This is the ordinary suit for infringement, based upon patent No. 1,090,992, issued March 24, 1914, to Kuch for an improved mercury vapor lamp. Appellant acquired title to the patent and was the plaintiff below. The bill was dismissed because the patent was thought to be invalid, and the plaintiff appeals.
The conclusion of invalidity is rested upon two grounds: Eirst, that new matter was introduced into the application and was not supported by new oath; and, second, that Kuch was anticipated by Bastían, and that this priority had been adjudicated in an interference between them. A due understanding of the questions thus presented requires a brief explanation of the invention! Lamps of this type consist essentially of a sealed glass tube, more or less exhausted, having a mercury electrode at each end. When the tube is filled with the mercury vapor and the electric current passed through, the vapor becomes luminous or incandescent. One of the recognized difficulties is that, since greater heat develops at the anode than at the cathode, there is at the anode a greater vaporization, and tire mercury is distilled over and accumulates at tire cathode, so as to destroy the proper balance. This difficulty had been, met by devices or constructions which caused or permitted the excess mercury at the cathode to run back occasionally along the tube to the anode; but this remedy brought new operating difficulties in the high pressure lamps. Kuch’s object was to find a better way to preserve an automatic balance between the two electrodes. He took that style of lamp which had a horizontal tube carrying the mercury at each end in a depending enlargement or bulb, and he accomplished a first or general automatic-balance by directly attacking the vaporization at the anode. His remedy for the excess was to make the anode bulb larger than the cathode bulb. It followed that the outer surface of the former bulb, which would radiate heat, would be larger, and would therefore by radiation subtract heat from that generated inside the bulb, and so diminish what may be called the net heat of the mercury, and the comparative resulting anode vaporization. He knew that the effect of a given current upon these electrodes and under given conditions could be computed, and his inventive thought was that, if the radiating surfaces of the two bulbs were proportioned to the amount of the heat to be generated in each, there would be corresponding radiation and the proper vaporization balance would be maintained.1 He thus brought about a rough and general equalization by attacking the cause of inequality; but changes of external temperature or other conditions would prevent this preliminary regulation from entirely and always stopping the accumulation of excessive mercury at the cathode, and so Kuch provided a further and final automatic regulation. Hé knew that there occurred at the cathode a phenomenon called the “cathodic aigret.” This was an agitation suggestive of boiling, and it extended a certain depth from the surface of the mercury in the cathode. Obviously, the bulb surface adjacent to the mercury thus agitated would radiate heat much *63more actively than the remaining and lower part of the bulb surface. Kuch utilized this effect of this phenomenon by interposing, between his horizontal illuminating tube and his depending cathode bulb, an intermediate upwardly-inclined tube of much less diameter than the illuminating tube. A small excess accumulation of mercury in'this intermediate tube would very considerably raise .the lowest point of this aigret agitation and so diminish radiation, increase the net heat, and promote vaporization. A slight drop in the mercury in this tube would have the converse magnified effects; and the mercury level would, automatically, maintain itself substantially constant. We here reproduce Pig. 1 of the Kuch patent as originally filed and as amended. Claims 2 and 3 are given in the margin.2 Claims 1 and 4, also sued upon, do not require separate consideration.
[1] The original drawing did not-show any anode bulb or passage thereto from the illuminating tube, nor did the specification contain any particular description of either. During the progress of the ap*64plication, an amended drawing was filed as above, and the specification was made to say, in so many words, that the anode bulb was larger than the cathode, and that the passage leading from the illuminating tube to the anode was larger than the passage or intermediate tubé leading to the cathode bulb. The claims, as issued, are made to depend in part upon these things not originally specified. Hence it is plausibly argued that the insertion was of new matter and was vital to the invention as patented; and thereupon it is said that the patent is void. Railroad v. Sayles, 97 U. S. 554, 563, 24 L. Ed. 1053; Railroad v. Consolidated Co. (C. C. A. 6) 67 Fed. 121, 129, 14 C. C. A. 232. This view overlooks the substance of the invention, as disclosed in the original specification and drawing. The rule is that insertions by way of amendment in the description or drawing, or both, do not hurt the patent, if the insertions are only in amplification and explanation of what was already reasonably indicated to be within the invention for which protection was sought — “something that might be fairly deduced from the original application.” Hobbs v. Beach, 180 U. S. 383, 395, 21 Sup. Ct. 409, 45 L. Ed. 586; Cleveland Co. v. Detroit Co. (C. C. A. 6) 131 Fed. 853, 857, 68 C. C. A. 233; Proudfit Co. v. Kalamazoo Co. (C. C. A. 6) 230 Fed. 120, 123, 144 C. C. A. 418; Cosper v. Gold, 36 App. D. C. 302. When we seek to apply this rule in this case, we first observe that the alleged new matter was not only permitted by the Patent Office, but was required, because an element claimed was not shown or sufficiently described. The Patent Office has a strict rule on this subject. It fully recognizes that new matter must not be permitted, and it is constantly engaged in defining what is and what is not new matter. The application of the rule must, of necessity, be more or less arbitrary, and the presumption of correctness which attends Patent Office rulings must apply with especial force to this class of ruling; and most peculiarly is that true when the applicant hás only complied with the demands which the Patent Office made.
*63
*64It is clear to us that Kuch’s original application did cover and include these features sufficiently to justify the later fuller description and drawing. After reciting that the evolution of heat at the anode was greater than at the cathode, so that the vaporization at the latter was less, he said that the first object of his invention was to so determine the sizes of the two electrode vessels that they should be in the same proportion as the heats developed in them, and that the second object was to connect the cathode with the illuminating tube by means of a narrow intermediate tube. He also says:
“It is easy to so proportion the positive and negative electrode vessels that the proportion of the heats given off at both electrodes is approximately the same as the proportion of heats developed at both electrodes.”
He also says:
“The size of the cathode vessel in proportion to that of the anode vessel is from the beginning so determined that” the desired results are effected.
His first claim was,
“In a gas or vapor electric lamp the combination with an illuminating tube ' of an anode vessel at the one end of said illuminating tube, a cathode vessel, a *65narrow intermediate tube between the other end of said illuminating tube and said cathode vessel, and a metal liquid doing work in said anode vessel, said cathode vessel .and said narrow intermediate tube.”
The second claim specified that the sizes of the anode and the cathode vessels should be in proportion to the developed heats. Eater claims had specific reference to the function of the narrow intermediate tube in controlling the cathode heat through the aigret action. Thus we find repeated and direct reference to'the fact that the anode bulb is to be larger than the cathode bulb, and in the very proportion which has always been specified in the claims. Since Kuch did not illustrate this anode vessel, he must he deemed to have referred to the well-known form in common use. Gosper v. Gold, supra. In this form, the illuminating tube entered the anode chamber without any constriction, but by an opening which was the full cross-section area of the illuminating tube. It follows that when Kuch specified that at the cathode end of the illuminating tube he had a narrower tube entering the cathode chamber, he naturally implied that this entrance was smaller than the entrance to the anode chamber at the other end of the illuminating tube.3 Tt surely is not the insertion of new matter to put into express words a structural characteristic already fairly implied. The attack upon the patent on this ground must fail. The situation is closely analogous to that in Abercrombie Co. v. Baldwin, 245 U. S. 198, 38 Sup. Ct. 104, 62 L. Ed.—, decided by the Supreme Court December 10, 1917.
[2, 3] After most of Kuch’s claims, about 20 in number, were allowed, the Office suggested to Bastían 16 of these claims. Bastían adopted them, and an interference was declared. Since both applicants were foreigners, and Bastian’s filing date was earlier, the Office awarded him priority. So far as the inventions pertained to a proportioning of the anode and cathode bulbs, this award of priority is the end of the matter; hut at the time of the declaration of interference Kuch’s application contained claims identical with 1 and 2 in his issued patent, and these claims were not included in the interference. like claims 3 and 4, added after the interference, they were characterized and limited by the provision that there should be between the illuminating tube and the cathode tube an “intermediate tube” or a “contracted passage” of less diameter than the main part or other end of the illuminating tube. Count 11 of the interference, which had been one of Kuch’s claims and which became claim 11 of the Bastían patent, was:
"In a vapor lamp, the combination with the horizontal illuminating tube of an anode, a cathode chamber, a narrow upwardly inclined intermediate lube connecting the illuminating tube and the cathode (Chamber, an active fluid in the cathode chamber and partly lllling the intermediate tube.”
Since this count plainly reads upon Kuch’s form, and priority was awarded to Bastían, the court took the view that the invalidity of Kuch’s patent necessarily followed, and assumed that a decree accordingly was in substantial conformity to the Patent Office decision.
*66We think this view does not give due force to the rule that a device may be at the same time an infringement upon the relatively generic claim of a prior patent and also be a patentable improvement thereon, and that there is no inconsistency — even presumptive — between these conclusions. Herman v. Youngstown (C. C. A. 6) 191 Fed. 579, 584, 112 C. C. A. 185). The intermediate tube of smaller diameter than the illuminating tube, or, in other words, the passage to the cathode chamber constructed so as to be smaller than the illuminating tube, was a feature certainly not shown by Bastían, and its addition to- or substitution in Bastian’s structure might well be a patentable improvement upon anything which Bastían had invented. In careful observance of this distinction and purported compliance therewith, the Patent Office reserved out of the declaration of interference two claiips which were limited to this additional feature, and later issued to Kuchthe patent in Suit, based on these and similar claims. To hold these claims invalid on account of the result of the interference is not to affirm but is to reverse the Patent Office conclusions on the subject; and the claims cannot be held invalid for that reason.
[4] This leaves only the question whether this so-called improvement, when considered with Bastían- put in the prior art (where, for the fourth defense, he must be placed — Lemley v. Dobson [C. C. A. 6] 243 Fed. 391, 395,-C. C. A.-), involved patentable invention. This is the controlling question in the case. It is not free from doubt, and very likely the decree below was intended to rest substantially upon a negative answer thereto.
Here, again, the patentee has the benefit, in rather unusual degree, of that presumption of correctness which attaches to the Patent Office rulings. After the interference was finished, the examiner rejected two of the remaining claims of Kuch upon the ground that they were identical with the interference issues, but allowed two others of those reserved from the interference (present 1 and 2). He also rejected, on the same ground, two new claims tendered (present 3 and 4). Kuch finally conceded that the interference adjudication covered the first, two just mentioned, but argued that it did not reach 3 and 4, for the same reason that 1 and 2 were not included. To these arguments the Patent Office yielded. Its action in this respect was therefore most deliberate and considered, and especial deference is given to such rulings. Canda v. Michigan Co. (C. C. A. 6) 124 Fed. 486, 493, 61 C. C. A. 194.
[5] Plainly,- when we consider a supposed invention with reference to a prior patent, to determine whether there was inventive advance, the claims of that prior patent are wholly immaterial — unless in some possible instances where the claims might interpret an obscurity. We must look, rather, to the structure disclosed by the specification and drawings, and the question is whether there has been any inventive advance upon that structure. If the claims seem to reveal the later form, when the specification and drawings do not, it must be because of the generic language of the claims, interpreted by our knowledge of the later device. We therefore cannot depend upon the language o-f Bastian’s claim 11 (and similar claims, 10 and 12) in which alone *67there was any reference to a “narrow intermediate tube” .as distinguished from the illuminating tube and the cathode chamber, and we refer to the specification and drawings to see whether Bastían has the equivalent of the “tube of lesser diameter than the illuminating tube” upon the presence of which the Kuch patent depends. Bastian’s specification does not mention any “narrow” tube or any “intermediate” tube. Since Bastian’s claim 11 declares that this “narrow intermediate tube” is found in connection with the horizontal illuminating tube, we may well look for it in his Fig. 1, showing his horizontal form and here reproduced.
Yielding to the Patent Office conclusion that Bastían could make this claim on his disclosure, it follows that this element must be there somewhere. It cannot be a part of the cathode bulb. It therefore must be that end of the illuminating tube which turns downwardly into the cathode bulb. It is rather an awkward conception that, of a unitary tube of constant diameter all of which seems to be the illuminating tube, part is iluminating tube and part is a “narrow intermediate tube” between the illuminating tube and the cathode; but there is no other solution. Bastian’s “narrow tube” is therefore an extension of the illuminating tube, and of precisely the same diameter, and is “nar ■ row” only with reference to the greater diameter of part of the cathode bulb, while Kuch’s tube is also narrow with reference to the illuminating tube. Is Kuch’s form only such a variation as any one skilled in the art might naturally make? Defendant insists that the vital relationship is between the size of the intermediate tube and the size o E the cathode bulb, and that, since it was common practice to increase the size of the illuminating tube proportionately to the increase in the current to be used, there was no invention in maintaining that part oí Bastian’s tube which served as the cathode extension at the small size shown by Bastían, and then expanding the remainder of the tube into any size which the current required.
We are not convinced that we should accept this conclusion. In the first place, Bastían certainly contemplated nothing of the kind. Not only is it evident that his entire illuminating tube must be of such very small diameter that superficial tension will prevent the flow of the mercury along the horizontal part, but in another patent, issued upon a divisional application, he expressly claimed this feature. Next, we observe that in Kuch, the mercury arc initially fills the main illumi*68nating tube and usually extends part of the length of the intermediate tube. So far as shown by the records in this and in the companion case (No. 3028) relating to the same art (249 Fed. 69,-C. C. A. -), there had never been a lamp with mercury electrodes in which the arc or its containing tube was constricted at or toward the cathode; on the contrary, the area of the active cathode surface had always been at least as great as the cross-section area of the arc tube; and we cannot now say that an innovation in this respect presented no new problems as to radiation, amperage or other elements. If the arc has become limited to a central “string,” then the constriction would be of the space between the arc and the tube walls, and there is no presumption that this would be immaterial.4 The Bastían device was inoperative, except with a small amperage. That its beneficial effects could be retained, and that it could be made to operate with a high amperage, by expanding part of the tube and leaving the remainder untouched, seems now entirely obvious to the defendant’s expert; but we think we have no right to say that this was ordinary skill in 1906. Such a change would involve, not only expanding part of the tube, but (probably) changes in the quantity of mercury to be vaporized and readjustment of the anode and cathode sizes — in fact, every factor involved would be disturbed and subject to recomputation if not to experimental research.
When to these considerations we add the fact that defendant, owner of the Bastían patent, has abandoned his specific form and adopted Kuch’s, thus paying the “tribute of imitation” (Diamond Co. v. Consolidated Co., 220 U. S. 428, 441, 31 Sup. Ct. 444, 55 L. Ed. 527), we conclude that the presumption of validity is not sufficiently overcome, and that there must be the usual decree for injunction and accounting.
Theories of heat transmission to the mass of mercury in the bulb are also involved, but do not need statement.
“Claim 2. The herein described gas or vapor electric lamp comprising a tube having electrode chambers at its ends, the cathode chamber being connected with the tube by a passage smaller than that connecting the tube and anode chamber, a body of mercury in the cathode chamber and partly filling the passage connecting said chamber with the body of the tube, and a body of mercury in the anode compartment.
“Claim 3. A vapor electric device comprising an anode bulb partly filled with mercury, a main illuminating tube communicating with said anode chamber and a tube of lesser diameter than said illuminating tube connecting the latter with the cathode chamber.”
When he made his new drawing, he showed a passage to the anode which may be slightly constricted, but his claims do not involve this feature.
We do not overlook Fig. 2 of Hewitt patent No. 682,690 (using one mercury electrode). Tiie inserted guide or collar which produced restriction was for a different purpose, and was of opaque and nonheat-conducting material. It is not more than a suggestion of one element of Kuch’s combination. If it were more, it would anticipate Bastian’s claim 11.