This is an appeal from the judgment of the United States District Court for the Western District of Tennessee, 597 F. Supp. 133, in a patent infringement suit holding U.S. Patent No. 3,960,006 ('006 patent) valid and infringed. The case was tried to the court, which decided only the liability and not the damages issue. We affirm.

The plaintiff below and the appellee here is Alco Standard Corporation (Alco), which acted through its Commercial Machine Works division (Commercial). The nominal defendant below and an appellant here is the Tennessee Valley Authority (TVA). The defendant real party in interest is the third party defendant, and co-appellant here, Westinghouse Electric Corporation (Westinghouse), which performed services involving the patented invention for TVA and agreed to indemnify TVA for any damages arising from those services.

Prior to trial, the court dismissed Alco's patent infringement claims against Westinghouse. Alco Standard Corp. v. Tennessee Valley Authority, 448 F. Supp. 1175 (W.D.Tenn.1978). Westinghouse remained in the case, as the indemnitor of TVA, and has conducted the defense of this suit.

* Background

A. The '006 patent covers a method of and apparatus for inspecting turbine rotors in electrical generators by the use of ultrasonic waves to detect discontinuities within the rotors. Discontinuities are cracks, flaws, or impurities and may result when the rotors are made or may develop later with use. Turbine rotors are large cylindrical metal forgings which can be as long as 30 feet and weigh as much as 100 tons. In operation, they may rotate at speeds as high as 3600 revolutions per minute and reach temperatures of 1000? F. The high speeds and temperatures subject the rotor to stress, so that discontinuities in the turbine forging may cause it to fly apart.

As early as 1946, turbine rotor forgings were being inspected with ultrasonic tests. Such tests were performed from the outside of the forgings of newly manufactured rotors. In 1957, Westinghouse undertook a development program to improve its ultrasonic inspections. In 1959, Westinghouse discovered that General Electric Corporation was producing a new type of ultrasonic inspection, a boresonic testing device, that could be inserted into the bore of the rotor as distinguished from prior inspection devices that scanned only the outside of the rotor. Upon learning this, Westinghouse abandoned its development program, purchased the General Electric device and used it until 1972. The General Electric device, although providing for a bore inspection, could be used to inspect only newly manufactured rotors.

In the 1950's it became clear that a system for inspecting rotors that were not newly manufactured was needed because such rotors otherwise had to be removed from the turbine and shipped to the inspection site, causing costly delays. Apparently, General Electric solved this problem, and Westinghouse attempted to buy the new General Electric device in 1972. General Electric, however, refused to sell the new device, and therefore Westinghouse again began its own independent development program, this time to produce an onsite inspection device. Westinghouse's development of an onsite inspection device did not go as planned, and soon fell behind schedule.

Also during 1972, Mr. Smith, an employee of Commercial, conceived the idea of producing a new type of boresonic test apparatus. In December 1972, Westinghouse officials, including a Mr. Ronca, visited Commercial to discuss Commercial's idea for such an apparatus. At that meeting, Mr. Ronca stated that he thought that Commercial's concept was not feasible because Westinghouse's own research and development in that area had been unsuccessful. Despite Westinghouse's skepticism, Commercial contracted to have the boresonic unit constructed.

In January 1974, Commercial demonstrated its boresonic device to Westinghouse. Among those Westinghouse representatives in attendance was Mr. Ronca, who, after witnessing the demonstration, wrote that the "system represent[ed] a significant advancement in the field of boresonics...." Based on the Commercial demonstration, Westinghouse discontinued its research on boresonic devices and pursued the possibility of purchasing Commercial's device.

In mid-1974, a rotor in TVA's Gallatin steam plant exploded, causing substantial property damage and, luckily, no bodily injuries. Commercial inspected the remaining rotor at the Gallatin plant and discovered that it had a discontinuity similar to the one that caused the first rotor to explode. Subsequent remedial measures were taken to remove the discontinuity in the remaining rotor.

By October 1974, Westinghouse had abandoned its plan to purchase Commercial's boresonic units and had begun developing its own boresonic unit. It developed such a unit. Alco alleged that Westinghouse infringed the '006 patent by using the Westinghouse device in inspecting TVA rotors.

B. The invention the '006 patent covers was made by Robert Smith when he was an employee of Commercial. The patent issued on June 1, 1976, and was assigned to Alco.

In the present suit, Commercial alleged infringement of six claims of the '006 patent. Three of these are apparatus claims and the other three are method claims.

Claim one, the broadest apparatus claim, is directed to a device for detecting discontinuities in a turbine rotor. The device has a probe that is inserted into the bore of a turbine rotor. Attached to this probe is an indexing means that determines the position of the probe within the rotor. The probe, itself, contains at least two ultrasonic sources (called "transducers") that simultaneously can send ultrasonic signals into the rotor. Any signals sent into the rotor that are reflected back to the probe are picked up by an ultrasonic pickup and recorded in such a manner that the "existence, position, nature, size and shape of the flaws in the rotor" can be determined.

The method claims generally describe a method for determining flaws within turbine rotors that utilizes the device described in the apparatus claims.

The claims are discussed in greater detail later in this opinion.

C. In a lengthy opinion, the district court held that the '006 patent was valid and that Westinghouse had infringed the patent in using its own boresonic device to inspect TVA rotors. Specifically, the district court ruled that the Commercial invention was novel under 35 U.S.C. Sec. 102 and non-obvious under 35 U.S.C. Sec. 103. The court also thoroughly examined whether Commercial had complied with the enabling, description, and definiteness requirements of 35 U.S.C. Sec. 112 and found that it had. Regarding infringement, the district court found that the "Westinghouse device contain[ed] every element found in claims 1, 2, 3, 7, 8 and 10 of the 006 patent."

Under 28 U.S.C. Sec. 1295(a)(1) (1982), we have jurisdiction over this appeal from the district court if its jurisdiction "was based, in whole or in part, on section 1338 of this title...." Section 1338(a) gives the district courts "jurisdiction of any civil action arising under any Act of Congress relating to patents...." The inquiry thus is whether the jurisdiction the district court exercised in this case was based upon an Act of Congress "relating to patents."

The Tennessee Valley Authority Act of 1933 contains a specific provision dealing with patents. Section 831r of title 16 (1982) is captioned "Patents; access to Patent and Trademark Office and right to copy patents; compensation to patentees." It gives the Tennessee Valley Authority (TVA) access to the Patent and Trademark Office

for the purpose of studying, ascertaining, and copying all methods, formulae, and scientific information (not including access to pending applications for patents) necessary to enable [TVA] to use and employ the most efficacious and economical process for ... any method of improving and cheapening the production of hydroelectric power.

any owner of a patent whose patent rights may have been thus in any way copied, used, infringed, or employed by the exercise of this authority by [TVA] shall have as the exclusive remedy a cause of action against [TVA] to be instituted and prosecuted on the equity side of the appropriate district court of the United States, for the recovery of reasonable compensation for such infringement.

The statute, in the second quoted passage, thus recognizes that the use by TVA of a patented invention constitutes infringement of the patent. Although the statute specifies that the patentee's "exclusive remedy" if its patent has been "thus ... infringed ..." is a civil suit against TVA on the equity side of the district court to recover "reasonable compensation for such infringement," that fact does not make the resulting suit any the less one for infringement of a patent. Thus, Sec. 831r specifies the conditions that govern the patentee's suit for infringement, while Sec. 1338(a) gives the district courts jurisdiction to hear that suit.

It would be anomalous if appeals in patent infringement suits against TVA were heard by the regional circuit, when all other appeals in patent infringement suits come to this court. (We have exclusive jurisdiction over appeals from the district courts in patent infringement suits generally, and over appeals from the Claims Court in suits for patent infringement against the United States, over which the Claims Court has exclusive jurisdiction under 28 U.S.C. Sec. 1498(a). 28 U.S.C. Sec. 1295(a)(1), (3).) Such a bifurcated jurisdiction would be inconsistent with the Congressional intent in enacting the Federal Courts Improvement Act of 1982 to "produce desirable uniformity in this area of the law." S.Rep. No. 275, 97th Cong., 2d Sess. 5 (1982), reprinted in 1982 U.S.Code Cong. & Ad.News 11, 15. There is no reason to believe that Congress intended the regional circuits rather than this court to hear appeals in this narrow category of infringement cases.

The earlier decision of the district court in this case, dismissing the suit against Westinghouse, is not inconsistent with this conclusion. The ground of that decision was that under 16 U.S.C. Sec. 831r, the patentee had no cause of action against Westinghouse because its "exclusive remedy with regard to the inspections for TVA is the claim for reasonable compensation asserted against TVA in this cause under Section [831r]." Alco Standard Corp. v. TVA, 448 F. Supp. 1175, 1181, 197 U.S.P.Q. (BNA) 671, 675 (W.D.Tenn.1978).

The earlier decision did not hold that the jurisdiction of the district court in this case was based upon any statute other than 28 U.S.C. Sec. 1338(a). The district court began both that opinion and its later opinion on the merits by describing this suit as "a patent infringement action" and an "action for patent infringement," respectively. Although the district court's characterization of the case does not bind us, we see no reason to reject it. In sum, the district court action in this case arose under an "Act of Congress relating to patents," and we therefore have jurisdiction over the appeal.

In challenging the district court's determination that the '006 patent is valid, the appellants mount a broad scale but largely unfocused attack upon the sufficiency of the evidence to support the district court's factual findings upon which its conclusion of validity rests. In effect, the appellants invite us to decide the case de novo. That is not our function, however; our role is to determine whether the district court committed any reversible errors, either in its factual findings or in its legal conclusions or rulings. Raytheon Co. v. Roper Corp., 724 F.2d 951, 956, 220 U.S.P.Q. (BNA) 592, 596 (Fed.Cir.1983), cert. denied, 469 U.S. 835, 105 S. Ct. 127, 83 L. Ed. 2d 69 (1984). The scattered and disorganized nature of the appellants' presentation has made the performance of our reviewing role more difficult.

In this appeal we do not specifically deal with the myriad of minor contentions that the appellants make, although we have considered them. Here we address only the major issues the parties have presented.

A. The Enabling and Description Requirements (35 U.S.C. Sec. 112 (1982)).

The first two paragraphs of 35 U.S.C. Sec. 112 provide:

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 district court ruled that the '006 patent met the requirements of section 112. The court found that "one skilled in the art of ultrasonic testing could have determined how to make and use the patented invention from reading the patent specification," that "the specification does adequately describe the claimed invention" because "[e]vidence established one skilled in the pertinent part would be aware of the various methods and various equipment that could be used to obtain and record the [ultrasonic] data generated by the invention's transducers," and that the specification of the '006 patent "supported" the claim. The district court further ruled that "after weighing all the evidence, the Court finds defendants have not proved the patented device was not able to achieve the results claimed in the patent."

On appeal, the contentions of both parties center around the language of the claims that pertains to correlating and combining the information derived from the ultrasonic scans. Claim 1 covers a "non-destructive test apparatus for detecting and providing three-dimensional analysis of flaws in a horizontally positioned generally cylindrical rotor" comprising, among other elements,

position determining means providing for the direct correlation of the information content of a transmitted or reflected ultrasonic signal in one mode with the information content of reflected ultrasonic signals and unreflected transmitted ultrasonic signals in another mode,

and

recording means for preserving the information content of said transmitted and said reflected ultrasonic signals in a fashion that permits combining the individual information content of each of said transmitted and said reflected ultrasonic signals in the different modes to derive an accurate indication of the existence, position, nature[,] size and shape of flaws in the rotor material.

The prosecution history of the '006 patent demonstrates that the words "correlating" and "combining" have specific and definite meanings, which one skilled in the art of the invention would have understood.

The word "correlate" as used in the '006 patent refers to the steps of (1) taking the raw data from the ultrasonic scans in the various modes, (2) calculating the position of the defects--circumferentially, longitudinally, and axially--by using the angle of the probe, its depth into the rotor and the time it took for the ultrasonic pulse to be reflected back off the defects (flight time), and (3) grouping those calculated positions with others in the same general area. This grouping is necessary because a single defect may generate more than one ultrasonic blip (reflected ultrasonic pulse) during an entire test run due to the spread of ultrasonic beams and the use of multiple transducers. In the '006 patent, to correlate means to group together all the ultrasonic blips that correspond to the same defect.

The word "combine" as used in the '006 patent refers to gathering all the ultrasonic blips that have been correlated for one defect and using those blips to derive information about the nature of the defect that any single blip might not have revealed. The inventor, speaking through his attorney, described "combining" best in an amendment filed with the Patent Office on October 19, 1975:

[S]uppose that a pair of parallel cracks existed at the same angle with respect to the axis of the rotor as that followed by the shear waves, with the ends of these cracks overlapping but spaced apart. Testing with a straight [longitudinal] beam would show the presence of the two cracks, but would be unable to determine if the cracks were connected. On the other hand, testing with a shear wave would provide no signal at all. Therefore, this lack of any indication by the shear wave test would provide no information at all without combining the results of this test with the results of the longitudinal mode test, such combination revealing the existence of two separate cracks rather than a single connected crack.

The district court's holding that one skilled in the art would have known how to correlate and combine the ultrasonic scan data from a reading of the patent disclosure is not erroneous. Raytheon, 724 F.2d at 960, 220 USPQ at 599.

The appellants also contend that the correlating and combining steps of the '006 patent are merely mental processes, and therefore, unpatentable. Under the meaning of correlating and combining used in the patent, these steps may be performed either by a person or by a machine. The record shows that the scan data are arranged and grouped using simple trigonometric calculations and graphic techniques. Any mental processes occur after the data has been subjected to calculation and graphing.

The inclusion in a patent of a process that may be performed by a person, but that also is capable of being performed by a machine, is not fatal to patentability. Diamond v. Diehr, 450 U.S. 175, 101 S. Ct. 1048, 67 L. Ed. 2d 155 (1981). The presence of the steps of correlating and combining, which a machine is capable of doing, does not invalidate the '006 patent.

B. Anticipation (35 U.S.C. Sec. 102(a) (1982)).

Under 35 U.S.C. Sec. 102(a), a patent is invalid if "the invention was ... described in a printed publication in this ... country ..., before the invention thereof by the applicant for patent." We have held that "[a]nticipation requires the disclosure in a single prior art reference of each element of the claim under consideration." W.L. Gore & Associates, Inc. v. Garlock, Inc., 721 F.2d 1540, 1554, 220 U.S.P.Q. (BNA) 303, 313 (Fed.Cir.1983). This is essentially the same standard the district court applied when it cited American Seating Co. v. National Seating Co., 586 F.2d 611, 618, 199 U.S.P.Q. (BNA) 257, 261 (6th Cir.1978), cert. denied, 441 U.S. 907, 99 S. Ct. 1999, 60 L. Ed. 2d 377 (1979), which stated that "all the elements of a patented device ... be found in a single pre-existing structure or description" (citations omitted).

The appellants contended in the district court, and repeat the contention here, that a publication entitled "Ultrasonic Inspection of the Nimrod Power Plant" (Nimrod), which described a method and device for boresonically inspecting a rotor, anticipated the '006 device patent. The district court held that the Nimrod publication did not anticipate the '006 patent because there were three elements in the '006 patent that were not disclosed in the Nimrod article:

1. "the Nimrod device does not teach simultaneous scanning with multiple transducers ...";

2. "[t]he Nimrod article ... does not teach combining or correlating the information content of ultrasonic signals from the various modes"; and

3. "the Nimrod device ... does not rotate around the internal circumference of the bore...."

The district court did not specify which claims of the '006 patent include the three elements that it found were not disclosed in the Nimrod article or whether each of the six disputed claims before us contains those three elements. In order to determine the correctness of the district court's findings of nonanticipation, therefore, a detailed comparison of the claims with the Nimrod article is necessary. The appellants have conceded that the Nimrod article does not anticipate claims 1, 2, and 3 of the '006 patent:

[T]he Nimrod article is anticipatory of method claims 7, 8 and 10. Nimrod differs from apparatus claims 1-3 in that only a single mode transducer at a time was placed upon the Nimrod probe so that "simultaneous" transmission of signals in various modes is not disclosed.

An examination of these claims, therefore, is not necessary. A comparison with the method claims, though, shows that each claim has at least one element not disclosed in the Nimrod article.

Claim 7 is directed at ultrasonic inspection. It includes the step of "combining the information [from the ultrasonic signals] to derive an accurate indication of the position, nature, size and shape of the flaws in the rotor material." Although the Nimrod article discloses three means for recording the output of the ultrasonic detector (Mk. VI flaw detector display unit, two channel flaw alarm unit, and an auxiliary display unit for "photographic recording purposes"), these recording means do not permit the combining of information.

Claim 8 describes in more detail the method described in claim 7. It includes the combining element of claim 7 not shown in the Nimrod article.

Claim 10, which depends upon claim 8, similarly contains the combining element of claim 7 that makes the Nimrod article not anticipatory.

In sum, claims 7, 8, and 10 are not anticipated by the Nimrod article. The district court's finding to that effect is not clearly erroneous. Carman Industries, Inc. v. Wahl, 724 F.2d 932, 220 U.S.P.Q. (BNA) 481 (Fed.Cir.1983).

C. Obviousness (35 U.S.C. Sec. 103).

1. The district court recognized the statutory presumption that a patent is valid (35 U.S.C. Sec. 282 (1982)). It ruled that the presumption of validity is weakened if there is prior art that is more pertinent than the art called to the attention of the Patent Office. It then concluded, however, that the "[d]efendants did not present any evidence demonstrating why the art not cited to the patent office was more relevant or more pertinent than the art cited to the patent examiner, and the Court has found no ground for such a conclusion," and that the presumption of validity had not been weakened.

We have repeatedly held that the existence of such uncited prior art does not weaken the presumption but merely makes it easier for the party challenging the validity of the patent to carry his burden of proof. Lindemann Maschinenfabrik GMBH v. American Hoist & Derrick Co., 730 F.2d 1452, 221 U.S.P.Q. (BNA) 481 (Fed.Cir.1984); Perkin-Elmer Corp. v. Computervision Corp., 732 F.2d 888, 221 U.S.P.Q. (BNA) 669 (Fed.Cir.), cert. denied, 469 U.S. 857, 105 S. Ct. 187, 83 L. Ed. 2d 120 (1984); Lear-Siegler, Inc. v. Aeroquip Corp., 733 F.2d 881, 221 U.S.P.Q. (BNA) 1025 (Fed.Cir.1984). The court's error in dealing with the effect of prior uncited art upon the presumption was harmless, however, in view of the court's holding that there was no prior art more pertinent than that called to the attention of the Patent Office. Cf. Stratoflex Inc. v. Aeroquip Corp., 713 F.2d 1530, 1534, 218 U.S.P.Q. (BNA) 871, 876 (Fed.Cir.1983).

The district court also erred in ruling that the presumption of validity could be overcome by a preponderance of the evidence. The presumption may be rebutted only by clear and convincing evidence. Jones v. Hardy, 727 F.2d 1524, 220 U.S.P.Q. (BNA) 1021 (Fed.Cir.1984). This error also was harmless, however, since the district court found that the appellants had not met the less exacting burden of proof the district court applied.

Portions of the district court's opinion do not focus upon the language of the claims of the '006 patent--both in comparing that patent with the prior art in determining obviousness and later in comparing it with the Westinghouse device in determining infringement--but more broadly refer to the "patent" itself. Other portions of the opinion, however, indicate that the district court's comparisons were based upon the language of the claims. Our review of the record satisfies us that the district court's conclusion of nonobviousness was correct, and that its findings of infringement, discussed later, were not clearly erroneous.

2. In holding that the invention of the '006 patent would not have been obvious to one of ordinary skill in the art at the time the invention was made, the district court applied the criteria for determining obviousness announced in Graham v. John Deere, 383 U.S. 1, 17-18, 86 S. Ct. 684, 693-94, 15 L. Ed. 2d 545 (1966). The court determined (i) "the scope and content of [the] prior art," (ii) "differences between the '006 claims and the prior art," and (iii) "the level of ordinary skill in the pertinent art." The court also discussed at length the "secondary considerations" relating to obviousness, including commercial success, long-felt but unsolved needs, and failure of others, an inquiry we have held is an essential and integral part of determining obviousness vel non. Jones v. Hardy, 727 F.2d 1524, 220 U.S.P.Q. (BNA) 1021 (Fed.Cir.1984); Medtronic, Inc. v. Cardiac Pacemakers, Inc., 721 F.2d 1563, 220 U.S.P.Q. (BNA) 97 (Fed.Cir.1984); Simmons Fastener Corp. v. Illinois Tool Works, Inc., 739 F.2d 1573, 222 U.S.P.Q. (BNA) 744 (Fed.Cir.1981), cert. denied, 471 U.S. 1065, 105 S. Ct. 2138, 85 L. Ed. 2d 497 (1985).

In determining the ordinary level of skill in the art, the district court found that Gilbert Ronca, a Westinghouse employee, was one of such skill. The appellants accuse the district court of not having made an adequate level-of-skill determination. We cannot say, however, based upon what the district court stated and what the record shows about Mr. Ronca's qualifications and his background and experience in the industry, that the district court's finding that the ordinary level of skill in the art was that of Mr. Ronca was clearly erroneous and did not constitute an adequate level-of-skill determination.

In evaluating the scope and content of the prior art and the differences between that art and the claims of the '006 patent, the district court erred in concluding that "none of the prior art not cited to the patent office teaches combining the information content of the various modes of reflected and unreflected signals, as does the '006 patent." U.S. Patent No. 3,221,544 (Gunkel), not cited to the Patent Office, discloses an electrical method for combining data from multiple ultrasonic scans. Moreover, the question is not simply whether the prior art "teaches" the particular element of the invention, but whether it would "suggest the desirability, and thus the obviousness, of making the combination." See, e.g., Interconnect Planning Corp. v. Feil, 774 F.2d 1132, 1143, 227 U.S.P.Q. (BNA) 543, 551 (Fed.Cir.1985); Lindemann Maschinenfabrik GMPH v. American Hoist & Derrick, 730 F.2d 1452, 1462, 221 U.S.P.Q. (BNA) 481, 488 (Fed.Cir.1984).

The district court described the Gunkel patent as a device which "detect[s] and analyze[s] defects in tubular articles such as pipes. The device inspect[s] pipes from the outer diameter. Multiple transducers transmit[ ] ultrasonic signals as the pipe [is] rotated spirally past the patented device." Although this statement was correct, the district court failed to recognize that the Gunkel device could electrically combine signals from multiple transducers.

The Gunkel patent is directed to an ultrasonic inspection system containing a "novel means for analyzing defect signals to determine the character of the defect." The Gunkel patent describes two methods of utilizing ultrasonic signals: by "primary" discrimination, and by "secondary" or "combinational" discrimination. Primary discrimination involves using only a single transducer to perform the study, while secondary or combinational discrimination utilizes multiple transducers and analyzes the combined signals from these transducers. The Gunkel patent further explains that

[i]nsofar as is known, the prior art systems for defect discrimination have been primary discrimination systems which recorded the signals from the various transducers independently and relied upon analysis by the operator to provide the secondary information. However, such systems rely heavily upon the accuracy and judgment of the operator and provide considerable opportunity for human error.

These disadvantages of the prior art systems are overcome with the present invention and novel means are provided for performing the secondary discrimination electrically in a manner which is substantially instantaneous and which completely eliminates the possibility of human error.

The advantages of the present invention are preferably attained by providing novel means for ultrasonic inspection comprising a plurality of transducers, means for performing a primary discrimination of the signals from each of the transducers, and electrical means for comparing and analyzing the results from all of the transducers to perform a secondary discrimination.

Although, as indicated, the district court noted that the Gunkel device used "[m]ultiple transducers [that] transmit[ ] ultrasonic signals as the pipe [is] rotated spirally past the patented device," apparently the court failed to recognize that a necessary element of the Gunkel device was the comparison of the data it obtained from the multiple transducers. In this respect, the Gunkel device performed a similar function to the correlation and combination functions of the '006 patent.

The district court also erred in distinguishing the '006 patent from the prior art on the ground that, unlike the prior art, "[t]he patented device transverse[d] [sic] the internal bore of the material being inspected both axially and circumferentially and accomplish[ed] the inspection from the internal diameter of the item being inspected." Claims 1, 2, 7, 8, and 10, however, do not require longitudinal or circumferential movement of the probe. Only claim 3 requires such movement, and the Nimrod article clearly shows an angular drive means and a longitudinal drive means which operate independently of each other.

One of the differences between the claims of the '006 patent and the Nimrod device was that the former, but not the latter, used multiple transducers. The district court found, however, and the finding is not challenged on appeal, that it was well known in the industry since 1960 that multiple transducers could be used simultaneously to scan an object. Additionally, the Gunkel patent indicates the utility of correlating and combining information from multiple modes, since it points out the disadvantages of using a single ultrasonic mode and the advantages of using multiple ultrasonic modes in a "secondary discrimination" system. Thus, standing alone, the prior art provides significant support for the appellants' contention that the '006 patent would have been obvious.

3. Prior art, however, cannot be evaluated in isolation, but must be considered in the light of the secondary considerations bearing on obviousness. As we have pointed out:

[E]vidence of secondary considerations may often be the most probative and cogent evidence of record. It may often establish that an invention appearing to have been obvious in light of the prior art was not. It is to be considered as part of all the evidence, not just when the decisionmaker remains in doubt after reviewing the art.

Stratoflex, 713 F.2d at 1538-39, 218 USPQ at 879; see also In re Piasecki, 745 F.2d 1468, 223 U.S.P.Q. (BNA) 785 (Fed.Cir.1984).

In its lengthy discussion of secondary considerations, the district court found (a) that "[t]he evidence is overwhelming that for well over a decade the industry had searched for a reliable method of detecting discontinuities in rotor forgings," (b) that "major turbine manufacturers had tried and failed to develop a reliable method of inspecting in-service turbine rotors," and (c) that "the patented device enjoyed commercial success and was perceived by Westinghouse and others as direct competition to Westinghouse's system of boresonic inspection." The appellants have not shown that those findings are clearly erroneous.

a. In 1963, ten years prior to the filing of the '006 patent application, Westinghouse engineers were seeking a method of rotor inspection through the bore. A letter from one Westinghouse engineer to another said that "search units to examine rotors from the bore surface are and have been a need for many years" and that "[t]here have been many times in the past when we could effectively use such units." At the time this letter was written, Westinghouse was using one of General Electric's boresonic test devices. Based on the letter, and testimony of Mr. Ronca and Mr. Renner of Westinghouse, the district court found that Westinghouse was not satisfied that the General Electric device was reliable.

The district court found that almost ten years later, in 1973, TVA asked Westinghouse to examine the possibility of performing rotor bore inspections. The district court based this finding on the testimony of Mr. Beck, a Westinghouse employee, that for a year following the Gallatin explosion, TVA insisted that Westinghouse obtain the capacity to perform rotor bore inspections.

b. The record shows that Westinghouse had attempted to discover a method of inspecting rotors from the bore, but had failed to find one that worked. Westinghouse department heads compared the device its own research and development efforts had produced, with Commercial's then existing device. Mr. Ronca's summary of the meeting states that some people present "expressed assessorial opinions regarding the technical superiority of the [Commercial] system. No attendee took exception to these assessments so acceptance of the [Commercial] system on a technical basis was considered unanimous." Later, Westinghouse had Commercial perform approximately 20 inspections for it using the device the '006 patent covered, until Westinghouse had perfected its own system.

The evidence fully supports the district court's finding that others in the industry were unable to solve the problem. Westinghouse, a large corporation working on this matter, had tried but failed. Indeed, Westinghouse had pursued other solutions to the problem, using such technology as variable angle transducers, acoustical holography, and emersion testing. In 1972, when Commercial explained to Westinghouse officials its concept of a boresonic test apparatus, Mr. Ronca of Westinghouse stated that he thought the idea was not feasible because Westinghouse's own research endeavors to create such a device had failed. When shown in 1974 the Commercial device that the '006 patent covered, Mr. Ronca expressed doubt that Commercial could have produced it. Yet two days later, Mr. Ronca wrote to Westinghouse department heads that "the [Commercial] system represents a significant advancement in the field of boresonics...."c. The district court found, and the record shows, that Commercial had performed close to 100 inspections using the patented '006 device, and that about twenty of those inspections had been performed for Westinghouse. The district court also cited a letter from Westinghouse to its field representatives urging them to discourage buyers from using Commercial's system and to push Westinghouse's system, from which the court inferred that Commercial was in active competition with Westinghouse in this technology. The court also cited a letter from an engineer at Detroit Edison stating that "[Commercial] is the forerunner and developer of onsite bore sonic [sic] and bottle bore equipment and are nationally recognized by EPRI [Electric Power Research Institute][,] Westinghouse, [General Electric] and others."

In this field of endeavor, where the number of bore inspections necessarily is relatively small, this is strong evidence of commercial success.

Westinghouse contends that these secondary considerations are irrelevant because Commercial's device did not operate effectively. Westinghouse cites the testimony of Dr. Gelhaus of EPRI describing Research Project RP502. In this project, Westinghouse and Commercial boresonic inspection devices were used to inspect a rotor that subsequently was cut apart and examined for defects. Dr. Gelhaus testified that the "techniques and the results were sorely wanting," and indicated that both test procedures produced inadequate results.

Batell Laboratories, one of the laboratories involved as an observer in the RP502 report, in reporting to the Consolidated Edison Company of New York on the analysis of stress and structural integrity it made of a generator rotor, stated that "[t]he ultrasonic inspection of the rotor is by necessity performed from the bore surface only. Hence, flaw orientation is inherently difficult to determine. The system used by Commercial Machine Works represents the best solution available to this problem. Nevertheless, there is still a degree of uncertainty." Although Commercial's '006 patented device might not have met Dr. Gelhaus' expectations, the industry apparently viewed the device as the best solution to the problem of boresonic inspection.

In light of the district court's findings and the evidence in the record, including the strong secondary considerations indicating nonobviousness, which weigh heavily in the determination of obviousness, In re Piasecki, supra, we agree with the district court's conclusion that the invention the '006 patent covers would not have been obvious to one of ordinary skill in the art. This is one of those cases where evidence of secondary considerations "may ... establish that an invention appearing to have been obvious in light of the prior art was not." Stratoflex, 713 F.2d at 1538.

A. In finding that the Westinghouse device infringed the '006 patent, the district court stated that Alco had presented "highly credible evidence, from two separate Westinghouse sources, that describe[d] Westinghouse's device for and method of boresonically inspecting turbine rotors," and that "[n]either defendant introduced evidence to discredit or rebut plaintiff's proof that the method and apparatus described in [the two sources] were the same method and apparatus used by Westinghouse to ultrasonically inspect TVA's rotors." The two Westinghouse sources were a Westinghouse Manual (called the "Blue Book") and a letter from Westinghouse's Steam Turbine Division Service Sales to one of its sales representatives that described the Westinghouse boresonic inspection system. The district court also noted that in an interrogatory Westinghouse had stated that it had used a boresonic system similar to the system described in the Blue Book to perform the TVA inspections.

The district court rejected the contention that Westinghouse's use of a gating procedure avoided infringement. The court stated that gating is a means of masking unwanted data from the recorder so that only the relevant portion of the collected data is displayed. The court found that Westinghouse inspected rotors using two separate scans in which one scan performed a shallow analysis by gating out the inner wall of the rotor and the portion of the rotor beyond about eight inches, and in which the other scan performed a deep analysis by gating out the inner wall, the outer wall, and that portion of the rotor within about eight inches from the probe. The district court found that the '006 patent did not require the inspection to be made on a single pass and that the Westinghouse method examined the entire rotor.

Finally, the district court rejected the contention that Westinghouse's device did not correlate or combine its data. Based upon its examination of the evidence, the court found that the Westinghouse device correlated and combined the information from its boresonic inspection.

The court stated that, "[b]ased on the evidence presented, the Court finds plaintiff ... has met its burden of proving infringement ... [and that] ... the accused Westinghouse device contains every element found in claims 1, 2, 3, 7, 8 and 10 of the 006 patent."

B. In their appeal the appellants challenge, on various grounds, the finding that the Westinghouse device correlates or combines the data it obtains from the boresonic inspection of the rotor. As in the case of the appellants' challenge to the factual determinations underlying the conclusion of nonobviousness, our review of the record satisfies us that the findings relating to infringement are not clearly erroneous and were based upon correct legal standards.

1. The appellants first contend that there is no evidence that Westinghouse correlates or combines transmitted data using unreflected signals, as the '006 patent requires. This argument rests upon a misconstruction of the patent claims. All of the claims specifically state that the correlation of the "information content" of the reflected and unreflected (transmitted) signals is accomplished "by precisely locating in a three-dimensional matrix the path through the rotor mass of each transmitted ultrasonic signal and the position of discontinuities in the rotor material evidenced by each reflected ultrasonic signal...." If there is no reflected signal, the information content is that there is no indication of a flaw when viewed from that ultrasonic mode at that angle. The district court correctly interpreted the claim language and applied it to Westinghouse's device.

2. The appellants next contend that the Westinghouse device uses a two-mode inspection only in the area from 3 to 8 inches from the bore and therefore does not operate "throughout the mass of the rotor material" as the patent claims require. Again, the appellants misread the claims.

The phrase in the claims, "the information content of said transmitted and said reflected signals relating to discontinuity characteristics throughout the mass of the rotor material," is a dependent clause which explains the meaning of the term "information content" used earlier in the claim. It requires not that the entire rotor be inspected with two modes, but only that the entire rotor must be inspected and the data correlated. Had the Westinghouse device inspected only a volume from three to eight inches from the bore with two ultrasonic modes, it would have avoided literal infringement. The Blue Book, however, states that "echos from about just beneath the bore surface to just beneath the exterior surface are recorded." Westinghouse searches the entire mass of the rotor and, therefore, literally employs this element of the '006 patent.

3. The appellants argue that the Blue Book and the interrogatory answer do not prove that the Westinghouse device correlates and combines its data. This argument, however, ignores the third item of evidence upon which the district court relied, the letter from Westinghouse to one of its sales representatives. That letter describes the correlating and combining aspects of the Westinghouse device in terms almost identical to the claim language: "a plot can be made which gives a spacial [sic] representation of the [defect's] size, shape, orientation and location."

4. Finally, the appellants contend that the district court inferred the existence in the Westinghouse device of an apparatus or means to correlate and combine data. With respect to the correlating feature, the Blue Book states that "[t]he amplitude and length of time required for an echo to return is recorded, along with the longitudinal position and the rotational orientation.... Thus, the recorded echo time, longitudinal position and rotational position permit an accurate location of any discontinuity." Elsewhere the Blue Book states that

[e]ach scan group is made with three ultrasonic transducers scanning simultaneously. Each transducer has a different orientation relative to the rotor interior. This not only permits a cross check of any ultrasonic findings by a single transducer but also permits the analyst to more precisely determine the location and orientation of any cracks or flaws that exist within the rotor interior.

As the district court found, since the Westinghouse device correlates the data, it must also have a means for doing so.

With respect to the combining aspect of the Westinghouse device, we have already noted the language in the letter from Westinghouse to one of its sales representatives which, as the district court said, "is an echo of the 006 patent's claims."

5. Although the evidence of infringement is circumstantial, that does not make it any less credible or persuasive. Moleculon Research Corp. v. CBS, Inc., 793 F.2d 1261, 229 U.S.P.Q. (BNA) 805 (Fed.Cir.1986). In view of the nature of the device and the function it performed, it is not surprising that Alco was unable to produce direct evidence of infringement. Alco hardly could have been expected to present eyewitnesses to the use and operation of the Westinghouse device who would compare the Westinghouse device to the '006 patent claims. The evidence Alco produced, however, is adequate to support the district court's finding that the Westinghouse device infringed the claims of the '006 patent.

The judgment of the district court that the claims of the '006 patent are valid and enforceable and that Westinghouse infringed those claims is affirmed.

AFFIRMED.

NIES, Circuit Judge, concurring.

I concur in affirming the judgment of the trial court. In my view, appellants have failed to carry the burden of showing reversible error.

On the issue of infringement, appellants argue that the district court erred in holding that the Westinghouse "device" possesses a means to correlate and combine. Alco correctly asserts that claims 1, 2, and 3 require no such means. Those claims cover a device which produces data which is capable of being correlated and combined. Westinghouse's equipment also produces such data. Thus, the district court's holding--which was made, in any event, only with respect to claim 3--is harmless error. With respect to claims 7, 8, and 10, the method claims, those claims do include a step of correlating and combining the data. Westinghouse conceded that it does on occasion itself correlate or combine information, not merely furnish the data to others. Thus, that limitation in the method claims is met. I agree with Judge Friedman's infringement analysis in other respects and agree that all claims were proved to be infringed.

I also concur with Judge Friedman that the district court's conclusion that the claims are not invalid under section 103 cannot be disturbed. I believe, however, Judge Friedman misreads the district court's differentiation of the subject claims over the prior art Gunkel patent. Precisely, the district court spoke of no prior art teaching the correlation of reflected and unreflected signals. That Gunkel combined signals from multiple transducers does not contradict the district court. Thus, I do not agree that the district court "failed to recognize" that Gunkel taught combining signals. Gunkel is irrelevant to the point the court was making. Indeed, I believe much labor is expended unnecessarily to uphold the patent on the record before us. Given the unique problems of on-site inspection of huge turbine rotors, Westinghouse's argument that the claims would have been obvious from the teachings of various prior art references appears to me to be a classic case of hindsight selection. Moreover, the reaction of Westinghouse personnel to the demonstration of Alco's device is persuasive as indicia of nonobviousness.

The dissent rejects the evidence of commercial success on the ground that the success must be attributed to Alco's services, not to the invention. The dissent considers that the examiner erroneously allowed amendments to the claims with respect to combining and correlating data which have no support in the specification. I disagree. Although raised below, there is no section 112 issue on appeal directed to the adequacy of the disclosure to support the claim language with respect to correlating and combining data. Such an assertion of error would be meritless. A fair reading of the specification supports the language of the claims, including the specific step in the method. The whole purpose of the acquisition of the data is to put the data together in a meaningful way so that the location and nature of flaws can be determined. The actual techniques for correlation and combining data are admitted by the parties to be well known in the prior art and are, thus, not required to be described in the specification. Thus, I depart from the dissent's analysis of this issue as a predicate for the denigration of the evidence of commercial success.

One final point, the panel has considered the question of this court's jurisdiction over the appeal, which depends upon the jurisdiction of the district court being based on 28 U.S.C. Sec. 1338(a). That section, in turn, requires that the claim arise under "an Act of Congress relating to patents." This suit is based on 16 U.S.C. Sec. 831r which is part of an Act "[t]o improve the navigability and to provide for the flood control of the Tennessee River" and similar purposes. Judges Friedman and Rich conclude that the claim arises under an Act relating to patents; I do not. In my view a 16 U.S.C. Sec. 831r claim against TVA is more comparable to a claim under 28 U.S.C. Sec. 1498 (1982) for reasonable compensation for use of a patented invention by the government, which does not arise under a patent statute, Motorola, Inc. v. United States, 729 F.2d 765, 768, 221 U.S.P.Q. (BNA) 297, 299 (Fed.Cir.1984), and cases cited therein, than to a claim for patent infringement under 35 U.S.C. Sec. 271. The only decision directly bearing on this question was rendered in this very case by Judge Brown (reported at 448 F. Supp. 1175 (1978)). In dismissing a claim against Westinghouse, Judge Brown held that Westinghouse could not be an infringer in view of the right of TVA to use the patented invention. Thus, this is a suit for compensation for a lawful use, not a suit for unlawful, i.e., infringing, use. That the word "infringed" as well as "used" appears in 16 U.S.C. Sec. 831r does not mean that Alco's claim is necessarily for patent infringement or arises under an act relating to patents.

RICH, Circuit Judge, dissenting.

The decision below is reported at 597 F. Supp. 133, 224 U.S.P.Q. (BNA) 577 (1984). Reference will be made to the opinion as there published.

I am constrained to dissent from the majority's holding that the six claims in suit are not invalid for obviousness in view of the prior art under 35 U.S.C. Sec. 103.

The majority opinion, after obviously careful consideration of the facts, preliminarily arrives at the conclusion that the claims in suit define only obvious subject matter and then reverses that conclusion because of the "secondary considerations," which I shall discuss further, concluding thatThis is one of those cases where the evidence of secondary considerations "may ... establish that an invention appearing to have been obvious in light of the prior art is not." Stratoflex, 713 F.2d at 1538 [218 U.S.P.Q. (BNA) 871, 879 (Fed.Cir.1983) ].

This is my point of disagreement. This is not one of these cases because, also per Stratoflex, a nexus is required between the invention disclosed in plaintiff's patent and the secondary considerations. No such nexus exists here.

The majority, like the trial judge, has been led astray and has assumed that the patent in suit is on some imaginary "system" for detecting flaws, which has enjoyed commercial success in the hands of Alco and was copied by Westinghouse, thus changing the prima facie obvious invention into a patentable invention.

What has happened here is that by a kind of magician's distracting patter, the purpose of which is to keep the viewer from observing what is actually happening, attention has been directed to the patent's claims to the exclusion of its disclosure. As the prosecution history shows, by a series of amendments the claims acquired a kind of life of their own, divorced from the disclosure of the patent's specification, including reference to what the majority opinion identifies as the "center" around which the contentions of the parties on appeal now revolve, namely, "the language of the claims that pertains to combining and correlating the information derived from the ultrasonic scans." (My emphasis.)

The sole function of claims is to delineate the scope of protection afforded by the patent, not to describe what the patentee has invented. Claim language, by law, is supposed to find support in the specification. Unsupported claim limitations should be ignored in appraising commercial success, though it could be a ground for invalidating the claim, but I am not discussing that issue. In determining what was invented, the specification must be considered in the state in which it was filed and the original claims, which are a part of it, may be regarded as a part of that disclosure. The patent and its file history, which are legal documents, are before us and it is necessary to consider them to determine what the invention disclosed therein truly is. In fact, that should be the first order of business. The salient fact that emerges from such examination is that there is no reference whatsoever in the specification and original claims, taken together, of combining or correlating anything, or of any means for or method of performing the function of combining or correlating. These are simply terms injected into the claims by the prosecuting attorney, without any support in the disclosure, in an effort to persuade the examiner that the claims patentably distinguished from the prior art he had cited against them. It does not appear to me that the examiner ever compared the claims he allowed with the disclosure to see if they are supported by it.

When the majority opinion refers to "the meaning of combining and correlating used in the patent" (in Part III, A) it must be borne in mind that those terms are not "used in the patent" insofar as the disclosure thereof is concerned, but merely appeared in the claims long after the application was filed. It must also be borne in mind that the majority has correctly held that "one skilled in the art would have known [before Smith's invention] how to combine and correlate the ultrasonic scan data"--a finding also made by the trial court--for which reason it is part of the prior art and therefore can be no part of what Smith invented. It was a skill of the art and a difficult one at that, comparable to the skill required of a radiologist in interpreting X-rays or CAT scans in deducing the existence of tumors.

The majority finds significant error by the district court in the following respects:

1. In failing to realize that the Gunkel patent reference, not considered by the PTO, discloses electrically combining ultrasonic signals from multiple transducers, in a flaw detection system, to compare and analyze the results, and thus "performed a similar function to the combination and correlation functions of the '006 patent [claims]." (Of course, as I have just pointed out, the '006 patent in suit does not disclose any such functions.)

2. In not understanding that the "Nimrod" article reference (Brooks et al. "Ultrasonic Inspection of the Nimrod Power Plant Alternator Rotors") discloses both longitudinal and angular drive means for the ultrasonic inspection of rotor bores. Those are the very same functions Smith's apparatus was designed to perform, in which, therefore, there is no novelty.

The majority also notes that the district court correctly found that it had been well known in the industry since 1960 that multiple transducers could be used simultaneously to scan an object and that this finding is not challenged on appeal.

On the basis of the foregoing, the majority holds that "the prior art provides significant support for appellants' contention that the '006 patent would have been obvious," (my emphasis) by which, I presume, the majority means that the inventions (apparatus and method) of the claims in suit would have been obvious. This is what I refer to as its holding of prima facie obviousness, with which I wholeheartedly agree. As a preliminary to discussing why the majority should not back away from that holding, let me summarize in one paragraph what the basis of that finding of prima facie obviousness is.

Long prior to Smith's supposed invention, the ultrasonic detection of flaws in metal parts was a highly developed and sophisticated art. Smith's filing date was Dec. 3, 1973. Over ten years before that, Gunkel taught the use of multiple transducers operating simultaneously, longitudinally of and around tubular articles such as pipe, but on the outside, to detect flaws with use of electrical analysis means to combine and correlate the data produced by the transducers and their related pickups, to detect and appraise the nature of flaws. Gunkel proposed using as many as four transducers generating as many different modes of signals to get as many different kinds of data. The Nimrod disclosure shows the use of ultrasonic testing conducted from the inside of a bore of a shaft, as in the '006 patent, through which the transducers are moved longitudinally and angularly rotated. Necessarily, some indexing means for accurately determining where the transducers are at the time when any given signal is received must be, and was, employed in all such apparatus or it would not be known where the flaws are. So the patentee did not invent that function.

What the '006 patent discloses as Smith's invention is simply a piece of mechanism for use in selectively positioning one or more transducers and pickups in a bore and simultaneously "indexing" to indicate or record the longitudinal and angular position, i.e., the exact location, of the transducers in the bore to show where they are when signals are received. The totality of Smith's disclosure of what he does with any signals received is contained in the underlined words in the following sentence, taken in conjunction with the single "schematic" (i.e., not a working) drawing in the patent, reproduced below:

If the propagated [ultrasonic] waves contact a flaw, there is a reflection to the pick-up means at the appropriate source 45 or 47, and an appropriate indication is sent back to a recording or display device 87. [Col. 5, lines 65-68.]

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLEThe "recording or display device 87" is not described further but merely indicated, as some alternative and presumed known device of the prior art, by the square at the lower left corner of the schematic figure containing the number 87. Furthermore, it will be seen in the drawing that box 87 is shown as connected to a line 13. Line 13 is described in the specification (col. 3, 11. 35-39) as an "air line" which carries pneumatic pressure into the device to actuate the legs 11 on supports 3 and 5, a function totally unrelated to recording or displaying signals. Signals presumably would be conducted by "electrical lines 77" (col. 5, 1. 34), not further described and not shown as connected to anything, since the signals are electrical. In sum, the patent disclosure is wholly devoid of any means or method for the "correlation of the information content of a transmitted or reflected ultrasonic signal" or for "combining the individual content" of signals (to quote claim language) or of any teaching of how to accomplish either of those functions. The claims in suit, therefore, were created out of the whole cloth by the imagination of the prosecuting attorney and based on knowledge of prior art or subsequent events, not on the teachings of the '006 patent. One must be wary of such expressions, used in the majority opinion, as "the device the '006 patent covered" and "the patented '006 device." They fail to distinguish between the invention disclosed in the patent and other things which may fall within the scope of the later-filed claims.

The majority opinion, page 1496, refers to "an amendment" as containing the inventor's best description of "combining." What is in fact referred to is attorney argument, in connection with an amendment, during prosecution and contained in "AMENDMENT 'B' " in the section headed "REMARKS" (Apx. 158, 166, 170). This is no part of the patent disclosure and cannot be relied on to supply deficiencies therein. Even if it had been an actual proposed amendment, it would not be part of the disclosure, any more than attorney argument in this court. Attention is further directed to the fact that "Amendment 'B' " was refused by the examiner and never entered (A-176) and was replaced by "Amendment 'C' " (A-195). In any event it is not the inventor's description.

It is also a significant fact that the inventor, Robert D. Smith, named in the patent issued to Alco, left the employ of Alco in October 1973 before the application was filed on December 3, 1973. He testified on deposition that after it was filed he never had any knowledge of the prosecution or any contact with the attorneys (A-1518, 1541-42).

When we come to examine whether there was commercial success of Smith's invention we are, of course, concerned only with the success of what he both invented and disclosed. If there was commercial success, that success does not count in appraising non-obviousness unless it can be attributed to what he both invented and disclosed. In light of all of the foregoing, I am convinced, as a matter of law, that unless there is a clear nexus between the disclosed subject matter of the patent (not just the vagueness of later-conceived broadly worded claims) and very significant secondary considerations the claims are invalid for obviousness.

Without nexus, of course, the invention that is disclosed in the patent does not enjoy the benefits of commercial success in appraising its nonobviousness. The district court discusses nonobviousness beginning at 596 F. Supp. at 147, 224 USPQ at 586, in part B of its opinion. Under sub-heading 3, "Ordinary skill in the art," the court several times refers to the invention of the patent as a boresonic inspection "system," or the "patented system." But no system is disclosed in the patent; the court must, therefore, have had in its mind something in addition to what the '006 patent discloses. In section 4(c), "Commercial success," it gets to a brief discussion of that subject (F.Supp. at 153, USPQ at 591). One will search in vain for any description of just what it was that enjoyed commercial success. The record makes clear, however, that the Commercial Machine Works (CMW) division of Alco, which is supposed to have had the commercial success, was operating a service business of testing turbine rotors in the field, moving to the site with its equipment and personnel. One such test may take as long as a week or ten days, hundreds or even thousands of readings being taken (A-1016). When it is over, the net result is a report (A-1470) from CMW to the owner of the rotor saying what it thinks of the soundness of the rotor, what if any probable flaws or "discontinuities" have been found, and a record of its tests, columns of numbers representing "A" scan data (A-1469), which can be used later for comparison with subsequent tests. Nobody will ever know whether the report is or is not accurate, or how accurate, unless the rotor is physically altered by machining (as by "bottle-boring") (A-1474) or cut to pieces as a check on the tests.

Now, the totality of the commercial success found by the district court was that Westinghouse, which is a manufacturer of turbines, as well as a field-testing competitor of CMW, employed CMW in a crisis situation to test some 20 rotors in the field because it thought CMW was, at the time, the most competent operator in the field, and that CMW may have made a total of "[close] to a hundred" such inspections altogether. The question is, what does this prove about commercial success of the invention disclosed in the Smith patent, which is a piece of apparatus for locating transducers and pick-ups in a rotor bore--nothing more. Is that apparatus the principal reason CMW was hired to make tests? Hardly. CMW was hired because it was considered competent to make tests. The apparatus disclosed in the patent is but part of the equipment necessary to obtain "A" scan data; it does not produce that data by itself, it only positions the transducers. The transducer must be powered by frequency generators, not shown but known to the prior art. The frequency generators must produce different frequencies. Different transducers must be used to produce different "modes" of ultrasonic waves, also not explained in the patent but known to the art. When the transducers are energized and send back data through the pick-ups, that electrical data must be recorded by undisclosed instruments (except for box 87), also known to the prior art. And even after that is done, the thusrecorded data must be interpreted by highly trained and skillful personnel in order to provide the information being sought. (Wallace A-1468 et seq.; Gelhaus A-1493 et seq.) Computers may be used which have to be properly programmed. And when all is said and done, there is no assurance that the deductions from the data are really true.

What it comes down to is that CMW's success in selling its testing services to rotor users depended on its customer's confidence in CMW's abilities as the best available testing service organization. Since the most important aspect of such confidence, as the majority has deduced from these adversary proceedings, was CMW's technical ability through its personnel to correlate and combine the information content of the ultrasonic signals (matters on which the patent gives no instructions whatsoever) to arrive, perhaps with the aid of a "fracture mechanic's analysis" (A-1472), at conclusions on the condition of the rotor tested. Therefore, it appears clearly to me that the only commercial success relied on here or below cannot be attributed to Smith's invention as disclosed in his patent but must have been due primarily to other factors. It follows that there has been no showing of nexus between Smith's prima facie obvious invention and the commercial success to take the invention out of the obviousness category.

I also emphasize that what has become the crux of this case as the supposed contribution of the patentee--correlating and combining data obtained from ultrasonic tests--has been thoroughly established by the record as a technical skill which existed in the prior art long before the patentee's invention. The mere presence of these correlating and combining limitations in the claims by reason of an attorney's effort to distinguish them from the disclosures of references is not a justification for treating them as part of Smith's invention as though they were his contribution to the art. True, they are a vital part of CMW's services, which have been successfully sold, but correlating and combining remains knowledge of the art free for all to practice. What is free to all cannot be attributed to the patentee.

I would therefore hold the claims in suit invalid for obviousness under Sec. 103.

The issue of nexus is definitely before us on appeal. Appellants' main brief devotes 10 pages to arguing lack of nexus, 20% of appellants' entire argument. Nexus is of the utmost significance in reaching the correct result on the obviousness issue because, as I have said, it is the only basis used by Judge Friedman to escape from his preliminary conclusion that the prior art makes the claims in suit obvious, which conclusion is clearly right.

Appellee has chosen to avoid answering the arguments that nexus was not established by almost totally ignoring them. Literally, its brief contains only a single short sentence on the question. It reads (p. 36):

It is hard to see how there was a lack of nexus when Westinghouse copied the patented invention and even arranged with CMW to conduct inspections for it on a subcontract basis by utilizing the apparatus and method that it had patented. [Emphasis mine.]

Now, it is "the patented invention" that I have been discussing and what I mean by it is the invention that Smith made and disclosed in his patent. What else could he have patented? I have tried to determine whether the so-called commercial success which led the lower court and Judge Friedman, and apparently Judge Nies also, to find commercial success and thus nonobviousness is properly attributable to what Smith disclosed as his invention so as possibly to tip the scales in favor of finding prima facie invalid claims valid. I am convinced it is not. CMW's business success is not shown to be success of the Smith invention.

Judge Nies is swayed in favor of nonobviousness by "the reaction of Westinghouse personnel to the demonstration of Alco's device." (My emphasis.) What device? The person from Westinghouse who was swayed was Gilbert E. Ronca. His testimony is not voluminous (A-1070-1135) and it clearly shows what impressed him. It was definitely not the disclosure of Mr. Smith's patent. I will summarize what the record shows.

Mr. Ronca was sent by Westinghouse to CMW to look at their boresonic system (A 1106). There was nothing new about boresonic systems as such; General Electric had one and the Nimrod apparatus was another. Smith did not invent, and his patent does not disclose, a "system," yet that is what everyone involved in this case, including the trial court, persistently talks about. A system is, of course, what one has to have to do boresonic testing. Mr. Ronca saw a demonstration of the CMW system, not the invention, and he was given information about it by two employees of Southwest Research Institute, which had built the system for Alco (CMW). The Southwest people were longtime experts in ultrasonic testing, research, and equipment design. In fact they appear to have designed the device shown in Smith's patent. In fact, the schematic drawing of the patent appears to have been copied from a submission to CMW from Southwest. Mr. Ronca wrote a report and he testified about his impressions of the CMW system. He said, "The system is very flexible and is functionally modular. By using the building block concept, the number and the type of examinations can be suited to a pre-established examination program.... The equipment used had a .001 inch reset capability, both in the axial and circumferential directions." (A 1125) "As I tried to describe, the system was described as accepting preprogrammed P.C. [printed circuit] boards which offered a selection of scanning sequences such as scanning axially over a pre-selected length and indexing clockwise.... It had several different options. It also had options to accept different scan modes and different scan--in a menu of scan programs, such as surface waves, any combination of shear waves and so forth. And these could be plugged into the system." (A 1127) Explaining the reference to "menu," Ronca said that there were also representations about the capability of "marrying" the CMW device to a mini-computer. Asked about that, he said "this was the reference I made on having a menu of protocols that could be inserted for scan purposes." (A 1130) On all of this, Mr. Ronca reported enthusiastically and recommended Westinghouse try to buy it.

One will search in the patent in suit in vain for any disclosure or teaching of the things that so impressed Mr. Ronca in the CMW system or for a disclosure of any system at all.

I reiterate that the only commercial success of the CMW system relied on below is that company's testing of a total of, perhaps, 100 rotors. That testing was necessarily done by using the system, not merely the device shown in the patent, which in itself cannot test anything. There is no showing that the tests were accurate and there is considerable evidence that they may not have been. All we know is that CMW was hired to do them because their customers considered them competent. To my mind, that is no proof that CMW's success in getting that much testing business was due to the invention disclosed in Smith's patent. All it can do is position transducers in a bore and tell you where they are. That had been done in the prior art. One always has to know where the transducers are or it is impossible to determine where the flaws, if any, are. The accuracy with which it does so depends entirely on the refinement of the machining by which the locator is manufactured. The Nimrod boresonic device was said to be just as accurate, if not more so. All the rest--the interpretation of the signals to acquire some intelligence and the apparatus and instrumentation essential thereto--was the common knowledge of the ultrasonic nondestructive testing art, as my colleagues point out whenever I suggest non-disclosure of some claim limitation in the patent.

One should not be overly impressed by the background recitation of nonspecific, arcane, ultrasonic terminology about wave propagation modes, all of which was knowledge in the public domain. Aside from Smith's piece of machinery, which is generally described without particularization on the basis of a mere schematic figure, Smith does not give a single example of how to practice any method. He merely recites a multitude of general possibilities. Neither does he explain how to achieve his much-touted thousandth of an inch location accuracy.

I hesitate to suggest that my learned colleagues have been gullible, but on the subject of "commercial success" it does seem to me that they have been unduly moved by mere words and innuendos. Just what was the supposedly successful CMW system? They do not say. I have seen no reference in the briefs to any clear evidence on the subject and have been unable to find it in the voluminous record.

It is asserted as another indicium of nonobviousness that Westinghouse "copied" the CMW system and used it to do testing for TVA. There is a picture of the transducer-carrying head used by Westinghouse at A 536.20 and at tab 7 of appellants' brief. It bears little resemblance to the schematic drawing of Smith's patent. Where, then, do we find the evidence of the commercial use of the invention disclosed in the patent in either the CMW system or Westinghouse's alleged infringing copy? Unless and until it has been shown that there has been commercial use of the piece of mechanism disclosed in Smith's patent (which is not a system) there has been no proof of commercial success of his invention and no evidence establishing the necessary nexus. Finding no nexus, I would hold the claims in suit obvious and invalid under 35 U.S.C. Sec. 103.

I join Judge Friedman in finding that this court has jurisdiction and find no merit in Judge Nies's grounds for questioning it. Notwithstanding the location of Sec. 831r in a title of the U.S.Code other than Title 35, the section is an Act of Congress relating to patents, as is 28 U.S.C. Sec. 1498 which is controlling on other suits against the government for the use of patented inventions. Motorola, Inc. v. United States, 729 F.2d 765, 221 U.S.P.Q. (BNA) 297 (Fed.Cir.1984) says nothing to the contrary.

Having given consideration to 16 U.S.C. Sec. 831r for the first time in connection with this appeal, I am impelled to say that it is, from the standpoint of patent law, one of the most ineptly drafted statutes I have seen, displaying a confusion between patents and patented inventions, and a total lack of comprehension of what the patent right is. One cannot copy a patent right, which is only the right to exclude, or use it, or employ it, all of which Sec. 831r mentions in meaningless confusion. Only the owner of a patent or an exclusive licensee with the right to sue, who is a virtual owner, can use or employ the right to exclude. In spite of these misfortunes, however, the intent is clear that when TVA uses a patented invention and thus infringes the patent, the exclusive remedy is a suit against TVA "on the equity side of the appropriate district court of the United States...." Since this is an act relating to patents, jurisdiction of the district court is under 28 U.S.C. Sec. 1338(a) and we have jurisdiction under 28 U.S.C. Sec. 1295(a)(1).