Glenwa, Inc. (“Glenwa”) appeals from a decision by the District Court for the District of Oregon granting summary judgment on behalf of the plaintiff, Old Town Canoe Company (“Old Town”). Old Town Canoe Company v. Glenwa, Inc., No. 00-1675-HA, slip op. (D.Or. Oct. 4, 2001). The district court granted both Old Town’s motion for summary judgment of validity and its motion for summary judgment of infringement. Because the district court erred in its claim construction and therefore arrived at erroneous conclusions with respect to validity and infringement, we vacate-in-part, reverse-in-part and remand.

Background

Plaintiff-appellee, Old Town sued defendant-appellant, Glenwa for infringement of three utility patents, alleging that each of three Glenwa kayak models (the Cobra Tandem, the Cobra Triple, and the Cobra Fish ’N Dive) infringed at least one, if not multiple claims in three of Old Town’s sit-on-top kayak patents. The patent claims actually at issue in this case are: claim 13 of U.S. Pat. No. 5,964,177 (“the ’177 pat*920ent”),1 issued October 1999, claims 23 and 24 of U.S. Pat. No. 6,152,063 (“the ’063 patent”),2 issued November 2000, and claims 1 and 2 of U.S. Pat. No. 6,178,912B1 (“the ’912 patent”),3 issued January 2001. All three patents issued from continuation applications sharing the same specification supporting a design patent on a small watercraft, U.S. Des. 377,473 (“the ’473 patent”), filed in 1994, and all three patents list Timothy Niemier as the sole inventor.

After construing the claims in the three patents, the district court completed sepa-

rate analyses of the validity of the patents and their possible infringement by the accused products upon cross motions for summary judgment. The district court granted summary judgment as to validity and infringement.

DISCUSSION

I. Claim Construction

The trial court made a concerted effort to construe all of the numerous disputed terms in this case and we leave much of its *921interpretation undisturbed. However, in its interpretation of the terms “sit-on-top kayak,” “hull,” “contoured,” and “seating surfaces” the trial court erred.

A.

The trial court correctly discerned that the preamble “sit-on-top kayak” should be understood to limit all of the claims to a category of watercrafts and “[i]n general, a preamble limits the invention if it recites essential structure or steps, or if it is ‘necessary to give life, meaning, and vitality’ to the claim.” Catalina Marketing Inti, Inc. v. Coolsavings.com, Inc., 289 F.3d 801, 808, 62 USPQ2d 1781, 1784 (Fed.Cir. 2002) (quoting Pitney Bowes, Inc. v. Hewlett-Packard Co., 182 F.3d 1298, 1305, 51 USPQ2d 1161, 1165 (Fed.Cir. 1999)). However, the trial court’s explanation of what this limitation is, namely that a sit-on-top kayak has “a generally V-shaped hull portion and a deeper draft than a surfboard to allow better tracking,” is not consistent with the plain meaning of the term as used in the patents. Old Town, 229 F.Supp.2d at 1158. To ascribe a hull shape to “sit-on-top kayak” would be redundant in view of the first element of the independent claims that refers specifically to the sit-on-top kayak as comprising “a hull defining a hull outer surface.” If “sit-on-top kayak” meant a craft with a particular hull, then it would be unnecessary to also claim a hull.

Similarly, the meaning of “hull” in the claims was improperly limited by the trial court. The written description is used to define but not limit terms in the claims. See Loctite Corp. v. Ultraseal Ltd., 781 F.2d 861, 867, 228 USPQ 90, 93 (Fed.Cir. 1985) (“That the specific examples set forth in the patents occur at room temperature does not mean that the claims are imbued with a room temperature limitation. Generally, particular limitations or embodiments appearing in the specification will not be read into the claims.”). The trial court relied on the written description’s statement: “[t]he sit-on-top kayak is somewhat similar to a surfboard in overall shape, but normally has a generally V-shaped hull portion and a deeper draft to allow better tracking.” Old Town, 229 F.Supp.2d at 1157 (quoting the ’177 patent, col. 1, II. 25-28). There are two flaws in the trial court’s reliance on this statement. First, the plain meaning of “hull” does not include a particular shape (V) because otherwise the phrase “V-shaped hull portion” would be redundant. Second, the contextual meaning of “hull” also cannot include a limitation on its shape because the qualifier “normally” precludes the term from always requiring a V-shape. In the claims, “hull” appears to be a generally descriptive term and as such the appellant is correct that it should be given its full meaning and no modifiers should be added. See York Prods., Inc. v. Central Tractor Farm & Family Cent., 99 F.3d 1568, 1572, 40 USPQ2d 1619, 1622 (Fed.Cir. 1996) (“Without an express intent to impart a novel meaning to claim terms, an inventor’s claim terms take on their ordinary meaning.”).

Moreover, the interpretation of “sit-on-top kayak” in the preamble as having a particular hull shape and of “hull” in the claims to be “generally V-shaped to create a deeper draft than a surfboard to obtain better tracking” is directly contradicted by statements in the written description of the ’912 patent describing particular embodiments. One such example is for the description of the fourth embodiment: *922’912 patent, at col. 9, 11. 17-20 (emphasis added). As the appellant asserts, this is incompatible with the idea that the terms “sit-on-top kayak” and “hull” in the ’912 patent require a particular shape. Our broader interpretation of “sit-on-top kayak” and “hull” is not solely for the ’912 patent. The terms should be construed similarly in the T77 patent and the ’063 patent because they use the same claim terms and, in fact, the preamble and the first line of the claims (referring to a “hull”) are identical in all three patents. Also, all three arise from the same original patent application. The terms must be construed in the same way in all three patents. See Jonsson v. Stanley Works, 903 F.2d 812, 818, 14 USPQ2d 1863, 1869 (Fed.Cir. 1990) (holding that when two patents use the same claim terms the prosecution histories of both are relevant).

*921The bottom portion 418 of the kayak hull 412 may be any size, shape, and/or configuration as appropriate to provide a stable and efficient platform for movement of the kayak 410 through the water.

*922We, therefore, adopt the district court’s construction of “sit-on-top kayak” and its function as a limitation of the claims with the exception of the clause “with a generally V-shaped hull portion and a deeper draft than a surfboard to allow better tracking.” That we reject. Old Town 229 F.Supp.2d at 1158. The proper definition includes crafts with both displacement and planing hulls.

We agree with the plaintiff that “hull” refers to a structure that forms “the bottom of the kayak” because it is consistent with the usage of the term in the patent itself. We also note that none of the extrinsic evidence refutes this definition.

B.

Next we turn to the construction of “seating surfaces” in all three patents. The term appears in all five claims at issue in this suit. In claim 13 of the T77 patent, a separate element of the claim asserts that the claimed seating surfaces must be contoured. In claim 1 of the ’912 patent the phrasing is also modified to read “seating discrete surfaces” instead of simply “seating surfaces.”

The trial court construed “seating surfaces” in the ’177 patent and the ’063 patent as areas or locations “on which passengers are intended to sit.” Old Town, 229 F.Supp.2d at 1158. Glenwa argues that this definition is overly broad in view of the prosecution history. We agree. Niemier avoided an anticipation rejection under French Pat. No. 2,686,055 (“Philippe”) by distinguishing the seating surfaces on the invention as discrete seats unlike the bench-style seating deck of Philippe on which passengers can sit anywhere. Clarifications by inventors that are made to avoid prior art prohibit contrary interpretations, whether by argument or narrowing amendment. Alpex Computer Corp. v. Nintendo Co., 102 F.3d 1214, 1220, 40 USPQ2d 1667, 1671 (Fed. Cir. 1996). The references to “seating surfaces” in all three patents must be construed as clearly delineated areas on which passengers are intended to sit. The delineation must be through shape, material, or other differentiation from the surrounding deck such that the “seating surfaces” are clearly different from the other surfaces on the top deck of the craft. In traversing the rejection of the claims, for example, the patentee argued that a “hatch surface” is described as “separate[ing] each of the seating surfaces,” implying that a “hatch surface” is distinct from a “seating surface” on the kayak deck.

Similarly the trial court’s construction of “contoured” in the ’177 patent must be adjusted. The trial court judge first construed it as meaning “[a] surface shaped to fit the outline or form of something.” Old Town, 229 F.Supp.2d at 1158. While this is not wrong, it is an incomplete definition. The full claim language is “con*923toured to provide comfortable seating areas.” The subsequent description of the purpose of the contouring suggests that it is not merely contouring to the shape of something, but rather, the specific shape or shapes that would be comfortable to sit on. As the trial court opinion mentions several times, the Detailed Description of the T77 patent explains that the seating surfaces are “slightly concave or contoured to provide a more comfortable seating area.” 177 patent, col. 4, 11. 10-11. While as a grammatical matter “slightly concave” could be understood as an alternative to “contoured” in the phrase, as a matter of logic — the Detailed Description of the patent is used to describe the invention claimed, not give alternatives that are not claimed — they should be understood more as synonyms or close approximations. Thus using the intrinsic evidence, “contoured to provide comfortable seating surfaces” must mean a surface shaped to fit the outline or form of something concave or similarly at least partially rounded.

This more complete and narrow definition is further supported by the prosecution history. For example, Niemier amended the claims in the 177 patent from “seating portions” to “seating surfaces” that “are contoured to provide comfortable seating areas” to overcome an anticipation rejection based on Philippe.

Finally, the addition of the word “discrete” in the ’912 patent actually does not act in the same way that the additional “contoured” requirement functions in claim 13 of the 177 patent. There is not really a separate limitation included because of the word “discrete,” rather the word makes explicit the limitation that the court construed is implicit in the other four claims in the very terms “seating surfaces” as a result of the prosecution history. The word “discrete” as a modifier is consistent with this court’s construction of “seating surfaces” as a seat structurally distinct from the deck surface.

II. Validity

Glenwa challenged the validity of all three of Old Town’s patents on the basis of prior art in the form of a commercially available watercraft called the Raider Double and one patent, the Philippe patent. The Raider Double is a self-proclaimed “waveski” designed for operation by two paddlers, but with three seats. It was made in Australia and sold in the United States in the 1980’s. The Philippe patent claims a kayak with a single bench seat, with multiple seating positions.

Glenwa argues on appeal that claims 23 and 24 of the ’063 patent and that claims 1 and 2 of the ’912 patent were anticipated by the Raider Double and by the Philippe patent. Glenwa also challenges the validity of claim 13 of the T77 patent on the basis of obviousness in light of the Raider Double. For the reasons explained below, this court concludes that none of the trial court’s analysis of any of these arguments was correct.

A. Invalidity for Anticipation in claim 23 of the ’063 Patent

The trial court’s holding of validity was based in part on the conclusion that the “Raider Double” does not anticipate any of the relevant claims of any of the three patents in question because the Raider Double is a “waveski.” The trial court found that waveskis, unlike sit-on-top kayaks, “do not employ a prominently V-shaped bow or hull to increase water displacement, and are more similar to surfboards, which are designed to plane on top of the water.” Old Town, 229 F.Supp.2d at 1161. In particular, the court emphasized that although today’s experts in the field may refer to waveskis as a type of sit-on-top kayak, this is a result of the *924changes in the market after the issuance of the relevant patents.

Clearly this conclusion about the Raider Double cannot stand, given in the face of this court’s new claim construction. The Niemier patents actually do not claim any particular hull shape. If the only thing that distinguished the Raider Double from the patents was that the Raider Double is a waveski and therefore does not have a V-shaped or displacement hull, then the Raider Double is actually covered by the patents. Thus they are invalid because designating it as a waveski would be a non-distinguishing description. Indeed, at oral argument Old Town agreed that claim 23 would be anticipated under the claim construction here adopted.

In the case of the ’063 patent, we simply reverse the trial court’s decision. Claim 23 is invalid as a matter of law. To invalidate by anticipation, a single prior art reference must show each and every limitation either expressly or inherently. 35 U.S.C. § 102(b). The Raider Double has all of the elements of claim 23, as we construe them, and Old Town does not dispute it.4

B. Invalidity for Obviousness in Claim 24 of the ’063 Patent

As for claim 24, it is identical to claim 23 except for an additional seating surface with corresponding footwells straddling the seat. At oral argument, Old Town conceded that claim 24 would be invalid under the claim construction here adopted. We agree. While the additional elements of claim 24 are not present in the prior art so there is no anticipation, their addition would be obvious in light of the Raider Double itself. Thus, claim 24 is equally invalid. The missing elements are obvious because it is uncontested that the Raider Double actually has a third seat and footwells straddle two out of the three seats. Three things are necessary for the Raider Double’s third seat and additional footwells to suggest the version of a third seat and footwells in claim 24. First, the disclosure of the third seat must teach a surface that is sufficiently distinct that it counts as a “seating surface.” Since it is not contested that the Raider Double has distinct seats, at least for two of its seats, this element appears to be taught. Second, the Raider Double’s second set of footwells apparently do not straddle a seat as Old Town correctly argues the claims require; rather it is the first set of foot-wells that straddle the second and third seats. On a craft that has footwells that straddle along with footwells that are merely in front of a seat without straddling, one configuration clearly teaches its substitution by the other configuration. The straddling footwells teach a substitution of the non-straddling footwells with straddling footwells. Third, claim 24 by reference to claim 23 requires that two of the three seats be “longitudinally spaced.” Not only does this requirement not claim any particular distance between the two seats,5 but if in claim 24 the third seat is injected between the first two seats from *925claim 23, the first two seats would remain “longitudinally spaced” (if they were so to start with). There is no indication in the patent that the third seat in claim 24 must be one that permits a third person to be seated in the craft while the other two seats are occupied, or that allows all three people to paddle at once. Such spacing is not required and therefore the absence of any such teaching in the Raider Double does not make claim 24 non-obvious.

C. Invalidity for Anticipation in the ’912 Patent

For the two claims in the ’912 patent, it is not clear whether there were any additional differences between the claimed elements and the Raider Double prior art because of the way the trial court truncated its analysis after finding a difference in hull shape. The record is not sufficiently complete to permit us to know whether all the elements of the claims can be found in the Raider Double. This court, therefore, vacates the trial court’s grant of summary judgment of validity based on the lack of anticipation by the Raider Double for claims 1 and 2 of the ’912 patent and remands it to the trial court to complete the comparison between the Raider Double and the claims as newly construed.

D. Invalidity Based on the Philippe Patent

Because the Philippe patent described a bench seat and we have construed the claims in the ’912 patent to require seats that are structurally distinct from their surroundings, we conclude that the Philippe patent does not anticipate any of the claims.

E. Invalidity for Obviousness in the ’177 Patent and the ’912 patent

In addition, also in its validity assessment, the trial court explained that it relied on a number of other pieces of evidence in arriving at its conclusion that none of the patents was obvious in light of the prior art. In particular, the court found that the Raider Double was not sufficiently analogous to a sit-on-top kayak to make it obvious how to “fashion a conveniently short multiple-rider kayak.” Old Town, slip op. at 16. As proof the court cited the fact that no one else combined the features of pre-existing watercraft to create a similar watercraft before Niemier did, the commercial success of Niemier’s models, and the fact that there is evidence that others in the field (including Glenwa’s president) think that Niemier was a pioneer in the industry.

This obviousness analysis by the trial court was incorrect for several reasons. To review an obviousness assessment we must review the trial court’s findings of fact in four areas for clear error, including: (1) the scope and content of the prior art; (2) the differences between the claims and the prior art; (3) the level of ordinary skill in the pertinent art; and (4) secondary considerations. Graham v. John Deere Co., 383 U.S. 1, 17-18, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966). Even though the trial court correctly stated the four factual inquiries necessary to determine obviousness, it did not complete such an analysis. In fact, the main analysis of obviousness that the trial court really did do was to assess several secondary factors — the commercial success of the invention, the satisfaction of a long-felt need, and the failure of others to solve the same problem.

The only other factual finding, presumably on the question of the scope of the relevant prior art, by the trial court was finding that the Raider Double was non-analogous art and therefore too remote to be considered relevant prior art. We disagree. A reference is analogous art *926if it is within the inventor’s field of endeav- or. In re Clay, 966 F.2d 656, 658-59, 23 USPQ2d 1058, 1060 (Fed.Cir. 1992). If the court finds the reference to be outside of the field, the court must determine whether the reference is “reasonably pertinent to the problem” facing the inventor. Id. In light of this court’s claim construction that there is no limit on hull shape in the patents, the claimed subject matter and what counts as analogous art are clearly also broader than the trial court determined. In view of what is claimed in the patents, the Raider Double is within the inventor’s field of endeavor. That means that a systematic comparison of the elements in claim 13 of the ’177 patent and claims 1 and 2 of the ’912 patent with the features of the Raider Double is necessary to determine whether each claim is obvious. For example, the trial court must determine whether the Raider Double taught towards a sit-on-top kayak, including (but not limited to) a hull, three contoured seats, planar spacing surfaces between the seats, and straddling footwells. Additionally, an inquiry into the level of ordinary skill of one in the art is also necessary as the record is incomplete on this question as well.

We remand for the district court to determine whether claim 13 of the T77 patent and claims 1 and 2 of the ’912 patent are obvious over the Raider Double in light of the correct claim construction and according to the legally required analysis as explained above.

III. Infringement

The trial court granted Old Town’s motion for summary judgment of infringement based on the determination that the T77 patent was infringed by the Cobra Tandem under the doctrine of equivalents (with respect to the “seating surfaces are contoured to provide comfortable seating areas” limitation) and for literal infringement of the other two patents by all three of Glenwa’s models. Glenwa appeals the grant of summary judgment on all of these grounds.

At the very least, in light of our claim construction, there are genuine issues of material fact as to whether any of the accused kayak models infringe any of the patents and the grant of summary judgment of infringement for all of the claims cannot stand. Because claims 23 and 24 of the ’063 patent are invalid, there is no need for us, or the trial court on remand, to consider whether any of Glenwa’s kayaks infringe those claims. For the other two patents, the answer is more complicated.

In its infringement analysis the trial court held that the Cobra Tandem literally infringes every element of claim 13 of the T77 patent “except for the provision for a third seat.” Old Town, 229 F.Supp.2d at 1165. That provision, the court concluded, is infringed by the Tandem under the doctrine of equivalents because a generally flat area of flexible molded plastic for sitting performs substantially the same function, in substantially the same way, to obtain the same result as the corresponding element in the T77 patent calling for a third seating surface “contoured to provide comfortable seating areas.” Id. The court also concluded Glenwa failed to establish prosecution history es-toppel that would prevent Old Town from “asserting that a generally planar, or less contoured, area can be a seating surface.” Id. at 1165.

Because of this court’s construction of the term “contoured” in the ’177 patent, the determination of whether a seat that is “generally planar” meets the limitation of “a third contoured seat” must be redone in order to know whether the Cobra Tandem infringes. We cannot determine on the *927record before us whether any genuine issues of material fact exist regarding infringement, either literal or under the doctrine of equivalents. We must, therefore, vacate and remand so that the district court can: (1) compare the Cobra Tandem with the literal requirements of the T77 patent under the correct claim construction; (2) ascertain, if there is no literal infringement, whether application of the doctrine of equivalents is estopped by amendment-based or argument-based prosecution history (see Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki, 535 U.S. 722, 122 S.Ct. 1831, 152 L.Ed.2d 944 (2002)); and (3) determine, if no prosecution history estoppel applies, whether the Cobra Tandem included equivalents of all of the limitations of the T77 patent under the correct claim construction that are otherwise missing.

The trial court also determined that all three Glenwa kayak models literally infringe claims 1 and 2 of the ’912 patent. This, however, was also based on an erroneous claim construction. As the trial court did not specify how the features of the accused models met the limitations of the patent, this court cannot determine whether the accused models infringe the ’912 patent as correctly construed. Therefore, the grant of summary judgment for infringement of claims 1 and 2 of the ’912 patent is vacated and remanded for the trial court to perform the same infringement analysis as is prescribed for the T77 patent for each of the accused models.

CONCLUSION

We therefore reverse-in-part and vacate-in-part the district court’s grants of summary judgment of validity and infringement. We remand the case for further proceedings consistent with this opinion.