Kaiser Industries Corp. v. McLouth Steel Corp.

CELEBREZZE, Circuit Judge. Appellants filed an action in the United States District Court for the Eastern District of Michigan complaining that Appellee had infringed the Appellants’ patented process for refining steel with pure oxygen. Appellee first defended by contending that a know-how agreement, for which Appellee paid $100,000 to Appellants, constituted a license to use the Appellants’ process. The District Court held against the Appellee on that point, Henry J. Kaiser Company v. McLouth Steel Corporation, 175 F.Supp. 743 (D.C.Mich.1959), and in an interlocutory appeal was affirmed by this Court. McLouth Steel Corporation v. Henry J. Kaiser Company, 277 F.2d 458 (6th Cir.1960). These rulings made it possible for the Appellee to contest the validity of the Appellants’ patent at the trial on infringement.1 The handling of the protracted trial on infringement by the District Judge was a monumental task, which he performed ably and well. After unusually long and complex litigation, involving several years of pretrial research, experimentation and discovery and over an entire calendar year of trial, the District Court concluded that the patentees were the inventors of a process that was not anticipated by the prior art and would not have been obvious to one skilled in the art at the time of the discovery. The Court held the patent invalid, however, for the reason that the specifications did not support the claims and the claims did not measure the invention as required by 35 U.S.C. § 112 and for the further reason that the claims covered the prior art.1a Henry J. Kaiser Co. v. McLouth Steel Corp., 257 F.Supp. 372 (D.C.Mich.1966). From that judgment the Appellants have perfected the instant appeal contesting the Court’s rulings covering the claims and specifications ; and in addition the Appellee has contested the District Court’s holdings on obviousness and inventorship. In reaching its decision, the District Court wrote a long and comprehensive decision ruling on the issues of anticipation, obviousness, inventorship, compliance of the specifications with Section 112, and compliance of the claims with Section 112. In addition the District Court ruled that the doctrine of file wrapper estoppel applied to prevent the Appellants from claiming the advance that they now allege is the gist of their invention. Since we conclude that the claims overclaim the invention as represented to the Patent Office by the patentees and that the claims cover the prior art, we find it unnecessary to consider the correctness of every ruling of the District Court. On the other hand, a consideration of the factual background of this case and of the errors we find in the District Court’s ruling on obviousness will aid in understanding the reasoning by which this Court has reached its decision. Plaintiffs-Appellants in this action are the owner, the exclusive world-wide licensing firm, and the exclusive licensing agent in the United States of the patent in suit. Vereinigte Oesterreieh-ische Eisen-und Stahlwerke Aktiengesell39 schaft (hereinafter VOEST), a steel company owned by the Austrian Government, is the present owner of the patent in suit by virtue of assignment from the paten-tees, who at the time of the experimentation that led to the application for the patent in suit were employees of VOEST. Brassert Oxygen Technik AG. (hereinafter BOT), a Swiss corporation, was formed in the early 1950’s by VOEST and several other European steel companies as the exclusive world-wide licensing agent for a pool of patent rights relating to the use of oxygen in steelmak-ing, which included the patent in suit after its issuance in 1957. In May of 1954 BOT and the Henry J. Kaiser Company (hereinafter Kaiser), a Nevada corporation, entered into an agreement by which Kaiser became the exclusive licensing agent for BOT in the United States and thus entitled to sue for infringement of the patent in suit when it became a part of the BOT pool of patents. Appellee McLouth Steel Corporation (hereinafter McLouth), is a Michigan corporation having a steel plant located in Trenton, Michigan, whose operations are alleged to infringe the patent in suit. In this action McLouth is accused of infringing Patent No. 2,800,631 (hereinafter the Suess Patent), which was issued July 23, 1957, to four Austrian steelmakers, Drs. Theodor Edward Suess, Herbert Trenkler, Hubert Hauttmann and Rudolf Rinesch on the basis of a process developed at Linz, Austria, in 1949, (hereinafter the L-D Process). Only two prior art patents are relied upon by Mc-Louth in this appeal: The Schwarz German Patent No. 735,196 dated July 3, 1943 (hereinafter the Schwarz Patent); and the Miles Belgian Patent No. 486,316 dated March 1, 1947 (hereinafter the Miles Patent). In addition, McLouth relies upon experimentation by Dr. Robert Durrer and Dr. Heinrich Hellbruegge, (hereinafter the Gerlaf ingen experiment) which was conducted in Gerla-fingen, Switzerland, in 1948 and the spring of 1949 and the results of which were fully disclosed to the patentees, as proving non-inventorship. Producing steel from iron ore involves at least two processes. Iron ore, which exists in nature in the form of oxides of iron, is first processed through a blast furnace where the oxygen is removed from the iron ore. This process reduces the iron oxide to relatively pure iron, but it does not remove the other impurities present in the iron ore. In fact, the blast furnace process increases the carbon content of the resultant product, which is known as pig iron. To produce steel from the pig iron, the amounts of these impurities, carbon, silicon, phosphorous, manganese, and sulphur, must be reduced to an acceptable level without adding any significant amount of oxygen or nitrogen. Accomplishing this result is the purpose of the second process involved in steelmaking: the refining process. The patents involved in this suit are concerned exclusively with the refining process, which involves essentially oxidizing the impurities and removing them from the molten pig iron. Carbon combines with oxygen to form carbon monoxide or carbon dioxide and is exhausted as a gas. Silicon, phosphorous, sulphur, and manganese form various oxides and are absorbed into a slag layer that is formed on top of the metal bath by the addition of lime and other materials before and during the operation. Since the various impurities react with oxygen at different rates, the problem of steelmak-ing is to control the refining process so that sufficient amounts of all the impurities are removed. Various methods have been devised to accomplish this refining process. The oldest method is the acid Bessemer process, which was developed in the 1850’s and involves blowing air into a molten bath of pig iron from below through tuyeres in the bottom of a converter. This process was unsatisfactory for most pig iron, however, because a basic slag for phosphorous removal could not be formed in the converter, which had an acid refractory lining. As a result, only pig iron with an already sufficiently low phosphorous content could be satisfactorily refined by this method. — 40 — The Thomas process, or basic Bessemer process, was devised during 1879 to permit the use of a basic slag capable of removing phosphorous and sulphur. This process also proved unsatisfactory: because air was blown, an excessive nitrogen content developed in the steel produced, and because the blowing was from below, insufficient iron oxide was formed in the slag for phosphorous removal, thus requiring an afterblow for adequate removal of phosphorous. Although the electric furnace has recently been used for refining, the high cost of the operation has limited its usefulness to the making of specialty steels. Therefore, since the 1880’s the principal method of refining steel has been the open hearth process. As late as 1953, approximately eighty-nine per cent of the steel produced in the United States came from open hearth furnaces, which produce steel of excellent quality. The process, however, has the disadvantage of high capital and operating costs and a relatively slow rate of production. An average open hearth heat requires eight to ten hours, while a blow in a Thomas operation requires only approximately eighteen minutes.2 The basic process combines the best qualities of the Thomas and the open hearth processes: it produces a steel of excellent quality in a relatively short blowing period; and it has low capital and operating costs. Since VOEST began the first commercial operation of the basic oxygen process at Linz in November 1952, the process has had a considerable impact on the steelmaking industry. According to the Appellants, more than seventy-five per cent of all new steel-making facilities constructed since 1954, or now in the planning stage use the basic oxygen process. Sixteen companies in the United States have begun or planned oxygen blowing facilities, and sixty-four companies in twenty-three countries have installed, or are installing, the new process. The process has demonstrated undoubted advantages over previous methods of refining steel. Its use has resulted in substantial savings to those companies now employing the process; and its use in the future will undoubtedly result in substantial savings and improved production for the steel industry as a whole. McLouth is admittedly using the basic oxygen process. When in 1953 McLouth decided to expand its operations, it learned of the basic oxygen process being used at Linz and sent one of its officers to Linz to observe the operations. Further information about the process was obtained from Kaiser, and in 1954 Kaiser agreed to furnish know-how, engineering knowledge, and consultation services to McLouth at a cost of $100,000. With this assistance McLouth constructed its OP-1 steelmaking shop, the design of the operation being copied from the one at Linz. Although McLouth interpreted the know-how agreement as a license to use the basic oxygen process, Kaiser brought suit for patent infringement. When the Suess Patent was finally granted, the complaint was amended to allege infringement of that patent. In the District Court, McLouth raised virtually all of the standard defenses in a patent case from invalidity and non-infringement to patent misuse.3 Although the District Court’s judgment for McLouth was based on the Section 112 defenses and although our affirmance is based on the same defense, a discussion of the District Court’s treatment of the defense of obviousness will better provide the necessary background. The primary thrust of McLouth’s defense appears to be this: when properly practiced the process taught in the prior art Schwarz and Miles Patents produces the same results in the same way as the L-D Process disclosed in the specifications of the Suess Patent. In asserting the defense McLouth evidenced some confusion — 41 — as to whether this contention should be raised under the defense of anticipation or under the defense of obviousness; and McLouth’s ambivalence concerning the proper defense seems to have contributed to the District Court’s failure to make certain essential findings of fact on the issue of obviousness. We share some of the Appellee’s perplexity concerning the proper defense,4 but in view of the disposition made of the case find it unnecessary to resolve that issue. Appellants contend that * * * it lies ill in the mouth of this copyist to assert * * * invalidity * * But when the defense is that the advance claimed by the patentees would have been obvious to one skilled in the art, more is involved than a weighing of the equities between the parties. The question becomes one of the public policy in granting patent monopolies and of applying the constitutional and statutory standards requisite to such a grant. Graham v. John Deere Co., 383 U.S. 1, 86 S.Ct. 684 (1966). In essence McLouth contends that the Linz experiments were nothing more than the development of know-how and engineering knowledge to facilitate the establishment of the first commercial basic oxygen process. If this contention is true, the Appellants had nothing to sell but know-how, which they sold to Mc-Louth for $100,000. To counter this contention, the Appellants contend that the patentee’s new concept “ * * * was the direct antithesis of the thinking then prevalent in the steelmaking art.” This penchant of the patentees for creating a verbal antithesis to distinguish every patent that might serve as an interference reference has raised other problems of divining exactly what novel element the patentees claim and how that novel element might be meaningfully distinguished from the prior art. Appellants say that the Suess Patent teaches “avoid-anee of deep penetration”, where the Schwarz and Miles Patents taught “deep penetration”. But “avoidance of deep penetration” as an abstract phrase means no more to one skilled in the art than does “deep penetration”. Depth of penetration can gain meaning only from the limitations placed upon it by the patents and from the knowledge that may be attributed to an ordinary worker skilled in the art. The nature of McLouth’s defense compels the Court to attempt to distinguish inventive experimentation from mere development of know-how in an art with which the Court has no familiarity. The ultimate distinction is one of law because it amounts to a determination of obviousness or nonobviousness. Great A. & P. Tea Co. v. Supermarket Equip. Corp., 340 U.S. 147, 71 S.Ct. 127, 95 L.Ed. 162 (1950). Several basic factual inquiries are essential, however, to establish a foundation upon which the legal conclusion can be based. The scope and content of. the prior art must be determined ; the differences between the prior art and the claims at issue must be ascertained ; and the level of ordinary skill in the pertinent art must be resolved. Graham v. John Deere Co., 383 U.S. 1,17, 86 S.Ct. 684 (1966). In the Graham case the United States Supreme Court noted that secondary considerations, such as longfelt but unsolved needs, commercial success, and the like, might have some relevance on the issue of obviousness. 383 U.S. at 17-18, 86 S.Ct. at 684. But the Court’s further analysis of the Scoggins Patent, 383 U.S. at 26-36, 86 S.Ct. at 684, shows clearly that these considerations are truly secondary and will not serve as a substitute for more direct evidence of obviousness. The difficulty we find with the District Court’s conclusion of non-obviousness is that the conclusion is based entirely upon these secondary eonsiderations. — 42 — 5 We do not have the benefit of a complete determination of the difference between the prior art and the patent in suit or of any findings concerning the level of skill in the pertinent art. In determining the prior art a court must consider only the art prior to the date of invention and not read into the prior art knowledge obtained from later disclosures. A foreign inventor, however, cannot establish a date of invention earlier than his first filing date in a foreign country, when that date is earlier than his first filing date in this country. 35 U.S.C. §§ 104 and 119. Patentees filed their first application for a patent on the L-D Process in Austria on January 31, 1950; so that is the earliest termination date for consideration of prior art. Prior to 1950 not a great deal had been done commercially with the use of pure oxygen in the refining process. McLouth contends that this lapse in the steel-making industry was due to the unavailability of a method for producing oxygen cheaply in commercial quantities. Appellants contend that the lapse was due to the lack of an operable method for refining steel with pure oxygen. In view of the conflicting evidence, the District Court found that it was unable to ascertain a date when the use of pure oxygen in commercial quantities became economically feasible but held that the unavailability of oxygen in commercial quantities would not necessarily have inhibited experimentation. 257 F.Supp. at 398. We agree with this finding of the District Court but must also note that the lack of commercial quantities of oxygen would affect the quality of the disclosures in the prior art on matters such as specific commercial operating conditions, demonstrative examples, and the like. Moreover, the commercial unavailability of oxygen would explain the absence of any prior attempts to apply the Schwarz and Miles Patents and would thus weaken any inferences that such a lapse in the industry might raise on the issue of obviousness. Thus, the importance of more direct evidence on the issue of obviousness becomes even more apparent. However, the District Court found three considerations dispositive of the issue of obviousness: (1) the presumption of validity of a patent; (2) the commercial success of the basic oxygen process; and (3) secondary considerations concerning prior attempts and failures at refining steel with oxygen. 257 F.Supp. at 406. Under the facts of this case, neither commercial success nor the presumption of validity will support a finding of nonobviousness. Commercial success is a relevant consideration only in a close case; it does not by itself establish the validity of a patent. Dow Chemical Co. v. Halliburton Oil Well Cementing Co., 324 U.S. 320, 65 S.Ct. 647, 89 L.Ed. 973 (1945); Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Co., 340 U.S. 147, 71 S.Ct. 127 (1950). Thus consideration of commer — 43 — cial success should be put aside until some conclusion is reached on the basis of more direct evidence of obviousness or non-obviousness. While it is well recognized that a patent will be presumed valid, 35 U.S.C. § 282, that presumption of validity is not conclusive and will be weakened when pertinent prior art was not considered by the Patent Office. Harvey v. Levine, 322 F.2d 481 (6th Cir. 1963); Aluminum Company of America v. Sperry Products Inc., 285 F.2d 911 (6th Cir. 1960); Tillotson Manufacturing Co. v. Textron, Inc., Homelite, 337 F.2d 833 (6th Cir. 1964); Felburn v. New York Central Railroad Company, 350 F.2d 416 (6th Cir. 1965). In the instant case one of the most pertinent prior art patents, the Schwarz Patent, was not cited as a reference by the Patent Office and, so far as the record indicates, was not considered by the Patent Office.6 Avoidance of the Miles Patent as a covering reference was accomplished only by the addition of the claim “avoidance of material agitation of the bath by the oxygen stream.” 257 F.Supp. at 404. The Schwarz Patent does not teach agitation of the bath by the oxygen stream; so that claim probably would not have distinguished the Schwarz Patent if it had been before the Patent Office. In similar circumstances this Court has considered that the presumption of validity was not a material factor in determining the validity of a patent. Harvey v. Levine, 322 F.2d 481 (6th Cir. 1963), supra. Although the Appellants emphasize the lack of any previous litigation concerning the validity of the Suess Patent and the lack of interferences in the United States Patent Office, the licensing structure of BOT negatives any favorable inferences that might ordinarily arise from such circumstances. When the patentees first developed their process, they became concerned as how best to exploit the process worldwide. They realized that the Schwarz and Miles Patents might serve as interferences or as covering references. So, in the Appellants’ words, “ * * * to resolve any conflicting claims * * * the exploiting firm of BOT was formed.” H. A. Bras-sert & Co., owner of the Schwarz Patent, and VOEST, along with several other European steel companies, were members of BOT from its inception. John Miles and Sons, Ltd., owner of the Miles Patent, became a member of BOT in 1952. The agreements with BOT were essentially license grant-backs: in exchange for placing in the BOT pool of patents their patents on certain uses of oxygen in steelmaking, the members obtained a license to utilize the processes of all other patents in the BOT pool. The agreements also contained provisions for the exchange of know-how and engineering knowledge. In 1954 Kaiser became a part of this arrangement. Meanwhile, Dr. Suess applied for an Austrian patent on the L-D Process on January 31, 1950; and on January 16, 1951, fifteen days before the one year limitation contained in 35 U.S.C. § 119 would have barred his application, Dr. Suess applied for a patent on the L-D Process in the United States Patent Office. The continuation-in-part application on which a patent finally issued was not filed until 1955. Aside from the antitrust implications that the BOT arrangement might have, see United States v. Singer Mfg. Co., 374 U.S. 174, 83 S.Ct. 1773, 10 L.Ed.2d 823 (1963), since by 1952 it owned all of the most pertinent prior art patents, a reasonable inference might be made that the prior holders of those patents had no great interest in contesting the Suess Patent once it was — 44 — issued.7 Also, the most pertinent prior art patents, Miles and Schwarz, were not the type for which an interference would be declared by the Patent Office. See 35 U.S.C. § 135. Hence, the prosecution in the Patent Office was not all that it could have been and would have been if the prior art patents had been owned by competitors of VOEST or had been pending or unexpired United States Patents. These considerations negative any favorable inferences that might be drawn from commercial success or a lack of previous litigation. We also find the secondary considerations that the District Court relied upon unpersuasive. The District Court found that there had been “ * * * [a] long-standing need for such an invention; [and] * * * many unsuccessful attempts to solve the problem.” 257 F.Supp. at 406. We find no evidence in the record to support that finding. As early as the 1850’s the Bessemer process taught that steel could be refined by jetting air into a molten bath of pig iron from tuyeres located in the bottom of a converter. Another line of patents going as far back as the Robert Patent in 1888 taught surface blowing with negligible penetration. 257 F.Supp. at 403. None of these processes proved successful when oxygen was substituted for air. Mere surface blowing causes a dangerous slopping condition, 257 F.Supp. at 400, and when oxygen is blown through tuyeres located in the bottom of a converter, the heat of chemical reaction at the gas outlets results in severe refractory damage and usually blows out the bottom of the converter. 257 F.Supp. at 378. But the apparently novel disclosure in the Schwarz Patent that gas could be blown from above with such a force that it would penetrate into a molten bath of pig iron “as a solid body” made all prior attempts to solve the problems confronted by the patentees wholly immaterial. Graham v. John Deere Co., 383 U.S. 1, 36, 86 S.Ct. 684 (1966). Since the Schwarz Patent was a German patent issued in 1943, it is reasonable to assume that because of World War II, it was available to the steelmaking industry as a whole by 1946 at the earliest. The Miles Patent issued in 1947 and made a disclosure similar to Schwarz’s. So any problems that might have been involved in deep blowing had existed for a relatively short period of time prior to the 1949 Linz experiments. See Dempster Bros., Inc. v. Buffalo Metal Container, 352 F.2d 420 (2d Cir. 1965). Other than the Gerlafingen experiments the record does not disclose any extensive attempts and failures to refine steel by deep blowing of oxygen between 1946 and 1949. The circumstances of this case do not allow an inference of such attempts and failures from the absence of any evidence that other companies had applied the process. As to whether other steelmakers independently arrived at the same results as VOEST, the BOT arrangement made it unnecessary for that part of the industry with experience in oxygen blowing to resolve the issue. Since that arrangement provided for the exchange of know-how and engineering knowledge, it would be impossible to separate the contribution of one from the contributions of the others to the basic oxygen process as it is practiced today. So whether the success of that process is attributable to the disclosure of the Suess Patent is unclear. In fact, as early as May of 1949, before the Linz experiments, VOEST entered into an agreement with several other European steel companies dividing the burden of experiments in the use — 45 — of oxygen steelmaking. To VOEST fell the fortuitous obligation of continuing the Gerlafingen experiments, while the other companies were to consider other uses of oxygen in steelmaking. This agreement eliminated several potential competitors of the patentees and negates any inference of extensive experimentation and failures by others skilled in the art. Why others did not avail themselves of the knowledge contained in the Schwarz and Miles Patents is not clear, but it is clear that such evidence is not conclusive on the issue of obviousness. Graham v. John Deere Co., 383 U.S. 1, 36, 86 S.Ct. 684 (1966). The failure to use a patent does not affect the validity of a patent or the persuasiveness of its teachings as a prior art reference. Til-lotson Manufacturing Co. v. Textron, Inc., Homelite, 337 F.2d 833, 837 (6th Cir. 1964). We also do not consider the District Court’s conclusion that the Schwarz and Miles Patents are inoperable as having much relevance on the issue of obviousness.8 This finding of the District Court was made in the section of the opinion concerned with the defense of anticipation and seems to have resulted from the stringent test applied in that section. Since the prior art patents did not specifically define the prescribed depths of penetration with reference to the maintenance of a fluid slag, the Court assumed that maintenance of a fluid slag was inconsistent with deep penetration as taught by Schwarz and Miles. Because a fluid slag is necessary in order to remove phosphorous, the Court concluded that the patents were inoperable. If the necessity of maintaining a fluid slag was obvious, however, so defining the depth of penetration would have been unnecessary. Therefore, to conclude from the lack of its inclusion in the definition of the depth of penetration that the characteristic itself is not comprehended within that depth of penetration begs the question when the issue is obviousness. The application for the continuation-in-part patent was not filed until 1955, and by that time the patentees had the benefit of almost six years of experience in using the basic oxygen process plus whatever information they might have obtained through the BOT arrangement. When compared with the specifications of the parent application, the specifications of the continuation-in-part application reflect the benefit of these years of experience in the extensive discussion of the chemical reactions and physical phenomena that bring about the good results achieved by the basic oxygen process and in the considerable number of demonstrative examples of proper operating conditions. In comparing the disclosures of this patent' with the disclosures of art prior to 1950, one must be careful to separate that part of the description containing well-known technical concepts from that part of the description that differentiates the alleged novel concept of the patentees from the prior art. So far as the District Court went in determining the differences between the disclosures of the Suess Patent and the disclosures of the prior art, we think it did an excellent job in separating the technical descriptions from the decisive distinctions. Therefore a concise statement of those findings will suffice. — 46 — Without considering the implications of the first claim of the Suess Patent,8a the Court found that patent’s teachings to be: (1) that pure oxygen should be used; (2) that it should be blown from above onto and below the surface of the molten bath; (3) that the blowing should be carried out by means of a vertical or almost vertical lance located so that the oxygen jet impinges on the surface of the molten bath at or near the center of the upper surface; (4) that the pressure and nozzle arrangement should be such as to avoid deep penetration of the jet into the bath; and (5) that sufficient heat for the process is produced autogenously in a limited center of chemical reaction spaced a substantial distance away from the refractory lining. The Schwarz Patent was found to teach: (1) that oxygen enriched air or oxygen should be used; (2) that it should be blown from above into the bath “as a solid body”; (3) that the pressure and nozzle arrangement should be such that the jet penetrates deeply into the bath; and (4) that sufficient heat for the process is produced autogenously in a limited center of chemical reaction located a substantial distance from the refractory lining.9 Schwarz contained no disclosure concerning the angle of the oxygen lance. The Miles Patent was found to teach: (1) that an oxygen-nitrogen mixture containing 70%-98% oxygen should be used; (2) that it should be blown from above with sufficient speed to penetrate into the bath; (3) that the blowing should be carried out by means of a lance placed at an angle of between 15 and 90 degrees to the bath surface, an angle of 30-75 degrees being recommended; (4) that the jet should agitate the bath; and (5) that sufficient heat for the process is produced autogenously in a limited center of chemical reaction spaced a substantial distance from the refractory lining.10 Miles contained no specific teaching of deep penetration, but the District Court implied such a teaching from the recommendation that the gas should be jetted at a supersonic speed. The use of pure oxygen is not claimed as inventive. Nor is the use of a vertical lance located so that the oxygen jet impinges on the molten bath at the center of the upper surface claimed as inventive in view of Miles’ suggestion of a 90 degree angle. And the District Court found that the autogenous production of heat in a limited center of chemical reaction located a substantial distance from the refractory lining could not be claimed as an advance over the prior art. In sum, the only difference between the Suess disclosure and the prior art is the teaching of avoidance of deep penetration as opposed to deep penetration. The difficulty arises in attempting to determine whether that distinction is merely verbal or whether it encompasses a concept that prior to 1950 would have been unobvious to a person skilled in the art. In determining the meaning of “avoidance of deep penetration,” the District Court found that the only manner in which that term could be defined was by characteristics of the process. 257 F. Supp. at 400-402. Expert opinion on the percentage of bath depth that was encompassed by “avoidance of deep penetration” varied from thirty per cent to anything less than ninety per cent. 257 F.Supp. at 402. Although depth of penetration is directly related to the pressure and nozzle configuration employed and inversely related to the distance of the lance tip from the bath surface, whether certain arrangements of pressure, nozzle configuration, and bath distance — 47 — would result in “avoidance of deep penetration” would depend upon the size and depth of the bath. The characteristics of the process that would place limitations on the depth of the penetration and that would meaningfully define the depth to one skilled in the art were found to be slopping, refractory damage, and fluidity of the slag. 257 F.Supp. at 402. ' The District Court found that the specifications of the Suess Patent defined “avoidance of deep penetration” as a penetration sufficient to avoid slopping but not deep enough to seriously impair refractory life or to prevent the formation of the fluid slag necessary for phosphorous removal. However, the Court found that both Schwarz and Miles similarly defined “deep penetration” as a penetration sufficient to avoid slopping but not deep enough to seriously impair refractory life. So it is apparent that the only meaningful distinction between the Suess Patent and the prior art patents is not precisely defined by the antithetical “avoidance of deep penetration” but by the teaching of the necessity of forming and maintaining an early fluid slag. At this point the District Court should have determined what the prior art taught concerning the necessity of a fluid slag, the method of obtaining a fluid slag, and the feasibility of the suggested method. Having concisely defined the difference between the depth of penetration taught by Suess and the limitation on depth of penetration as “clearly defined” by Schwarz and Miles, the District Court failed to take this further step. Instead, the Court fell back upon the otherwise meaningless antithesis and apparently assumed, since deep penetration was not clearly limited by the teaching of formation of a fluid slag, that maintenance of a fluid slag could not be comprehended by that depth of penetration. Persuasive evidence in the record, however, indicates that the prior art was aware of the necessity of maintaining a fluid slag. As indicated, the purpose of the fluid slag is to remove phosphorous; and on this point the Miles Patent teaches that “the silicon and the phosphorous pass into the slag.” In explaining the functioning of the slag, Miles further teaches that “the high speed jet causes a vigorous agitation and an entrainment of the slag in the body of the metal, thus assuring an intimate mixture and rapid reactions * * * [and] in each of these cases, the oxygen of the jet reacts not only with the constituent elements of the metal but also with the slag, thus increasing the oxygen content of the latter and therefore the oxidizing reactions in general * * * [and further] the jet actually strikes the slag at the surface of the molten metal and acts on the metal by passing through the slag and reaching the metal and/or by renewing the slag so as to regenerate or increase its power to act on the metal.” (Emphasis added.) Miles also teaches that the speed of the jet of oxygen can be varied, which to one skilled in the art is the same as teaching that the amount of penetration can be varied, and that “[t]he course of the refining can be controlled * * * by varying the velocity of the gases * * These disclosures appear clearly to teach that phosphorous is removed by reactions with the slag, that the oxygen jet reacts with the slag and increases the slag’s reaction capabilities, and that the course of the refining can be controlled by varying the amount of penetration of the jet. In addition, Dr. Chipman, the patentees’ expert, testified that a knowledgeable person skilled in the art would know what he should do to obtain an active slag. On the other hand, Appellants contend that persons skilled in the art prior to 1950 were so concerned with stirring the bath by the oxygen stream that they would not have thought refining possible at a depth of penetration where a fluid slag is maintained. The recommendation of supersonic speeds in the prior art patents lends some credence to this assertion. Also, the Gerlafingen experiments show that at least one experimenter had not solved the refractory damage prob — 48 — lem. The failures at Gerlafingen, however, are more readily explainable by the cumbersome nature of the equipment used11 than by any “blind spot” concerning the stirring of the bath. In any event, such an isolated incidence can hardly prove that persons skilled in the art in the entire steelmaking industry had a “blind spot” concerning the depth of penetration that would cause them to ignore the other teachings of Schwarz and Miles. The District Court’s conclusion of nonobviousness could have been affirmed if other evidence, whether relied upon or not, had supported that conclusion. Cold Metal Process Co. v. McLouth Steel Corporation, 126 F.2d 185 (6th Cir. 1942). Other evidence, however, leaves that conclusion considerably in doubt. On the other hand, if the disclosures of the Miles Patent were the only evidence on obviousness, this Court could reverse, making a finding of obviousness on the disclosures of that document. See Shumaker v. Groboski Industries, Inc., 352 F.2d 837 (7th Cir. 1965); Letcher County, Ky. v. De Foe, 151 F.2d 987 (6th Cir. 1946). Conflicts in the evidence, however, require factual determinations which are best entrusted to the District Court in the first instance, and which the District Court did not make in this case. However, the case need not be remanded for further findings because the District Court can be affirmed on another ground. Assuming that “avoidance of deep penetration” defines some meaningful concept, the claim of the Suess Patent does not describe that concept as the subject matter of the invention and does not distinguish the alleged invention from the prior art. Apparently the patentees were aware that, when it is once admitted that the oxygen jet penetrates into the bath to any not insubstantial degree, it becomes almost impossible to distinguish “deep penetration” as taught by Schwarz and Miles from “avoidance of deep penetration” as taught by the pat-entees. The various patent applications on the L-D Process reflect this difficulty. When confronted by an interference by the Schwarz Patent in making application for a German patent, Dr. Suess replied : “In contradistinction to * * * [Schwarz], in the case of the present invention the molten bath is blown with so little kinetic energy that the oxygen does not penetrate. * * * ” And later to further opposition Dr. Suess again replied: “ * * * the oxygen is not blown into the molten bath, but rather onto the surface of the bath and reacts at the surface. * * * The jet of oxygen * * * acts only on the surface. * * * ” Likewise in their Australian application for a patent the patentees stated that “ * * * the gas jet does not penetrate into the bath of molten metal * * * [and further] an essential characteristic of the invention is that the gas jet does not penetrate into the bath * * In other applications the amount of penetration was not even urged as the gist of the invention; rather the patentees contended that the good results of the L-D Process were due to the maintenance of a proper ratio between the bath surface and the bath depth.12 Standing alone, this evidence of statements in foreign applications for patents would not necessarily bind the patentees to any particular description of avoidance of deep penetration. The same ar — 49 — guments, however, were used in the prosecution of the United States applications. When confronted by interference from the Miles Patent, the patentees stated: “[the] * * * conditions of. blast. nozzle operation * * * distinguish the applicant’s method from the Miles et al. method in emphasizing a much lower impingement pressure and a practically negligible penetration of the surface of the molten metal by the oxygen jet * * * [and also] it will be evident that the blast velocity at the surface of the bath will be insufficient to cause any substantial penetration of the bath by the jet.” Also in an attempt to obtain a patent after a protracted prosecution in the Patent Office, Dr. Rinesch and Dr. Hauttmann, two of the patentees, filed affidavits to distinguish the Miles and Schwarz methods of deep blowing. To support their claim that “avoidance of deep penetration” was “just the opposite of the methods disclosed in the Schwarz and Miles et al. patents * * Dr. Rinesch stated that the jet did not penetrate “appreciably” into the bath and further limited the center of reaction to the “surface layer”. Likewise, Dr. Hauttmann limited the reactions caused by the oxygen jet to “the surface layer of the bath”. On the basis of this evidence the District Court found: “Clearly the patentees and their attorneys defined the avoidance of deep penetration as practically no penetration at all in order to distinguish the prior art for purposes of prosecuting their own patent applications.” Having consistently and repeatedly limited “avoidance of deep penetration” to a “practically negligible” penetration or to a “surface layer” penetration or to no penetration when faced with an interference from the Schwarz and Miles Patents, both in the United States Patent Office and in foreign patent offices, the patentees now seek a broader description of that term encompassing any penetration of one half the bath depth or less. Dr. Trenkler testified that “avoidance of deep penetration” meant a penetration of from one-third to one-half the bath depth. Dr. Hauttmann testified that the L-D Process involved a penetration of not more than one-half of the bath depth. Appellants’ experts also testified concerning an “avoidance of deep penetration” that was considerably more than negligible. 257 F.Supp. at 402. We held in Thabet Mfg. Co. v. Kool Vent Metal Awning Corp., 226 F.2d 207 (6th Cir. 1955), that, when a patentee defines a term and the Patent Office accepts that interpretation, the Court will give the same interpretation to the term. In prosecuting their patent application, the patentees clearly limited “avoidance of deep penetration” to a “practically negligible” penetration. The Patent Office accepted that interpretation of the term; so the patentees are not now entitled to an interpretation of “avoidance of deep penetration” encompassing more than the limitation accepted by the Patent Office. Westinghouse Electric Corp. v. Hanovia Chem. & Mfg. Co., 179 F.2d 293 (3rd Cir. 1949); Doran Coffee Roasting Co. v. Wyott Manufacturing Co., 267 F.2d 200 (10th Cir. 1959). Appellants have insisted throughout this case that they regard the subject matter of the patentees’ invention to be “avoidance of deep penetration”. The proposition is too well settled to need citation that the claims measure the invention. Congress has provided in 35 U.S.C. § 112: “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 Suess Patent concluded with two claims, only one of which need be considered. Claim 1 of that Patent states: “1. Method of refining molten impure iron in the presence of a slag in a vessel having a refractory lining by blowing with oxygen, which comprises discharging a stream of oxygen vertically downwardly through the slag layer onto and below the surface of the bath — 50 — at the central portion thereof, to an extent to avoid material agitation of the bath by the oxygen stream, the contact of the oxygen with the bath resulting in reaction of the oxygen with a portion of the iron and with the oxidizable impurities of the bath in a localized reaction zone spaced a substantial distance from the refractory lining, the reactions in said zone resulting in refining of the iron and gas evolution and in the production of high temperature in the reaction zone away from the refractory lining, said reaction producing a circulatory movement in the molten metal, which circulatory movement brings those portions of the molten metal bath which are remote from the reaction zone into the reaction zone whereby those portions are subjected to said reaction with minimum injury to the refractory lining.” No express reference to “avoidance of deep penetration” is made in the claim. Appellants attempt to incorporate the term in the claim by relating it to the phrase “ * * * to an extent to avoid material agitation of the bath by the oxygen stream * * * ” and to the phrase “ * * * contact of the oxygen with the bath resulting in * * * a localized reaction zone spaced a substantial distance from the refractory lining * * Yet neither of those phrases can be interpreted as limiting the penetration of the oxygen stream to a “practically negligible” penetration. In the case of an improvement patent, the United States Supreme Court stated in Halliburton Oil Well Cementing Co. v. Walker, 329 U.S. 1, 67 S.Ct. 6, 91 L.Ed. 3 (1946), that a patentee cannot obtain greater coverage for his patent by failing to describe his alleged invention than by describing it as the statute commands. We find that rule just as applicable to a process patent. In the specifications of the Suess Patent no description of the process refers to material agitation of the bath by the oxygen stream, and no description relates agitation or the avoidance of agitation to the depth of penetration of the oxygen jet. The District Court found that the phrase “material agitation” had no meaning to one skilled in the art, and we find that ruling amply supported by the evidence. Appellants contend, however, that the phrase must be given the meaning attributed to it by the patentees as reflected in the affidavits of Dr. Rinesch and Dr. Hauttmann. In so contending, the Appellants urge this Court to place an unreasonable construction on the rule stated in Kool Vent. That case in effect held that when a pat-entee places a certain limitation upon a term used in his patent he cannot later expand the scope of the term by ignoring his earlier restrictive interpretation. Appellants would have us read the rule as permitting a patentee to give meaning to a meaningless term through affidavits submitted to the Patent Office and filed somewhere obscurely in the file wrapper history of the patent. The patent law does not require a prospective inventor to search so far in attempting to determine the scope of a patent and the areas left open for inventive inquiry. Section 112 provides that it is the claims that shall particularly point out and distinctly claim the subject matter of the patentee's invention. Certainly, terms used in the claims can gain meaning from the specifications or from the knowledge attributable to one skilled in the art. But a prospective inventor is required to go no further in attempting to determine the invention claimed and the areas foreclosed to future enterprise. When read in light of the specifications, these claims would not describe a “practically negligible” penetration to one skilled in the art. Hence, the patentees failed to describe their invention. The only meaningful limitation placed on the depth of penetration of the oxygen stream is the phrase “ * * * a localized reaction zone spaced a substantial distance from the refractory lining.” The District Court found that both Schwarz and Miles placed the same limitation on depth of penetration. So if the only difference between the patentees’ process — 51 — and the process taught by Schwarz and Miles is the depth of penetration of the oxygen stream, the patentees have failed to distinguish their depth of penetration in any way from the depth of penetration taught by the prior art. We recently held in Holstensson v. V-M Corporation, 325 F.2d 109 (6th Cir. 1963), that a policy in the patent law “ * * * visits total forfeiture upon a patent which attempts to extend its monopoly to something already patented.” 325 F.2d at 122. That case concerned an improvement patent, but we find the rule just as applicable to the patent in suit. In our view of this case, the Appellants ask this Court to ignore the precision of definition required of them by the statute. As the Supreme Court noted in United Carbon Co. v. Binney & Smith Co., 317 U.S. 228, 63 S.Ct. 165, 87 L.Ed. 232 (1942), to allow claims “so indefinite as not to give the notice required by the statute would be in direct contravention of the public interest which Congress * * * recognized and sought to protect.” In that case the Court considered a product patent where the distinction from the prior art was a matter of degree. The Court rejected terms, such as “substantially pure”, “commercially uniform”, and “comparatively small”, that gave no standard for comparison or were so indefinite as to have no established meaning to one skilled in the art. Likewise, in the instant case we are concerned with a patent that varies from the prior art only in the alleged degree of penetration of the oxygen stream. But the only term allegedly defining that difference has no established meaning. When patentees’ experts were asked when agitation became material, they uniformly replied, “I don’t know.” As indicated earlier the specifications nowhere establish a standard for determining when agitation becomes material and nowhere relates the amount of agitation to the depth of penetration. The knowledge attributable to one skilled in the art will also not aid the patentees in reading their alleged invention into the claim. Whatever interpretation one skilled in the art might have placed on “avoidance of material agitation”, it would not have been the “practically negligible” penetration that the patentees represented to the Patent Office as “avoidance of deep penetration.” The District Court found that continued shallow blowing causes a turbulent slopping condition. Although a slopping condition might appear inconsistent with what the specifications would teach one skilled in the art concerning the depth of penetration of the oxygen stream, the inconsistency arises because of the inconsistent positions that the patentees assumed before the Patent Office. If the patentees had contended for the amount of penetration before the Patent Office that they now urge upon this Court, their patent would probably not have passed muster. The inconsistency could have resulted from the rather indeterminate nature of the patentees’ advance over the prior art, or it could have resulted from an attempt to get an artfully worded restatement of the prior art through the Patent Office. Under similar circumstances in the Binney Co. case the Supreme Court stated: “Whether the vagueness of the claim has its source in the language employed or in the somewhat indeterminate character of the advance claimed to have been made in the art is not material. An invention must be capable of accurate definition, and it must be accurately defined, to be patentable.” 317 U.S. at 237, 63 S.Ct. at 170. We find that the patentees failed to define their alleged advance over the prior art and that in failing to properly limit their claims they attempted to extend their monopoly to something already patented. For either reason, or for both reasons, the patent must be held invalid. Even if “avoidance of deep penetration” could be given a broader meaning than “a practically negligible penetration”, another reason prevents the claims from being read so as to contain that limitation. The District Court found that the doctrine of file wrapper estoppel prevented the patentees from incorporat — 52 — ing the limitation of “avoidance of deep penetration” into their Claim 1. If the patentees cannot limit their claim to this sole patentable advance over the prior art, the patent is invalid regardless of any other findings of the District Court. Resolution of the question requires certain factual inquiries into the file wrapper history of continuation-in-part Application No. 547,233. In attempting to apply the doctrine of file wrapper estoppel to the facts of the instant case, two considerations must be borne in mind: (1) 35 U.S.C. § 282 specifically provides that a patent shall be presumed to be valid and the burden of proving invalidity shall rest upon the person asserting that the patent is a nullity; thus, in the present case McLouth has the burden of proving that the patent is void; and (2) the estoppel is invoked in connection with the plaintiff’s attempt to satisfy that requirement of 35 U.S.C. §112 which provides, “the specifications shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.” The basic principle underlying file wrapper estoppel is that “a party cannot take inconsistent positions concerning the same subject matter in different transactions.” 4 Walker, Patents § 234, p. 87 (Deller 2d ed. 1965). In making its holding concerning file wrapper estoppel, the District Court relied heavily upon Schriber-Schroth Co. v. Cleveland Trust Co., 311 U.S. 211, 61 S.Ct. 235, 85 L.Ed. 132 (1940). There the Court phrased the doctrine in these words: “But the particular invention to which the patentee has made claim in conformity to the statute is not always to be ascertained from an inspection of the specifications and claims of the patent alone. Where the patentee in the course of his application in the patent office has, by amendment, can-celled or surrendered claims, those which are allowed are to be read in the light of those abandoned and an abandoned claim cannot be revived and restored to the patent by reading it by construction into the claims which are allowed.” 311 U.S. at 218, 61 S.Ct. at 238. Hence, the determinative question on this issue is whether the Appellants did in fact cancel, surrender, or abandon the claim of “avoidance of deep penetration by the oxygen stream” during the prosecution of Patent Application No. 547,-233. In Schriber, cited by this Court in Bo-bertz v. General Motors Corporation, 228 F.2d 94 (6th Cir. 1955), the Supreme Court said: “It is a rule of patent construction consistently observed that a claim in a patent as allowed must be read and interpreted with reference to claims that have been. cancelled or rejected and the claims allowed cannot by construction be read to cover what was thus eliminated from the patent. (Citing cases.) The patentee may not, by resort to the doctrine of equivalents, give to an allowed claim a scope which it might have had without the amendments, the cancellation of which amounts to a disclaimer. (Citing cases.) The injurious consequences to the public and to inventors and patent applicants if patentees were thus permitted to revise cancelled or rejected claims and restore them to their patents are manifest. See Leggett v. Avery, 101 U.S. 256, 259, 25 L.Ed. 865. “True, the rule is most frequently invoked when the original and cancelled claim is broader than that allowed, but the rule and the reason for it are the same if the cancelled or rejected claim be narrower. (Citing cases, Cf. Smith v. Magic City Kennel Club, Incorporated, 282 U.S. 784, 790, 51 S.Ct. 291, 75 L.Ed. 707).” 311 U.S. at 220-221, 61 S.Ct. at 239. In making this determination the only pertinent material in the file wrapper history of Application No. 547,233 would appear to be the parent patent Application No. 206,147 together with Patent Office correspondence, the present patent — 53 — and answers of the Examiner concerning prior rejected claims, the patents against which Application No. 547,233 was compared,13 and the affidavits submitted with the final change in claims.14 When making the parent Application, No. 206,147, in 1951, Dr. Suess did not emphasize avoidance of deep penetration as of any great significance in the process. Only one off-hand reference was made in the description that a certain pressure and nozzle arrangement would bring “about a sufficiently high reaction velocity without deep penetration of the gas into the molten bath.” No claim referred to avoidance of deep penetration as an element of the process. After final rejection of No. 206,147, the Appellants on March 20, 1956, cancelled all prior claims and submitted two claims to place “the case in better form for consideration on appeal.” In one of those claims, Claim 14, a reference was made to avoid-anee of “deep penetration of the oxygen jet into the molten metal.”15 This entire Application was abandoned before any Patent Office action was taken; so none of the Patent Office correspondence on No. 206,147 has much relevance on the issue of file wrapper estoppel. Application No. 547,233, however, does make references to “avoiding deep penetration by the oxygen jet.” That phrase is incorporated into each of Claims 1 through 4. In the Patent Office rejection of these claims the Examiner held them unpatentable over Brassert and Cheswiek and Kalling. Brassert was said to disclose “the use of a blast of oxygen into the surface of the molten metal * * * ” and “regulating the pressure and amount of the compressed air or blast * * Cheswiek was said to disclose “the use of an oxygen blast upon the surface of molten steel for refining purposes,” and Kalling was — 54 — said to make the use of surface blowing “to prevent local attacks upon the furnace lining” old art. Claims 1 through 15 were rejected over the Miles French Patent and Cuscoleca. Miles was said to disclose surface blowing “with a pressure of 7 to 8 atmosphere.” Most of the indicated conflicts relate in some manner to the penetration of the oxygen stream: (1) surface blowing implies very little, if any penetration; and (2) the depth of penetration is directly related to the pressure with which the oxygen is blown. On the other hand, the Examiner did not mention avoidance of deep penetration directly as a conflicting element. The Appellants, in response to the rejection, cancelled Claims 1 through 10 but amended Claim 12 to include the statement “to avoid substantial penetration of the jet into said bath.” Also Claim 16 was added, which included in it the statement, “the pressure and nozzle spacing being related to avoid deep penetration of the oxygen jet into the molten metal.” The remarks following the submission of these amendments make it clear that the Appellants considered avoidance of deep penetration an important part of the invention, specifically emphasizing “the very important particular of depth of penetration of the molten bath by the jet of oxygen.” The Patent Office never took action on these claims. Pursuant to a discussion with the Examiner, however, the amended Claims were cancelled, and Claims 17 and 18 submitted in their place. These new Claims, which with minor amendments became the final Claims, did not contain any direct reference to avoidance of deep penetration. Also the remarks accompanying the submission of the new Claims contained no reference to avoidance of deep penetration. The only references were contained in affidavits submitted by two of the patentees, Dr. Rinesch and Dr. Hautt-mann. We believe that sound inferences can be drawn from the Patent Office’s rejections of earlier claims which contained the element of avoidance of deep penetration and its reference to the Cuscoleca Patent. Although the record is void as to the content of the discussion with the Examiner on May 1, 1957, we believe that an inference of abandonment can be drawn from the omission of “avoidance of deep penetration” from the final claims. In our opinion, if the element was as important as the affidavits imply, it would have been explicitly included in the claims; the Patentees were well aware that our patent laws require that the claims “distinctly” claim the alleged invention. The Appellants cannot regain the phrase or qualify the effect of their acquiescence through the affidavits. Parke, Davis & Co. v. American Cyanamid Co., 207 F.2d 571, 574 (6th Cir. 1953). In the face of prior rejections, the claim allowed, cannot, by construction or affidavit, be read to cover that which was specifically abandoned by the patentee so as to give the allowed claim a scope it might have had without the omission of the language. In this case the issue of file wrapper estoppel hinges entirely upon the factual determination of abandonment or non-abandonment.16 While the — 55 — record does not indicate whether the District Court made any extensive factual inquiry into this problem, we believe the Court was correct in its view that abandonment is established if a phrase used in an abandoned or rejected claim is not used in a later accepted claim. The determination being essentially a factual one, prior cases do not offer any clear-cut guide lines concerning what constitutes abandonment or non-abandonment. Most of the cases involved a limitation specifically imposed upon the claims by the Applicant and the Applicant was precluded in a later infringement suit from placing a broader interpretation on the Claims.17 Here the process was reversed; the Applicant was refused a narrow claim and granted a broader claim. Although this difference does not affect the application of the doctrine of file wrapper estoppel, it does make the factual determination more difficult. In relying upon Schriber-Schroth Co. v. Cleveland Trust Co., 311 U.S. 211, 61 S.Ct. 235 (1940), the District Court chose the case most directly in point factually with the instant situation. Many of the facts in the Schriber Co. case are identical to the instant case. The question in Schriber Co. was whether the Jardine Patent for a piston contained a claim of a flexible web feature. 311 U.S. at 218, 61 S.Ct. at 235. As the Applicants in the instant case, Jardine did not claim flexible webs as a feature in the claims originally filed; and while the application was pending, Jardine amended the claims to supply this omission; here the patentees filed a continuation-in-part application to supply the omissions. As in the instant case, these claims were rejected as reading upon other patents; as a result the claims were cancelled and new claims submitted that did not contain a claim for flexible webs. Also as here, the claims that were finally allowed were broader than if the omitted claim had been included. Appellants attempt to distinguish Schriber Co. on the grounds that in that case the limiting phrase was omitted as the result of an interferencé proceeding. That fact cannot be a distinguishing feature of the case, however, see Graham v. John Deere Co., 383 U.S. 1, 32-33, 86 S.Ct. 684 (1966); otherwise, the doctrine of file wrapper estop-pel would never apply to a patent where, as in the instant case, the only pertinent prior art consisted of foreign patents, because interference proceedings will not be — 56 — declared on such patents by the Patent Office. 35 U.S.C. § 135. The District Court concluded that Schriber Co. controlled the result concerning file wrapper estoppel. Judge Weick and I agree. Judge Edwards joins in the opinion in all respects except that he does not feel that the defense of file wrapper estoppel is established by the record. Affirmed. Footnotes [1] Only in exceptional circumstances can a licensee contest the validity of the patent that he is licensed to use. Cf. Sola Electric Co. v. Jefferson Electric Co., 317 U.S. 173, 63 S.Ct. 172, 87 L.Ed. 165 (1942). [1a] The apparent inconsistency of the District Court’s holding that the invention was nonobvious and the holding that the claims cover the prior art results from the manner in which the Court decided the question of obviousness. In fact, the holdings are not necessarily inconsistent since a patentee can disclose a nonobvious invention in his specifications yet claim too broadly, covering not only what is new but also what is old in the art. Holstensson v. V-M Corp., 325 E.2d 109 (6th Cir. 1963). [2] For a more comprehensive discussion of steelmaking and the history of its development see the District Court’s Opinion, 257 F.Supp. at 378-379. [3] 257 F.Supp. at 386. [4] In Graham v. John Deere Co., 383 U.S. 1, 10-12, 86 S.Ct. 684 (1966), the United States Supreme Court noted the confusion that resulted from the early development of the term “invention” as a word of legal art. Although Graham went far towards clearing up the confusion, certain areas still remain perplexing. [5] The District Court did not extensively analyze the prior art patents in the section of the Opinion concerned with obviousness. The most extensive analysis of the prior art patents was made in the section of the Opinion concerned with the defense of anticipation, 257 F.Supp. at 387-395, where the test applied was that the disclosures of the prior patents must be so full and complete as to require no further experimentation. Whatever the proper test for anticipation may be, that is not the test for obviousness, where the inquiry is whether the experimentation involved was within the ingenuity of a workman having ordinary skills in the art. Reiner v. I. Leon Co., Inc., 285 F.2d 501 (2d Cir. 1960). Because of the stringent test applied, the findings of the District Court in the section concerning anticipation cannot be applied to a determination of obviousness without some qualifications. The same qualifications must be applied to the findings in the section of the Opinion concerned with inventorship, 257 F.Supp. at 406-419, where the same stringent test of “no further experimentation” was employed. Since the District Court did not rely upon these findings in concluding that the patentees’ invention was nonobvious, we consider them applicable only after a proper evaluation with the test of obviousness in mind. [6] The only mention of the Schwarz Patent in the file wrapper is contained in the affidavits of Rinesch and Hauttmann. These affidavits were submitted only about two weeks before the issuance of the patent, and the record does not indicate what, if any, consideration the Patent Office gave to the affidavits. Such an off-hand reference does not amount to Patent Office consideration of the prior art patent. [7] This inference is strengthened by the agreements between BOT and its members. For example, the agreement with Brassert concerning the Schwarz Patent concludes: “For the duration of this agreement Brassert is engaged to enter no corn-plaint against any patent, which is owned or controlled by BOT and which concerned the same subject of the patent transferred by this agreement and no protest against applications of such protective rights.” [8] The only evidence to support the holding was the testimony of the patentees’ experts, who had made no experimentations and who testified on the basis of their interpretation of deep penetration as taught by Schwarz and Miles as a penetration that damaged the refractory lining. The District Court found that Schwarz and Miles limited the penetration to a depth that would not damage the refractory lining. 257 F.Supp. at 397-398. So it is doubtful that this evidence supports a finding of inoperability. For the same reason the Gerlafingen experiments do not prove inoperability. The District Court found that, at Gerla-fingen, Dr. Durrer had not solved the refractory damage problem. 257 F.Supp. at 414. This is not a case like Adams v. United States, 383 U.S. 39, 86 S.Ct. 708, 15 L.Ed.2d 572 (1966), where an expert for the plaintiff testified without contradiction that he had attempted to build a battery by prior art teachings and the battery had exploded. [8a] Since the invention is measured by the claims, perhaps the District Court should have first determined what the claims described as the invention, which the Court later found was not “advoidance of deep penetration”. Although the approach causes some confusion, we follow it for the purpose of analyzing the District Court’s conclusion on obviousness. See also footnote la, supra. [9] Disclosure (4) is a paraphrase of the finding of the District Court at page 404 of its Opinion, rather than a paraphrase of the patent. [10] See footnote 9 concerning disclosure (5). [11] The gist of the negotiations preceding tlie transfer of the oxygen blowing experiments from Gerlafingen to Linz was that the transfer was made because Ger-infingon did not have an oxj'gon plant and because other factors made experimentation cumbersome. In addition Dr. Trenk-ler testified that the lance used at Ger-iafingen could not be raised or lowered during the process. Also Dr. Durrer testified that the experiments were moved to Linz because of the better conditions there. [12] This ratio theory was also the basis of the earlier Suess application, which was rejected by the United States Patent Office and finally abandoned. [13] In the prosecution of Application No. 547,233 perhaps the Examiner considered some patents as prior art, which should not have been so considered, (e. g. the Cuscoleca Patent, filed on December 22, 1954). But whether the patentee could have been legally required to give up the claim surrendered is immaterial, the question in applying the doctrine of file wrapper estoppel is whether the patentee did in fact give up the claim. If the patentee disagreed with the Examiner, his remedy was appeal; he cannot give up the claim and then attempt to recover it through collateral infringement suits. Schriber-Schroth Co. v. Cleveland Trust Co., 311 U.S. 211, 221, 61 S.Ct. 235 (1940). Therefore, in determining what was abandoned in the rejected claims, all of the patents with which No. 547,233 was compared must be considered. Those patents were: (foreign patents are designated as such) Miles Patents (British and French) Potts Patent (British) Cuscoleca Patent Cheswiek Patent Kalling Patent Brassert Patent Only the Cuscoleca Patent is included in the record on appeal. The assumed conflict with the other patents must be determined from correspondence in the file wrapper history of No. 547,233 and No. 206,147. [14] Although these affidavits might not be pertinent in determining other issues, such as whether the specifications support the claims, Of. Tillotson Mfg. Co. v. Textron, Inc., Homelite, 337 F.2d 833 (6th Cir. 1964), it seems that they would be appropriate considerations in determining exactly what elements of the prior claims were abandoned on submission of the final claims, in absence of other more direct evidence. [15] Section 120 of Title 35 of the United States Code grants the benefit of the earlier filing date of the parent patent, to a continuation-in-part patent when the invention was “disclosed in the manner provided by the first paragraph of Section 112 * * * in * * * ” the earlier “application previously filed * * * by the same inventor * * * ” Since the continuation-in-part application was made in 1955 and since the only claim containing any reference to the subject matter of avoidance of deep penetration was not made until 1956, apparently as an afterthought, some question might be raised whether the alleged invention of avoidance of deep penetration is entitled to the benefit of the earlier filing date. Merry Manufacturing Company v. Burns Tool Company, 335 F.2d 239 (5th Cir. 1964). If it was not so entitled, the claim would be barred by Dr. Suess’s 1950 Austrian patent. 35 U.S.C. § 119. That issue, however, has not been raised in this case. [16] Certain assertions made by the Appellants deserve comment at this point. The argument that “avoidance of deep penetration” was included in Amended Claim 12 and added Claim 16 and that these Claims were never rejected is immaterial except for the factual inferences that might be drawn. The doctrine is equally applicable to claims given up informally pursuant to a discussion or conference as to claims given up after a formal rejection. Graham v. John Deere Co., 383 TJ.S. 1, 32-33, 86 S.Ct. 684 (1966). That the claims allowed were broader than they would have been if the omitted phrase had been included is likewise immaterial. Schriber-Schroth Co. v. Cleveland Trust Co., 311 TJ.S. 211, 221, 61 S. Ct. 235 (1940). As pointed out earlier, the accompanying affidavits cannot be utilized to recover the abandoned claim if its relinquishment was in fact required by the Examiner during the “discussion”. Parke, Davis & Co. v. American Cyanamid Co., 207 F.2d 571 (6th Cir. 1953). The only disputable finding, therefore, is the factual determination of abandonment. [17] In these cases the factual determination was easier, because it is easier to prove the purpose of a positive occurrence such as an inclusion of an element than to prove the purpose of a negative occurrence such as the omission of an element. Was the omission made purposely or was it made under the assumption that an included phrase was the omitted phrase’s equivalent? The difficulty of the determination is compounded when the final claim is broader than the cancelled claim. See Graham v. John Deere Co., 383 U.S. 1, 32-33, 86 S.Ot. 684 (1966) (limitations were specifically spelled out in remarks accompanying the final submission); Exhibit Supply Co. v. Ace Patents Corp., 315 U.S. 126, 136-137, 62 S.Ct. 513, 86 L.Ed. 736 (1942) (claims limited to conductors embedded in table clear from prior rejected claims); Keystone Driller Co. v. Northwest Eng. Corp., 294 U.S. 42, 47-49, 55 S.Ct. 262, 79 L.Ed. 747 (1935) (limitation clear from letter from applicant’s solicitor); Smith v. Magic City Kennel Club, Inc., 282 U.S. 784, 788-789, 51 S.Ct. 291, 75 L.Ed. 707 (1931) ; (limitation clear from prior rejection references) ; I. T. S. Rubber Co. v. Essex Rubber Co., 272 U.S. 429, 439, 47 S.Ct. 136, 71 L.Ed. 335 (1926) (limitation clear from prior rejection reference); Parke, Davis & Co. v. American Cyanamid Co., 207 F.2d 571, 574 (6th Cir. 1953) (limitation clear from statements of Patent Office); Lyon, Inc. v. Clayton & Lambert Mfg. Co., 127 F.2d 164 (6th Cir. 1942) (facts not clear); Valjean v. Perfection Stove Co., 103 F. 2d 60, 61-62 (6th Cir. 1939) (limitations clear from rejection reference and statements of Patent Office); Bishop & Babcock Mfg. Co. v. Western Auto Supply Co., 105 F.2d 886, 889 (6th Cir. 1939) (limitation clear from representations to Patent Office); Firestone Tire & Rubber Co. v. United States Rubber Co., 79 F.2d 948, 954-955 (6th Cir. 1935) (limitation clear from rejection reference).