Appellants, Glass Fibers, Inc., as assignee* and Everett J. Cook as claimant-inventor, by application, Serial No. 628,781, sought a patent for an 'Apparatus for Producing Glass Fibers.' The Commissioner of Patents allowed claims 3, 13, 14, 15 and 21 but disallowed claims 1, 2, 9, 23 and 24 as unpatentable over the prior art as disclosed in two Barnard patents No. 2,229,489 and No. 2,294,266 and the Staelin patent, No. 2,335,135. The two Barnard patents are substantially identical, and for our purposes may be treated as one. The Patent Office Board of Appeals, in sustaining the Examiner's rejection, treated and described appellants' claim 1 as 'representative.' The District judge read claim 1 into the record as follows:

'In a heating apparatus for heating and melting glass the combination of, a substantially cylindrical metal crucible having one wall thereof containing a plurality of substantially concentrically circularly arranged apertures, a substantially circularly arranged high frequency induction coil substantially equidistantly encircling said crucible at the glass melting level in the crucible, said crucible having means cooperating therewith to establish a pressure differential upon a body of glass within the crucible to cause the glass to exude through the said apertures at a uniform rate, and an opening in said crucible through which glass stock is adapted to be fed.'

It was established that allowed claim 3 materially differed from rejected claim 1 only because the former added the words '. . . and means for controlling the feeding of the glass stock into said crucible in response to the level of the molten glass in the crucible.' The last quoted language spelled invention, Patent Office counsel assured the court, although otherwise the structure defined in allowed claim 3 was identical with that in rejected claim 1. Colloquy among respective counsel and the trial judge developed conclusively that the Patent Office did not apply to claim 3 either the Barnard or Staelin patents. Notwithstanding, Patent Office Counsel said 'we are satisfied that claim 3 is patentable. We are satisfied that claim 1 is not.' Appellants' counsel for the record said: 'I will put it this way: If we could get from the Patent Office, or from Your Honor, claim 1, I'd forget about claims 3 and all claims that they have allowed.' The trial went forward on that representation, and the evidence, exhibits and argument were focussed upon the patentability of claim 1.

We, too, will treat claim 1 as 'representative' and discuss the case on that basis, adopting appellee's statement of the question 'whether the structure set forth in claims 1, 2, 9, 23 and 24 of the appellants' application differs from the structure shown in the Barnard and Staelin patents in such a way 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.'

The District judge found that 'claims 1, 2 and 9 do not define invention over the disclosures of the Barnard patents,' and that 'claims 23 and 24 do not define invention over the Barnard patents in view of the Staelin teaching.' He concluded that appellants must be denied a patent containing any of the mentioned claims and that the complaint must be dismissed.

'The principal reference relied upon by both the Patent Office and the District Court is the device shown in the Barnard patents. This point should be emphasized since a substantial portion of the appellants' brief is devoted to a discussion of the [97 U.S.App.D.C. 71] insufficiency of Staelin as an anticipation. Actually the latter patent was used merely to show that attenuating devices and circular crucibles are old in the art.' (Emphasis supplied.)

R. H. Barnard, president of Glass Fibers, Inc., 'grew up in a glass factory,' where he first conceived the idea that glass shapes could be extruded under artificial pressure. The Barnard patents basically embodies that belief. Before leaving Owens-Illinois Glass Company with which he had served twenty-two years, Barnard offered his development to that company which turned it down as impractical. Barnard in 1944 then went to Toledo University and hired appellant Cook, an electronics expert, to work with him in perfecting the Barnard idea of extruding glass. He was still trying to 'work glass under pressure.' Barnard then was using a square induction heating coil parallel to the contour of and outside a square pot to heat and melt glass, with additional heat possibly to be supplied by two electrodes inside the pot, using the glass itself as a resistance element. Under internal pressure, Barnard planned to force out, to extrude, the molten glass through orifices drilled into the cooled bottom plate in such fashion that streams of glass emanated as heavy, thick, non-uniform fibers, producing 'glass wool' for use as heat or sound insulation. Such short-fiber glass, uneven in thickness and in length, resulted in a matted output, due in large part to uneven heating in the rectangular pot with consequent uneven flow of glass through the unevenly cooled bottom plate. Efforts to produce smaller diameter fiber from Barnard's apparatus would encounter difficulties due to (a) in part, inability to drill smaller holes and (b) inability in any event, to drill such holes without weakening or buckling the bottom plate under the high pressure required to extrude the fiber. The basic elements of the first Barnard patent, No. 2,229,489, were disclosed in the second, No. 2,294,266, with the chief difference, immaterial here, that the second provided a different method of feeding glass into the pot and certain controls were omitted. In sum, Barnard's patents, issued January 21, 1941; and August 25, 1942, both disclosed inductance outside heating, both included electrodes in the center of the pot to augment internal heating, both employed a rectangular pot, and both relied upon fluid pressure on the glass as it was melted to extrude glass fiber through the flat, perforated, cooled bottom plate. Barnard with this apparatus sought to induce first the General Electric Company and then the Westinghouse Corporation to run experiments utilizing his patents, but, not receiving assurance that success could be expected, he declined to pay the substantial charges they required for their research. Then he engaged Cook.

The Staelin patent issued November 23, 1943. Staelin, unlike Barnard, was not trying to produce wool, he sought glass fibers. He disclosed a substantially circular two-pot arrangement, one pot being above and within the other. The lower part of the latter is tapered into conical shape in a series of circular planes, each of successively smaller diameter, and with each receding level perforated with apertures arranged in concentric rows. Each such perforation penetrated a downwardly-projecting 'nipple' through which the melted glass could be passed, the nipples and the out-flowing glass being cooled by a diffused current of gas or air. Staelin used his inner crucible as a heating resistor, passing ordinary, not high frequency (as in Cook) current thereto by means of conductors. He makes no mention of induction heating, indeed, he relied solely upon resistance heating, so applied that his glass flowed from the inner pot to the cooler outer container, and thereafter through his orifice plate, the cone shaped, nipple bearing appendage. Staelin's glass, cooled by jets of gas or air, had irregular temperatures which yielded a fibrous product, which could, however, be drawn off mechanically. Staelin clearly believed that a flat aperture plate is impractical, but he realized he must seek to cool his glass evenly at [97 U.S.App.D.C. 72] each nipple in an effort to avoid relatively different temperatures. To the extent that he fails, he is unable to 'maintain proper resistance for satisfactory attenuation of the fibers,' as he, expressly, says. In any event, Staelin's teaching differs markedly from Cook's, and appellee's counsel tells us Staelin has been cited 'merely to show that attenuating devices and circular crucibles are old in the art.'

Substantially, then, when Barnard engaged Cook, the Barnard patents were believed to be impractical and had been rejected by Barnard's company, one of the largest in the industry, were deemed unworthy of risk by General Electric and Westinghouse, and were later abandoned by their owner.

As Cook went forward, he could have perceived from Barnard's teaching that a rectangular pot surrounded by a heating coil parallel to its surfaces would melt glass, but with uneven temperatures due to hot spots in the corners. He could have tried, as Barnard was willing to try, to adduce supplemental heat by the use of electrodes, but he still produced unequal heat because of hot spots in the center of the pot. He could see that Barnard extruded glass under pressure through straight line apertures from which the glass fibers, cooled as they fell in batches, produced, not yarn, or fibers capable of yielding yarn, but glass wool. And that was what Barnard said he was trying to make, wool.

Cook could have perceived from Staelin that a round pot with resistance heating might meet some of Barnard's heating problems, but he still had greater heat, uneven heat, in the center. He could have realized that Staelin's conical, recessed-plane, discharge apertures inevitably produced unequally heated glass streams, for the outer rows were successively cooler than the center, and the apparatus would fail to yield filaments of uniform diameters. He could have been that Staelin made clear he believed that a flat orifice plate with a multiplicity of rows of apertures was not possible, although he had long been associated with developments in the industry through his service with the Owens-Illinois Glass Company and the Owens-Corning Company.

There must have been many experts and engineers in the glass industry like Barnard and Staelin, but the devices developed and the principles taught by the two latter were the only references here. When we consider what Cook did, and attempt to appraise it, not only ordinary skill in the art, but the best there was had failed to disclose a structure the subject matter of which, taken as a whole, was obvious. A mere exercise of the skill of the calling, read against Barnard's teachings, might have taught Staelin to go to a circular pot and then to embody Staelin's additional features, but there he stopped. His expertness failed to suggest to him a further 'advance plainly indicated by the prior art.'1 If there were to be an addition to the sum of useful knowledge, there had to be creative talent, uncommon talent, yielding a structure not wanting in unusual or surprising consequences, or where 'the whole in some way exceeds the sum of its parts . . ..'2 Therein is the standard of invention to be met.

But it can be done, and it often is done, and invention is not to be denied simply because, when 'Viewed after the event, the means . . . (may) seem simple and such as should have been obvious to those who worked in the field . . ..'3 And here the means had been neither simple nor obvious to the best in the industry. Cook, like prehistoric man knew that heat would melt viscous substances. Like the Egyptians of old, he knew that a cooking pot could be round or cylindrical. Like countless physicists [97 U.S.App.D.C. 73] he knew that a molten liquid was subject to atmospheric pressure as well as that produced by its own head, and that such pressure could be increased mechanically. He could have learned from electronic developments during and since the war that high frequency heat induced by an induction coil follows a circle, and he would so install his coil equidistantly at all points to surround his cylindrical crucible with even heat. He must have known that heat evenly applied to the pot, having melted the glass and using the glass itself as a conductor, would achieve uniform internal heat and, in turn, uniform viscosity of the glass. But, granting the importance of each of these elements by itself, he still had to get his glass from the pot so as to produce commercial continuous-length fiber of uniform diameters. Here he required, first, the same amount of evenly heated glass flowing through each of the orifices, and second, that the glass be drawn away from all orifices at the same rate of speed. By his pressure differential above the molten glass plus the head of glass itself, the uniformly heated, uniformly viscous body of molten glass at a uniform rate exuded through the concentrically arranged apertures in the flat bottom plate, not cooled as in Barnard and Staelin, but heated alike at all points by the glass above it. Appearing, first, as a bead or a bubble, and then as each bead uniformly falls away from each orifice, as a continuous stream 60 thousandths of an inch in diameter, the glass filaments may be attenuated following adaptation by the Cook method patent, No. 2,495,956, founded upon and devised for use with the subject structure, in continuous, uniform diameters at a uniform rate of more than 2 miles per minute. The microscopic diameters of the filaments exhibited in court were of the order of 2 ten-thousandths of an inch, with fiber containing 200 filaments of continuous lengths up to 90,000 feet per pound, with a variance of a minute 10 per cent as the maximum permitted. While Staelin can produce glass fibers, uniformity is lacking. While Cook can produce wool, as does Barnard, Barnard cannot possibly produce yarn. On the other hand, yarn made by Glass Fibers, Inc., is in demand by industry for ignition cables, aircraft wire and cable, heating pads and comparable uses, with $254,000 volume in 1949, $1,013,000 in 1950 and $1,434,000 in 1951.4 'This is a 24 hours a day, seven-day-a-week, 365 days-of-the-year type of operation,' as was testified, and the commercial yarn was produced by apparatus responding to and measured by the Cook claim 1.

At about this point in the testimony the record discloses that the trial judge said (J.A. 154): 'In other words, as I said yesterday,5 I am convinced that the [97 U.S.App.D.C. 74] result of the Cook application, carried out into actual practice, is much better, in the absence of evidence to the contrary, than any other type of manufacture of this type of yarn. So this evidence, while it is perfectly all right for the record, is not helpful to me, because you have convinced me of that aspect of the matter.' Plaintiffs' counsel accepted, as he well might in the light of the record which has been made, the assurances of conviction on the part of the trial judge who continued his observations thus (J.A. 154):

'The Court: I want to get back to the basic issue. The basic issue is whether or not a person skilled in the art, such as Dr. Plummer is, would have readily suggested to him the innovations made in the Cook application by reading Barnard and Staelin. That's the point.

'And, of course, I may say parenthetically, the only way I can determine that is to read the transcript and go over the patents, individually.' (Emphasis supplied.)

Who is 'skilled in the art'? The trial judge concluded he must be one 'such as Dr. Plummer is'6 but Dr. Plummer had risen to the post of general manager of Glass Fibers, Inc., practicing the Cook disclosure. He was far more than a person having 'ordinary skill in the art.' Barnard himself, with all his skill, found no way from his own teachings to achieve the Cook apparatus and result. Staelin devised no such improvements over Barnard, indeed Staelin's teaching was sufficiently different that his apparatus was found patentable. The plain fact is that neither Cook nor any other person possessed of ordinary skill in the art found the Barnard and Staelin disclosures to render obvious the steps taken by Cook. Even men of extraordinary skill found no obvious next step or series of steps pointing to Cook's combination. It is clear that the trial judge thought of Dr. Plummer as one who 'would have readily suggested to him the innovations made in the Cook application by reading Barnard and Staelin,'7 but Dr. Plummer was taught by Cook. It may be noted, Dr. Plummer, despite his qualifications and skill and his years in the industry-- even working with Staelin-- was not the one who conceived the subject device, it was Cook who did. And the record is barren of fact, circumstance, or detailed findings to demonstrate that either Barnard or Staelin made such disclosure that Cook's claim 1 reflected their teaching. The finding that Cook's claim 1 is unpatentable over the prior art of Barnard and Staelin stems from the application by the trial judge of a higher standard than the law demands with the result that his conclusion lacks substantial support in the evidence. It was not a case of what should have been-- but was not-- obvious to Dr. Plummer, and enough has been said to demonstrate that there was no 'obviousness' to the other experts.

We would not 'lightly' reverse the able trial judge,8 but we, too, have read the transcript and examined the exhibits. While there is evidence upon which, in part, the findings can be supported, on the entire evidence we are [97 U.S.App.D.C. 75] left with the definite and firm conviction that a mistake has been committed.9 On the record made and with the exhibits before us, we are permitted a 'subjective opinion or formulation of a judgment,' which when so predicated, makes clear that the 'combination and improvement admittedly brought about are the product of creative talent' which lifts the Cook claim 1 to the level of patentability over and above the prior art cited.10 The trial judge failed to recognize that Cook's combination of parts, working in cylindrical concentric symmetry in the application of heat at the glass melting level of the crucible, when operated with the other features claimed, each contributing its function to the production of the end result, caused 'the glass to exude through the apertures at a uniform rate.' This simultaneous cooperation of all elements was the gist, the heart of Cook's teaching, already in allowed claim 3 (recognized by the Patent Office as invention when coupled with the control),11 and without it, there could be no uniform filaments and no continuous length fibers of the type desired by the trade. Unless all of the elements were so cooperatively combined as to produce the desired result, the apparatus failed of its purpose. Thus the claim as a whole outlined a limitation upon and not an extension of the monopoly of the invention. The combination simply had to work, and operate correctly, or it would not work at all. Appellants set forth in their claim a series of steps which, as cooperatively combined, 'were not taught by the prior art and which do give an unobvious result.'12 While a round pot or an induction coil may be old or well-known, or a conical aperture discharge device may yield fibrous glass of a certain type, it is not correct to say because Barnard produces wool or Staelin produces fibrous glass, that either makes Cook unpatentable over the prior art.

'It is plain from the evidence, and from the very fact that it was not sooner adopted and used, that it did not, for years, occur in this light to even the most skilful persons. It may have been under their very eyes, they may almost be said to have stumbled over it; but they certainly failed to see it, to estimate its value, and to bring it into notice . . .. Now that it has succeeded, it may seem very plain to any one that he could have done it as well. this is often the case with inventions of the greatest merit. It may be laid down as a general rule, though perhaps not an invariable one, that if a new combination and arrangement of known elements produce a new and beneficial result, never attained before, it is evidence of invention. It was certainly a new and useful result to make a loom produce fifty yards a day when it never before had produced more than forty; and we think that the elements by which this was effected, even if those elements were separately known before, was invention sufficient to form the basis of a patent.'13

Here the Cook apparatus produced glass fiber of minute diameter, uniform throughout its entire length, at the rate of more than 2 miles per minute. 'It may be safely said that if those skilled in the mechanical arts are working in a given field, and have failed, after repeated [97 U.S.App.D.C. 76] efforts, to discover a certain new and useful improvement, that he who first makes the discovery has done more than make the obvious improvement which would suggest itself to a mechanic skilled in the art, and is entitled to protection as an inventor.'14

We are bound first to look into the prior art to ascertain what may be the real merit of the invention and to decide whether or not it has substantially advanced the art. 'If it has done so, then the court is liberal in its construction of the patent, to secure to the inventor the reward he deserves.'15 As did Elbel, Cook made a 'very useful discovery which has substantially advanced the art.' Engaged professionally by Barnard himself, to perfect Barnard's patents, he abandoned them, as did Barnard, and came up with an apparatus almost entirely different, concededly and unquestionably yielding a totally different result. 'As in our view the court could properly under the uncontradicted evidence have made no finding except one which reflected the facts which we have stated, it would be idle to return the case for further findings.'16 The trial court should have authorized the issuance of a patent for claim 1.

Without ourselves attempting either definition or rule, with Justice Jackson we agree we must be able to see that 'the whole in some way exceeds the sum of its parts . . ..'17 The 'whole' of Cook exceeds all expectations and yields the 'surprising consequences' which even the experts failed to find 'obvious.' Here was invention.

The case will be remanded for the entry of a judgment authorizing the Commissioner of Patents to allow claim 1 and for further proceedings in conformity with this opinion.

Reversed.

BAZELON, Circuit Judge (dissenting).

My brethren of the majority reverse the judgment below because 'on the entire evidence (they) are left with the definite and firm conviction that a mistake has been committed.'1 I do not share their conviction even though, as they correctly state, 'we are permitted a 'subjective opinion or formulation of a judgment."2 I would therefore affirm.