BRYSON, Circuit Judge.

Honeywell Inc. is the owner of U.S. Patent No. 4,425,501 (“the ’501 patent”), which is entitled “Light Aperture for a Lenslet-Photodetector Array.” The patent claims an aperture mask for an autofo-cus system used in cameras and videocam-eras. Honeywell asserted the ’501 patent against Victor Company of Japan, Ltd., and U.S. JVC Corp. (collectively, “JVC”) in the United States District Court for the District of Minnesota. On JVC’s motion, the district court granted partial summary judgment, holding that JVC did not infringe the patent. In addition, the court ruled that certain imaging chips purchased by JVC from a subsidiary of Matsushita Electric Industrial Co., Ltd. (“MEI”), fell under the terms of an agreement licensing the ’501 patent to MEI and that Honeywell’s infringement claim against JVC could not succeed with respect to those chips. At the request of the parties, the court entered a judgment pursuant to Fed. R.Civ.P. 54(b) as to the claim of infringement and stayed further proceedings with respect to JVC’s counterclaim for a declaratory judgment of invalidity. We reverse in part, affirm in part, and remand.

An autofocus camera automatically adjusts the focus of its lens in order to obtain the sharpest possible photographic image. There are two main types of autofocus cameras. “Active” autofocus cameras operate by emitting either sound or light energy that is reflected off the object being photographed. That energy is then received back at the camera and processed in such a way as to determine the distance to the object and adjust the focus of the lens accordingly. “Passive” autofocus cameras, on the other hand, generate images of the object being photographed and determine whether the object is in focus based on the positions of the radiation patterns of the image on detector arrays within the camera.

The ’501 patent is directed to an improvement to a particular type of passive autofocus system that includes a set of small lenses, or lenslets, and a circuit chip containing light-sensitive detectors. During the operation of the system, light from a remote object is projected onto the detectors after first passing through the camera’s objective lens and the lenslets, which are embedded in a transparent member. Integrated circuitry coupled to the detectors then determines whether the image is in focus by comparing the light *1320 patterns appearing on the various detectors.

The ’501 patent explains that there are two problems with this type of autofocus system. First, because the member that includes the lenslets is transparent, light from the objective lens reaches not only the detectors, but also the adjacent circuit areas. Because certain circuit elements are sensitive to radiation in the visible range of the electromagnetic spectrum, allowing light to strike circuit elements may result in “spurious signals that can produce error in the circuit operation.” ’501 patent, col. 1, 11. 36-37. Second, surface defects in the transparent member can cause a scattering of light so that some light reaches the detector areas without first passing through the lenslets, causing erroneous readings by the detectors. We refer to these two problems as the “exposed circuit area problem” and the “surface imperfections problem,” respectively.

The improvement claimed in the ’501 patent is the use of an “opaque member,” or mask, to prevent the passage of some of the light coming from the object at which the camera is pointed. The mask is indicated by reference numeral 30 in Figure 1 of the ’501 patent shown below. Figure 1 also depicts transparent member 20, len-slets 21 and 22, circuit chip 10, and detector areas 11 and 12. Although not depicted in Figure 1, the detector areas are typically surrounded by circuit components. ’501 patent, col. 1,1. 67 to col. 2,1. 2.

*1321 [[Image here]]

The positioning of the mask in the embodiment of the invention shown in Figure 1 solves both the exposed circuit area problem and the surface imperfections problem because it ensures that light does not reach the exposed circuit area and that any light that passes through the transparent member passes only through the len-slets. Honeywell and JVC dispute whether claim 1 of the ’501 patent requires that the opaque member be positioned so that the device solves both problems or whether it is sufficient for the mask to be positioned to solve only the exposed circuit area problem. Claim 1, which has been reformatted for clarity, provides as follows:

1. Apparatus for use with an auto focus camera including

a circuit chip having a plurality of radiation responsive detector areas and a circuit area adjacent the detector areas which circuit area may produce error signals for the auto focus camera if exposed to light and including

a transparent member having a plurality of lenslets formed therein and placed contiguous the chip so that each lenslet directs radiation from a remote source onto one of the detector areas,

the improvement comprising:

*1322 an opaque member having a plurality of transparent portions therein placed contiguous the transparent member so as to permit radiation through the len-slets to reach the detector areas and to prevent radiation through the transparent member from reaching the circuit area.

Dependent claim 2 recites the apparatus of claim 1, in which the opaque member “has a plurality of holes therein corresponding to the plurality of lenslets.”

Honeywell brought this action against JVC, alleging that one of the components in JVC’s videocameras, known as a charge coupled device imaging chip, was covered by claims 1 and 2 of the ’501 patent. In response, JVC filed a motion for summary judgment of noninfringement, contending that its chips did not fall within the scope of those claims. In addition, JVC argued that some of the chips used in the accused products were sold to JVC by a supplier, MEI, that had licensed the ’501 patent from Honeywell, so that JVC was not liable for infringement based on its use of those chips. Honeywell filed a cross-motion for summary judgment on the latter ground, contending that the MEI chips were not covered by the Honeywell-MEI license agreement.

The district court first addressed the license defense and ruled in favor of JVC on that issue. Because JVC had purchased some imaging chips from suppliers other than MEI, however, that ruling did not entirely dispose of Honeywell’s infringement claim, but merely limited the number of accused devices to which the claim applied.

The court then addressed the question whether JVC’s remaining chips fell within the scope of the asserted claims. In addressing that issue, the district court construed the term “contiguous,” as used in claim 1, to mean “ ‘next in succession’ without any intervening structure.” As a result, the court ruled,

the claim limitations that the transparent member be contiguous the circuit chip and the transparent member be contiguous the opaque member require that the claim be limited to devices with the opaque member/transparent member/circuit chip structural order; a device which includes an opaque member between the transparent member and circuit chip falls outside of the scope of the claim.

Because the court construed claim 1 as being limited to devices in which the circuit chip is next to the transparent member (in which the lenslets are formed) without any intervening structure, the court granted summary judgment of non-infringement to JVC. The court held that the structure it had interpreted as being required by claim 1 was not satisfied by the accused device because the opaque member in the JVC imaging chips was not on top of the transparent member. Because, as the court described it, the JVC imaging chip “has a mask which is essentially under the lenslet — an embodiment expressly criticized in the patent specification,” the court held that the accused JVC imaging chips did not literally infringe the asserted claims of the ’501 patent and could not be held to infringe under the doctrine of equivalents.

Ill

Honeywell’s infringement claim turns on the construction of claim 1 of the ’501 patent. JVC argues that because the claim language requires that the mask be “contiguous” to the transparent member and that the transparent member be “contiguous” to the chip, claim 1 reads only on devices having a mask-transparent member-chip sequence. Honeywell, on the oth *1323 er hand, contends that neither the term “contiguous” nor anything else in the patent dictates that the elements must be arranged in that order.

This seemingly straightforward issue of claim construction is made difficult by the fact that the patent is so poorly drafted that it does not give a readily discernible answer to what would seem to be a basic question regarding the structure of the claimed invention. The district court struggled with the issue, as have we, and while we cannot fault the district court’s efforts to make sense of the patent language, we conclude on the basis of all the evidence available to us that the district court’s claim construction was too narrow.

Claim 1 recites that the transparent member containing the lenslets is “placed contiguous the chip,” and that the opaque member having a plurality of transparent portions (i.e., the mask) is “placed contiguous the transparent member.” The district court construed that language to require the mask to be placed on top of the transparent member and the transparent member to be placed on top of the chip. In large part, the court premised that conclusion on its view that the term “contiguous” implied the absence of intervening structure. Thus, according to the district court, if the mask were positioned between the transparent member and the chip, the transparent member and the chip would not be “contiguous.” From the court’s remark that the transparent member would not be contiguous to the chip even if “the circuit chip is in contact with or next to the transparent member at the points of the holes,” it appears that the court would regard the transparent member and the chip as not contiguous in such a structure, even if the two were touching.

Honeywell argues that the district court’s claim construction is too restrictive. Honeywell correctly notes that claim 1 does not expressly require any particular ordering of elements, i.e., the claim does not explicitly require that the mask be on top of the transparent member and that the transparent member be on top of the chip. Honeywell further contends that the use of the term “contiguous” in claim 1 does not implicitly require that ordering of elements. In fact, as Honeywell interprets the term “contiguous,” there need not even be contact between the elements in question. Rather, Honeywell argues, one element is contiguous to another if the two elements are simply near one another. In Honeywell’s view, then, claim 1 reads on a device in which the mask is positioned between the transparent member and the chip, as long as there is close proximity between the mask and the transparent member, and between the transparent member and the chip.

Honeywell’s proposed construction of “contiguous” is based on a definition of the term appearing in the prosecution history of the application that matured into the ’501 patent. The definition offered by the inventor was as follows: “in actual contact; touching; also, near, though not in contact; neighboring, adjoining; near in succession.” The inventor further explained that two elements would be “contiguous” if they were placed “in actual contact” or if they were “near, though not in contact.”

The district court erred in not according more weight to the inventor’s definition. It is well settled that a paten-tee may define a claim term either in the written description of the patent or, as in the present case, in the prosecution history. Mycogen Plant Science v. Monsanto Co., 243 F.3d 1316, 1327, 58 USPQ2d 1030, 1039 (Fed.Cir.2001). Frequently, a definition offered during prosecution is made in response to a rejection, and is entered in *1324 conjunction with a narrowing amendment. See, e.g., Southwall Techs., Inc. v. Cardinal IG Co., 54 F.3d 1570, 1576, 34 USPQ2d 1673, 1677 (Fed.Cir.1995). Such a definition limits the scope of the claim, preventing the patentee from later recapturing what was previously surrendered. Although the inventor’s definition does not have a narrowing effect, it is nonetheless relevant in indicating the meaning that the inventor ascribed to the term. See Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582, 39 USPQ2d 1573, 1577 (Fed. Cir.1996) (“[Tjhe record before the Patent and Trademark Office is often of critical significance in determining the meaning of the claims.”); E.I. du Pont de Nemours & Co. v. Phillips Petroleum Co., 849 F.2d 1430, 1438, 7 USPQ2d 1129, 1135 (Fed.Cir.1988) (prosecution history “must be examined to ascertain the true meaning of what the inventor intended to convey in the claims”).

The definition of “contiguous” set forth in the prosecution history-“in actual contact” or “near, though not in eontact”-is taken from a standard English dictionary. Although the definition is broad-the most common meaning of “contiguous” connotes actual contact, not general proximity-that fact is not disqualifying, because the inventor made clear that he intended to invoke the broadest meaning of the term “contiguous,” not just the most common meaning. We therefore conclude that the term “contiguous,” as used in claim 1, does not require actual contact between elements that are described as “contiguous,” and that the use of that term in claim 1 does not necessarily confine the scope of the claim to devices in which the elements are positioned in the mask-transparent member-chip order.

Admittedly, accepting Honeywell’s definition of “contiguous” does not wholly resolve the issue of infringement, particularly with respect to the problem of intervening structure. The court rejected defining “contiguous” as “near” because it would “lead to the conclusion that the letter ‘A’ is contiguous the letter ‘C’ “ a proposition that the district court found untenable. The court then offered a second hypothetical to illustrate what it regarded as the limits imposed by the term “contiguous”: “For example, if one makes a sandwich placing a piece of cheese between the bread and the bologna, the bread and the bologna are not contiguous regardless of whether the cheese is American or Swiss.”

Because the structure of the accused device is not entirely clear from the record on appeal, we are unsure whether the “Swiss cheese” problem will be implicated in the determination of infringement on remand. We feel compelled, however, to offer some guidance on this issue, particularly since the district court focused on it so prominently.

To take the easiest case first, we think it evident that the bread and the bologna in the district court’s example would be contiguous if they were touching through the holes in the Swiss cheese. It also seems obvious that the bread and the bologna would not be contiguous if American cheese instead of Swiss cheese separated them so that the cheese intervened between the bologna and the bread at all points. We therefore reject what seems to be the broadest implication of Honeywell’s argument-that two elements can be considered contiguous if they are near one another even if they are separated by an intervening structure at every point. The absence of intervening structure is inherent even in the broad definition of “contiguous.” See, e.g., Webster’s Third International Dictionary 492 (1966) (“contiguous” defined as “next or adjoining with nothing similar intervening ... immediately preceding or following in time or sequence: *1325 without intervening interval or item”). The district court, however, seemed to be principally concerned with the thornier question whether the bread and the bologna are contiguous if they are not touching but there is nothing between them at the places where there are holes in the Swiss cheese. Under the definition of “contiguous” offered by Honeywell, we believe the answer to that question depends on whether, in context, it can be said that the bread and the bologna are “near” one another. That is, the question whether the two are contiguous depends in part on the thickness of the layer of Swiss cheese. Under Honeywell’s definition of “contiguous” as “near, though not in contact,” the bread and the bologna would be contiguous at the points where the holes in the Swiss cheese were located if the bread and the bologna were close to one another at those points.

By analogy, the transparent member would be contiguous to the circuit chip within the meaning of Honeywell’s definition if the transparent member extended through the holes in the mask to touch or almost touch the circuit chip. On remand, then, the district court must determine whether there is some point at which the transparent member and the chip are not separated by some intervening structure, and whether the distance between the transparent member and the chip at that point is sufficiently small, in the context of the structure as a whole, to satisfy the requirements of contiguity recited in claim 1.

The district court’s claim construction was heavily influenced by its interpretation of the written description of the ’501 patent. The court interpreted the written description to require the claimed invention to solve both problems identified in the prior art: the exposed surface area problem and the surface imperfections problem. Because both problems are solved by the mask-transparent member-chip structure, but not by the transparent member-mask-chip structure, the district court concluded that the written description dictated that the claim must be construed to cover only the mask-transparent member-chip structure. In reaching that conclusion, the district court placed particular importance on the following sentence from the 'written description: “While the mask 30 could be placed adjacent the circuit board 10 and under the lenslet member 20 to likewise mask out the circuit areas on the chip 10, this arrangement would not help the problem of imperfections in the surface of member 20 causing scattering of light to strike other' detectors.” ’501 patent, col. 2,1. 65 to col. 3,1. 2. The court found that this passage constituted a “clear case of disclaimer of subject matter,” analogizing to SciMed Life Systems, Inc. v. Advanced Cardiovascular Systems, Inc., 242 F.3d 1337, 58 USPQ2d 1059 (Fed.Cir.2001).

In SciMed, which involved patents on a type of medical catheter used in angioplasty, the common specification of the patents at issue stated that the catheters of “all embodiments of the present invention contemplated and disclosed herein” employed a coaxial lumen. Id. at 1343, 242 F.3d 1337, 58 USPQ2d at 1065. We held that the “broad and unequivocal” language of the specification defined SciMed’s invention in a way that excluded lumens that were not coaxial. Id. at 1344, 242 F.3d 1337, 58 USPQ2d at 1065.

There is no similar disclaimer of subject matter in the ’501 patent. The language in the ’501 patent on which the district court relied is far less explicit than the “all embodiments” language in the patents at issue in SciMed. Moreover, the language in claim 1 of the ’501 patent makes reasonably clear that the inventor did not intend *1326 to limit the scope of the claim to structures that would solve both of the prior art problems identified in the background section of the written description of the ’501 patent.

Claim 1 recites that the opaque member is “placed contiguous the transparent member so as to permit radiation through the lenslets to reach the detector areas and to prevent radiation through the transparent member from reaching the circuit area.” That language expressly requires that the opaque member must be positioned to solve the exposed circuit area problem. The claim does not, however, state that the opaque member must be positioned to prevent radiation from reaching the detector areas except through the lenslets. Thus, claim 1 does not require that the opaque member be situated to solve the surface imperfections problem.

That omission is significant. The fact that the patentee chose to include language in claim 1 relating to only one of the two cited prior art problems is persuasive evidence that the claim does not require the solution of both problems. Instead, the natural implication is that embodiments within the scope of claim 1 must solve only the recited problem of radiation reaching exposed circuit areas. We therefore reject JVC’s argument that the term “contiguous” must be construed in a manner that would lead to the solution of both prior art problems discussed in the written description, when claim 1 requires the solution of only one of those problems.

That conclusion is reinforced by the context of the sentence on which JVC relies. The sentence in the written description immediately following the purported disclaimer states:

It is therefore seen that I have provided a means for overcoming the problem of light passing through areas of a lenslet array in such a way as to be received by the circuit apparatus on an auto focus module thus overcoming the problem of spuriously generated signals.

’501 patent, col. 3,11. 3-7.

That sentence is plainly intended to serve as a general characterization of the invention. Significantly, the sentence refers only to the exposed circuit area problem, i.e., the problem of light being “received by circuit apparatus,” which can lead to “spuriously generated signals.” There is no suggestion in that characterization that the invention must also solve the surface imperfections problem. We therefore conclude that the description of the mask-transparent member-chip structure,-when considered in context, is not a disclaimer of subject matter, but simply an explanation of why the “mask on top” structure is a preferred embodiment of the invention. Such a description of a preferred embodiment does not, of course, limit the scope of the claims. See Electro Med. Sys. SA v. Cooper Life Sciences, 34 F.3d 1048, 1054, 32 USPQ2d 1017, 1021 (Fed.Cir.1994).

Two other portions of the written description provide further support for Honeywell’s argument that the invention of claim 1 is not required to solve both cited prior art problems. Like the language of claim 1, the Abstract mentions only the exposed circuit area problem. It states that the mask is positioned to “permit[ ] radiation to pass through the len-slets and be focused on the detector areas but prevent[ ] light from passing through transparent areas adjacent the lenslets and striking the circuit areas.” The fact that the Abstract mentions only the exposed circuit area problem supports the argument that claim 1 should not be read as requiring a structure that will solve both that problem and the surface imperfections problem.

The Summary of the Invention, although not free from ambiguity, also contains lan *1327 guage that supports Honeywell’s proposed construction of the claim language. The Summary consists of a single sentence, which reads as follows:

The apparatus of the present invention utilizes a light aperture or masking means placed in the area adjacent the lenslets or adjacent the detector areas of the circuit chip which mask is opaque to light except in the areas of the aperture and thus prevents light from passing through the translucent member except through the lenslets and thus prevents light from reaching the circuit areas of the chip other than the detectors.

’501 patent, col. 1, 11. 47-54.

The first portion of that sentence, which states that the mask may be placed either “in the area adjacent the lenslets or adjacent the detector areas of the chip,” is squarely contrary to JVC’s argument that claim 1 requires the mask to be adjacent to the lenslets but not adjacent to the circuit chip, and it buttresses Honeywell’s argument that the mask'could either be placed above the lenslets and the chip or between the lenslets and the chip.

To be sure, the persuasive force of the Summary is undercut somewhat by the fact that the sentence contains an internal inconsistency. The second portion of the sentence states that the positioning of the mask “prevents light from passing through the transparent member.” That would happen only if the mask were placed above the lenslets. Thus, while the first portion of the sentence suggests that the mask may be placed adjacent to the detector areas underneath the lenslets, the second portion describes a condition that would apply only if the mask were placed above the lenslets.

In light of the inconsistency between the two portions of the sentence quoted above, it appears that the Summary of the Invention was not well thought' out. For that reason, we do not attach conclusive weight to the description of the invention that appears in the first portion of the Summary. Nonetheless, the first portion of the sentence so plainly contemplates placement of the mask either next to the len-slets or next to the detector areas of the chip that it is difficult to imagine that the author of that passage could have intended claim l to be as narrow as JVC suggests.

JVC also contends that during the prosecution of the patent the inventor made a representation that the mask-transparent member-chip order was required by the claims. JVC argues that in view of that admission it would be improper to read the claims as broadly as Honeywell urges. As we read the prosecution history, however, there was no such admission.

When the application that matured into the ’501 patent was filed, the inventor submitted several references, including a published Japanese patent application. Although the inventor stated that his invention “solves a somewhat different problem in a somewhat different way,” he nevertheless indicated an intent to swear behind the Japanese reference. , To that end, the inventor submitted affidavits from himself and two others describing the activity that predated the effective date of the Japanese reference.

The examiner found the affidavits insufficient to swear behind the Japanese reference. Specifically, the examiner complained that the affidavits, and the accompanying exhibits did not show that the mask was placed contiguous to the transparent member, and that the transparent member was placed contiguous to the circuit chip, as required by claim 1. The inventor responded by arguing that the affidavits did show that he had reduced to practice a device having “the *1328 proper positioning as required by the claims.”

JVC contends that the statement by the inventor requires a construction of claim 1 in which the mask is situated above the lenslets. We disagree. The prosecution history contains no representation by the inventor that claim 1 requires a particular ordering of elements. Instead, the inventor made a much different representation-that an embodiment of the invention having a mask-transparent member-chip structural order satisfies the requirement of claim 1 that both the mask and the transparent member, and the transparent member and the chip be “placed contiguous” to one another.

The inventor was arguing that he had reduced to practice an embodiment of his invention that was within the scope of claim 1 prior to the effective date of the Japanese reference. In making that argument, the inventor took issue with the examiner’s contention that the affidavits the inventor submitted did not demonstrate that the mask was contiguous to the transparent member and that the transparent member was contiguous to the circuit chip. In response, the inventor stated that in the embodiment of the invention referenced by his affidavits, the mask and the lenslets were contiguous because the mask was placed “on a lenslet and detector assembly.” Similarly, the inventor stated that the lenslets and the chip were contiguous because the lenslets and detectors were referred to as “an assembly.”

In making the latter point, the inventor recited the dictionary definition of “contiguous” referred to above and concluded that “[placing the mask on the lenslet and detector assembly assures that these elements are either ‘in actual contact’ or are ‘near, though not in contact.’ “ The inventor thus contended that his earlier embodiment of the invention had elements that were contiguous within the meaning of that term as used in the ’501 patent. However, arguing that a particular embodiment satisfies the definition of “contiguous” is quite different from arguing that “contiguous” is limited to the order of elements found in that embodiment. When the inventor stated that he had demonstrated that the elements of the invention were “mounted as required by the claims,” he was explaining that his assembly satisfied the requirement of contiguity; he was not arguing that the claims required that the elements of the claimed device be arranged in a specific order. For that reason, the prosecution history does not limit the express definition of the term offered by the inventor.

In sum, we conclude that the district court’s claim construction was unduly restrictive. There are several possible structures that permit the transparent member to be contiguous to the chip and the mask to be contiguous to the transparent member even though the mask is not located on top of the transparent member. Although we decline Honeywell’s invitation to decide that the broader claim construction that we adopt dictates a finding of infringement by JVC’s imaging chips, Honeywell is entitled to an opportunity to show that JVC’s chips infringe under this construction. We therefore reverse the judgment of nonin-fringement and remand for further proceedings on that issue.

Because Honeywell’s infringement claim is not foreclosed by our construction of claim 1 of the ’501 patent, it is necessary for us to reach the second issue in this case-whether JVC is immune from liability for infringement to the extent that Honeywell’s claim is based on imaging chips that JVC purchased from MEI. The answer to that question turns on whether the imaging chips are “Products” under the license *1329 agreement entered into by Honeywell and MEI.

In relevant part, the license agreement defines the term “Products” as follows:

1.3 “Products” means (i) an apparatus or combination of apparatus for use with a camera, to provide optical images of objects which are converted into video signals....

Honeywell contends that the district court erred in construing that provision to require only that a device “provide optical images of objects.” According to Honeywell, a device within the meaning of “Products” in the license agreement must also convert optical images into video signals. We agree with the district court that Honeywell’s argument is contradicted by the plain language of the agreement, which specifies that the conversion function may be performed by another component in the camera. Thus, a device need only “provide optical images of objects” to be within the scope of the license agreement.

Honeywell contends, however, that the district court erred by failing to consider the specific meaning of that phrase within the optics industry. Before the district court, Honeywell offered testimony that the phrase “provide optical images” refers to “the function of the objective lens.” In Honeywell's view, because the imaging chips that MEI sold to JVC do not include an objective lens, they are not covered by the license agreement.

We cannot agree with that interpretation. Even if the phrase “optical images of objects” refers to the output of the objective lens of a camera, the definition of “Products” does not require that the “apparatus” in question actually create the optical image of the object. Instead, we agree with the district court that the definition requires only that the apparatus “provide” such images, that is, the apparatus need only act as a conduit in conveying a signal representing that image to other components of the camera. The imaging chips in question indisputably meet that requirement. Accordingly, we hold that Honeywell’s evidence of trade usage does not make it reasonable to interpret the definition of “Products” within the license agreement as applying only to those devices that have an objective lens.

Honeywell’s proposed interpretation also conflicts with the context in which the license agreement was entered. The agreement was signed because of a dispute between Honeywell and MEI over the ’501 patent and another of Honeywell’s patents. The parties therefore entered into an agreement under which Honeywell waived all claims relating to the “manufacture, use, sale, lease and other disposition of ... Products.” Additionally, Honeywell granted MEI a nonexclusive license “to make, have made, use, sell, lease and otherwise dispose of ... Products” covered by the ’501 patent.

The agreement was plainly intended to cover devices within the scope of the ’501 patent. Because the ’501 patent covers devices without regard to whether they include an objective lens, we cannot accept Honeywell’s narrower definition of “Products,” which would exclude devices squarely covered by the ’501 patent, the very devices that led the parties to enter into the agreement in the first place. Accordingly, we agree with the district court that the imaging chips that JVC purchased from MEI are encompassed within the license agreement and therefore are not subject to Honeywell’s claim of infringement.

AFFIRMED IN PART, REVERSED IN PART, and REMANDED.