MANSFIELD, Circuit Judge:

Mayer, Rothkopf Industries, Inc. and Mayer Cie, GMBH Co. ("Mayer"), defendants below, appeal from a judgment entered by Judge Eugene H. Nickerson in the District Court for the Eastern District of New York after a non-jury trial, holding that they had infringed plaintiff Morris Philip's patent for a Knitting Machine Cam Mechanism, U.S. Pat. No. 3,026,695, and rejecting their claim that the Philip patent was invalid. We affirm, substantially for the reasons stated by Judge Nickerson in his decision that the Philip patent adequately claimed a novel and non-obvious patent in a cam mechanism for interlock knitting machines. However, our affirmance does not represent approval of his alternative holding that the Philip patent would have been valid even if directed at all circular knitting machines in general.

Circular knitting machines, whose mechanics are at the heart of this dispute, come in two types: cylinder-only, and cylinder-and-dial (interlock). Both types of machines have alternating long and short needles through which yarn passes. As these needles move up and down or in and out, the yarn passing through them becomes looped. The knitting process involves passing successive loops through each other.

Cylinder-only machines are less complicated than cylinder-and-dial machines. In cylinder-only machines, the long and short needles project vertically downward. The up and down motion of the long and short needles is regulated by devices called cams, with one cam track for the long needles and one for the short ones. The cams are essentially a pattern of grooves in a cylinder, which roll from peaks to valleys as the machine turns and thereby control whether the needles are in a fully extended (knitting) position or a fully retracted (clear) position. Cylinder-only machines are useful for turning out a simple knitted stitch.

Cylinder-and-dial machines, by contrast, are used to produce an interlock stitch or other related stitches that involve the interweaving of two stitches knitting simultaneously (e. g., doubleknits). In these machines, a dial with horizontal short and long needles projects into the same knitting area as a cylinder with vertical needles and intermeshes its stitch with the stitch produced by the cylinder. Like the cylinder, the dial has cams which regulate the in-and-out movement of its long and short needles. However, the timing of movements in these interlock machines is significantly more delicate than that in cylinder-only machines, because the needles from the dial can easily run into the needles from the cylinder. For many years knitting machines avoided the danger of smashing needles into each other by structuring the cams on the cylinder and dial so that only one set of needles at a time moved in each needle bank. If the short needles were moving to clear and knit, the long needles would remain basically inactive, and vice versa. This alternating stoppage of half the needles resulted in a loss of efficiency that was offset by the increase in quality of the knits produced.

For a long time inventors prior to Philip sought ways to make knitting machines operate more productively. Though the parties are in dispute over exactly what advances had been accomplished, it seems clear that inventors had managed to increase the number of feeds in some interlock machines (such as the Stibbe machine) by decreasing the width of the needle gauge and by changing the angle at which the dial and the cylinder met each other. In addition, some cylinder-only machines (such as the Sinfra machine) had employed cams which led needle movement to overlap — that is, which placed long needles in motion before the short needles had stopped, and vice versa. But no one had proposed or built an interlock machine that employed overlapping needles, because of an apparently widespread belief that such overlapping motion would lead to smashing of needles.

Philip developed a method of carving the cams in the cylinder and dial of an interlock knitting machine so that needle movements would overlap. His invention enabled the number of feeds that could be fit into a 30-inch knitting machine to be increased by 50% (from 32 to 48 feeds) without danger of smashing the needles. The patterns of these cams were quite complex, even as compared to the patterns of the cams on the machines that employed overlapping needle action in a cylinder-only knit. The cam mechanism from an overlapping cylinder-only machine could not have been employed in an interlock machine without leading to the smashing of needles.

Philip sought a patent for his cam mechanism beginning in 1958, which was issued in 1962. While the patent was pending, he at first had difficulty convincing anyone to try his invention, because it ran against the accepted learning with respect to interlock machines. Finally, he induced the Supreme Knitting Co. to build a prototype from his drawings. Supreme was so skeptical of the likelihood of success that it made Philip accept full financial responsibility for the cost of the machine it built. The machine turned out to work even better than Philip had anticipated. For some unexplained reason the act of knitting an interlock stitch before the alternating needle has fully retracted leads to a higher quality knit, and makes it possible for the machine to adjust automatically when a needle drops a stitch. Philip sold his invention to Singer Co., which employed it in all new interlock machines it constructed and which paid him over a million dollars during the life of the patent.

Mayer is a manufacturer of knitting machines. There is no question that if the Philip patent is valid, Mayer is guilty of infringement. However, Mayer claims the patent to be invalid on two alternative theories. First, it contends that the Philip patent attempted to embrace all circular knitting machines, including cylinder-only machines, not just cylinder-and-dial machines designed for knitting the interlock stitch. As such, Mayer suggests, the invention represents a restatement of prior art, or at least represents only an obvious advance, since overlapping needles existed in cylinder-only machines before 1962 and since other interlock machines had more than 40 needles. Second, Mayer claims that if the patent is directed at interlock machines only, it is invalid under 35 U.S.C. § 112, which provides in relevant part:

"The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.

"The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention."

The Philip patent, cast in what is called the Jepson form, claimed an improvement "[i]n a camming mechanism for a circular knitting machine employing two sets of needles and having two groups of cams one for each of said needle sets." The words "interlock stitch," "cylinder," or "cylinder and dial" do not appear in the claims. The specifications, however, begin with a paragraph that states:

"This invention relates to circular knitting machines and especially to machines for knitting the interlock stitch, or other stitch employing two sets of needles in each bank. The two sets of needles may be mounted on either a conventional cylinder or a dial, or on both cylinder and dial, and each set has its own group of operating cams. Machines for knitting the interlock stitch have both cylinder and dial needles."

Although interlock stitches and cylinder-and-dial machines are not specifically referred to elsewhere, the language of the specifications discusses the relationship between the cylinder needle movement and the dial needle movement "in knitting machines such as this." The drawings of the cylinder and dial cams attached to the specifications did not indicate how the cylinder and the dial cams were to be intermeshed. But no prior interlock machine patent had made such an indication in its drawings, either. Judge Nickerson accepted the trial testimony of Philip's expert, Dr. Peter Brown, who asserted that one skilled in the knitting machine art would recognize that the specifications for the Philip patent were aimed at construction of cam mechanisms for an interlock machine.

The file wrapper on the Philip patent, which contains all of the correspondence between Philip and the Patent Office, reveals, in response to a series of queries by the Patent Examiner, continuing efforts by Philip to make clear that he claimed a camming mechanism solely for interlock knitting machines.

At the outset the Examiner's suggestion that the application was drawn toward a knitting machine rather than a set of cams was met by Philip's narrowing the claims to a cam mechanism useful in such machines and pointing out that in a well-known textbook on interlock machines "the dial and cylinder cams of a circular interlock machine are illustrated in a manner similar to the drawing of the present application." The Examiner then objected on the ground that if Philip wanted to put his application in the context of interlock knitting machines alone, as the Examiner speculated to be the case, Philip could have done so by referring to the Morley or Scott patents on interlock machines in his original specifications. To this Philip responded with a submission making it clear that he intended his application to be directed toward interlock machines by referring to a chapter on interlock knitting from a recognized textbook on knitting machines and to the expired patents of Scott and Morley, which demonstrated that the elements of an interlock machine were well known in the art. Although the Patent Office repeatedly found problems with Philip's claiming methods, it ultimately granted his application. Its decision allowing the patent did not specify whether the patent covered cylinder-only machines as well as cylinder-and-dial machines, or just cylinder-and-dial machines alone.

Judge Nickerson upheld the validity of the patent against Mayer's claims that it was not novel, was obvious, and was insufficiently set forth in the claims and specifications. Reading the claims and specifications as fairly indicating an invention directed specifically at interlock machines, he saw no basis for an assertion that this invention was unworthy of a patent. Then, assuming for the sake of argument that the patent application was directed at all circular knitting machines, he held alternatively that the Philip patent represented a novel and non-obvious development even for cylinder-only machines. Later, when Mayer moved for a new trial claiming newly discovered evidence, Judge Nickerson denied the request, saying that the evidence was available to Mayer at trial and that the proffered new evidence would not have changed his decision anyway.

DISCUSSION

The record evidence is clear that Philip designed a useful, novel, and non-obvious invention for a camming mechanism in interlock knitting machines. Prior to his invention no interlock machines had overlapping needle movement. Popular opinion among those skilled in the art maintained that such overlapping was impossible notwithstanding its existence on a limited and withstanding its existence on a limited and unimpressive basis in cylinder-only machines. All other increases in production by interlock machines had been accomplished through changes in the needle gauge or in the angle of alignment between the cylinder and the dial. Philip's invention, which increased yarn feeds by 50% with a better quality resulting product, represented a sufficient development over prior art to be both novel and non-obvious, as required by 35 U.S.C. §§ 102, 103.

Of course, the recognition that Philip produced a novel and non-obvious invention by no means ends our inquiry into the validity of the Philip patent. We have long held that a patent is invalid, no matter how useful and original the invention it protects, if the applicant for the patent claims the invention so broadly that it encompasses already established prior art. Maclaren v. B-I-W Group Inc., 535 F.2d 1367, 1372-73 (2d Cir.), cert. denied, 429 U.S. 1001, 97 S.Ct. 531, 50 L.Ed.2d 612 (1976); Bishop and Babcock Mfg. Co. v. Fedders-Quigan Corp., 270 F.2d 102, 106 (2d Cir. 1959); Foxboro Co. v. Taylor Instrument Co., 157 F.2d 226, 232 (2d Cir. 1946), cert. denied, 329 U.S. 800, 67 S.Ct. 494, 91 L.Ed. 684 (1947). As Judge Learned Hand observed in Foxboro:

"A patentee who claims broadly must prove broadly; he may not claim broadly, and recede as he later finds that the art unknown to him has limited his invention. That is the chance he must take in making broad claims; if he has claimed more than he was entitled to, the statute does not give him a locus poenitentiae.... "

Id.

The Supreme Court explained the reasons for this rule in Permutit Co. v. Graver Corp., 284 U.S. 52, 60, 52 S.Ct. 53, 55, 76 L.Ed. 163 (1931), where it noted that a patentee must fully "inform the public during the life of the patent of the limits of the monopoly asserted, so that it may be known which features may be safely used or manufactured without a license and which may not." When a patentee takes out a broad patent, he discourages efforts to produce inventions and obtain patents in all areas at the periphery of his claims. Inventors cannot be permitted to reap the fruits of excessive exclusionary monopoly power when they make overbroad claims, and at the same time to narrow their claims, after the issuance of an overbroad patent when it is revealed that the excessive claims include some ideas already embraced within the prior art. The validity of a patent must be judged according to the invention claimed, not merely according to the invention actually designed.

Our recent decision in Maclaren v. B-I-W Group Inc., supra, thoroughly explores the proper methods for discerning the exact nature of the invention claimed in a patent application. Appellants come into this court faced with the burden of overcoming two hurdles. First, patents issued by the Patent Office must be presumed valid. 35 U.S.C. § 282; Champion Spark Plug Co. v. The Gyromat Corp., 603 F.2d 361 (2d Cir. 1979). Second, although the question of a patent's validity is one of law, Graham v. John Deere Co., 383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966), findings of fact by the district court must be credited unless clearly erroneous. Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 123, 89 S.Ct. 1562, 1576, 23 L.Ed.2d 129 (1969). When the trial judge has explicitly relied on expert testimony as to matters which depended on such testimony, we have limited latitude to question his determinations. On the other hand,

"where ... the findings are based on documentary evidence which we are as competent to appraise as the district court, we have not hesitated to reject a determination of validity, either because the significance of the facts had not yet been fully perceived by the district court or there has been an error in the application of legal principles." Maclaren v. B-I-W Group Inc., supra, 535 F.2d at 1371 (citations omitted).

Maclaren dictates that with respect to the documentary evidence of Philip's patent application, it is the "claims which define the boundaries of a patent monopoly." Id. at 1372, quoting Great A P Tea Co. v. Supermarket Equipment Corp., 340 U.S. 147, 149, 71 S.Ct. 127, 129, 95 L.Ed. 162 (1950). However,

"[t]his rule does not preclude the court from looking to the specifications and drawings for the purpose of understanding and interpreting the claims of the patent. Indeed, to do otherwise would be to proceed in a vacuum. Neither accuracy nor fairness of decision would be furthered by demanding that a court conceptualize complex patent claims simply in the abstract, without resort to clarifying descriptions. Instead, `[w]hile the claims of a patent limit the invention, and specification cannot be utilized to expand the patent monopoly, . . . it is fundamental that claims are to be construed in light of the specifications and both are to be read with a view to ascertaining the invention.'" Maclaren, supra, 535 F.2d at 1372-73, quoting United States v. Adams, 383 U.S. 39, 48-49, 86 S.Ct. 708, 712-713, 15 L.Ed.2d 572 (1966).

Read literally the claims of the Philip patent are not expressly limited to cylinder-and-dial machines for knitting the interlock stitch. On the other hand the claims, standing alone, are ambiguous to the extent that they are unclear as to whether a "camming mechanism for a circular knitting machine employing two sets of needles and having two groups of cams one for each of said needle sets" necessarily refers to cylinder-and-dial machines or can also refer to cylinder-only machines. After reviewing the claims in the light of the specifications, however, as would be done by a reader skilled in the art of building knitting machines, we are satisfied that the specifications redeem any uncertainty caused by the lack of clarity in the claims. Dr. Brown, an expert in the precise field, who was credited by Judge Nickerson, testified that one skilled in the knitting machine art would read the claim and specifications as directed to interlock knitting machines. We need not spell out the modifying language in detail except to point out that it begins by noting that the invention relates "especially to machines for knitting the interlock stitch," and seems clearly directed at explaining how the machine works when cylinders and dials intermesh. The camming mechanism is described as one for a machine employing two sets of needles and having two groups or banks of cams for each of the needle sets, which describes an arrangement for interlock machines, i. e., one group of two cams for the cylinder needles and another group of two dial cams. Although there could be one set of two cam groups for the cylinder alone, with only one cam mechanism for the dial, the machine would still be an interlock one. Since overlapping had existed to some degree but without any marked success in a non-interlock machine, Philip had no reason to extend his claimed cam mechanism to such a machine, where (unlike the incorporation of the mechanism in interlock machines) it would not be useful. Moreover, the file history shows a consistent position by Philip that his invention was aimed at interlock knitting machines, including repeated reference to textbooks and chapters on interlock stitching as well as attempts to prove the employment of the invention in the Supreme Knitting Co. prototype and to cite prior interlock machine art as a clarification device.

The situation before us is distinguishable in significant respects from that which we faced in Maclaren, heavily relied on by Mayer, in which the challenged patent was invalidated. First, in Maclaren the claims and specifications, unlike those here, failed entirely to mention the new and useful improvement upon which the patentee was relying in his defense to the allegation of patent invalidity, i. e., the one-hand, instant-folding capability of a new collapsible baby carriage frame. Instead, the patentee claimed a collapsible support assembly in general which embodied many features of the prior art. We there noted:

"Even though the specifications and drawings might be used for the purposes of construing the claims, there is simply no claim language on the subject of simultaneous folding, ambiguous or otherwise, to be construed. Even if we went further and treated the two paragraphs of the specifications as part of the claims, they would not suggest simultaneous folding. On the contrary, the statement is made that ` Continued folding, beyond the condition dotted in FIGURE 2, although possible in theory by continuing to bring the handles 7D, 8D together, is in fact not practicable, so the user assists folding by directly raising the forward end of the structure (or pressing down the handles toward the ground) until the folded condition of FIGURE 3 is reached.' (Emphasis added) (E 82). This language seems to point to the opposite conclusion, indicating that at least two steps are required to collapse the structure. To discover that it might be simultaneously folded one would have to build the collapsible support assembly described in the patent and test it. Accordingly, we must conclude that, despite the patent's broad claims with respect to a collapsible support assembly, it does not include simultaneous folding as part of those claims." 535 F.2d at 1374.

The Maclaren patentee's failure to claim the new invention made recognition of that new invention impossible. Here, by contrast, Philip fully disclosed the new camming mechanism that he had invented, in his claims, specifications, drawings, and file correspondence with the Patent Office. The sole question is whether that fully described invention was aimed at all circular knitting machines or at only cylinder-and-dial machines.

Second, the Maclaren application revealed a recurring effort by the patentee to broaden the basis for its claims, in the apparent hope that an increased segment of the industry producing collapsible support assemblies would come under the new patent's shadow. This calculated overclaiming that uses the invention in an attempt to embrace the prior art represents one of the primary techniques of unjustifiably chilling competition that the patent laws requiring specific claims were designed to prevent. Permutit Co. v. Graver Corp., supra.

In this case, unlike Maclaren, Philip could not gain any advantage by overclaiming in order to embrace dial-only or cylinder-only machines. First, there is no such thing as a "dial-only" knitting machine. Second, other inventors already knew how to produce some overlapping movement in cylinder-only machines by using camming mechanisms that were considerably less complicated than the Philip invention's mechanism. As Judge Nickerson aptly suggested below, "so long as the patentee has not left a scarecrow to frighten away others from the broader area he should not be deprived of the fruits of his invention." Since Philip's invention had practical value only when employed in cylinder-and-dial machines, no chilling of inventiveness with reference to any other type of machine could have resulted from his patent's claims.

For these reasons we affirm the district court's decision that the Philip patent must be upheld as claiming only an interlock knitting machine camming mechanism, which was new and useful. However, we do not agree with the trial court's alternative holding that the patent would be valid even if directed at circular knitting machines in general, including cylinder-only machines. The record reveals that camming mechanisms which provided some overlapping of needles in a cylinder-only machine had existed in the prior art. The Sinfra machine, for instance, disclosed overlapping needle movement in a cylinder-only machine as early as 1952, several years before Philip applied for and obtained his patent. We therefore do not adopt this portion of the district court's otherwise persuasive decision.

Appellant's next claim, that the district court erred in finding that the Philip patent met the requirement of 35 U.S.C. § 112 that it enable one skilled in the art to make and use the invention, requires little discussion. The requirement is met by disclosures that are reasonably clear to one skilled in the art, Loom Co. v. Higgins, 105 U.S. 580, 585-86, 26 L.Ed. 1177 (1881); Ansul Co. v. Uniroyal, Inc., 448 F.2d 872, 879 (2d Cir. 1971), even though some experimentation may be necessary, Minerals Separation, Ltd. v. Hyde, 242 U.S. 261, 270-71, 37 S.Ct. 82, 86, 61 L.Ed. 286 (1916); Ansul Co., supra, 448 F.2d at 878. Dr. Brown, an expert found credible by the trial court, testified that the patent's specifications were sufficient to enable one skilled in the art to construct the invention in a knitting machine. The Supreme Knitting Machine Co. confirmed this by building an interlock machine in which the Philip invention was incorporated according to the patent's drawing. In view of this proof we cannot label Judge Nickerson's finding clearly erroneous.

Finally, we affirm the trial court's refusal to order a new trial under Fed.R.Civ.P. 59 in response to appellants' asserted discovery of new evidence. Such decisions are reversible only if the trial judge clearly abused his discretion. Diapulse Corp. of America v. Birtcher Corp., 362 F.2d 736, 744 (2d Cir. 1966). The Sinfra machine, described in appellants' summary of the "new" evidence, is merely a cylinder-only machine with overlapping needle movement. The Stibbe article, the other piece of new evidence proffered by appellants, showed only an interlock machine that increased the number of needles by narrowing the gauge of the needles and adjusting the angle at which the cylinder and dial met. None of these pieces of evidence could change our result here. Furthermore, we see no reason to disagree with Judge Nickerson's finding that this evidence was available to Mayer before the trial ended.

Affirmed.