reject that argument. There was evidence before the district court that the VPRo I had features that mechanically secured the movable arm in place in both the operating position and the storage position. In the operating position, the evidence showed, the arm of the VPRo I would not be capable of movement, even in the absence of a wheelchair and the securing belts, unless steps were taken to disengage the locking mechanism. We therefore reverse the district court’s ruling, on summary judgment, that the VPRo I did not infringe the ’038 patent.
B
With respect to the VPRo II, the court based its summary judgment ruling on two grounds: that the movable arm of the [674]*674VPRo II is not locked directly to the vehicle, and that the movable arm of the VPRo II is not immobilized until the straps connecting the arm to the wheelchair are attached. We have already addressed the first point in connection with the USSC Group’s argument regarding the VPRo I. For the reasons given above, we therefore reject the court’s conclusion that the “movable securing element” limitation must be construed to require that the movable arm be locked directly to the vehicle.
The district court’s second ground for holding that the VPRo II did not infringe was that it lacks a “lock” as defined in the ’038 patent. The court defined the term “locked” to mean “made immobile with respect to the vehicle by use of a mechanical means, the means engaging the vehicle in some manner, which means can be disengaged by mechanical force applied by an operator or user.” The court further ruled that, in light of the requirement that the housing be locked to the vehicle, claim 1 does not include systems in which “the locking of the movable securing means occurs by virtue of the presence of the wheelchair.”1 We agree with the court’s definition of “locked,” which is consistent with the definition we approved for the same limitation of the same patent in a previous action brought by American Seating, American Seating Co. v. Transportation Seating, Inc., 62 Fed.Appx. 344 (Fed.Cir.2003). In that case, we held that the claim in question excludes locking mechanisms in which the assembly is “held in position by virtue of its connection to the wheelchair, not to the vehicle.” 62 Fed. Appx. at 350. The dispositive issue in this case is the same-whether the movable arm is secured in place by a mechanical locking system or is secured only when the belts from the housing on the movable arm are attached to the wheelchair under tension.
The district court rejected American Seating’s argument that the VPRo II had a mechanical lock that secured the movable arm and housing in place even in the absence of the wheelchair and tensioned belts. We uphold the court’s ruling in that regard. For its evidence on this issue, American Seating relied principally on the declaration of its expert, Charles F. Reinholtz, who examined the VPRo II in detail and concluded that its movable arm was locked to the vehicle when in the operating position. Dr. Reinholtz’s declaration, however, is more helpful to the USSC Group than to American Seating. His description of the VPRo II makes it clear that even after the VPRo II is placed in the operating position it is not locked in place so as to prevent any movement. Rather, it is only when a lifting force is applied to the end of the arm, such as by tension from a strap hooked to a wheelchair, that a pin attached to the movable arm of the VPRo II device is held in a position within the pivot housing that locks the arm against all motion. As Dr. Reinholtz explained, once each arm of the VPRo II is positioned horizontally, “a lifting force, e.g., due to strap tension or as applied by an operator, biases a pin into a V-shaped slot causing interference with the surface of the slot, and the arm is locked against all motion.” Thus, the tension in the strap is necessary to hold the pivot pin in a position in which the arm is prevented from moving. That evidence is consistent with the evidence offered by the USSC Group to the effect that the arms of [675]*675the VPRo II are not locked into place until and unless the tensioned straps are attached to the arms so that the arms are pulled up sufficiently to bias the pin into the slot that locks the arms in place. We therefore agree with the district court that the evidence did not show that the movable arms of the VPRo II were locked to the vehicle as required by claim 1 of the ’038 patent. For that reason, we uphold the court’s grant of summary judgment as to the claim of literal infringement with respect to the VPRo II.
American Seating has raised a claim that the VPRo II infringes under the doctrine of equivalents. Because American Seating raised that issue before the district court in its response to the USSC Group’s motion for summary judgment, we do not agree with the USSC Group that the claim has been waived. However, because we have rejected in part the district court’s claim construction as applied to the VPRo II, our revised claim construction may affect the proper analysis of the doctrine of equivalents issue. For that reason, the parties and the district court should address that issue in light of the claim construction adopted herein, and we therefore remand that issue to the district court for further consideration.
II
American Seating also appeals the district court’s grant of summary judgment of noninfringement with respect to the ’931 and ’539 patents, which claim a vandal-resistant seating insert. The invention consists of an upright vertical pile fabric affixed to a hard substrate by an adhesive. There are two principal issues on appeal. First, whether the district court correctly construed the term “high strength adhesive” in both patents to require an adhesive with a peel strength of at least 25 pounds per inch and second, whether American Seating disclaimed vertical pile materials of the sort used in the USSC Group’s T2C product.
With regard to the first issue, the district court erred as to both patents. Because the patents contain slightly different language we will treat them separately.
Claim 1 of the ’539 patent requires:
a layer of high strength adhesive interposed between the pile fabric and the substrate, the adhesive at least partially permeating the backing of the fabric without materially impairing the visual appearance of the fabric from the other side, the adhesive bonding the fabric to the substrate with a bond strength sufficient to resist peeling of the fabric from the substrate, the bond strength being such that if upright fibers are grasped in an effort to peel the fabric from the substrate, the fibers will break before the backing will peel from the substrate.
The district court concluded that the claim term “high strength adhesive” was explained in the specification to mean an adhesive with “a stripping strength exceeding 25 pounds per square inch.” Contrary to the district judge’s holding, however, the language of the claim itself is sufficient to set forth the bonding requirement of the adhesive. The claim states that the adhesive bonds the “fabric to the substrate with a bond strength sufficient to resist peeling of the fabric from the substrate, the bond strength being such that if upright fibers are grasped in an effort to peel the fabric from the substrate, the fibers will break before the backing will peel from the substrate.”
The district court stated that using the breaking of fibers as a measure of [676]*676bond strength “is not a reasonable definition when the carpet is weak, because one cannot legitimately describe an adhesive used in that context as a ‘high strength adhesive.’ ” Claim 1, however, sets forth the bonding requirements of the adhesive by requiring that it be greater than the tensile strength of the fibers. There is nothing improper about defining the requisite strength of the adhesive by reference to the material that is being bonded, and in light of that definition in the claim, it would be improper to import another, narrower definition based on the notion that the definition in the claim would include some adhesives that would not normally be considered “high strength.”
The district court also remarked that the functional definition of “high strength adhesive” lacked utility because “the invention would not be useful if the carpeting readily disintegrated, even though the pulling of the substrate would be avoided.” We reject that conclusion because the specification makes clear that the benefit of the invention is that when individual fibers break before the fabric backing separates from the substrate it is more difficult to vandalize the seat. ’539 patent, col. 2, II. 57-63; col. 3, II. 3-6.
The district court’s adoption of the 25-pound per inch requirement was unwarranted for three additional reasons. First, the specification described the 25-pound per inch peeling strength as preferred and desired, not as a necessary feature of the claimed invention. See ’539 patent, col. 2, II. 31-34 (“Preferably, the fabric to substrate bond either exceeds the tensile strength of the fabric itself or has a peeling strength of at least twenty-five pounds per inch width of fabric.”); ’539 patent, col. 4, 1. 66 to col. 5, 1. 4 (“As a minimum standard, it is desired that the fabric to substrate bond exhibit a peel (stripping) strength after being fully cured ... exceeding the tensile strength of the fabric or a minimum bond strength of at least twenty-five (25) pounds per inch width of the fabric.”). While the specification taught that that level of strength was desirable in the preferred embodiment, there is nothing in the specification to indicate that it was meant to be incorporated into the claims as an absolute limitation.
Second, and more importantly, the references in the specification to the strength of the adhesive bond were both couched in alternative form; that is, both described the desired bond strength as being either 25 pounds per inch of fabric width, or greater than the tensile strength of the pile fibers. Thus, even if the language of the specification is imported into the claim, and even if the language of preference is treated as language of absolute requirement, the language on which the district court relied does not support the court’s conclusion that a “high strength adhesive” must have a peel strength of at least 25 pounds per inch of fabric width.
Finally, claim 3 of the ’539 patent, which was dependent on claims 1 and 2, added the requirement that the “bond strength of the fabric to the substrate is at least twenty-five (25) pounds per inch width of fabric” as an alternative means of satisfying the “high strength adhesive” limitation. The absence of that alternative means from the definition of bond strength in claim 1 is a strong indication that claim 1 should not be construed to incorporate that limitation by implication and thereby exclude any adhesive with a bond strength of less than 25 pounds per inch of fabric width. See Karlin Tech., Inc. v. Surgical Dynamics, Inc., 177 F.3d 968, 972 (Fed.Cir.1999) (“normally ... limitations stated in dependent claims are not to be read into the independent claim from which they depend”). For all these reasons, we conclude that it was error for the district [677]*677court to interpret the term “high strength adhesive” in claim 1 of the ’539 patent to mean an adhesive having a bond strength of at least 25 pounds per inch.
Interpreting claim 1 of the ’931 patent presents more of a challenge because the claim language simply states that the bond should be sufficient to “resist peeling.” The relevant claim language reads:
a layer of high strength adhesive interposed between the vertical pile fabric and the substrate and affixing the vertical pile fabric to the substrate, the adhesive bonding the fabric to the substrate with a bond strength that is sufficient to resist peeling of the fabric from the substrate.
As in the case of the ’539 patent, the district court held that the claim language was ambiguous and that in light of the specification, the claim had to be limited to adhesives with a bond strength greater than 25 pounds per inch of fabric width.
Although the language of the ’931 patent is broader than that of the ’539 patent, there is still no justification for importing the 25-pound bond strength limitation from the specification into the claims. The specification of the two patents is identical, and the 25-pound strength requirement is expressed in the ’931 patent, as in the ’539 patent, as a preferred or desired feature, and as only an alternative measure of bond strength. In addition, as in the case of the ’539 patent, a separate claim of the ’931 patent (claim 2) recites the requirement that the bond strength be at least 25 pound per inch of fabric width. As a matter of claim construction, it was therefore improper for the district court to interpret the term “high strength adhesive” in the ’931 patent to be limited to an adhesive with a bond strength greater than 25 pounds per inch.
The USSC Group is correct that the requirement that the bond strength only be “sufficient to resist peeling of the fabric from the substrate” is a minimal requirement that renders the claim broad. But we conclude that the claim was intended to be broad. While the breadth of the claim might create an issue as to validity, it does not create an issue as to its construction. It was thus error for the district court to construe the claim to incorporate 25-pound bond strength as a limitation.
The district court also held that the ’539 and ’931 patents do not cover vertical pile fabrics consisting of “conventional carpet materials with latex coatings.” Because the court concluded that the USSC Group’s accused product “utilized conventional carpeting with a latex coating,” the court invoked the coating issue as a second ground to support its grant of summary judgment of noninfringement of the vandal-resistant seating patents. American Seating argues that the court’s analysis was wrong because the patents exclude “only those fabrics with a coating that is impermeable to the adhesive that is used to adhere the fabric to the seat.” According to American Seating, the evidence in the summary judgment proceedings was at least sufficient to raise a genuine issue of fact as to whether the coating on the T2C product was permeable to the adhesive, and the court therefore erred by upholding the summary judgment of noninfringement on this ground.
The patents do not give the terms “vertical pile fabric” or “pile fabric” a special meaning different from the ordinary meaning of those terms. The specification explains that a permeable backing for the pile fabric allows the adhesive to penetrate the backing and adhere directly to the fibers, thus providing protection against efforts to strip the fabric from the seat [678]*678backing. That feature of the invention is incorporated in all the claims of the ’539 patent, which provide that the pile fabric comprises “a permeable backing material.” The same feature is incorporated in three of the independent claims of the ’931 patent, which provide that the backing of the vertical pile fabric comprises “an uncoated permeable fibrous material” (’931 patent, claim 3), or has a backing that is “permeable by the adhesive” (’931 patent, claims 7, 8). Independent claims 1, 2, and 11 of the ’931 patent are silent with respect to the backing or coating issue.
Relying on the prosecution history and common specifications of the ’539 and ’931 patents, the USSC Group argues, in support of the district court’s ruling, that the asserted claims further exclude from their coverage all conventional, latex-backed pile fabrics. In particular, the USSC Group points to a portion of the common specification that emphasizes the reason for and importance of a permeable backing for the fabric. The pertinent portion of the specification, which is a part of the “Detailed Description of Preferred Embodiment,” reads as follows:
It is important that the lower end extremities of the pile fibers be accessible and contactible by an adhesive (preferably, a sprayable liquid adhesive) applied on the inner side of the backing material, so that the fibers and backing material are bonded together and both are bonded to the substrate. In this connection, it is important that the vertical pile fabric be different from conventionally available vertical pile fabrics. Conventional vertical pile fabrics are provided with a latex or other impermeable coating on the underside thereof.... However, in the present invention, the adhesive must adhere directly to the backing itself and to the pile fibers in order to bond the fabric to the substrate with sufficient bonding strength, and the presence of a coating impairs this function and prevents adequate bonding of the fabric to the substrate.
With the requirement that the fabric be uncoated, a number of convention vertical pile fabrics ... can be used.
’539 patent, col. 4,11. 9-24; ’931 patent, col. 4,11. 4-19.
It is true that the portion of the specification relied on by the USSC Group criticizes conventional coated pile fabrics, but it does so on the ground that they possess an impermeable coating, whereas the invention requires that the adhesive permeate the backing of the pile fabric and adhere to the fibers. We do not interpret the portion of the specification quoted above as excluding all conventional coated, or even all latex-coated pile fabrics. Instead, that passage focuses on the issue of permeability and thus is properly read to exclude only those coatings that are impermeable to the adhesive.
The prosecution history is to the same effect and supports the argument made by American Seating. In response to a prior art rejection, the inventors advised the examiner that the vertical pile fabric of the invention “has a permeable backing material that the adhesive can penetrate. The use of a conventional pile fabric upholstery material, which customarily is produced with a penetration resistant backing, is unsatisfactory.”
Accordingly, neither the claim language, nor the common specification, nor the prosecution history supports the USSC Group’s argument that a conventional pile fabric with a latex coating is necessarily outside the scope of the asserted claims of the ’539 and ’931 patents. Instead, the extent to which the claims cover coated fabrics depends on whether the coating of [679]*679those fabrics is permeable by the adhesive used in the particular application.
American Seating contends that the evidence offered in the summary judgment proceedings made clear that the coatings used in the accused T2C were penetrated by the adhesive; the USSC Group contends that the evidence introduced by American Seating does not show that the T2C employed permeable coatings. Because the district court based its ruling on the conclusion that the asserted claims did not cover “conventional carpeting with a latex coating,” the court did not explore whether the evidence before it was sufficient to create a disputed issue as to the permeability of the coating used in the T2C product. In light of the claim construction we have adopted, we remand to the district court to address that issue as part of its determination whether the USSC Group’s products infringe the patents asserted by American Seating.