OPINION OF THE COURT
This appeal presents the question whether a tamperproof plastic milk-bottle closure, Crisci U.S. Patent No. 3,504,818 (1970), is obvious in light of prior art. The patent was assigned to Northern Engineering and Plastics Corporation (NEPCO). Roger Eddy was employed by NEPCO in 1966-67 and contributed to development of the closure. In 1976, Eddy, having terminated his employment with NEPCO some years before, started manufacturing Crisci-type closures without a patent license. In October 1978, NEPCO sued Eddy for infringement. Eddy defended and counterclaimed on a number of grounds. After a trial, the district court held the Crisci patent invalid on the ground that its subject matter was obvious under 35 U.S.C. § 103 (1976). NEPCO appealed.1 We affirm.
I.
Since at least 1958 those who design plastic bottle closures have been aware of the tamperproofing function that can be served by bottles with ratchet-toothed necks and caps flimsily connected to ratchet-toothed rings.2 The two sets of ratchet-toothing, one set on the bottle neck and the other on an annular ring, will engage only when the cap is unscrewed, thus exerting a force on the ring sufficient to break the “frangible” (/. e., breakable) interconnecting “elements” between ring and cap. However, the tam-perproof feature is useless if the ring breaks off of the cap during screw-on, because the only function of the connection between cap and ring is to provide assurance to the consumer that the milk bottle has not been opened since it left the bottler.3
The closure at issue here, like others that have been developed, cannot be screwed off *335without breaking the frangible elements between the cap and the ring. As is apparent from examination of the patents listed in note 2, supra, the feature of the closure that requires breakage of the connection in order to unscrew the cap was well-developed prior to the invention embodied in Crisci. The advance embodied in Crisci, however, is in the method by which it allowed the cap simply to be screwed onto the bottle without breaking the connection.
The force exerted by the engaging of the ratchet teeth only operates when the cap is being unscrewed. However, the ratchet teeth also create a force that tends to cause breakage of the frangible elements when the cap is being screwed onto the bottle. This force that operates during screw-on can be analyzed into two components. On one hand, the ring is subject to expansionary pressure during screw-on. The ratchet arrangement only works because the diameter of the ring is sufficiently small so that the teeth on the ring fit into the spaces between the teeth on the bottle when the cap is tinscrewed. But when the ring is rotated relative to the bottle during screw-on, these teeth must pass over each other. The gliding of one set of teeth over the other tends to cause the ring to expand, in the sense of having its diameter enlarged, when the cap is screwed onto the bottle.4 Moreover, because the ring resists expansion, there is an inevitable drag on the ring during screw-on. In other words, while the screwing on of the cap tends to pull the ring around in the same clockwise circular motion as the cap, friction, engendered by the squeezing of the tooth-ends on the ring over the tooth-ends on the neck of the bottle, tends to impede this clockwise motion of the ring.
The success of any ratchet-toothed closure will depend on the relationship between (a) the strength of the frangible connectors between cap and ring, and (b) the magnitude of the forces of “expansion” and “drag” that are transmitted, during screw-on, to these frangible points or elements. Roughly speaking, the former must exceed the sum of the latter. Other constraints on closure design exist. Obviously, the strength of the connectors must be limited so that the consumer can break them by screwing the cap off of the bottle. Furthermore, expansion and drag must not be reduced beyond the point at which such reduction impairs the ability of the ratchet teeth to effectively engage upon screw-off. Thus, while the strength of the connection between cap and ring must exceed the strength of the combined expansionary and dragging forces exerted on the connection during screw-on, the former is subject to a ceiling and the latter is subject to a floor.
Fisehbach (the German patent) attacks this problem by wholly eliminating transmission of drag from the ring to the frangible connectors. This is accomplished by means of an extra rib/cam arrangement.5 None of the other patents really addresses the problem of relieving stress upon the frangible connectors, although Babiol, like Crisci, uses flexibility to relieve stress on the ring and ring “teeth” (in Babiol, the ring teeth are referred to as “pawls”).
The Crisci patent attacks the problem in a simpler, more elegant manner. Expansion and drag are absorbed by the irregularity of the outward expansion of the ring itself. Flexibility of the ring allows it to expand at certain points along its circumference without exerting a great amount of *336expansionary pressure at others. The crucial feature is “circumferential spacing” — i. e., separation of the necessary expansion points (/. e., the points along the circumference of the ring at which teeth are placed) from those places along the circumference of the ring at which the frangible connectors are attached to the ring. For this reason, it is significant that, as NEPCO points out, the other patents fail to disclose
angular projections [i. e., teeth] on the inside of the ring which are circumferen-tially spaced from the frangible elements, and . .. outward flexing of the annular ring between the frangible elements to permit slipover engagement of the angular projections with ratchet sections on the bottle neck.
Reply Br. at 7 (emphasis added).
It appears that, in effect, Crisci involves nothing more than the prior known designs (a) absent alternate ring-teeth and connecting points between ring and cap, and (b) incorporating a slight rotation of the remaining ring-teeth relative to the connecting points. Also, to enable these changes to improve the screw-on performance of the cap, the ring is made (perhaps) more flexible.6 The question is whether this alteration constituted a patentable invention; more particularly, whether it “would have been obvious ... to a person having ordinary skill in the art to which said subject matter pertains.” 35 U.S.C. § 103 (1976).
II.
Although the question of obviousness is ultimately a legal one, Sakraida v. Ag Pro, Inc., 425 U.S. 273, 280, 96 S.Ct. 1532, 1536, 47 L.Ed.2d 784 (1976), its resolution must be predicated on the factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966). A court must, in a case such as this, determine (1) the scope and content of the prior art; (2) the differences between the prior art and the claims at issue; and (3) the level of ordinary skill in the pertinent art. Id. at 17, 86 S.Ct. at 693.
We look first to the prior art. As did the district court, we take the prior art to consist in those patents listed in note 2 supra. Those patents (with the exception of Fischbach) make no provision for reducing, during screw-on, the stress on the frangible elements connecting the cap and the ring. Fischbach does incorporate means for overcoming this problem, although it does so by means of an additional structure that increases the production cost of the device.7 Crisci, however, attacks the problem in a conceptually opposite manner, simply reducing the number of elements in the cap and ring.8
Having determined the prior art and its significance to the Crisci innovation, we inquire into the ordinary level of skill in the pertinent art. The district court correctly identified the pertinent art as the design and manufacture of closures but it did not ascertain the level of ordinary skill in that art at the time Crisci applied for his patent. On this record, however, we believe that we are in as good a position as the district court to determine that level and we see no need to remand for this purpose. The number of patents issued for bottle closures over the years between 1958 and 1968, in addition to the seminal patent issued in 1942 to Sharp, would seem to indicate an intense interest and sophisticated level of skill in the art. The various types of clo*337sures then currently available and the materials used to manufacture them were, as found by the district court, familiar to those possessing ordinary skill in the art. Elimination of unwanted breakage without utilizing the costly production step (needed in Fischbach) of separating the ribs and cams was a typical problem for which closure manufacturers were apt to redesign their products. Cheevely, an engineer with considerable experience in the injection mold industry and in the design of tools for caps, closures, and special projects in the plastics industry, testified that he too would have looked for ways to eliminate the ribs and cams that made the Fischbach patent expensive to produce and that it would have been natural “to have gone the direction Mr. Eddy did to produce a cap” less expensively which functioned effectively.
III.
We proceed from the factual questions to the legal question whether, to those skilled in the art of tamperproof closures, the jump from the prior art to Crisci entailed the work of the “skillful mechanic” or that of the “inventor.” Hotchkiss v. Greenwood, 52 U.S. (11 How.) 248, 267, 13 L.Ed. 683 (1851). As the Supreme Court has recognized, the difficulties in deciding this question “are comparable to those encountered daily by the courts in such frames of reference as negligence and scienter.” Graham v. John Deere Co., 383 U.S. at 18, 86 S.Ct. at 694. “[Wjhether the variation relied upon in any particular case is anything more than ordinary mechanical skill is a question which cannot be answered by applying the test of any general definition.” McClain v. Ortmayer, 141 U.S. 419, 427, 12 S.Ct. 76, 78, 35 L.Ed. 800 (1891), quoted in Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., 340 U.S. 147, 151 n.6, 71 S.Ct. 127, 129, 130 n.6, 95 L.Ed. 162 (1950). Although the answer to the question must be sought by probing the hypothetical frame of mind of one whose aptitude for spotting improvements to the products of his art are presumably greater than ours, we must also guard against that hindsight which renders every pure concept natural, intuitive, and “obvious” just because it is fundamentally simple.9
Perhaps because of these difficulties, compounded by the fact that the judiciary “is most ill-fitted to discharge the technological duties cast upon it by patent legislation,” 10 the law embodies a requirement that courts defer, in at least two ways, to determinations of validity by the Patent Office.
It is a fundamental canon governing judicial consideration in this field that a presumption of validity attaches to patents issued by the United States Patent Office. Not only has a unanimous Supreme Court noted that “patentees are heavily favored as a class of litigants by the patent statute,” but this court has stated on several occasions that the burden of proving patent invalidity is a heavy one. Moreover, such invalidity must be demonstrated by “clear and convincing proof.”
Universal Athletic Sales Co. v. American Gym, Recreational & Athletic Equipment Corp., 546 F.2d 530, 540 (3d Cir. 1976), cert. denied, 430 U.S. 984, 97 S.Ct. 1681, 52 L.Ed.2d 378 (1977) (footnotes omitted); 35 U.S.C. § 282 (1976).
The presumption of validity that attaches to a patent issued by the U.S. Patent *338Office, however, is “weakened” when the party asserting obviousness comes forward with “significant prior art not considered by the Patent Office.” U.S. Expansion Bolt Co. v. Jordan Industries, Inc., 488 F.2d 566, 569 (3d Cir. 1973); accord, Aluminum Co. of America v. Amerola Products Corp., 552 F.2d 1020, 1024 & n.7 (3d Cir. 1977); Escoa Fintube Corp. v. Tranter, Inc., 631 F.2d 682, 691 (10th Cir. 1980). Eddy claims that he has done so, because the prior art presented to the Patent Office, comprised of the claims to which Bardell and Homma were drawn, was characterized at the time by Crisci’s attorney as including only rings secured to the cap by “longitudinally extending frangible portions.” Crisci’s frangible elements, on the other hand, are “radially extending,” and so, says Eddy, are the frangible elements on Fischbach and Fapex. NEPCO, on the other hand, argues that Homma and Bardell effectively incorporated all of the significant prior art disclosed in the patents considered by the district court.
The district court did not purport to rely on any supposed deficiency in the showing made before the Patent Office.11 Moreover, the question whether the frangible portions of the prior art closures extended radially or longitudinally is not, in itself, crucial to the relevance of the patents claimed in the district court below to constitute prior art. Rather, the relevant issue is whether the prior art before the Patent Office disclosed ways of ensuring that the caps could be screwed on to bottles without breaking off the rings.12 Whether or not the frangible elements were radial or longitudinal was not, in the context of this case, relevant to that question. It appears that insofar as Bardell and Homma dealt with the ability of closures to afford evidence of tampering, they were, in effect, indistinguishable from Crisci or the prior art raised in the district court.13
Thus we are not wholly persuaded that this is a case in which the presumption of patent validity has been “weakened.” However, we need not resolve this presumption issue because we believe that Eddy has sustained the rigorous burden of proving that the Patent Office erred in granting the *339Crisci patent, even assuming that the references submitted to the Patent Office constituted, as Crisci argues, all of the significant prior art. We recognize that inventions perform an exceedingly important and useful function, especially in our complex, industrial society; but to justify a patent monopoly, an improvement must demonstrate an inventive genius which makes a distinctive contribution to “the Progress of . .. useful Arts.”14 In this case the change in the mechanical construction by reducing the thickness of the annular ring (thereby increasing its resiliency), reducing the number of angular projections, and spacing the projections apart from the frangible elements on the inner surface of the ring did not make the admittedly improved device patentable.15 It was not the creation of an “inventive” faculty, even though it realized better results, but rather displayed the sort of useful intelligence that any “skilled mechanic” would have been called upon to exert daily in the course of exercising his skill. Unlike the invention in Eibel Process Co. v. Minnesota & Ontario Paper Co., 261 U.S. 45, 43 S.Ct. 322, 67 L.Ed. 523 (1923), Mr. Crisci’s rearrangement of the “old” elements produced no unanticipated result; Crisci did not discover the source of a previously impenetrable difficulty in tamper-proof cap design. Cf. id. at 67-68, 43 S.Ct. at 329-30. Instead, Crisci deduced, from the rather simple and totally apparent geometrical properties of his product, a logical means for minimizing the otherwise necessary stress on the frangible elements of his closures during screw-on.16
Moreover, we see no evidence that closure designers were struggling unsuccessfully to achieve the results that Crisci achieved. In fact, it appears that many successful tam-perproof plastic bottle closures have been designed and manufactured without the “circumferential spacing” that forms the heart of Crisci’s improvement. We draw from this the conclusion that in all likelihood, the only reason that the Crisci design had not been “invented” prior to Mr. Cris-ci’s effort was not so much to be found in the elusiveness of the concept he grasped, but rather in the relative triviality (in practical terms) of his refinement. Thus, we agree with the district court that based on the level of the prior art, that improvement to the bottle closure described in the Crisci patent “would have been obvious to a person having ordinary skill in the art of manufacturing closures.”
IV.
The judgment of the district court will be affirmed. Each side to bear its own costs.