Appellee Smith-Blair, Inc., brought this action to obtain a declaratory judgment declaring invalid claims 1 to 8 of United States Patent No. 2,897,568, the Hoke patent, relating to a flexible band-type repair clamp used primarily for repairing breaks and other leaks in underground water distribution systems, owned by Dresser Industries, Inc., a corporation. Appellant counterclaimed alleging infringement by appellee of the same alleged valid Hoke patent.

The trial court, having jurisdiction under 28 U.S.C. §§ 1338 (a), 1400(b) and 2201, rendered a declaratory judgment holding claims 1 to 8 of the Hoke patent invalid, and dismissing appellant’s counterclaim for infringement.

Jurisdiction on appeal is based on 28 U.S.C. § 1291 .

I FACTS

Certain facts are undisputed:

Claims 1 to 8 of the 1959 Hoke patent describe the following: a “conduit clamp” comprising a thin, flexible metal band, split so that it may be extended around a “conduit”; a pair of lugs, i. e., bars, secured respectively to the two ends of the band, the lugs facing each other and having aligned bolt holes; bolts extending through the holes and operative to draw the lugs toward one another, the pull of the bolts as they are tightened tending to roll the lugs toward each other; each of the lugs having a plurality of bearing surfaces and a plurality of arms alternately arranged; the arms projecting beyond the lug on which they are placed, overlapping the bearing surfaces of the other lug so as to resist the rolling of the lugs by the pull of the bolts; the arms of each lug being slidable on the bearing surfaces of the other lug in a direction parallel to the bolts to guide the lugs as they are drawn toward each other by the bolts. Close to each bolt are three arms, one on one lug, and two others on the opposite lug; which two arms slide on either side of the opposing one lug as the bolt is tightened. (R. 847.)

Claims 8 through 12 of the Hoke patent, though not held invalid by the trial court, are mentioned in certain disputed findings of fact and therefore must be set forth here.

They describe the means used to secure the lug, i. e., a bar, to one end of the flexible band: each of the lugs having an elongated lip adjacent the conduit and an elongated recess between the lip and the body portion of the lug, each of the ends of the band passing between the conduit and a lug and being bent back over a lip with the end of the band in the recess, a wire being provided along the end of the band, each of the lugs also having a malleable finger projecting forwardly from the body portion of the lug and bent down over the end portion of the band to grip the end portion between the finger and the lip and secure the wire in the recess, the end portion of the band being bent at least part way around the wire. (R. 847.)

The Hoke patent device (claims 1 through 12) operates to press a lining of rubber or other material, called a gasket, against the pipe to seal off a break or leak.

Early flexible band-type pipe repair clamps were subject to a similar disability; i. e., when the bolts were tightened too tight, a tendency existed for the lugs to “roll,” or tilt toward each other, and bend the bolts; this frequently made it impossible to sufficiently tighten the band. The Hoke device was not subject to this disability.

Appellant lists a number of alleged errors on the part of the trial court, set forth under three headings: “Points of Appeal,” “Errors in * * * Findings of Fact,” “Errors in Conclusions of Law.”

II POINTS OF APPEAL OR ALLEGED ERRORS

The sixteen “Points of Appeal” are the following:

1. The failure to apply the proper tests and standards to determine as a matter of law the patentability of the invention and the validity of the claims in suit, and in applying improper tests.

2. The failure to make findings as to the facts which preceded, attended and succeeded the appearance of the invention of the patent in suit, and which support the presumption of validity.

3. The making of erroneous findings on contradicted and unsupported opinion testimony of plaintiff’s witnesses through failure to test such testimony by recognized legal standards as to the qualifications, expertness, interest and prejudice of such witnesses, and through an erroneous view of the law.

4. In holding that the presumption of validity attaching to the grant of the patent in suit was rebutted by the absence of any citation of the Lindsay patent No. 512,406, issued by the British Patent Office, where (a) the British patent was in a nonanalogous art, and (b) the plaintiff did not satisfy its burden of proof in showing that the British patent had not been considered by the Patent Office Examiner.

5. In attaching any anticipatory significance to the Lindsay patent or to the Lindsay catalogs, both relating to adapters for converting ordinary bolts into hook bolts and in no way relating to the full circumference pipe repair clamps claimed in the patent in suit.

6. In considering Exhibits 5 and 6 alleged to have been made by plaintiff in 1953 to be a part of the prior art, where testimony of plaintiff’s own witnesses established that there was but an unsuccessful experiment and no public use or sale of the device.

7. In holding the claims of the patent in suit invalid under Section 102(b) of Title 35 United States Code, on the grounds that they were anticipated by tapping tees sold by plaintiff to Dominguez Water Company over a year before the Hoke application was filed.

8. In holding that plaintiff had satisfied its burden of proof in establishing that the patent in suit was invalid, where testimony by witnesses of both parties reinforced the presumption of validity attaching to the grant of the patent.

9. In admitting into evidence plaintiff’s Exhibits 63, 64, 14, 15, 17, 80 and 81 over the defendant’s objection that these drawings and devices were irrelevant, since they are not prior art; and if they purport to represent prior art, they are not the best evidence.

10. In combining the teachings of the Lindsay patent and “Lindapter catalogs” with testimony of plaintiff’s witnesses as to activities alleged to constitute prior art; and thereby finding invalidity of the patent in suit, where the combination was taught only by defendant’s patent and not by the prior art.

11. In finding that the invention of the patent in suit had copied plaintiff’s “2-F” form of connection, where it is clear from an inspection of the devices that this is not true.

12. In denying defendant’s motions to produce documents (R. 12, 15-27, 37) and thereby precluding defendant from offering evidence relevant to the validity of the patent in suit.

13. In adopting without modification findings submitted by plaintiff which were contrary to the evidence and law.

14. In making conclusions of law not supported by the findings.

15. In holding that claims of the Hoke patent cannot be infringed by plaintiff’s pipe repair clamp allegedly copied from an identical clamp in public domain since 1939, where evidence of infringement was clear and unrebutted and comparisons of devices and history of the development show that the accused device was copied from defendant’s pipe repair clamp and not from a prior art device.

16. In denying defendant’s motion to amend the judgment.

III FINDINGS OF FACT

Appellant challenges twenty-four of the findings of fact. Under Fed.R.Civ.P. 52(a), “ * * * Findings of fact shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses. * * >>

A finding will be held “clearly erroneous” when, althoug-h there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed. Hoppe v. Rittenhouse, 9 Cir. 1960, 279 F.2d 3 .

In considering the record in this case it becomes immediately clear that much of the dispute between the parties as to the findings made in the trial court resolves itself into a question of semantics —what is a “pipe clamp,” and what are the parts or components making it up? It seemed necessary to this court, on oral argument, (as it does here) to refer to three separate components of any flexible, band-type pipe repair clamp as: (a) the split band, (b) the clamping component, and (c) the connecting component.

When thus analyzed we discover in the findings objected to some language which might well be characterized as ambiguous. Yet when fairly interpreted and made reasonably applicable to the facts existing before the trial court, we find no reversible error in the findings. We consider each such alleged error in the margin.

IV PRIOR ART

Appellant urges that one “basic question” is whether the district court, “in -determining patentability should accept the self-serving opinion testimony of ■plaintiff’s president or should be guided 'by uncontroverted facts preceding, attending and succeeding the emergence of the invention which clearly show that it was not obvious.” Such “argument,” with the two pages that follow it in defendant-appellant’s brief (Brf. pp. 16-17), reads nicely, but aids the appellate •court little in defining or meeting the actual issues.

Appellant continually refers to “the problem” which it asserts the Hoke patent solved and which Telford Smith was unable to solve. No specific definition of the problem appears in the brief, •unless we take it “the problem” in repairing all leaks in pipe lines of water distribution systems is the “holding power” •of the clamp — i. e., is this power sufficient to produce a gasket pressure sufficient to render the pipe repair clamp leakproof?

Such pipe repair clamps are admittedly old in the art (Finding 8). Early pipe repair clamps were deficient in that when the bolts between the side bars at the respective ends of the clamps were tightened, they would tend to bend. To prevent such bending was the object of the Schustack patents Nos. 2,713,352 and 2,826,799 (Exs. Q and R). (Filed November 9, 1953.)

This was solved, says appellant, (Brief p. 24) by Hoke in supplying the “side bars” or “lugs” or “clamping component” with “a plurality of integral arms which project circumferentially beyond the face of the respective lugs and slide on bearing surfaces provided on the opposite lug to resist the rolling” of the lugs caused by the pull of the bolts.

The court found (Finding 11) that Hoke admitted “all he did was to add the sliding arm or finger to the plain lug of the prior art, and that this was the essence of his alleged invention.” The court then found that the prior Lindsay patent showed the sliding fingers clamp. (1939 Lindsay Patent, British No. 512, 406, R. 83.)

Appellant first urges that the Lindsay patent shows but one view of each of two forms; that at least two views are needed to determine the structure, and that appellee “by judicious fudging” used freehand sketches instead of the patent drawings themselves. Likewise, appellant objects to the use of “made up” lugs of the Lindapter patent. (Ex. 14.) These were not prior art, says appellant, but were made in 1961. Whether they were accurate reproductions of the lugs was a question of their admissibility as “mock ups” — a matter for the trial court to pass upon. No cases are cited to prove their introduction was error, merely that such evidence “misled” the court. We cannot conceive this to be a fact.

IV-B ANALOGOUS ART

Based on its findings, the trial court, among other things, concluded that the Lindsay patent is in “the clamp art as is the Hoke patent in suit.” Appellant disputes this.

In Stearns, the claims in suit related to a device to detect flaws in the coal tar or asphalt enamel coatings which are applied to steel pipe lines prior to burial in the ground. One component of this device was a coiled spring electrode that was wrapped around a pipe and rolled along its length. A prior patent, not cited by the Patent Office, was for an electric snap-switch which utilized a rolling coiled spring in its operation. The court in Stearns held that, assuming the elements to be similar, the purpose of the snap-switch art, i. e., to open and close an electric circuit, “has no relation or similarity to the purpose” of the art of flaw detection, i. e., to determine the condition of pipe coating by subjecting it to electrical inspection.

Appellant contends that an inventor working in full-circumference band-type pipe repair devices cannot reasonably be expected to naturally look to bolt adapters to find an answer to his problem.

The basic elements of the Lindsay patent are lugs, bolts, a bearing surface on one lug, and a finger or fingers on the other lug sliding across and onto the first lug to prevent rolling. These are the same elements found in the clamping component of the Hoke patent device.

And, the purposes of the two are neither remote, nor dissimilar. The purpose of the pipe repair clamp art is to clamp a split-band around a pipe; and, the purpose of the art of “bolt adapters” is to provide a general purpose clamp.

The trial court was not in error in holding “bolt adapter” analogous to “pipe repair clamp” art.

IV-C PRESUMPTIONS OF VALIDITY

The trial court further concluded that the failure of the Patent Office Examiner to cite the Lindsay prior art patent rebuts any presumption of validity attaching to the grant of the Hoke patent.

Even one prior art reference, which has not been considered by the Patent Office, may overthrow the presumption of validity. Jacuzzi Bros. v. Berkeley Pump Co., 9 Cir. 1951, 191 F.2d 632 .

Appellant refers to Artmoore Co. v. Dayless Mfg. Co., 7 Cir. 1953, 208 F.2d 1 , in which the court said:

In the Artmoore case, two prior art patents not 'cited by the Patent Office had “long since expired,” and the others were deemed not pertinent. But the Lindsay patent was pertinent. Appellant’s own expert Newell, as indicated in the discussion under Finding 13, did testify to the similarity in parts and functions between the Lindsay patent device and the clamping component of the Hoke patent device.

The term “clamping component” is used because only Hoke patent claims 1 to 8 are in issue. The component that secured the ends of the band to the lugs is described in claims 8 through 12.

Thus, since the Lindsay device is a general purpose clamp, and since Hoke patent claims 1 through 7 describe a similar clamping component except for mentioning the specific object the Hoke device clamps onto, the Lindsay patent seems obviously pertinent.

Appellant sets forth three facts to bolster its argument that the Lindsay patent was not pertinent. The first is that in 1961, Telford Smith obtained a patent on a pipe repair clamp modeled after his unsuccessful 1953 experimental clamp, but embodying a different connecting component. The Patent Office did not cite the Lindsay patent against the application for this patent.

The second is that Smith did not call the Lindsay patent to the attention of the United States Patent Office in connection with this application.

Third, that the Patent Office did cite the Lindsay patent against the 1956 “Risley patent” (R. 823).

The first and second facts were not before the trial court; they were presented in appellant’s brief for purposes of showing prejudicial error in denial of appellant’s motions to produce (to be discussed infra) and should be considered on appeal only on the question of prejudicial error in denial of such mtF tions.

That Lindsay was cited with respect to the 1956 Risley patent (which deals with a pipe coupling with multipart clamp), does not preclude the possibility that the Lindsay patent was overlooked in 1959, when the Hoke patent was issued.

V LACK OF INVENTION

The trial court concluded that claims 1 to 8 of the Hoke patent were invalid for lack of invention. The court relied on Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., 1950, 340 U.S. 147 , 71 S.Ct. 127 , 95 L.Ed. 162 .

The A. & P. case sets forth a test for determining whether or not there has been invention when there is a conjunction of known elements: The unification of the elements must result in unusual or surprising consequences; the old elements which made up the device must perform some additional or different function in the combination than they perform out of it. In combination, the Hoke clamping component still performs the same function other clamping components have performed, as does the connecting component, and the split band.

The unification of these three elements results in no unusual or surprising consequences. Certainly it is not, as appellant contends, the prevention of lug tilting — the Lindsay device had that consequence; and, certainly it is not a more effective connection — even if we were to read into Hoke patent claims 1 through 7 the connecting component.

There does indeed appear to this court, there was a lack of invention.

Appellant argues that the unusual or surprising consequence of unification was increased pulling power. (Appellant’s Reply Brief, p. 13.) But, this increase is only what follows from eliminating the tilting of the lugs that resulted in a bending stress on the bolts; the critical consequence is prevention of lug tilting.

Appellant further meets the A. & P. case with two arguments: That other standards of invention control, and that § 103 of the 1952 Patent Act (quoted supra), modifies the A. & P. ease to the extent that the latter is inconsistent with § 103.

Both arguments fall when it is recognized that in cases involving combinations of old elements, the A. & P. case is not modified by the 1952 Patent Act, but continues to be the law. General Motors Corp. v. Estate Stove Co., 6 Cir. 1953, 203 F.2d 912 ; Harvey v. Levine, N.D. Ohio 1962, 204 F.Supp. 947 .

Not having met the strict standard of the A. & P. ease, the Hoke patent claims in suit cannot somehow be deemed valid under the obviousness test of § 103. In cases involving combinations of old elements, it is the A. & P. standard that determines whether there is validity under § 103 of the Patent Act.

Thus, the conclusion of the trial court that Hoke did not exercise that degree of inventive skill necessary to comply with the requirements of § 103 of the Patent Act is supported by the facts of this case, and not error.

VI INFRINGEMENT

The conclusion that the claims of the Hoke patent in suit cannot be infringed by plaintiff’s accused pipe repair clamp is not error. An invalid patent cannot be infringed. See Bergman v. Aluminum Lock Shingle Corp. of America, 9 Cir. 1957, 251 F.2d 801 .

VII ADDITIONAL ERRORS

Appellant also contends that the trial court committed a number of subsidiary errors:

First, that it was error for the court to rely on the drawings in the Lindsay patent as prior art. Appellant’s contention is that Figures 9 and 10 in the Lindsay patent drawings are not two views of the same device, but are separate views of two different devices; the patent contains only one view of each of the two devices; therefore, it can only be conjectured what a plain view of the Figure 9 device would look like, and what a cross-section of the Figure 10 device would look like.

Since other views of the Lindsay device were available to the court in the Lindapter brochures, there is no error. (See Record, pp. 817, 838.)

Second, the Lindsay patent is allegedly deficient in that, being a foreign patent, it fails to contain and exhibit [such] a substantial representation of the patented improvement that anyone skilled in the applicable art can construct or practice the invention.

The case cited by appellant in support of this proposition indicates that in determining whether or not there is a substantial representation, both the description and drawings are considered. See National Latex Products Co. v. Sun Rubber Co., 6 Cir. 1959, 274 F.2d 224 . Here, the drawings and description in combination are, in our opinion, fully sufficient.. (See Lindsay patent, R. 863.)

Third, certain sketches (Exhibits 64,. 65, 69) are not prior art and not proper-evidence.

Appellant’s own counsel participated in-making at least one of these drawings. (R. 250, 345.) The objection is without, merit.

Fourth, Exhibit 63, a drawing of the-Lindsay device, was a recent drawing' and not prior art.

Since the court had the Lindsay patent and Lindapter brochures before it, it had' valid and reliable sources in assessing-this particular prior art, as it previously existed.

Fifth, Exhibits 14, 15, 80 and 81 are-not prior art, but were made for the-trial.

. Exhibits 14 and 15 were scale models-of Lindsay, and represented as such. No-misrepresentation occurred, and, again the court had other ample evidence on. this prior art.

Exhibits 80 and 81 are Lindapter lugs-that were shaped for purposes of trial; again, the court could not have been, misled. (See Record, pp. 383-85.)

Sixth, since neither Telford Smith, “nor any of the witnesses for that matter-were patent experts,” but were rendering opinions on the prior art, an appellate-court can review the weight to be given the testimony of such witnesses who are-not patent experts.

The type of expert indicated is not a. “patent expert” in the abstract; rather, he is an expert in the particular art to-which the device relates.

In this case, there was no paucity of the latter type of expert.

Seventh, appellant urges that there was prejudicial error in denial of its motions to produce two of appellee’s patent application files. Appellant urges that they were relevant as to such matters as Telford Smith’s credibility, and the issue of “obviousness.”

Although these motions were made before the Law and Motion judge, appellant did not renew its requests at the trial.

The case cited by appellant as an example of the rule of secrecy as to patent applications being set aside involved the issue of who was the first inventor. See James B. Clow & Sons, Inc. v. United States Pipe & Foundry Co., 5 Cir. 1963, 313 F.2d 46 .

No such fundamental issue relating to the patent files is involved here.

Eighth, there was error alleged in denying defendant’s motion to amend the judgment. That motion proposed this amendment: “That claims 1, 3, 4, 5, 6 and 8 of Patent No. 2,897,568, if valid, would be infringed by the accused device manufactured and sold by plaintiff.”

In view of the differences between the Hoke patent connecting component, and the connecting component of the accused device, and the similarity of the Hoke patent connecting component with the Lindsay patent, such a denial by the trial ■court was not unreasonable.

Ninth: Apparently, appellant tried to 'have the judgment amended to substitute “claims 1 to 12” for “claims 1 to 8,” its •argument being, there is no authority holding that a person can be the owner ■of certain claims of a patent and not of ■others. One case cited by appellee in answer to appellant’s motion to amend would, at least by implication, indicate that appellate courts will allow the question of validity to be directed to some and not other claims in a patent.

But in this suit, if appellee prewails, with claims 9 through 12 left valid, .just what does appellant have left? Would appellant be the owner of .% an invention ? We think not. The patent is invalid.

Finally, appellant argues that the 1953 experimental pipe repair clamp of Tel-ford Smith was not a part of the prior art, but merely an abandoned experiment.

It has been held that a prior unsuccessful experiment does not constitute invention. Pacific Contact Laboratories v. Solex Laboratories, 9 Cir. 1953, 209 F.2d 529 .

The trial court, however, did have a prior art invention before it sufficient to sustain its decision — the Lindsay patent. That was all it needed.

The judgment must be and is affirmed.

. (a) Binding 8 states, in part: “A pipe damp is made up of a split band * * * and some form of a damp * * * ” Appellant’s witness, John Joseph Connelly, a pipe foreman, testified that he referred to “the complete assembly” as a damp, and that he didn’t refer to just the lug part as being the clamp. (R. 703.) On cross-examination, appellant’s chief engineer, Frederick T. Newell, indicated that at the bottom of the main body portion of the Hoke patent lug was the “damping projection.” (R. 131.) (b) Finding 11 refers to the “art of damps,” and Finding 12 to the “damp art.” Telford Smith, president of appellee testified that he considered a clamp to be “something that is coming together — coming together with pressure and to be able to put something in between it, and either squeeze it or pull it together.” (R. 531.) A consulting engineer called by appellant, Charles H. Capen, when asked whether or not the term “clamping art” meant anything to him, replied: “No, I have seen pictures of damps used around the human heart, and I certainly wouldn’t go looking for that to solve a water works problem.” (R. 723.) (e) Finding 13, in part, refers to the 1939 Lindsay British Patent No. 512,406, which relates to “hook adapters” for bolts (the Lindsay patent). Finding 13, in part, states: “Lindsay calls attention in his specification to the advantage of adding the fingers or ‘axial projections’ to the clamp lugs, as he calls them * * * ” Appellant argues that there is “no reference” in the Lindsay specification to “clamp lugs.” (Appellant’s Brief, p. 10.) But the Finding 13 phrase “as he calls them” could refer either to the “axial projections” or to the clamp lugs. If it refers to the former, it was no mistake, since mark 1. on any of the Lindsay patent figures might be called an “axial projection” or a “finger.” (See Record, p. 863.) However, if this phrase refers to clamp lugs, it is not accurate since the Lindsay patent does not use the term “clamp lug.” In the Lindsay patent, the “axial projection.” extends from the hook portion. (See Record, p. 863.) These terms mentioned in (a), (b) and (c) above, are somewhat ambiguous, but, if the presumption in favor of the trial court’s findings is to be maintained, where one possible meaning is valid and not a mistake, the findings in question should not be held clearly erroneous, and we will not so hold them. (d) Finding 13 also states, in part: “ [Witnesses on both sides were in full agreement that regardless of names used, the Lindsay clamp lugs and the Hoke clamp lugs were part for part and function for function identical one to the other.” Appellee’s president, Telford Smith, when asked about the original Dresser “360” Repair Clamp bulletin (R. 839), describing a commercial model of the Hoke patent device, and the Lindapter F 900 bolt adapter bulletin (R. 840), describing several commercial models of the Lindsay patent device, stated that “there is not any difference” in either the functioning of one or the other, or the parts. (R. 344.) Frederick T. Newell, appellant’s chief engineer, in the course of cross-examination involving comparison of Exhibit No. 62, a sketch of the Hoke patent device (ft. 830), with Exhibit No. 64, a sketch of -the Lindsay patent device (R. 832) and with Exhibit No. 65, a sketch of the Lindapter F 900 (R. 833), when asked whether the parts are the same replied, “They are not.” (R. 133.) However, in the course of Newell’s ■cross-examination on these three exhibits, ■he testified that each exhibit shows a pair of lugs having a main body portion with holes through them through which a bolt extends, each shows a clamping projection at the bottom of the main body portion of the lug, a torque arm secured to the top of the main body portion, a bearing surface at the top of the main body portion of each lug on which the torque ■arm from the opposite lug bears, and •a bolt projecting through the holes in the main body portion so that by tightening the nut you can apply a clamping pressure. Further, he testified that as the nut is tightened, the clamping projections move •toward each other; as the clamping projection begins to encounter resistance, the further tightening of the bolt puts the reaction force of the torque arm on the opposing lug; if the torque arms were not there, as you tighten the nut, you would get the rolling of the lugs in toward each other; and with the rolling of the lugs in toward each other there would be a stress on the bolt, which would •cause it to take a bow shape. In the face of this testimony, it cannot be said that the “part for part and function for function” language of Finding 13 is clearly erroneous. There was ample •evidence (though disputed) to support the finding. (e) Finding 14 states in part: “The ■same clamp the Hoke patent described as supposedly novel with Hoke is found in the Lindsay 1939 patent * * * ” Again, there is ambiguity. If the word “clamp” means the whole patent device, then a comparison of the Hoke patent, with the Lindsay patent, will indicate a clear difference. But, if the word “clamp” means “clamping component,” then the discussion under the “part for part and function for function” part of Finding 13 applies here. This alternative meaning could not, therefore, on the state of the record, be deemed clearly erroneous. Again, where one of the possible meanings is logical and sound, the finding, or the part thereof in question, should not be held clearly erroneous. (f) Finding 16 states that the accused sliding finger pipe clamp, appellee’s Series 220 (the accused clamp), “bears a close kinship to the Lindsay sliding finger clamp as it is directly descended from the Lindsay clamp, which has been in the public domain since 1939.” Finding 16 suffers from the same ambiguity: Does “clamp” mean the whole device or the clamping component? If it means the whole device, a comparison of the Lindsay patent with a circular, drawings and photographs of the accused clamp would indicate an error. If it means clamping component, this same comparison would result in holding that Finding 16 is not clearly erroneous. We hold there was no clear error. (g) Finding 32 states, in part: “The 'accused device descends (a) from the Lindsay sliding finger clamp lugs * * * ” The same reasoning and result which we have followed heretofore should be applied to Finding 32, only here the possible ambiguity would be whether “clamp lugs” refers to the clamping component, or the clamping component plus the connecting component. (h) Finding 19 states, in part, that Tel-ford Smith, appellee’s president, in April 1953, decided to test the clamping power of the Lindsay sliding finger clamp; that he folded the ends of a split band back over the lugs and laid the ends under the sliding fingers to make a “bent back form of connection” to the lug; that the pulling power of the sliding finger lugs was so great it caused the ends of the split band to slide around the lugs; that Telford Smith was satisfied he had a clamp with much greater pulling power than the plain lug; that the problem to be solved was to find a production lino method for connecting the split band to each lug; that there is no dispute in the evidence about the making of this sliding finger clamp, “nor about its being tested successfully as described above * * # V Appellant argues that Finding 19 is in error in stating that this 1953 test clamp “was tested successfully,” and cites testimony to the effect that “the clamp failed by the band slipping off the lug without any measurement of pull being made,” and to the effect that the device was not hydrostatically tested, field tested or sold. (Appellant’s Brief, pp. 10-11.) In Finding 19, phrase “tested successfully” is modified by the phrase that follows it: “as described above.” And, the testing described is that of pulling power, alone. Therefore, the only question is, was there error in finding a successful pulling power test, in view of what the test wms, i. e., its limited nature? Telford Smith when asked whether he performed any tests on the 1953 device, replied: “The tests were to put it on the pipe and to pull it in and to see what —how much pressure it would hold, if it had tightenability; and it pulled the metal back and buckled it up, but I was quite satisfied that it was an effective way of tightening the band.” (R. 563.) And, on being asked how much it pulled, he answered: “Well, it took some muscle in the arm. I was kind of strong in those days, but it did take- considerable pressure, up to where I could realize that it was tightening very well.” (R. 563.) In view of this testimony, it is not reasonable for us to conclude that Finding 19 is clearly erroneous. (i) Finding 20 states: “It was obvious to Telford Smith in 1953, soon after having seen the Lindsay advertisement of the sliding finger clamp, to assemble the pipe damp of Exhibit 5 [the 1953 test clamp) by using the Lindsay sliding finger clamp and connecting to its clamping lugs a prior art split band by the prior art bent back form of connection.” Telford Smith testified that the first time he heard of fingers on lugs was in 1937 or 1938; that the next time was either the last part of 1952 or the early part of 1953 when he saw an illustration in a Lindapter circular showing a lug with fingers; and that in late 1952 or early 1953, “I was trying to find a way to get clamps to hold on those large diameters * * * we started working on clamps, and I suddenly got the thought of the fingers, of applying it * * * ” (R. 304-305.) This testimony supports the facts found, hence it cannot be said that an error has been made in finding of fact 20. But, does the word “obvious” make Finding 20 a conclusion of law, thereby stripping it of its presumption of correctness ? By Patent Act § 103 ( 35 U.S.C. § 103 (1952)), a patent may not be obtained if “the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains. * :¡í * The fact that the ultimate finding is a conclusion drawn from undisputed or established subsidiary facts does not change such a finding from one of fact to a conclusion of law. Being a finding of fact,, it is not to be set aside unless clearly erroneous. C. I. R. v. Spermacet Whaling & Shipping, 6 Cir. 1960, 281 F.2d 646 . (j) Finding 21 states, in part, that in September 1953, appellee had completed the fabrication of a “split band” tapping sleeve for the Dominguez Water Corporation of Long Beach, and that as this device was being tested, Telford Smith noticed that the lugs tilted as the bolts were being tightened “to damp the split band around the test pipe * * * ” Appellant’s chief engineer, Frederick T. Newell, in comparing the Dominguez split tapping sleeve with the Dresser 360 Repair Clamp, a commercial model of the Hoke patent device, testified that on the Dominguez tapping sleeve “we also find side-bolts, or tangential bolts, and as these-bolts are tightened, the two halves are-joined together. But the two halves-do not clamp against the conduit or pipe. So in the one case we have a device which damps the conduit and in the other we have a tapping sleeve which does not damp the pipe.” (R. 716.) This testimony is supported by a diagram opposite page 21 of appellee’s brief. This diagram indicates that the clamping action is not along the whole length of the conduit, but only at the ends of the split tapping sleeve, and along the line where the two halves come together. And, in the course of Telford Smith’s testimony on the Dominguez device, his references to the testing of the Dominguez device do not expressly indicate that the bolts were being tightened “to-damp the split band around the test pipe * * * ” (See Record, pp. 320, 330, 585-588.) If, therefore, this part of Finding 21 refers to clamping action along the whole length of the conduit, there has been an error in the finding, but if it means. clamping action at the ends of the split tapping sleeve, then it cannot be said to be clearly erroneous. Again, we are required to rely on the rule that where there is ambiguity in a finding, or a part of a finding in question, and one possible meaning is not an. error, that finding or part should not: be held clearly erroneous. We will not so hold it. (k) Finding 21 also states, in part: “Telford Smith directed Dunmire to add two Lindsay sliding fingers to each lug.” Telford Smith testified on this point: “Then I told them to put fingers on it * * * ” And, he indicated that on the Dominguez device there are fingers on opposite sides of the split coupling, each of which bears on the bearing surface of the other side. But, throughout his testimony on the Dominguez device, Telford Smith does not refer to the Dominguez fingers as “Lindsay sliding fingers.” (R. 319-336; 579-598.) A comparison of the Lindsay patent with appellee’s own blue print of the Dominguez device, indicates that the Lindsay fingers and the Dominguez fingers are not precisely similar. The Lindsay fingers slide over the bearing surface on the opposite lug; they are not welded to the surface of the band. The Dominguez fingers are welded to the band and do not slide across the bearing surface of the opposite lug. (See right hand portion of blueprint, Record, p. 893.) But both fingers and lugs are welded to sleeves, and while the method of attachment of the lugs differs, the result achieved is similar; the action sought by pressure the same. We hold the finding not clearly erroneous. (l) Finding 23 states, in part: “A tapping sleeve comprises a split band with lugs and bolts to draw it tight around a pipe.” Appellant urges that this part of Finding 23 is subject to the same error that was pointed to in the Finding 21 phrase: “to clamp the split band around the test pipe * * * ” The same testimony, reasons, and conclusion discussed as to a similar part of Finding 21 apply with equal force to this part of Finding 23. We hold it is not, therefore, clearly erroneous. (m) Finding 23 also states, in part: “The function of the split sleeve is to support whatever form of gasketing is placed between the pipe and the band.” Telford Smith, referring to the Dominguez Water Corporation testified: “The customer asked us if we could supply a fitting which is to be clamped to the pipe so they could make a tap.” (R. 580.) And, “what they [Dominguez Water Corporation] needed was a fitting that could be clamped on the pipe and would seal all around and would have provisions for an outlet.” (R. 580.) If “the function” refers to the only function, there would be error, but, another possible meaning is one of two or more functions, which based on the testimony quoted would not be a clearly erroneous meaning. (n) Finding 24 states: “The testimony is uncontradicted that the Dominguez sliding fingers function identically with the Hoke sliding fingers in preventing the lugs from rolling.” Telford Smith indicated that the Dominguez fingers were welded on one side of the split sleeve, and slid over onto the other side. This is contradictory testimony; the Hoke fingers are not welded onto the sleeve, and do not bear on the opposite sleeve. (See Hoke patent, Record, p. 847.) But see our discussion with respect to Finding 21. The same reasoning here applies. (o) Finding 25, in referring to the Dominguez device, states, in part: “[T]he lug is made up of a portion of the split band, the gusset, the crossbar with the bolt holes, and the sliding fingers.” Frederick T. Newell, appellant’s chief engineer, in referring to the Hoke patent device, indicated that the term “lug” could be used interchangeably with the term “bar.” Telford Smith, on cross-examination, stated that the gussets supported the slide bar, and that “the combination of all of them would form a lug”; and in referring to a diagram showing the Dominguez fingers, he said: “It shows here — this would be a lug of metal on the opposing side welded in place which would act as a finger or be a finger.” It cannot be said that this part of Finding 25 is clearly erroneous. (p) Finding 26 states, in part: “The claims in suit describe exactly the structure of the split tapping sleeve as sold to Dominguez in 1953.” A comparison of the Hoke patent with appellee’s blueprint of the Dominguez split tapping sleeve indicates some minor differences: As opposed to the Dominguez device, the Hoke patent device does not have fingers welded to the band, does not have a bar welded to the band, does not have gussets supporting the bar-, and does not have its fingers sliding onto the surface of the opposite half of the band. We find the quoted portion of Finding 26 to be erroneous, but we cannot hold this error so prejudicial as to require reversal. (q) Finding 27, in referring to the Dresser 360, the commercial model of the Hoke patent device states in part: “This #360 device is a combination of the Lindsay sliding fingers and the cavity-wedge connection Smith-Blair had on the market since 1954 as its 2F line of clamps.” Appellant argues that this finding is “in error in stating that Hoke copied the cavity wedge connection of Plaintiff’s 2-F line of clamps.” (Appellant’s Brief, p. 12.) Howard Hoke, in testifying about the disadvantages of appellee’s 2-F clamp, demonstrated that he had studied the 2-F; but, he did not mention the 2-F connecting component. On cross-examination, Telford Smith, in talking about the Dresser 360 pipe repair clamp, said: “I could obviously see that the Lindapter clamp was being used and that an attempt was being made to copy the wedge that we were using in our 2 F½ clamp * * * ” (R. 536-537.) A comparison of the Hoke patent with an example of the 2-F connection in one of appellee’s patents indicates that the two are not dissimilar. Based on this evidence, it cannot be said that this part of Finding 27 is clearly erroneous. (r) Finding 33 states, in part, that Hoke solved the problem of the connecting component “the easy way by copying the essentials of plaintiff’s 2 F form of connection.” What was set forth in (q), supra, with reference to that part of Finding 27 in question is applicable here, and this part of Finding 33 cannot be held clearly erroneous. (s) Finding 34 states: “Structurally speaking, all that the accused device amounts to, so far as the structure called for by the five elements of the Hoke claims in suit [split sleeve, lugs, bolts, bearing surface, and finger (See Finding 37) ] are concerned, is the Lindsay sliding finger clamp with the prior art split band connected to it.” This is true. A comparison of the Lindsay patent with a circular, drawings and photographs of the accused clamp would indicate no error on the part of the trial court. This finding, therefore, cannot be held clearly erroneous. (t) Appellant contends that Finding 35 “is an argumentative ‘characterization’ or conclusion and is not a statement of fact,” and Findings 36 through 41 are “argumentative and are in the nature of conclusions rather than facts.” The meaning of argumentative that appellant apparently had in mind is that the statement of fact is based on an inference. If the inference arrived at is an ultimate finding and a conclusion, the fact that the ultimate finding is a conclusion drawn from undisputed, or established subsidiary facts does not change such a finding from one of fact to a conclusion of law; and, being a finding of fact, it is not to be set aside unless clearly erroneous. C. I. R. v. Spermacet Whaling & Shipping, supra. In Spermacet, the finding was that petitioner was not engaged in any substantial, regular or continuous ordinary business activity in the United States; and the conclusion of the lower court was that the taxpayer was not “engaged in trade or business within the United States” within the meaning of Section 231(b), Internal Revenue Code of 1939. Appellant cites National Lead Co. v. Western Lead Products, 9 Cir. 1961, 291 F.2d 447 , for the proposition that findings of fact “better denominated conclusions of law” are subject to review without the restrictions of Rule 52(a) Federal Rules of Civil Procedure. In National Lead the findings included these: “ ‘7. Claims 1 and 2 * * * each defines an assemblage of process steps, all of which were old in the prior art * * * and the subject matter of each as a whole at the time the alleged invention was made would have been obvious * * * any person having ordinary skill in the art * * *. “ ‘8. Claims 1 and 2 * * • is each wanting in invention. “ ‘9. Claim 2 of said Letters Patent No. 2,235,487, if valid, has been infringed by the defendant * * * “ ‘10. The plaintiff * * * was not guilty of laches * * ” The court pointed out that Findings 7 and 8 were little more than a paraphrase of 35 U.S.C. § 103 , and a legal conclusion of want of invention. The court vacated and set aside the judgment of the district court, and remanded the cause to the district court to make appropriate findings of fact on such matters as what the prior art was, what plaintiff did to improve upon it, etc. And the court, even in the face of such wholly inadequate findings did not hold that findings “better denominated conclusions” should be subject to review without the restrictions of Rule 52(a). Are Findings 35 through 41 clearly erroneous ? Finding 35 states: “A corollary structural characterization of the accused clamp in relation to the structure of Hoke’s claims in suit is that it is the plain prior art clamping lugs and bolt with the sliding fingers of Lindsay transferred from the Lindsay lugs to the plain prior art lugs where the fingers function to prevent the lugs from rolling, exactly as they functioned in Lindsay.” A comparison of the Lindsay patent with a circular, drawings and photographs of the accused damp, and with a circular of an early pipe repair clamp would not indicate any error on the part of the trial court. This finding, therefore, cannot be held clearly erroneous. Finding 36 states: “The evidence is clear that the accused device has copied from the public domain the Lindsay sliding finger clamp lugs and has connected to its lugs the split band that formerly was fastened to clamp lugs which had no sliding fingers. This is just an aggregation or congregation of prior art elements.” The comparison and conclusion in the discussion under Finding 35 is applicable to Finding 36; it, too, is not clearly erroneous. Finding 37 states: “The five elements in the Hoke claims in suit do not perform any additional or different function in combination than they perform out of it. The split sleeve tightens the gasketing around the pipe, as a split sleeve has always done. The lugs pull the split sleeve tight on the pipe, as lugs have always done. The bolts tighten the lugs, as bolts have always done. The bearing surface on one lug and the finger on the other lug which slides on it prevents the lugs from rolling or the bolt from bending, as fingers have done since 1939 in the Lindsay prior art sliding finger clamp lugs.” A comparison of the Hoke patent with a circular of an early pipe repair clamp, and with the Lindsay patent, would not indicate that the trial court is in error. This finding, therefore, cannot be held clearly erroneous. Finding 38 states: "The operative result of the five elements combined in the Hoke claims in suit is just the sum of its parts, with each old part doing what it did in the prior art.” The comparison and conclusion in the discussion of Finding 37 is applicable to Finding 38; it, too, is not clearly erroneous. Finding 39 states: “There is no unusual or surprising consequence of the unification of the five elements in the Hoke claims in suit. Hoke only claimed to have added the sliding fingers to the plain lug of the prior art to prevent the lugs from rolling, but the Lindsay clamp from which the Hoke device descended had the same operative characteristic, namely, its lugs did not roll.” A comparison of the Lindsay patent with the Hoke patent, and a review of testimony given by Howard Hoke, in which he indicated the No. 1 problem his patent in suit solved was keeping the lugs from rolling in, would indicate that this finding is not clearly erroneous. Finding 40 states: “The Lindsay sliding finger clamp is in the clamp art, as is the pipe clamp described in the Hoke patent in suit. The dictionary meaning of clamp includes in its terms the Lindsay sliding finger clamp and the prior art pipe clamps. The testimony of experts on both sides is to the effect that if one were going to tighten a band around a pipe, the obvious thing to do would be to get a clamp to pull it together. Plaintiff keeps in its engineering offices a file of catalogs including clamp catalogs. The Lindsay sales literature makes an appeal to engineers to buy the sliding finger clamp as ‘a general purpose clamp useful to engineering welders, joiners and others.’ Dunmire, prior to 1955, ordered for plaintiff’s engineering file the Lindsay catalog on clamps because they applied to the pipe clamp devices plaintiff was manufacturing. When in 1953 Telford Smith saw the Lindsay sliding finger clamp advertised for clamping the flanges on pipe, it was obvious to him to turn the clamp around to where it was clamping the flanges on a band around the same piece of pipe.” To the part of Finding 40 dealing with “clamp art,” what was said with reference to Findings 11 and 12 is applicable. If “clamp art” refers to the art of clamping components or split bands, then the court committed no error. Mr. Charles H. Capen, a consulting engineer called by appellant, and Paul G. Dunmire, a consulting engineer called by appellee, indicated that they would look to a clamp to tighten a band around a pipe. Paul G. Dunmire did indicate he saw a picture of the Lindapter in a publication, and “recognizing it as a clamping device of the type that we were utilizing, we sent for a catalog * * * ” The Lindsay sales literature does make an appeal to engineers to buy the Bind-apter as “a general purpose clamp.” And, as for what was “obvious” to Tel-ford Smith in 1953, what was discussed under Finding 20 is equally applicable here. Finding 40, therefore, cannot be held ■dearly erroneous. After a careful consideration of all alleged errors in findings of fact, we con-dude, with one exception (Finding 26) that no error existed. But while we cannot say the claims in suit “exactly” describe the structure of the tapping sleeve as sold to Dominguez, this does not render the court’s findings of invalidity on other grounds invalid, and is not sufficient to constitute grounds for reversal. . Telford Smith was an officer and guid-' ing light of the appellee corporation. . While not word for word, the court found the two patents disclosed identical parts and identical functions in the two finger respective lugs used in the two patents (Finding 13). . E.g., “Moreover, since tlie A. & P. case only stands for the proposition that there must be a combination as distinguished from a mere aggregation, Steams sets out the real test, namely, that the combination itself must be patentable.” (Appellant’s Reply Brief, p. 18.) . In view of these considerations, it is unnecessary to discuss two other conclusions of the district court: That the claims of the Hoke patent in suit are invalid under § 102(b) of the Patent Act; and that these claims are invalid under the Supreme Court’s rule in Knapp v. Morss, 1893, 150 U.S. 221 , 14 S.Ct. 81 , 37 L.Ed. 1059 . . In Brown & Sharpe Mfg. Co. v. Kar Engineering Co., 1 Cir. 1946, 154 F.2d 48 , a judgment dismissing a complaint in an action brought for the infringement of claims 1, 5, 7 and 14 of a certain patent was reversed; the trial court had found these claims invalid, but the appellate court found them valid.