On summary judgment, the United States District Court for the Southern District of New York declared claims 1-5, 7-11, 13, and 15 of U.S. Patent No. 4,452,053 (the ’053 patent) invalid for obviousness. Because the trial court improperly resolved genuine issues of material fact, this court vacates the judgment and remands for further proceedings.
I.
The ’053 patent covers thin metal needles for machines that automatically knit yarn into fabric. The knitting machine industry *879refers to the ’053 needles as “low-profile” needles. A figure from the ’053 patent, reproduced below, shows a “low-profile” needle. The labels accompanying- the figure denote various portions of the needle,
[[Image here]]
As illustrated, a low-profile needle has a “head” that is bent like a hook, a shaft (or shank) that has at least one “bridge” (the bridge is a “stem” supported by two or more “guidepieces” that extend to the “back” of the needle), and at least one “butt” attached to the shaft in the region of the bridge. The portion of the stem that runs parallel to the needle back is called a “segment.” The “first segment” lies between the head and the butt.
Depending on its size, a knitting machine requires up to 4000 or more needles. Each needle rests with its back against the bottom of a groove. A cam strikes and releases the butt to move the needle forward and back within the groove. This reciprocating needle motion within the groove accomplishes the knitting.
To optimize output, fabric manufacturers, of course, wish to run the machine as fast as possible. The major limit on the speed of automated knitting machines, however, has been the tendency of needle hooks to break. Even a single broken needle leaves a long disfiguring line in the knitted fabric. To avoid the downtime required to replace a broken needle, machine owners often must run the machines at slower speeds in order to minimize needle breakage.
In 1976, one of the appellants, Sulzer Mo-rat GmbH (Sulzer Morat), commissioned a study to determine the causes of hook breakage. The researchers observed the effects of various design changes on the vibration of knitting needles. For example, when the cam’s surface strikes the butt of a needle, it causes shock waves and vibrations in the needle. After using high-speed photography to observe the vibrations, the researchers determined that a smaller first segment height produced superior results. Sulzer Morat filed a German patent application on May 12, 1978, and a U.S. patent application on May 11, 1979, claiming priority from the filing date of the German application. The ’053 patent issued on June 5,1984.
The only independent claim of the ’053 patent, claim 1, states:
A stamped knitting-tool which can be employed in knitting machines and has a head bent like a hook, a shaft having at least one bridge which consists of two gui-depieees extending as far as the back of the knitting-needle and a stem arranged above the back to bridge across the gui-depieces, and at least one butt coupled to the shaft in the region of the bridge, where the stem includes at least one first segment arranged between the head and the butt, said tool including means for controlling vibration within the needle and preventing tool breakage at increasing knitting speeds, said means including the shape of said needle which has said first segment ... having a length (l) of at least eight millimeters (8 mm) and a height (h) of at most eleven tenths of a millimeter (1.1 mm).
(emphasis added). During the prosecution of the ’053 patent, the examiner rejected claim *8801 as obvious in view of U.S. Patent No. 3,464,237, issued to Kohorn in 1969. -The Kohorn ’237 patent taught the use of cutouts at the back of the needle to reduce vibration, “thereby substantially lowering the frequency of needle breakage” and increasing the speed of knitting machines “by twenty percent.” Kohorn ’237 patent, col. 2, lines 5-12. Kohorn taught that the cutouts should extend approximately one-half the distance from the back of the needle towards the front. The examiner reasoned that one of ordinary skill in this art, knowing of solid shank needles with a shank height of 2.2 mm and applying the teachings of the Ko-horn ’237 patent, would create a needle with the dimensions of claim 1. In response, Sul-zer Morat pointed out that no known prior art needles had a shank height of less than 2.85 mm, let alone 2.2 mm. The examiner allowed the application, and the ’053 patent issued. Sulzer Morat was the initial assign-ee.
In 1985, Monarch Knitting Machinery Corp. (Monarch) filed a declaratory judgment action against Sulzer Morat in the United States District Court for the Southern District of New York. Monarch asserted, inter alia, that the ’053 patent was invalid for obviousness. Shortly thereafter, the district court stayed the proceedings pending a reexamination by the Patent and Trademark Office '(PTO) in light of newly discovered prior art.
The new prior art was a Japanese publication that disclosed two needles with shank heights of 1.92 mm and 2.0 mm. Once again, the examiner rejected the ’053 patent in view of the Kohorn ’237 patent, this time in combination with the Japanese publication. On appeal, the Board of Patent Appeals and Interferences concluded that the examiner had not shown a prima facie case of obviousness, because the needles disclosed in the Japanese publication had first segment lengths of less than 8 mm. A reexamination certificate confirming all of the original claims issued on November 13,1990.
Meanwhile, Sulzer Morat had assigned all of its rights in the ’053 patent to Theodor Groz & Sohne and Ernst Beckert Nadelfa-brik Commandit-Gesellshaft (Groz-Beekert). After the reexamination, Groz-Beckert commenced an infringement action against Monarch and its suppliers. The district court consolidated the two actions.
On September 7, 1994, Monarch filed a motion seeking summary judgment that all claims of the ’053 patent were invalid for obviousness. The district court granted the motion as to claim 1, but concluded that the record did not contain sufficient evidence to rule on the other claims. On April 19, 1996, Monarch filed a similar summary judgment motion for claims 2-5, 7-11, 13, and 15 on a supplemented record. The district court granted the motion and entered final judgment on the declaratory judgment claim for invalidity under Fed.R.Civ.P. 54(b). This appeal followed.
This court concludes that genuine issues of material fact remain. Although Groz-Bec-kert later supplemented the record with additional evidence of objective indicia of non-obviousness, this evidence was not before the trial court for the first motion for summary judgment. Except as otherwise stated, this court relies only on the evidence available to the district court when it entered summary judgment on claim 1. Thus, this court’s conclusion necessarily follows for all the claims at issue, regardless of the fact that the district court invalidated all other claims in a second motion on a supplemented record.
II.
This court reviews a district court’s grant of summary judgment de novo by reapplying the standard applicable at the district court. See Conroy v. Reebok Int’l, Ltd., 14 F.3d 1570, 1575, 29 USPQ2d 1373, 1377 (Fed.Cir.1994). Summary judgment is appropriate only when “there is no genuine issue as to any material fact and ... the moving party is entitled to a judgment as a matter of law.” Fed.R.Civ.P. 56(c). In its review, this court draws all reasonable inferences in favor of the nonmovant. See United States v. Diebold, Inc., 369 U.S. 654, 655, 82 S.Ct. 993, 994, 8 L.Ed.2d 176 (1962). Moreover, in rendering a decision on a motion for summary judgment, a court must “view the evidence presented through the prism of the substantive evidentiary burden” that would *881inhere at trial. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 254, 106 S.Ct. 2505, 2513, 91 L.Ed.2d 202 (1986). The moving party “bears the burden of demonstrating the absence of genuine issues of material fact.” Conroy, 14 F.3d at 1575.
In this case, the district court rendered summary judgment on the question of obviousness. Obviousness is ultimately a determination of law based on underlying determinations of fact. See Richardson-Vicks Inc. v. Upjohn Co., 122 F.3d 1476, 1479, 44 USPQ2d 1181, 1183 (Fed.Cir.1997). These underlying factual determinations include: (1) the scope and content of the prior art; (2) the level of ordinary skill in the art; (3) the differences between the claimed invention and the prior art; and (4) the extent of any proffered objective indicia of nonobviousness. See Graham v. John Deere Co., 383 U.S. 1, 17-18, 86 S.Ct. 684, 693-94, 15 L.Ed.2d 545 (1966). Thus, to review a summary judgment of obviousness, this court first determines anew whether the.record raises any genuine issues about these critical facts. In doing so, this court remains cognizant of the statutory presumption of validity, see 35 U.S.C. § 282 (1994), and of the movant’s burden to show invalidity of an issued patent by clear and convincing evidence, see Ryko Mfg. Co. v. Nu-Star, Inc., 950 F.2d 714, 716, 21 USPQ2d 1053, 1055 (Fed.Cir.1991). If facts remain in dispute, this court weighs the materiality of the dispute, i.e., whether resolution of the dispute one way or the other makes a difference to the final determination of obviousness.
Given the occasional use of archaic terminology in the district court’s opinion, this court also emphasizes that the standard for patentability is the statutory standard. The inquiry is not whether there was a “real discovery of merit” or whether the claimed invention offered a “new solution,” but whether the claimed subject matter as a whole “would have been obvious at the time the invention was made to a person having ordinary skill in the art.” 35 U.S.C. § 103(a) (Supp. 1 1995).
III.
To ascertain the scope of the prior art, a court examines “the field of the inventor’s endeavor,” Shatterproof Glass Corp. v. Libbey-Owens Ford Co., 758 F.2d 613, 620, 225 USPQ 634, 638 (Fed.Cir.1985), and “ ‘the particular problem with which the inventor was involved,’” Stratoflex, Inc. v. Aeroquip Corp., 713 F.2d 1530, 1535, 218 USPQ 871, 876 (Fed.Cir.1983) (quoting In re Wood, 599 F.2d 1032, 1036, 202 USPQ 171, 174 (CCPA 1979)), at the “time the invention was made,” see 35 U.S.C. § 103(a). The district court defined the problem as “designing the stem segment of a knitting needle ... [to] minimize[ ] needle head breakage and thus maximize[ ] the operating speed of an industrial knitting machine.” (emphasis added). The ’053 patent, on the other hand, describes the inventor’s problem as “providing [knitting needles] with a means which avoids head breakages or lets [breakages] start to an extent worth mentioning only at higher knitting speeds.” ’053 patent, col. 1, lines 48-51. The district court’s formulation of the problem confronting the ’053 inventors presumes the solution to the problem — modification of the stem segment. Defining the problem in terms of its solution reveals improper hindsight in the selection of the prior art relevant to obviousness. See, e.g., In re Antle, 58 C.C.P.A. 1382, 444 F.2d 1168, 1171-72, 170 USPQ 285, 287-88 (CCPA 1971) (warning against selection of prior art with hindsight). By importing the ultimate solution into the problem facing the inventor, the district court adopted an overly narrow view of the scope of the prior art. It also infected the district court’s determinations about the content of the prior art.
The district court based its conclusion of obviousness heavily on its determination that the prior art showed a “trend” towards increasingly lower stem segment heights. A “trend” might very well constitute a suggestion or teaching to one of ordinary skill in the art to make “minor” changes from the prior art in accordance with that trend to produce the claimed invention. Cf. In re Chu, 66 F.3d 292, 298, 36 USPQ2d 1089, 1094 (Fed.Cir.1995) (stating that even when changes from the prior art are “minor” or “simple,” *882an inquiry must be made as to whether “ ‘the prior art provides any teaching or suggestion to one of ordinary skill in the art to make the changes’ ” (quoting Northern Telecom, Inc. v. Datapoint Corp., 908 F.2d 931, 935, 15 USPQ2d 1321, 1324 (Fed.Cir.1990))). Whether the prior art discloses a “trend” is a question of fact. The existence of a trend depends on the content of the prior art, i.e., what the prior art would have taught one of ordinary skill in this art at the time of this invention.
The district court identified four prior art needles as establishing the “trend.” All four needles were made by Groz-Beekert: The first was made in 1965 and included “some cutouts” in the shank. The second was made in 1967 and had a cutout that produced a segment height of 1.7 mm. In 1972, Groz-Beekert made the third needle with a segment height of 1.6 mm, and in 1973, it made the “meander needle” with a segment height of 1.4 mm. The district court reasoned that the differences between the ’053 needle and the fourth needle in the district court’s series — the meander needle — was' only one of degree.
Before the trial court, however, may examine the existence of a trend towards decreasing first segment heights among those needles with a first segment configuration, it must resolve an antecedent question. By defining the inventor’s problem in terms of its solution, the district court missed this necessary antecedent ' question, namely, whether the prior art contains a suggestion or motivation to combine references to form a trend. See Carella v. Starlight Archery, 804 F.2d 135, 140, 231 USPQ 644, 647 (Fed.Cir.1986) (“Obviousness cannot be established by combining the teachings of the prior art ... absent some teaching, suggestion or incentive supporting the combination.”); Orthopedic Equip. Co. v. United States, 702 F.2d 1005, 1012, 217 USPQ 193, 199 (Fed.Cir.1983) (“It is wrong to use the patent in suit as a guide through the maze of prior art references, combining the right references in the right way so as to achieve [a desired result].”). Stated otherwise, what would have impelled one of ordinary skill to recognize a relationship between stem segment height and the hook breakage problem? If those of ordinary skill would have recognized a relationship, then, and only then, does the trial court proceed to examine whether the prior art in fact contains a coherent teaching about that relationship. Thus, before proceeding to find a trend, the trial court must discern whether one of ordinary skill would have had a motivation to combine references to form a trend.
In this regard, the Kohom ’237 patent contains a suggestion to use “cutouts” as a solution to the hook breakage problem. On the other hand, the record also contains evidence suggesting that one of ordinary skill would not have had a motivation to combine references to form a trend. Instead, those of skill in the art were experimenting with many different methods for reducing hook breakage. For example, the record shows that in 1966, the problem of hook breakage at high speeds was recognized in a German patent. It ascribed this problem to shock waves propagating from the butt to the hook. The patent’s proposed solution was to dampen these shock waves by putting several small notches at the back and front of the shaft of the needle or, alternatively, to have a portion of the needle shaft made of a different mechanical structure. Several other pri- or art references recognized the problem of hook breakage at high speeds, theorizing other causes and/or prop.osing other solutions. These solutions included forming a triangular-shaped, spring-acting portion in the shank to absorb oscillation (1971), coating the shank with shock-absorbing polymeric material (1972), forming curved, undulating portions in the shank to absorb longitudinal shock (1974), placing vertical offsets in the form of bridges in the shank (1974), using scalloped cutouts rather than rectangular cutouts (1975), and constructing the shaft of the needle out of several shock-absorbing intermediate members (1976). All of these references stated the problem as preventing hook breakage at high speeds. Each of these references proposed a different solution. Thus, this evidence creates a genuine issue as to whether those of ordinary skill would have had- a motivation to combine needles with varying stem segment heights to form a trend.
*883Beyond the motivation to combine question, the trial court must also determine whether the prior art forms a trend. The record contains evidence calling into question whether, even among needles employing a first segment configuration, the entire content of the prior art shows a trend towards decreasing first segment heights. The record shows that the four needles selected by the district court were made for four different knitting machines, not replacements for the same machine over time. Given this evidence, a reasonable inference can — -and on summary judgment must — be drawn in favor of the patentee that the four examples chosen by the trial court do not show a trend. Moreover, as early as 1957, Groz-Beckert made a needle having a segment height of 1.5 mm. Because this 1957 needle “bucks the trend” perceived in the four selected by the trial court, a reasonable fact finder could conclude that the entire body of prior art does not evince a trend toward decreasing segment heights. In sum, after properly ascertaining the scope of the prior art, genuine issues of fact remain about the content of that art. Specifically, a further genuine issue of fact remains as to whether a trend toward decreasing first segment heights ex-, isted in the prior art.
IV.
Next this court examines the trial court’s determination on the level of ordinary skill in the art. The district court properly determined that “a person of ordinary skill in the art of needle design ... was a mechanical engineer with a working knowledge of knitting machinery and needle design.” Neither party disputes this conclusion. In conjunction with this determination, the district court discussed two alleged instances of independent development of the same invention. Although this court has noted the relevance of contemporaneous independent invention to the level of ordinary knowledge or skill in the art, see In re Merck & Co., 800 F.2d 1091, 1098, 231 USPQ 375, 380 (Fed.Cir.1986), it has also acknowledged the view that this evidence is relevant as a secondary consideration, see Medtronic, Inc. v. Daig Corp., 789 F.2d 903, 906, 229 USPQ 664, 667 (Fed.Cir.1986). Therefore, this court will address the evidence of contemporaneous invention in that context.
V.
Turning to the objective criteria or the secondary considerations, this court examines first the trial court’s conclusions about commercial success. The district court concluded — and none of the parties disputes— that the low-profile needle enjoys commercial success and that this success is prima facie due to the needle’s differences from the prior art. The undisputed evidence shows that low-profile needles meet more than 70% of the requirements of original equipment knitting machine manufacturers. Moreover, low-profile needles command a sales price that is greater than 1.4 mm meander needles by about fifteen to thirty percent. The district court correctly determined that the record disclosed no genuine issue of fact about the existence and degree of commercial success.
The record also contains evidence of a long-felt need. However, the district court concluded that it was not an unsolved need because at least two companies had already designed low-profile needles before the ’053 priority filing date. Because Title 35 provides for interference proceedings, it implicitly recognizes that contemporaneous independent invention may not alone show obviousness. See Lindemann Maschinenfabrik GMBH v. American Hoist & Derrick Co., 730 F.2d 1452, 1460, 221 USPQ 481, 487 (Fed.Cir.1984). Accordingly, this court weighs evidence of contemporaneous invention “in light of all the circumstances,” id., especially in light of evidence of long-felt need.
The district court cited evidence of two contemporaneous independent developments. First, the record shows that while working with Mayer & Cie, a knitting machine manufacturer, Groz-Beckert designed a knitting needle having dimensions required by claim 1. A diagram of this needle design, dated March 1, 1978, discloses a needle with a first segment height of 1.0 mm and a length of 22 mm. In addition, the record shows that Fukuhara, one of Monarch’s sup*884pliers and an appellee in this case, independently designed a similar needle. An engineering drawing dated December 8, 1976, shows a needle with a first segment height of 1.1 mm over a length of 11 mm. These needles do not qualify as “prior art” under 35 U.S.C. § 102 or § 103(a), but are relevant to obviousness as a secondary consideration.
To rebut this evidence, appellants argue that the 1.0 mm needle was designed by Ernst Beck, the chief of Groz-Beckert’s needle design department. As such, appellants assert that he possessed extraordinary skill in the art of needle design and his labors do not aid the inquiry about obviousness to one of ordinary skill in the art. Without question, the focus for obviousness is on the level of ordinary skill in the art, but appellants point to no evidence other than Beck’s title that he possessed extraordinary skill.
Further, appellants also argue that the uninterrupted existence of a long-felt need in the art renders immaterial any evidence of contemporaneous development. Appellants reason that a long-felt need in the art is inherently in tension with concluding that an invention is obvious on the basis of independent contemporaneous development. Once again the appellants’ correct assertion sharpens the focus. The relevant secondary consideration is “long-felt but unsolved need,” not long-felt need in isolation. The question, then, is whether this evidence of contemporaneous development solved the need. On that point, the record discloses genuine issues of material fact.
For example, the record shows that Groz-Beckert had doubts about the ’053 needle design, even though Groz-Beckert had already itself designed a low-profile needle. This evidence justifies a reasonable inference that Groz-Beckert did not recognize the benefit of its needle design. Assuming the accuracy of these asserted facts, it is difficult to see how a solution solves long-felt need if its developer does not recognize that it does so.
Furthermore, the parties do not dispute that the Fukuhara needle was never produced. As for the Groz-Beckert needle, it was not produced until after the ’053 priority filing date. Even then, the Groz-Beckert needle was made initially only for Mayer & Cie. Because it was an entire industry that had endured the long-felt need, these facts— if proven — raise a reasonable inference that these two designs were isolated and failed to satisfy the long-felt need that existed at large. Although Monarch may ultimately prevail at trial on this issue, the district court erred in summarily concluding that the record showed no long-felt but unsolved need.
The district court did not address the existence of unexpected results. The record shows that smaller first segment heights produced less hook breakage and increases in knitting machine speeds. For example, of the needles comprising the alleged “trend,” the 1.6 mm needle permitted machine speeds to increase by 14%. The 1.4 mm needle permitted speeds to increase by 12%. The record also suggests that cutting metal away from a needle stem is “normal” and gave results that were expected “from an engineering point of view.” This evidence tends to suggest that decreasing the first segment height would result in machine speed increases.
Groz-Beckert does not point to any evidence of unexpected results that was before .the district court on the first motion for summary judgment. However, Groz-Bec-kert supplemented the record with evidence of unexpected results in response to the second motion. Thus, this evidence is only applicable to the summary judgment for claims 2-5, 7-11, 13,- and 15. For example, the record shows that ’053 needles increased machine speeds on the average about 15-25%. Other evidence, however, shows that some machines saw improvements of up to 40%. Moreover, there is evidence that hook breakage was no longer the limiting factor to increasing knitting machine speed. Thus, although previous needles had allowed machine speeds to increase, a reasonable fact finder could conclude that the needle allowed machine speeds to increase by 40% — an increase that was larger in magnitude than any previously attained. More significantly, regardless of the actual percentage increase in machine speeds, none of the previous needles had suddenly removed hook breakage as the major limit on knitting machine speeds. Both of these results — if proven — could sup*885port an inference of unexpected results. The evidence that some of ordinary skill were skeptical of the advantages of the ’053 needle design may also help show unexpected results. Given the possibility that the record could support these inferences, this objective criterion also remains a disputed question of fact.
A prior art reference may be considered to teach away when “a person of ordinary skill, upon reading the reference, would be discouraged from following the páth set out in the reference, or would be led in a direction divergent from the path that was taken by the applicant.” In re Gurley, 27 F.3d 551, 553, 31 USPQ2d 1130, 1131 (Fed. Cir.1994). General skepticism of those in the art — not amounting to teaching away — is also “relevant and persuasive evidence” of nonobviousness. Gillette Co. v. S.C. Johnson & Son, Inc., 919 F.2d 720, 726, 16 USPQ2d 1923, 1929 (Fed.Cir.1990). In effect, “teaching away” is a more pointed and probative form of skepticism expressed in the prior art. In any case, the presence of either of these indicia gives insight into the question of obviousness.
The district court concluded that there was no genuine issue as to this consideration. This conclusion, however, discounts record evidence to the contrary. The record contains ample evidence from which a reasonable faet finder could conclude that the art taught away from decreasing segment height. For example, U.S. Patent No. 4,036,-036, filed in March 1976, stated:
We must also be sure that we have not made the needle too light, which might be the case with needles which incorporate large cutouts or notches. While the ordinary engineer might think that the lighter needle with its reduced mass would always be beneficial in such a reciprocating motion mechanism, that is not always the case with knitting needles.
’036 patent, col. 2, lines 22-35.
The record also contains evidence of skepticism in the art. A sworn affidavit by Ger-hard Miiller, an engineer at Terrot, a knitting machine manufacturer, states: “[I]t was our common opinion that a needle should not be more ‘elastic’ than a meander needle with a stem height of 1.4 mm. Especially, we dreaded that such needles would stretch or lengthen....”
In addition, the declaration of Groz-Bec-kert’s expert, George Tay, stated: “I saw the flexing and bending of the needle portion and the consequent vibrating of the needle head as a phenomenon which it would be counterproductive to increase.” In a puzzling passage, the district court concluded that this declaration did not raise a contested issue of faet, because it was “refuted” by an “admission” of one of the ’053 inventors. However, even if there was such an “admission,” the testimony of an inventor is simply strong refuting evidence, not dispositive evidence at the summary judgment stage of an obviousness proceeding.
Groz-Beckert supplemented the record with other evidence of skepticism in response to the second motion for summary judgment. Thus, this evidence is only applicable to the summary judgment for claims 2-5, 7-11, 13, and 15. For example, William Ross, a needle designer for Torrington Company, stated: “Far from seeing larger cutouts as a solution to hook breakage, Torrington’s needle designers, and I myself, believed that larger cutouts would lead to increased needle failures (including increased hook breakage)_” Kurt Wiednhofer, a needle designer for Groz-Beckert, stated: “[Ujntil 1979, [Groz-Beckert’s] needle designers, including myself, apprehended that reduction below the 1.4 mm stem height would result in a whippy needle susceptible, inter alia, to butt jamming and undesirable needle lengthening.” Finally, Groz-Beckert submitted evidence of a computer modeling analysis that Sulzer Morat had performed before the ’053 invention. The analysis had predicted that reducing the height of stem segments would cause increased, hook breakage. Thus, when the ’053 inventors informed Martin Elsasser, a knitting engineer in charge of development at Sulzer Morat, of their prototypes and the results, he told them that “the suggestion of reducing the height of the stem segment, as their experiments suggested, was not a good idea.” Based on the computer modeling analysis, he believed that “reducing the height of the stem segment would cause in*886creased needle hook breakage, contrary to the results of the experiments.”
In sum, factual issues remain about skepticism and teaching away at the time of invention. These genuine issues of material fact preclude resolution on summary judgment of the question of the existence of these secondary considerations.
VI.
The evidence in the record raises genuine issues of fact concerning the content of the prior art as well as several secondary considerations. This court concludes that these issues are material to an obviousness determination in this case. Accordingly, this court vacates the district court’s grant of summary judgment' and remands for further proceedings.
COSTS
Each party shall bear its own costs.
VACATED and REMANDED.