From a decree of the District Court holding valid and infringed two patents relating to devices for aerating water, particularly adapted for use in household faucets, this appeal was taken. The patents are 2,210,846, issued to Elie Aghnides, August 6, 1940, and 2,316,832 for an improvement thereon, issued to him April 29, 1943. The suit was originally filed by the patentee, but later *720Chase Brass & Copper Company, Inc., holder of a license from him, took an assignment of the patents and, by leave of Court and with the consent of the defendant, became an additional party-plaintiff.
The purpose of the patented devices is to obtain a coherent, aerated stream of water that will not splash as it emerges from a domestic faucet. The device is attached to a kitchen water tap and consists essentially of a chamber into which both water under ordinary pressure and air are delivered and intimately mixed so that the stream issuing therefrom is a mixture of water and entrained bubbles. Because of the great volume of air in the mixture, the water is soft and produces a large volume of suds when mixed with soap. The result is effected by projecting water through small apertures in a diaphragm at the upper or inlet end of the chamber, thereby breaking the water into multiple jets as it enters into the chamber. Air is drawn into the chamber through openings in the side, and it is mixed with the water therein as a result of the turbulence of the water as it strikes resistance elements in the chamber preventing the free through-flow of the water. At the outlet of the chamber a fine screen or screens are located. The combination is typically described in the first claim of the basic patent ’846, which is quoted in the margin.1
The improvement patent, ’832, combines the basic elements of the above with a second perforated diaphragm or restricted orifice, or both, at the upstream end of the aerator, to obtain a more perfect stream of soft, splashless, bubbly water.
The District Judge held both patents in suit valid and infringed, in spite of the fact that the first patent, '846, had previously been declared invalid by the United States Court of Appeals for the Seventh Circuit because anticipated by the Felten English patent. Aghnides v. Goodrie, 210 F.2d 859. The District Judge’s opinion observed that in approaching his task he felt constrained to follow the Seventh Circuit if the evidence warranted it. Expressing proper deference for the opinion of that Court he, however, found himself unable to agree with it. As the Court of Appeals itself declared in Aghnides v. Holden, 226 F.2d 949, its Goodrie decision is not res adjudicata, even in the same Circuit, since there was no privity between the defendants in the two actions.
After a long trial and many experiments conducted in his presence in the courtroom, District Judge Hayes concluded that Felten operated on a different principle and did not achieve the result of Aghnides. One after another of the observations made in the Goodrie opinion, suggesting similarities between Felten and Aghnides, were fully answered by highly persuasive visual evidence to the contrary.
The same Rule of Practice (Rule 52 (a), Fed.Rules Civ.Proc. 28 U.S.C.A.) upon which our brethren in the Seventh Circuit relied in upholding the findings of their District Judge as not clearly erroneous would warrant our putting equal reliance in the presumptive correctness of District Judge Hayes’ contrary conclusions. However, to avoid unnecessary conflict between the Circuits, we deemed it our duty to address our*721selves independently and with special care to an examination of the District Court’s findings. Of course, our review is of the record made in the District Court below, which is different from the Goodrie record, and appears to be more detailed and comprehensive.
Turning more particularly to the ’846 patent, it may be noted that in the instant case, further supporting this presumption of correctness is the provision of the Patent Act (35 U.S.C.A. § 282), which puts the burden of establishing invalidity upon the party asserting it; and the usual presumption of validity is strengthened by the history of this patent in the Patent Office. It was granted only after considerable controversy and after review by the Board of Appeals, where the references relied upon by the defendant were all closely considered by administrative officials possessing special competence in the field.
It is noteworthy that 21,000,000 units embodying the patented device have been sold in this country, and licenses have been taken by large concerns like Chase Brass & Copper Co., Firestone Tire & Rubber Co., and Chicago Faucet Co. Chase Brass alone had, up to the time of the trial, paid royalties on more than 16,000,000 of its “Spring-Flo” units made according to the Aghnides teaching. Before taking the license, it had conducted an extensive search of the art and made unsuccessful attempts to attain the results of the Aghnides invention by other means. By the payment of substantial royalties, such business concerns attest their opinion of the merit of the patent. While it is true that commercial success does not determine either invention or the validity of a patent, nevertheless, in a doubtful case it may be a convincing factor. Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., 340 U.S. 147, 153, 71 S.Ct. 127, 95 L.Ed. 162; Textile Machine Works v. Louis Hirsch Textile Machines, Inc., 302 U.S. 490, 498, 58 S.Ct. 291, 82 L.Ed. 382; Smith v. Hall, 301 U.S. 216, 233, 57 S.Ct. 711, 81 L.Ed. 1049. Before this invention, no water aerator that could be used in connection with a household faucet had been successfully marketed. The wide acceptance and use of this item, which has become standard equipment in the plumbing supply industry and is used by leading faucet manufacturers and has been installed in millions of homes, are circumstances not without significance.
Like the District Court, we are unable to agree that this patent was anticipated by the Felten British patent No. 6012 of 1909, or the corresponding United States patent No. 1,090,044, issued to Fuss. The drawings of the Felten patent have a superficial resemblance to the patent in suit, but on closer examination Felten will be found to be different. Felten operates on a different principle, is designed to achieve a different result, and has never been utilized, as has the plaintiffs’ patent, to supply the householder with a soft, coherent, splashless stream of aerated water. The purpose of the Felten patent was to add ozone to water to sterilize it, and it is not shown that it was ever actually used even for this purpose. Judge Hayes had the benefit of witnessing the operation of demonstration equipment—a tank with faucets for the attachment of aerators, with a pump for circulating water from the tanks through the faucets, and with a pressure gauge and a vacuum gauge. The testimony is singularly free from the confusions which often creep into the record of courtroom demonstrations, and is clear and understandable. With respect to the Felten patent, the Court found (Finding XXVII (f) ):
“This device is also designed to secure a vacuum. Models thereof were demonstrated to the Court by plaintiff from which Felten’s construction, mode of operation and results were made evident. Felten’s device does not produce the fully aerated, soft, stream characteristic of Aghnides and does not operate on the same principle or use the same combination of elements. * * *. Felten is so constructed that the energy of the water is spent *722in creating a vacuum and not in securing aeration. Demonstration showed that it adds less than 5/10 of a cubic inch of air to 1 cubic inch of water whereas Aghnides’ aerator adds 6 cubic inches or 12 times as much.
“Not only was it demonstrated that Felten’s result is different but that he secures his result by different elements and by a different mode of operation. The patent states that the result is secured by means of a ‘conically widening chamber e, the widening of which produces in a well known manner the suction of the gas from the perforation d owing to the flow of the liquid from the perforation c into the widened chamber e’. (Felten Patent, p. 1, line 28). This has no counterpart in Aghnides. Felten, to secure his result, must keep the chamber flooded—the exact opposite of Aghnides. This was demonstrated by a transparent model (Exh. 44) which made it possible to see that the widening chamber was at all times flooded and that there were no individual jets striking Felten’s baskets. The flooding of the chamber was also demonstrated by drilling a hole into a widening chamber whereupon water squirted out and, when the air through the inlet provided by Felten above the chamber was cut off no air entered any place and the device became inoperative even as a vacuum pump. Felten’s necessity for flooding was also demonstrated by eliminating the flooding (by removing the outlet baskets) under which circumstances the device produced a mere jet of water with neither vacuum nor aeration. This is a fundamentally different mode of operation than Aghnides as could be seen from the demonstration of an Aghnides’ aerator with the sides cut out (P. Exh. 39). The chamber in Aghnides is not a widening one as in Felten and was not flooded but on the contrary contained only the naked water jets in free air striking his resistance screens.
“The fundamental nature of the difference between the Felten device and the Aghnides’ aerator was also visually demonstrated to this Court by providing a Felten device (P. Exh. 42) with greatly increased resistance below his baskets—in the amount of 21 screens! The device none the less created substantially the same amount of vacuum as it did without the screens, showing that the intended function of the device was not changed even by a great change in the amount of resistance. On the other hand, when the resistance of an Aghnides’ aerator (P. Exh. 42) was increased by adding only three screens its proper proportioning was destroyed, the device flooded and it became useless for its intended purpose. It was also demonstrated that Felten’s baskets are not essential to his operation and are not elements corresponding to Aghnides’ essential resistance screens or shoulders; when Felten’s baskets were removed, his device would still operate with a flooded chamber and produce his high vacuum.”
The foregoing findings are supported by substantial evidence, pointed and convincing, and we are persuaded that they should not be set aside. They show that while in Felten some air was present, increase in the amount of air would defeat the operation of the device, since what was desired was a vacuum; whereas in Aghnides the presence of a large quantity of air was desired to effect aeration, which is the heart of the invention. Unless these solidly grounded findings are to be disregarded completely, we must conclude that the Felten British patent No. 6012 of 1909 and the corresponding United States patent, Fuss, No. 1,090,044, do not anticipate the patent and do not indicate invalidity for lack of invention thereover.
*723The other patents cited by the defendants also failed to produce the objective of the Aghnides device. They were never applied to the purpose of the Aghnides patent. In Court these earlier patents emitted violent, splashing, divergent streams without the desired aeration. None of them refers to the problem to which Aghnides addressed himself. The prior art patents, other than Felten, which were chiefly relied on by the defendant, are Solomon and Hyatt. The Solomon British patent No. 405,316, for example, intended to aerate beer, produces no bubbly stream comparable to that produced by the Aghnides device. The admission of a slight quantity of air increases froth on beer sufficient for the desired purpose. It will not perform the same function as Aghnides, as was demonstrated before the Trial Judge. The British patent Hyatt, No. 326,221, is a massive device for purifying water for cities or large industrial installations, and is not addressed to the objective of delivering an aerated stream of water at the point of use.
Other patent references are even wider of the mark. None is the counterpart of patent No. 2,210,846 in principle, operation, or results, as the demonstrations proved step by step and in detail. Even the speculations as to how certain modifications in pre-existing devices might adapt them to achieve the desired result were refuted by actual experiments in open court. The demonstrations in the District Courtroom are entitled to special weight, because resemblances and differences not readily noted from verbal descriptions or graphic illustrations were made vivid by use of models constructed of transparent material to make the operation of the various devices of the experiments clearly visible.
The usual defense that the result would have been obvious to one skilled in the art cannot avail the defendant here. Obviousness does not mean that one skilled in the art can perceive the solution after it has been found and pointed out by someone else; the test of obviousness is as of an earlier time, when the search is on. If the answer to the riddle were in fact discoverable as readily as the defendant now claims, it seems incredible that with a latent demand strong enough to result in the sale of 21,000,000 units of this device, the persistent researchers in the industry who were engrossed by the problem would have been blind to the solution found by Aghnides.
We conclude that the '846 Aghnides device was inventive and not merely within the ordinary skill of the art, and that it was not anticipated by the prior art. The District Judge’s findings are not only adequately supported, but accord with our own appraisal of the evidence.
Defendant contends that because plaintiff admits that Figure 8 of the ’846 patent2 is in the prior art, the claims in the patent are invalid. The complete answer is that the claims are not directed to this figure and do not read upon it any more than they read upon Felten. The scope of the claims is controlling, not the figures in the drawing. A form shown in a drawing is not necessarily covered by the granted claims. When an application for a patent is filed, many forms may be shown and an attempt made to cover them all in the claims as originally presented. During the prosecution of the application, however, the claims may be restricted in the light of the prior art. In the end, only certain of the forms shown may be covered by the restricted claims. Whether the forms not covered by the surviving claims remain in the patent or are cancelled from the drawing is immaterial, for the claims, and not the drawings, control. We consider it immaterial that Figure 8, which is not covered by any of the claims, was not *724omitted in the granted patent. Denominational Envelope Co. v. Duplex Envelope Co., 4 Cir., 80 F.2d 186, 192; Bassick Co. v. Faultless Caster Corp., 7 Cir., 105 F.2d 228, 231; see also Freedman v. Friedman, 4 Cir., 242 F.2d 364.
The opinion of the Seventh Circuit commented on the fact that one form of Aghnides’ device may be used as a vacuum pump and, apparently, largely for this reason, invalidated the Aghnides patent as an equivalent of Feltens prior patent. The reference must have been to Figure 8, the only form of Aghnides that can function as a pump; but this form is outside the claims, because it does not respond to their terms or produce the desired result of an aerated stream.
Contesting the charge of infringement, the defendant undertook to distinguish the accused device, called “Mel-O-Flo,” manufactured by Melard Manufacturing Co., from that of the plaintiffs. The defendant’s aerator has, at the top of the chamber, a diaphragm with small orifices through which the water from the faucet goes, and below the diaphragm is a casing with lateral holes for the admission of air. The diaphragm is formed by a ring and a plate that is seated in the ring. The plate is provided with teeth. The plate itself is not perforated, but when the parts are brought together, there are openings around the border through which the water can escape downward. This is substantially like the apertured plate or plates at the upper end of the chamber in the plaintiffs’ device.
The defendant’s contention of noninfrisgement is that the accused device, unlike that of the plaintiffs, does not utilize “resistance screens” to effect break-up or mixing of the water and air. The claims of the patent, however, are broader and are not limited to such screens. As pointed out above, claim one simply calls for “means in the path of the stream of water * * * for finely breaking up the water and for offering sufficient resistance for thoroughly mixing it with air * * *.” Clearly, this language applies as well to the accused structure as to numerous figures of the ’846 patent. The fact that the defendant’s structure uses an “onion” and the inner surfaces in the chamber and a single outlet screen to create the necessary resistance, whereas Aghnides uses a plurality of screens, is an immaterial circumstance. They are equivalents in every sense. The claims of the ’846 patent clearly read on the defendant’s structure. See Graver Mfg. Co. v. Linde, 339 U.S. 605, 607, 70 S.Ct. 854, 94 L.Ed. 1097.
The second patent, '832, which is claimed to be an improvement upon the first, covers the same general combination as the basic patent, plus an added element, an additional apertured diaphragm in the upper part of the chamber, or a restricted orifice upstream of the chamber, or both. This is set forth in claim one as a “foraminous member upstream of the apertured diaphragm”; in claim two as a “restricted orifice upstream of the apertured diaphragm”; in claim three as a “restricted orifice and foraminous member upstream the apertured diaphragm”; and in claim four as “superposed apertured diaphragms at the upstream end of the chamber through which water is adapted to be forced into the chamber with substantial velocity.”
But the Aghnides original, or ’846 patent, had already taught the use of a perforated diaphragm at the upstream end of the chamber to force water into the chamber with substantial velocity. Merely adding another perforated diaphragm or diaphragms at the upstream end, or another screen or screens at the outlet end, introduces nothing new, especially as the earlier patent had shown that perforated plates or screens could be used in either position, and that “even better results can be obtained if two or more superposed screens are used in the outlet discharge.” Adding interchangeable screens or diaphragms at the other end is merely a means of maintaining the proportion between inlet and outgo, and does not rise to the level of invention over the earlier patent. Nor *725does the second feature of the ’832 patent introduce novelty in the device or in the result to be achieved by it. Merely noting that “when the water is discharged [into the chamber] through a restricted orifice [a narrow pipe] that the flow through the orifices in the upstream diaphragm is improved,” is—in the context of all that had gone before, including Aghnides’ own contribution—too meagre a discovery to be regarded as invention. The changes seem to be the normal and obvious result of skill and experience gained in the practice of the ’846 invention. They are entitled to no new patent, for they are variations in the application of the principle already disclosed, and are of no important consequence mechanically and should be accorded no significance legally. We conclude that the ’832 patent is invalid. See Atlantic Works v. Brady, 107 U.S. 192, 200, 2 S.Ct. 225, 27 L.Ed. 438; Wine Ry. Appliance Co. v. Baltimore & O. R. Co., 4 Cir., 78 F.2d 312, 319, and others there cited.
A considerable portion of the defendant’s briefs and the argument before us was concerned with matter not in the record, but attempted to be raised by motion to reopen the case many weeks after the conclusion of the trial, which lasted over a long period of months. Therein the defendant offered to show that Aghnides had represented to the Canadian Patent Office that the inventions of his Canadian applications corresponding to the patents in suit were not the same inventions as those covered by a prior Belgian application made by Aghnides; whereas in the United States, to get the benefit of an earlier date, he had previously represented that the invention of the United States application was the same as that of the prior Belgian application.
Ordinarily, whether or not a case shall be reopened for the reception of additional testimony is in the discretion of the Trial Judge, which discretion will not be disturbed on appeal. Particularly is this so where the party moving to reopen has had long and full opportunity to develop the information he now seeks to present, but did not do so during the pendency of the trial. In this case, we are the more strongly disinclined to interfere with Judge Hayes’ exercise of discretion, because on the documents before us, we fully agree with his comment that even if the case were reopened, the proffered testimony would not alter the result. From our own examination of the United States patent ’846 and the Belgian patent, we think it unquestionably true that they are, as Aghnides told the United States Patent Office, the same invention. Whether or not on another occasion Aghnides made an inconsistent claim in Canada to obtain a patent there, is an issue that we need not decide. If he made a misstatement there, this might bear upon the validity of his Canadian patent. It would not alter the fact that what he told the United States Patent Office was the truth and that the patent here was obtained without deception or fraud. Even if we were to assume what we are not prepared to adjudicate, that the inventor’s representations to the Canadian Patent Office were untrue, this misconduct would not be so closely related to the proceedings in this country and the issuance of the patent to him here as to invalidate the grant or to constitute such unclean hands as to disentitle him to the relief he asks in this case.
Defendant complains of the trial court’s failure to dismiss the complaint as to Chase Brass and Copper Company, Inc. The argument is that although Chase became a party plaintiff with the consent—almost insistence—of the defendant, and defendant does net complain of the intervention of Chase, nevertheless, no relief should be entered in favor of this plaintiff and the case should be dismissed as to it, because Chase failed to prove infringement after the assignment of the patent to it. Yet the defendant argued that no injunction should be granted at the instance of Aghnides, because he no longer is the owner of the patent. Accordingly, although the Court has before it Aghnides, who was *726the owner when the infringement occurred, and Chase, his assignee, injunctive relief, it is said, is not to be allowed. We find no merit in the defendant’s contention, raised for the first time on this appeal, especially as it consented to the admission of Chase and announced its intention to utilize against Chase any decree that might be rendered against it. With all parties in interest before the Court, it was in a position to do complete justice.
Affirmed in part and reversed in part, and remanded for the passage of a decree in conformity with the views herein expressed.