(after stating the facts as above). This court is not required to give, and the parties may be spared, any description or history of the art of radio communication, as a preface to the consideration of this appeal. The case at bar does not raise for decision questions affecting the fundamentals of that art; and in so far as it does attack defendant’s right to use a certain method of producing ether waves, the historical relation of the method in question to the art in general, and many details of commercial practice said to prove or disprove infringement, are sufficiently treated in the opinion of Judge Mayer. In that careful and illuminating review of a record unusually extended, the efforts of counsel have pointed out no omissions, however divergent their arguments as to the force or probative value of the facts there summarized.
[1] In expressing agreement with the result, below we shall confine ourselves to (1) a consideration of the scope and meaning of the patent in suit, (a) from its own language and office history, (b) from the circumstances surrounding the efforts or experiments said to have given it birth; and (2) a comparison of the statement or theory of invention now insisted upon, with (a) that urged in the Third circuit and our own views derived from the above described sources, and (b) the proven acts of others, antedating the patentee’s invention, whether such acts consisted in the practical use of radio appliances or prior patents or publications.
[2] 1. (a) The specification, drawings, and claims of the patent, considered alone, offer no difficulties to the investigator, and the patentee is conclusively presumed to have known what he invented or discovered, better than did any one else, at the time he applied for a patent. This is true, even though subsequent students may perceive, through the imperfect or ignorant language of the applicant, that he disclosed methods, means, or processes having capabilities surpassing the inventor’s dreams at the time attempt was made to put achievements into words.
Fessenden describes and pictures an arrangement of radio appliances not unfamiliar at date of application, and for which he claims no patentable novelty; he then states as the problem successfully attacked by him the “annulment” or “prevention of atmospheric and other disturbances,” which in the “tropics” and “even in the more northern climates” cause “great difficulty * * * in wireless signaling.”
*718This means that he had overcome “static,” and is plain enough, though inaccurate; for no one has prevented, or as yet can prevent, the electrical phenomena accompanying, e. g., a thunderstorm, in sultry weather. The conditions are inherent in nature; what Mr. Fessenden desired was to minimize or overcome the observed effects of such disturbances upon radio communication.
To accomplish his purpose he suggests as a voltage source an “alternating current dynamo” giving its voltage “at a frequency of 500 periods per second,” and both justifies the suggestion, and explains his method, by disclosing that while “experimenting with a high frequency alternator” he had noted that, when “the trains [of waves] succeeded each other at a frequency above the normal frequencies used for alternating current work, the signals became mote distibct in the presence of atmospheric disturbances”; and especially had he observed that signals given with a spark frequency of 250 per second were useless under conditions that permitted the operator to “easily read” them “when the spark frequency was raised to 900.”
It being understood that “reading signals” presupposes the production of sounds, depending for their nature and existence upon a chain of natural phenomena starting with the generation of ether waves by the sparks resulting from condenser discharges, the statement quoted' from the specification plainly means that, the more sparks per second were produced, the more easily the resulting sounds were heard.
To explain this ease, it is further taken as understood that “the low1 er noises” are made by “atmospheric disturbances,” while “higher notes” resulted from “higher, frequencies,” which means that high frequency in sparks, when translated or transmuted into sound waves, gave a higher pitch than low frequency similarly translated, _ for the equally understood reason that rapid oscillations of ether became (so to speak) correspondingly rapid vibrations of the atmosphere, and the more rapid such vibration the higher the note, within the limitations of the auditory nerves of man.
The patentee then assigns, as a reason for the asserted ease in hearing the sounds which represented sparks with a frequency of 900, the “physiological phenomenon” that a hearer could “concentrate” his “attention” on higher notes, so that the lower noises of static “ceased to-affect consciousness.” This is called a “novel physiological effect, discovered by the applicant”; but, however dressed out in long words, it is no more than an observation in acoustics familiar to any one who has heard the whistle of a fire engine dominating and piercing through the rumble of a city and the roar of a crowded street.
This “freedom from disturbance” the patentee calls “selectivity,” and states that his “physiological method of selectivity” renders tuning, either electrical or mechanical, “not always advantageous” and he expressly excludes it from his “present invention”; that is, he uses for hearing or receiving purposes an ordinary telephone receiver.
Upon these quite simple disclosures, Of which the inaccuracies or overstatements above noted would not affect the substantial facts, are based the two claims in suit, which may fairly be compressed into the following statement of invention, viz.: A method of overcoming or *719dominating static consisting of signal production by ether waves of-definite frequency, having á definite group frequency1 of not less than 250 per second, and preferably approximately 1,000 per second, and then reading said signals with or by an ordinary telephone receiver. The file wrapper shows that the original application nowhere used the word “definite,” expressed a preference for a receiver tuned both electrically and mechanically, and asked to cover by claim “the production of signals by groups of impulses having a group frequency higher than commercially used alternating current frequencies,” which means (as explained by the specification) that Mr. Fessenden thought himself the first to “get through” static by means of frequencies rising to about 900 per second, or at any rate exceeding 250.
The application stood substantially thus until March 13, 1909, when, in order to “bring out more clearly the distinction between applicant’s invention and the method” of the references cited by the examiner, the specification and claims took their final form, the patentee arguing that the receivers of the references accomplished “selectivity by rendering the receiver responsive to only one definite group frequency” (that is, they were tuned, at least partially, as Fessenden preferred to have his tuned when he filed the application); but, as changed by the amendment of 1909, the patentee declared that he attained “selectivity by utilizing the physiological discovery referred to in the specification.”
[3] Of these undoubted occurrences it is to be noted that, if definiteness was always a part of Fessenden’s method, the belated appearance of the word “definite1 in the description or definition thereof does not deprive him of the benefit of the same; but if, in truth, the requirement of “definite group frequency”2 constituted a departure from an addition to the original disclosure, it was not supported by the statutory oath and the claims are unlawful. National Elec. Signaling Co. v. Telefunken W. Tel. Co., 208 Fed. 692, 125 C. C. A. 647. The solution of this problem obviously depends on the meaning of “definite.”
It is further observable that nowhere in the patent as issued, or tire file wrapper, is there mention of any special spark gap, or arrangement therefor. Fessenden pointed to his drawing and in the application said, “37 is a spark gap”; the drawing indicates (if it means anything definite) a fixed gap; so the matter has stood since application filed.
Finally, we note that not until March, 1909, did the patentee in argument with the office describe the sound producible by ether waves as a “musical note,” and he then speaks of the “musical notes produced by. atmospheric disturbances,” as well as. those produced “by waves from the sending station.” His assertion was that, because the ''pitch” of the two musical notes was “different,” his physiological selectivity enabled the operator, using an untuned receiver, to distinguish signals from static.
(b) Mr. Fessenden conducted his experiments, and evolved the method under consideration, while employed by the National Electric Sig*720naling Company, at Brant Rock, Mass.; this is undenied. The primary purpose of the Brant Rock station, as early as 1905, was to establish radio communication with Scotland. To that end it was provided with the best apparatus then procurable, and the patentee was striving to improve that best.
The testimony as to what this station could and did do has been so thoroughly considered in the court below that we find it unnecessary to go into details. It is enough to say that in 1905 the station had no appliances capable of producing 250 sparks per second, and it used a synchronous rotary gap, a device not common among radio workers, not invented by Fessenden, not new, and known as producing mechanical regularity in timing discharges. This apparatus did produce or ultimately cause an even persisting series of sounds, which were noted as musical as early as December, 1905, by an operator who heard them at San Juan, Porto Rico. .
Mr. Fessenden was anxious, long before he filed his application, to get and use dynamos capable of at least 500 sparks per second; he evidently wanted high frequency, but there is no evidence that prior, to the date of his application he spoke or wrote of the desirability of having high frequency with spark regularity, or of a resulting musical quality in sound producible by such a method. There is, however, ample proof that, substantially at the date of application, he discovered his principle of physiological selectivity, and we consider it shown beyond a doubt (dehors the patent itself) that his belief in tire value of that idea impelled him to file application. Not only the patent as it exists, but tíre wording of the original application, fortifies this conclusion.
2. The tenor of plaintiff’s evidence (especially Kintner’s), the argument below so plainly reflected in the opinion of the District Court, and the presentation of this appeal, all compel belief that throughout the present litigation appellants have thought themselves entitled to succeed only if it could be held that the “gist of the patent is the generation of waves having a definite — i. e., regular — group frequency, above 250 per second, preferably of the order of approximately 1,000 per second, and the reception’ of such groups with an indicator resonantly unresponsive to said group frequency.” To quote further from plaintiff’s final brief:
“The invention overcomes static by generating waves in groups of definite recurring period, and at such high frequency that the sound produced thereby in the receiving telephone will be a musical tone or note, clearly distinguished from the noises” of static.
We.accept as-the final position of plaintiffs that the patent discloses one inventive thought dominating and distinguishing the method claimed, viz., regularity or uniform periodicity in spark production, at frequencies so high as to make the resultant sound musical because of the regularity, and high-pitched because of the high frequency. If the claims in suit do not embody and describe this thought, they cover nothing worth contest.
[4] (a) A patent, like any other instrument of conveyance or grant, may be of doubtful meaning, or it may contain several meanings suit*721able to varying expected conditions; but it cannot have two meanings inconsistent with each other.
[5] This patent was sustained in the Third circuit because, as was declared by Buffington, Circuit Judge, it disclosed, from date of application filed, high frequency, an untuned receiver, and physiological selectivity. 208 Fed. 685, 693. The union of these ideas or elements produced the method. This reading of the claims is now totally abandoned ; it was wrong, because, as is now admitted, the physiological phenomenon of the specification was well known before the application, if not before radio communication itself, while high frequencies in radio work also antedated Fessenden, and ordinary receivers were used before 1905.3
The present reading of the claims makes the word “definite” contain the one vitalizing idea. It is said to mean “regular,” in the sense of uniform recurrence (of wave groups) in point of time. On this point the patent itself, its office history, and the evidence surrounding its evolution lead us to observe that there is no reason why the same word should have unexplained differences of meaning in the same document. Fessenden speaks of the “definite periodicity” of the vibrations of a tuning fork when expounding his supposed discovery in physiology (page 2, line 119), and in both the claims both “wave frequency” and ^ “group frequency” are described as “definite.” No reason at all appears why he should be held to have assigned different meanings to the same word, thus used. Yet such must be the holding on appellants’ contention.
[6] Further, if words are to be taken in their usual acceptation as reflected in lexicons of approved authority, definite does not mean regular, nor indicate uniform recurrence in point of time; it does signify fixed and determinate, but without reference to regularity in recurrence. This meaning is entirely appropriate to its use in respect of waves and tuning fork vibrations.
Finally, there is no evidence that the patentee had any idea that definite — i. e., regularly and uniformly recurring — wave groups, were a part of his invention when he filed application, nor until March, 1909.
But it is said that Fessenden used a synchronous rotary spark gap, which insured definiteness in the sense urged, and from this fact is built up the argument that he invented his method as early as 1905. Let the fact of use be assumed; but, if so, then there is no excuse for the omission in the specification of the only piece of apparatus, even now suggested, for accomplishing the patentee’s method, for the reasons and by the system said to be wrapped up in the word “definite.” We decline to believe that, when Fessenden wrote “37 is a spark gap,” he meant “37 is a synchronous rotary spark gap producing uniformly recurring impulses, which I describe as definite group frequencies.”
Unless this can be done, however, the conclusion is irresistible that, if “definite” means what appellants insist upon, the Fessenden method *722was never completed, nor applied for, until March, 1909, and was unlawfully patented. It follows with equal certainty that, if the patentee meant in 1907 what is now claimed, his patent never meant what these plaintiffs urged for it in 208 Fed. Men may justly admit error in matters of opinion, but when'two histories of facts are successively put forward under the circumstances here shown, it can afford little cause of complaint if those charged with the duty of decision end in believing neither.
(b) Examination of prior patents and publications relating to radio communication shows that before this application was filed the word “definite,” as applied to spark production, was by no means unknown. The main citation against Fessenden (Blondel, No. 824,682) uses it repeatedly, and always in its normal sense of fixed or predetermined; there are other instances of such use, but no case is called to our attention, in the whole prior history of the_art, wherein the word was used with the meaning we are now asked to give it.
By overwhelming evidence it is proven that, before San Juan heard the Brant Rock signal, the United States Navy, using a fixed gap and a mercury turbine interrupter, accomplished high frequencies, resulting in a high pitched signal, which in the case of the U. S. S. Kentucky became quite literally famous in the service; nor do we consider the testimony on this point open to criticism, because much of it was given by the wireless operators of the Navy who attended court in uniform. We are aware of no reason why the dress of their profession should be avoided by witnesses.
Whether regularity in timing discharges was or could be attained by the Navy method is a point on which no decision is necessary. It is enough that in 1905 there was no novelty in the high note resulting from high frequencies.4
Summarizing the foregoing observations, we hold, from a consideration of the application resulting in this patent, that, when filed, .regularity in point of timing sparks was not, and was not intended to be, any part or portion of the method applied for. That idea, now said to be the foundation stone of the patented system, appeared (if it ever did appear) in 1909, constituted “a radical change of base,” and is now represented in unlawful claims. National, etc., Co. v. Telefunken, etc., Co. (D. C.) 209 Fed. 856.
Further, it is plain from the known Brant Rock practice and experiments, plus the paper history aforesaid, that what Mr. Fessenden had thought out in 1907 (not before), and the reason for filing application, is correctly described in the opinion of Buffington, J., and now that the theory of validity .then successful is abandoned by those who urged it, no other meaning can be ascribed to the language of the patent consistently with the very facts that overturn the original contention.
It is therefore concluded that Fessenden never patented the method *723now ascribed to him, for in 190S he had not the spark on which the method depended — his frequency was not high enough'; if he had conceived such method in 1907, he filed an application which disclosed a wholly different one; and if he ever disclosed a method of which definite (i. e., regular) spark production was the essential or an essential thought, he did it in 1909, without complying with the law that authorized a patent to issue.
But we also conclude that he never intended to patent, or supposed he had patented, the method of appellants’ argument; on the contrary, Fessenden’s method depended for vitality on that physiological selectivity which has been abandoned in the face of facts put in evidence for the first time in this case.
The claims being invalid, the decree below is affirmed, with costs.
Or “spark frequency” or “wave train frequency.”
The definite wave frequency of the claims may be disregarded; by uncon-tradieted evidence, any given circuit arrangement produces waves both fixed, regular, and predetermined.
These points are fully elaborated in the opinion of Mayer, J. The evidence thereon, here first adduced, renders the result reached in 208 Fed. impossible, by the reasoning there used.
It is not thought necessary to prolong the matter by consideration of the De Forest methods, especially at Galilee, N. J., nor to consider whether infringement can result from use of the quenched gap, because (and only because) decision is not to be rested on these matters.