MANTON, Circuit Judge.

These appeals by.both plaintiffs and defendant involve an infringement of five patents. The American Telephone & Telegraph Company owns the patents; the Western Electric Company, Inc., a subsidiary of the owning company, manufactures under the patents; the Electrical Research Products, Inc., a wholly owned talking motion picture subsidiary of the Western Electric Company, and the Western Electric Company, are exclusive licensees and proper parties plaintiffs. Western Electric Co. v. Pacent Reproducer Corp’n, 42 F.(2d) 116 (C. C. A. 2). The defendant is "an exhibitor of sound pictures, and owns a theater in Buffalo, N. Y.

The patents relate to sound transmission and communication, particularly as now found in the development and reproduction of sound pictures. In improving its telephone service on transcontinental lines, the American Telephone & Telegraph Company used amplifiers, or repeaters, as they are sometimes termed, and met its best success by greatly improving the De Forest three-electrode vacuum tube. De Forest had added the third element or grid to the Fleming valve in 1906, and had used that device to some extent as a detector of radio waves; he had made little progress with it as an amplifier prior to 1912. The problem of amplification was to reproduce without distortion the input of the amplifier in its output, greatly increased in energy. Speech has been found to be extremely complicated; each of the ordinary sounds being composed of many frequencies which must he accurately reproduced. The range of audible frequencies is from 16 to 16,000 vibrations per second. For sound pictures and publie address systems, substantially from 30 to 5,000 cycles must be reproduced without distortion in order to give the high quality which is known to-day. Greater power is needed to drive the loud speakers of these systems than is required for radio or telephone. The lack of faithful reproduction would be so accentuated under certain eonditions as to destroy the utility of the systems. The success of sound pictures was due solely to high quality reproduction; that is, without distortions.

Each of the patents in suit relates to circuits in which the three-electrode vacuum tube is used. It is claimed that all of the inventions may be used to advantage in a single amplifier in which the final output contains the integrated result and combined advantages of all the inventions.

Lowenstein Patent.

The Lowenstein patent, No. 1,231,764, for a telephone relay, was filed April 24,1912, and issued July 3, 1917. Claims 1, 2, 4, 5, 6, and 7 are relied' upon, and claims 1 and 7 have been selected as typical. The court below found this patent invalid.

This invention consisted essentially in the discovery that, if the grid of a vacuum tube amplifier is made to operate by potential and not by current, the distortion produced in the input circuit by current (and reproduced in amplified form in the output circuit) could be prevented, and that this restriction to potential operation could be attained by the application to the grid of a suitable initial negative potential on which the signals to he amplified are superposed.

Lowenstein disclosed a telephone system connected to a vacuum tube amplifier by a transformer 12-13. One terminal of the transformer 13 is connected to the grid 18, of the vacuum tube and the other terminal to a battery 21. This is the input circuit. The telephone receiver 24 is connected to the vacuum tube by the transformer 22-23. The transformer is connected to the plate 17 and to the battery 21. This is the output circuit. The filament 16 of the vacuum tube is heated by the battery 19, and the temperature of the filament is adjusted by a resistance 20. The battery 19 is called the “A battery.” The battery which supplies current for the plate eir-*725cuit is called the “B battery.” The inventiou claimed for Lowenstein is that he inserted a third battery in the grid circuit and connected its negative terminal to the grid in order to make the grid at all times more negative than any part of the filament. This prevents the grid from becoming positive and causing current to flow in the grid circuit. This negative potential, applied to the grid by the C battery bad to be greater than the most positive signal voltage that would be impressed on the grid. The negative potential applied 'to the grid by the C battery had also to be maintained more negative than the potential of any part of the filament. If the grid were more positive than any part of the filament, it would still compete with the plate by attracting some of the electrons which should pass to the plate, and would thus fail to prevent the flow of current and harmful consumption of energy in the external or grid circuit.

The reason for its use is stated in an article by the defendant’s expert Bowles in Popular Radio, published before this controversy arose, in which he said: “Much distortion is introduced when the grid of a valve is not properly biased (negatively) due to the current taken by the grid. A reversed C battery will cause a set to be almost inoperative.”

Thus the primary object of Lowenstein was to eliminate distortion. His specifications show that the distortion which he sought to eliminate was that which arises in the circuits as distinguished from that arising in the tube itself. He bad ascertained by experiment that relays which function according to current values cannot remedy this distortion. Therefore he sought a way to make Kis relay function without permitting the (low of any current; i. e., according to terminal voltages as distinguished from current values. The court below stated: “The controversy as to the scope and meaning of this patent arises out of the construction to be given to the following paragraph of the specification: ‘The potentials created in secondary 13 are made to control the current flowing through the ionic field and originating in battery 21 by connecting the one terminal of coil 13 to the modulating member 18 and the other terminal of said coil to a point on the battery 21, which is located ultranegatively relative to the negative point of the battery connected to the filament.’ The plaintiffs’ position is that these; words must be interpreted as meaning that the grid was connected to a point on the battery 21, more negative than any part of the filament. The defendant’s position is that a grid biasing potential, more negative than the most negative potential point of the filament, is not disclosed.”

This controversy arises because the diagram of Lowenstein does not show which way the filament of the A battery is connected, and therefore it cannot be determined which is the negative end of the filament. If the battery 21 is connected to the negative side of battery 19, then the grid would necessarily be negative to all parts of the filament. If the battery 21 is connected to the positive side of battery 19, then the grid would be negative with respect to all parts of the filament only if the C battery had a higher voltage than the A battery. The diagram does not help in determining these things. In tlie specification as originally filed, Lowenstein stated that the terminal of the transformer 13 was connected to a point on the battery 21, preferably located ultranegatively relative to the negative point of the battery connected to the filament. The meaning which Lowenstein intended to give to the term “filtran egative” appears in the prosecution of the case in a response under date of August 6, 1914, by Lowenstein to an action by tbe Patent Office, citing Von lieben as an anticipation. Low-enstein said: “Von Lieben grid H is positive as against the negative portion of the filament; therefore, a negative current actually flows from that portion of the filament to the grid H. Therefore this is not a disclosure of a potential operated device. As to claim 4, it seems only necessary to point out that the disclosure cited does not permit of making the grid H. ultranegative, that is, more negative than any part of the filament; while claim 4 expressly sets forth this feature.”

The Examiner finally acquiesced in this view. Thus it is established that Lowenstein intended his grid to be negative with respect to all parts of the filament.

The pertinent prior art referred to below, and to which we should refer, is the Von Lieben tube in which the grid potential is adjustable by a resistance connected across the filament terminals. With this arrangement, the grid potential can be varied, but it can never be made more negative than the negative end of the filament. Current always flows in the control circuit, and Gherardi and Jewett (Defendant’s Exhibit, p. 1278) point out that serious distortion results from this. The defendant argues that Von Lieben taught the necessity of an adjustable grid potential, but nothing in Von Lieben suggests that the grid should be negative. There is no such teaching because the Von Lieben tube re*726quires a positive potential on the grid and it would not work at all if the grid were maintained more negative than any part of the filament.

Lowenstein defines the meaning of the term “ultranegative” as being more negative than any part of the filament, and it is unnecessary to consider the question, What bias was suitable for the earlier De Forest tubes ? Lowenstein could only have made his invention by observation, and it is clear that the tubes he produced had the effect which he described.

Apparently Von Lieben’s is a device of an entirely different character from the audion. It operates by virtue of the very thing that renders the audion wholly inoperative; the potential of its control electrode was positive and could not be made ultranegative; distortion producing current always flows in the control circuit, and cannot be avoided. The positive potential was essential to its operation; the tube would cease to operate if an ultranegative potential were used. It does not anticipate.

The claims were rewritten and substituted for the ultranegative grid claims previously submitted, and there is no doubt that they are clearly distinguished from the references. In defining his ultranegative grid in the claims, he refers to “means for impressing upon said modulating device a potential more negative than that of said cathode” (claim 1); “means for energizing said modulating device; * * * from a source of potential which is negative with respect to said cathode” (claim 2); “a modifying or controlling device * * * subject to an impressed potential more negative than that of said cathode” (claim 4) ; “a modifying or controlling device * * * connected to a source of potential more negative than said cathode” (claims 5 and 6); and “means for impressing upon the audion grid a potential more negative than that of the audion cathode” (claim 7). These claims, as expressed, define the invention, point out the use of the word “ul-tranegative,” and distinguish Von Lieben, by stating that the grid was “more negative than any part of the filament.”

This contribution to the art has been of great merit. It has worked well, and has resulted in the practical and everyday operation of the vacuum tube. No prior art suggests the invention. Marconi Wireless Tel. Co. v. De Forest Radio Tel. & Tel. Co., 243 F. 560 (C. C. A. 2). It was held valid in the District Court in Radio Corp’n v. Bunnell & Co. (D. C.) 22 F.(2d) 847.

Reference is made to the prior knowledge of high vacuum tubes. De Forest Radio Co. v. General Elec. Co., 283 U. S. 664, 51 S. Ct. 563, 566, 75 L. Ed. 1339. In the De Forest Case, reference is made to the so-called high vacuum tubes as having no practical use in 1913 and that the audion was not then in general use, and the court below said that this offered additional reason for believing that Lowenstein was' referring in his affidavit (filed in the Patent Office), to a low vacuum tube which the commercial art then knew, and which the court below erroneously assumed was incapable of giving Lowenstein’s result. But Lowenstein’s affidavit conclusively shows that he is referring to the vacuum tubes which he had made and which were not only suitable for, but actually did give this new result in operating with a negative grid bias. The Supreme Court in the high vacuum tube ease pointed out that knowledge of the so-called high vacuum tubes, and how to make them, had existed for many years before Lowenstein filed his application. The court said: “It suffices to say that an examination of the pri- or art discloses that long before the earliest date claimed for Langmuir, the necessity for removing occluded gas from tubes or other electrical discharge devices in order to procure a high vacuum, and the methods of doing it by heating an electronic bombardment were well known, as was the procedure for constructing the high vacuum tube by expelling occluded gas while evacuating the tube.”

The court cited publications and patents as early as 1896 showing clearly that men skilled in the art had knowledge of the so-called high vacuum tubes. It was known pri- or to the filing day how to make the vacuum tube which was suitable for Lowenstein’s purposes.

The court below stated that the language employed by Lowenstein in his specifications describing the operation of the audion would justify the inference that Lowenstein was dealing with the vacuum tube at low evacuation containing ionizable gas. When Lowen-stein filed his application, some gas was believed to be necessary to make the current between the filament and plate conductive. The expressions “ionic field,” meaning the space, and the “ionization of the gas,” “ionization power of the filament,” and “ionized gas particles,” were used. But we think the court below failed to distinguish between the ordinary use of gas ionization and what was known as gas ionization by impact or collision. The latter is evidenced by the blue glow which renders the audion wholly inop*727erative. When the gas is sufficient, as in Von Lichen’s tube, electrons collide with the gas atoms and knock them to pieces. This produces in the spgee positive as well as negative ions. It is established that Lowenstein was not speaking of ionization by impact, but merely ionization of the space which rendered it conductive.

Demonstrations were made by the defendant’s experts which intended to show that no tubes were available to Lowenstein which would give Ids result. "What the test showed was that larger negative potentials can be used on the grids of the vacuum tubes to-day than could be used in most of the tubes of Lowenstein’s time. The record shows that negative grid bias reduced the plate current. It increases the resistance of plate circuit, and makes it harder for the electrons to pass across the spare from filament to plate. To obtain a. given plate current (output capacity) with a negative bias, a higher plate voltage must therefore be used to offset the effect of the C battery in reducing the plate current. The tubes of to-day permit the use of high piafe voltage, and consequently it is possible to apply larger negative potentials to their grids without reducing the plate current to a nonuseful extent. These modern tubes have greater' power capacity than the tube of Lowenstein’s time. But this does not change the result.

It is not a sufficient answer that with after-acquired knowledge one may explain the characteristics of vacuum tubes as they are known to-day and to reason therefrom that Lowenstein’s invention was obvious. The characteristic curves of the vacuum lube operation which Professor Bowles introduced were not thought of when Lowenstein made his invention; the information and data upon which they are based having been acquired since that time. Defendant points to the static grid voltage plate current curves, and reasons therefrom that the grid must he made more negative as the plate voltage is increased in order to keep the tube operating on the so-called straight portion of the curve. But the curves are grid voltage curves. Lowenstein had no such curves, nor did any one else at that time, lie was not interested in operating on any particular part of any curve; he was aiming at the elimination of external circuit distortion. The internal characteristics of tube operation were unknown. With the knowledge of to-day and the aid of curves actually obtained by using the negative grid voltage, it is now easy to reason that a negative grid bias should be used. But that is not the test of invention. Diamond Rubber Co. v. Consolidated Rubber Tire Co., 220 U. S. 428, 31 S. Ct. 444, 55 L. Ed. 527. The test is what was known to the art in 1912 when Lowenstein made Ms invention.

The defendant has failed to prove that Lowenstoin’s patent is inoperative or that invention is not involved in what ho accomplished. Lowenstein made a- valuable contribution which has proven of groat aid to mankind, particularly in the art here under consideration. The defendant has adopted it, and thereby infringes. This patent is therefore held valid and infringed.

Blattner Patent.

The Blattner patent, No. 1,4-83,273, is for a circuit for heating the filaments of audions. It was filed October 4, 1920, and issued February 12,1924. Claims 6 and 8 are relied upon. The patent was held not to have been infringed.

The patent granted to White, No. 1,195,-632, shows how a source of alternating current may be employed for heating the filamentary cathode of an electron discharge device without introducing into the circuit fluctuations of a frequency corresponding to the cathode heating frequency. The grid and the anode electrodes are connected, in this arrangement, to the mid-point of the secondary winding of the transformer which connects the heating source to the cathode. By this patent, the fluctuations of the cathode heating frequency are suppressed without necessitating the use of a transformer and without requiring that a connection he made to the mid-point of a transformer winding. It relates to the act of heating audion filaments by alternating current without introducing into the circuit fluctuations of frequency corresponding to the alternating current frequency. The infringement alleged is for lighting the filaments of the two vacuum tubes in its last stage or push-pull amplifier from an alternating current circuit.

Claims 6 and 8 cover any and all arrangements for heating by alternating current the filaments of audions arranged in push-pull relation, as is illustrated in the Colpitts patent hereafter considered. As so interpreted, the claims covered, not a means, but a result. They disclosed a particular means previously shown in the Swedish patent No. 44,872 to C. R. Gertz, patented and published prior to Blattner’s filing date. The Gertz patent “refers to an amplification arrangement for variable electric currents by means of a,n electron-relay, and its purpose is to make pos*728sible the use of direct current from a distribution system for heating the cathode of the relay as well as for its anode voltage without being troubled by disturbances due to the fact.that the alternating current due to commutation is superimposed on the direct current of the distribution system.” Gertz says his invention is principally characterized by the fact that the circuit of the in-coming current is connected to the grid and the circuit for the amplified current is connected to the anodes, and by the fact that for giving the latter a positive charge a neutral point of the circuit connected to the anodes is connected to a point on a potentiometer shunting an illumination network which also feeds the hot cathode. Thus a single section of a shunting impedance is used to light the fila^ ments of a five-element audion arranged in push-pull circuit. The specifications of'the Swedish patent state that two separate three-element tubes may be used in place of the single five-element tubes with the same effect, and it refers to a back-to-back arrangement as compensating for the fluctuation of current from ordinary commercial sources supplied by machine generators. He clearly describes the balancing effect of the push-pull arrangement as follows: “The input current goes through the winding 10-11 and induces a current in the winding 12-14 by which the grids 2 and 3 become oppositely charged. If at a certain moment the grid 2 is becoming positively charged and 3 negatively charged the thermionic current 4-1 increases, that is, a current impulse is obtained in direction 16-15 in the winding 15-17 of the transformer B, and the current 5-1 decreases which means a current impulse in the direction 17-16. Both these current impulses are added in the winding 18-19 between the terminals of which the amplified current is taken off. It remains to be shown that the alternating current superimposed on the lighting current does not interfere. As 16 is the midpoint of the winding 15-17 the current branches in the directions 16-15 and 16-17. These two branch currents balance each other’s interference in the winding 18-19 and consequently they will not be heard.”

With this state of the prior art, Blattner’s claims do not disclose inventive thought. In the arrangements described in the patent to Gertz, everything that Blattner suggests is patented.

The claim allowed to Gertz provides for “arrangement for amplification of variable electric currents by means of electron relays consisting of a hot cathode, grids and anodes, characterized by the fact that the circuit of the incoming current is connected to the grids and the circuit of the amplified current is connected to the anodes, and by the fact that for the positive charging of the latter a neutral point on the circuit, connected to the anodes is connected to a point on the potentiometer across the current source which is feeding the hot cathode, all for the purpose of eliminating the disturbing effects of a possible alternating current produced in the source, as for instance the commutator ripple-in dynamos.”

In view of what is thus disclosed in the-patent to Gertz, claims 6 and 8 of the Blatt-ner patent are invalid.

Colpitts Fatent.

Colpitts patent, No. 1,128,292, for an-electric wave amplifier, was filed January 3,. 1914, and issued February 16, 1915. This-patent was held valid and infringed below as to claims 1 and 5. It relates to electric wave-repeating apparatus and particularly to the-use of vacuum discharge repeaters as exemplified by an audion.

The plaintiffs contend that it covers the-push-pull circuit with three-electrode vacuum, tubes such as the De Forest audions. The defendant’s claim is that this alleged invention-devised no new circuit arrangement, but merely connected two De Forest audions in the-already well-known push-pull circuit, and it. is said that there was no invention involved in substituting audions for old repeaters in. this old circuit. The push-pull circuit is one-in which two similar performing devices are-oppositely connected so that their fundamentally like operation produces an additive-result, but, being opposed, the defects of one-cure the defects of the other. The basis of this is that the fundamental frequency is amplified and preserved while the even harmonies which tend to produce distortion oppose each other and are thereby reduced. The-push-pull circuit is well known (Dean, 1895,. patent, No. 549,477). Dean described it in his claim as “the combination with a local transmitter circuit comprising two parallel branches, of two induction coils having their-primaries included one in each of the branches of said transmitter circuit and their secondaries included in series in the telephone-line, a microphone included in each of the-branches of said transmitter circuit, means for increasing the resistance through one of' the microphones simultaneously with a decrease of resistance in the other microphone, and a source of electricity connected witln *729said telephone dine adapted to direct current through said branches in parallel; substantially as described.”

In the Dean patent, the telephone voice currents to be transmitted arose from the movement of the microphone diaphragm by direct impact thereon of sound waves in air, set up by the voice of the speaker. The push-pull circuit was later applied with different forms of telephone repeaters and amplifiers, When the telephone repeater or amplifier is used, the sound waves do not impinge directly on the diaphragm, and the diaphragm (a carbon button repeater) is moved by the push and pull of the electromagnets connected to the input circuit of the repeater. The carbon microphone amplifier repeater was known in another form prior to Colpitts (Kitsee, No. 770,296, 1903; Grissinger, No. 1,198,212, 1916). Defendant’s amplifier has a pusli-pull circuit as an amplifying repeater. The repeater elements are not actuated directly by the voice, hut are actuated electromagnetieally by the energy coming to the repeaters from the input circuit. It is clearly established that the mode of operation of these push-pull circuits is the same, and it is as described in the claim of the Dean patent. It is also clear that the defendant used the push-pull circuit of the Doan patent with the De Forest audion. There was a well-known advantage of using the push-pull circuit with an amplifier. The defendant maintains that it was not invention to use the De Forest audion in this old telephone circuit. The De Forest audion was known in 1912. At that time the telephone engineers were seeking an improved telephone repeater to be used in the then known telephone circuits including the old push-pull eireuit in place of the older forms of repeaters such as the carbon button which was known as the Shreeve repeater. It was admitted by the plaintiffs’ expert Waterman that “the mere placing of devices in a parallel feed arrangement of this sort [Dean, Kitsee, Stragiotti and Grissinger] was, of course, very, very old. It is, in effect, only supplying two devices from a common battery, and it was common to do that in these back-to-back arrangements and in carbon buttons for the purpose of getting added sensitivity.” He could not sa.y whether the use of a push-pull circuit with carbon button repeaters and ionic repeaters was familiar to telephone engineers of 1912, but admitted that after Dean the telephone man was presumably advised of the use of these double opposed devices. It required no greater engineering skill and no greater inventive thought to make the substitution of the audion in the old push-pull telephone circuit than was required to mate the same substitution in the old two-way long-distance telephone repeater circuits. The use of De Forest audions by Colpitts in the old push-pull amplifier circuit was another obvious engineering use. It was something which had been furnished to telephone engineer by He Forest in October, 1912. With such simplicity to engineers, expert testimony cannot help to bring it to the height of invention. General Electric Co. v. Steinberger, 214 F. 781 (C. C. A. 2). With this knowledge extending over a long period of years, matters of specific structural difference and consequent relative difference of operation between the audion telephone repeater and other types of prior art telephone repeaters that have been used in the push-pull circuits are not of importance. Hewitt v. Amer. Tel. & Tel. Co., 272 F. 392 (C. C. A. 2); Marconi Wireless Tel. Co. v. De Forest Radio Tel. & Tel. Co., 243 F. 560 (C. C. A. 2).

It was well known in 1912 how to apply the energy that is to be amplified to the eon-trol element (grid) of an audion. The Do Forest audion amplifier (patent No. 841,384, 1907) showed the way to apply energy to be amplified to the three-electrode audion and also the relation of that type of connection to the connection that is used when energy to be amplified is applied to an ionic relay eon-trolled by a magnet. It did not require in-ventivo thought for an engineer to know that in connecting two audions in parallel circuits you must apply the input energy to the grid filament circuits of both of them, and that, if the parallel circuits are to act in push-pull re-ktion, you must reversely connect the input circuit to the two audion grid filament eir-euits. This is all that is meant by the divided input circuit referred to in claim 1 of the Col-pitts patent. He put at the end of the input eirCuit two input transformer coils just as the carbon button repeaters and the ionic repeaters had put at the end of the input circuit two magnet coils. The secondary coils of the two i21put transformers were then connected reversely, one to each of the two audion grid filament circuits. In this way the straight audion hook-up was substituted for each branch of the prior art push-pull circuit, Thich was of common knowledge,

Plaintiffs’ expert said it was a fact that there was no other way which two audions could connected in a push-pull circuit than this well-known and indispensable way.

^ There was no invention in what Colpitts accomplished, and it was error below to sustain the validity of this patent.

*730Arnold Patent.

This patent No. 1,504,537, for power limiting amplifying device, filed September 3, 1915, and issued nine years later, August 12, 1924, contains two groups of claims: (1) 17, 18, and 20; (2) 33, 34, 35, and 36 — all of which were held valid and infringed below. Claims 17,18, and 20 are referred to as the high impedance claims, and 33, 34, 35, and 36 as push-pull loading claims.

The high impedance claims define an amplifier comprising two electrical discharge devices; input and output circuits therefor, the devices being symmetrically and oppositely disposed with respect to each of the circuits (push-pull arrangement); space current paths for said devices known as plate-filament circuits, and "said paths having a common external portion of high impedance.”

The object of the Arnold patent and these claims was stated in the original specification to be as follows: "This invention relates to receiving systems for radio communication, particularly to devices for limiting the electrical power which may be transmitted to a receiving instrument in such a system, and more particularly to devices in which such limiting action is obtained by employing electric currents in an evacuated vessel.”

The inventor disclosed a high electrical resistance, and the defendant uses a high electrical inductance. The term "impedance” is inclusive of an electrical inductance as well as of an electrical resistance, and the purpose and effect of high resistance in the common path is power limiting, very different from the stabilizing purpose and effect of a flywheel inductance. Arnold accomplished this limiting by placing a resistance in the common path of the push-pull amplifier and by adjusting this resistance to the strength of the signal, producing what he called a stopper. The object of the resistance was to cut down or limit the amount of output which the tubes would give when subject to static discharges to the antennas, and the adjustment was so made that “all signals lo.uder than the speech signals will be cut off and hence the accidental noises cannot be louder than the speech signals.”

The defendant uses a push-pull amplifier which has in the common path in the plate circuit an inductance coil. The plaintiffs-contend that the function of the inductance coil is the same as the resistance in Arnold. The defendant says that the purpose of the inductance coil is the same as in the prior art telephone repeaters; that is, to enable them to give greater amplification without distortion. But there are several prior uses shown.

The prior patent to Stragiotti, No. 958,-961, granted in 1910, states the problem Arnold sought to cover, and presents its accomplishment. Stragiotti says (page 3, line 8) : “The principal feature of this application is the use of an auto-induction coil in series with the differential microphone which on this account becomes more éfficient as will now appear. The changes of resistance which originate in the two branches of the microphone as well as the induction in the two transformers never compensate each other. The balance acts as to make the main current oscillate while it is intended to remain constant; or in other words, together, with the constant current a new oscillatory current flows. If this current divides in the two branches of the microphone into two equal parts no current is originated in the secondaries of the transformers; but as the two branches are never in exactly the same condition this current does not divide into two equal parts, consequently a troubling current is originated in the secondaries together with the telephonic current. The advantages of the auto-induction coil consequently becomes apparent, for by it the current flowing to the central electrode of the microphone is maintained exactly constant thus obviating all troubles.”

The claim allowed Stragiotti was for the combination of a telephone line with a bridge circuit comprising a differential microphone in series with an impedance coil. This patent shows the arrangement and the effect of the stabilizing, or flywheel inductance, used in the defendant’s amplifier, and justifies its use. Stragiotti’s stabilizing inductance became well known in the push-pull repeater circuits, and using Stragiotti’s flywheel inductance in a push-pull circuit containing audions was not a new combination for which Arnold was entitled to a patent.

The charge of infringement is based upon defendant’s use of this, which was old, and it makes little difference whether the high impedance claims of the Arnold patent are or are not based upon a disclosure out of which the defendant’s uses could have emerged. Burr v. Duryee, 1 Wall. (68 U. S.) 531, 17 L. Ed. 650; Grubman Engineering & Mfg. Co. v. Goldberger, 47 F.(2d) 151 (C. C. A. 2). Even in the push-pull amplifying circuit such as Colpitts, there could be no invention in inserting an inductance in the common path to cause the current to divide equally between the two tubes by means of the Stragiotti flywheel inductance. Nor can this in-*731ductanee be regarded as tbe equivalent oí the resistance in the Arnold arrangement. The purpose of the Arnold arrangement was to cut down the amplifying power of the tubes to be just equivalent to a signal. The object of the inductance, in the defendant’s appara- , tus, is to prevent the nonbalanee of the currents in the two halves in the push-pull circuit and to make it a more effective, not a less effective, amplifier. Apparently Arnold’s purpose was to limit the output of a device under exceptionally heavy atmospheric disturbances. Defendant’s purpose was to eliminate the, unbalanced currents in the two circuits on the jiusli-pull sides of the systems.

The push-pull loading claims, 33, 34, 35, and 36, broadly cover the two stage amplifier in which the first stage was a single vacuum tube and the second stago'is two vacuum tubes connected in push-pull relation — -the De Forest audion feeding into a push-pull amplifier of Colpitts. It was old to connect two single amplifiers in cascade (Yon Lichen and De Forest patents). It was not invention to use a two-stage amplifier in which each si age consisted of two tubes in push-pull relation (Colpitts) so that the advantages of the push-pull circuit were obtained in each stage. It was not invention to dispense with one tube and thereby .leave out the advantages of the push-pull relations in one of the stages. The illustration of Arnold in which a single tube feeds into a push-pull circuit is the engineering way of putting a small amplifier first and a large amplifier afterward. Nor is there merit in the argument as to loading to the greatest possible point. This amounts to nothing more than to say that an engineer puts a little amplifier first but a big amplifier of whatever type it may be, after it. It is an inherent characteristic of a two-tube amplifier that it has a greater amplifying power than a single tube. We think this patent should he held invalid in view of the prior art.

Mathes Patent.

This patent No. 1,426,754, circuits for electron discharge devices filed October 25, 1916, and issued August 22, 1922, was held valid and infringed as to claim 25. It relates to the source of the direct current potential for the input circuit of the elect ron discharge device of the audion type, and it is referred to also as a grid biasing potential. It discloses the arrangement of a biasing resistance in a filament — grid circuit of the audion and a method and means for compensating for fluctuations in the potential of ihe output circuit battery of the vacuum tube. The defendant says that the portion of its amplifier charged to infringe this claim is its grid biasing resistance IÍ3 (see Diagram, Record, p. 1394), through which the current from the power supply in the plate-filament circuit passes on its way to the filament. The potential drop across this resistance, due to the passage of the plate current there-through is applied to the grid as a grid biasing potential. In tbe defendant’s system the grid biasing potential is derived from the plate current source. It is not of the variable source. A storage battery runs down and requires re-' charging. If there is any variation in the power source, there will be a corresponding compensation of the grid biasing potential, since tbe grid biasing potential is derived from the power source. The compensated impulse of the Mathes patent is present whenever the grid biasing potential is derived from the plate current source. This was commonly used, prior to the patent, as the source of derivation of gj'id biasing potential. A direct derivation of the grid biasing potential from tlie plate current source used by the defendant is shown in the Colpitts and Arnold patent. Plaintiffs’ expert admits that the Arnold and Colpitts patent would be a complete anticipation of Mathes if it indicated a condenser around the biasing resistance. The difference between this arrangement and the one used by the defendant is that the defendant’s biasing resistance is shunted by the capacity C3 which permits the high frequency signal energy to pass around the biasing resistance without going through it, whereas the direct current power supply cannot pass through the eondsenser and must pass through the resistance Eg. By by-passing the signal around the resistance R3 the defendant avoids that amount of loss of signal power that would result from passing the signal through the resistance as in Colpitts and Arnold. It is on this difference from Colpitts and Arnold that plaintiffs assert the right to enjoin the defendant.

Mathes’ disclosure does not make direct use of the output circuit battery or plate circuit battery as a source of grid biasing potential; the defendant does. When the plate circuit battery is used directly as a source of grid biasing voltage, it is necessarily compensated when any variations in the plate current source occurs. Mathes had in mind the problem of effecting an equivalent compensation when the grid biasing voltage was derived indirectly from any oilier source such as the filament heating battery or the sep*732arate grid battery. He explains that the output (plate circuit) -will undergo fluctuations in the ease of sudden or gradual internal changes within the power battery and that such fluctuations are troublesome; that the method by which his invention overcomes this error assumes that the current from the input circuit comes from a battery in series with the output circuit and if the input and output batteries are of similar nature this change will in all probability also be subjected to like changes, and in this way a fluctuation in the plate circuit will be accompanied by compensating variations of the grid biasing potential. But it is too much to say that the compensating variation is derived from an effect produced by variations of this .source, when the only derivation disclosed is the expectation that variations of the plate battery will in all probability be accompanied by like variations in the other battery (A or G battery).

But in the patent in suit the grid biasing potential is derived from a battery.separate from the plate circuit battery and in tne specifications there is nowhere any suggestion of supplying the grid biasing voltage from a plate current source. It therefore follows that the defendant has no't adopted Mathes’ system, but has gone to the prior art. The use of the condenser around the grid biasing resistance is common to Mathes and defendant, and it is the one thing that is common to both. The use of the resistance as a means for effecting grid bias was invented shortly before Mathes’ application. The defendant should not be held to have derived the use of this shunting capacity from Mathes. The use of a condenser shunting the grid biasing resistance is not a part or claimed as a part of Mathes’ invention. These differentiations make it clear that by no liberality of interpretation can we stretch claim 25 to read upon the defendant’s device, if, indeed, it is not anticipated by the patent to Colpitts and Arnold above referred to.

Since we have sustained the Lowenstein patent and held it to have been infringed, it becomes necessary to construe the defense interposed that producers were licensed to make use of that patent with the others in suit. The defendant is charged with infringement on account of the use of reproducing apparatus in the commercial reproduction of sound from phonograph records. Four producers of talking motion pictures had licenses under the Lowenstein patent, as well as others, from the plaintiffs. It is claimed that not only did they make the sound phonograph; records, but also distributed them among exhibitors for commercial reproduction. It is conceded that the defendant was licensed. He paid a rental to the Vitaphone Company, who in turn paid a license fee to the plaintiffs. The provision of the license relied upon by the defendant refers to the reproducing apparatus which was not licensed to the defendant by any of the plaintiffs. The plaintiffs, retained the right to prevent the producers from distributing their records to. any one-who would not reproduce them satisfactorily. There appears to have been no intent, either expressed of implied, to grant any exhibitor-the license to use any reproducing apparatus whether its reproducing qualities were good or otherwise. The producers acquired no right under the license to grant reproducing licenses. Their rights were confined to recording. The royalty is paid by the producer.. The defendant has no license by agreement or acquiescence of the plaintiffs. The contractual provisions of a license contract to insure-good quality of reproduction does not authorize the producers to grant licenses to this-defendant in terms relieving him from the charge of infringement of the patent. There-is a vast difference between recording and reproducing. The licensor, products company, received no rights from the telephone company which would have enabled it to grant a license to the defendant, except to-use the plaintiffs’ own apparatus when furnished by the products company. This defense has no merit.

The decree below will be modified as indicated above, holding the Lowenstein patent valid and- infringed, and holding the other-patents invalid and not infringed, as indicated in this opinion.

Decree modified.