General Electric Co. v. De Forest Radio Co. , 23 F.2d 698 ( 1928 )

MORRIS, District Judge.

These two suits, of General Electric Company against De Forest Radio Company, were tried together. In the first suit all the claims of patent No. 1,558,436, granted to Irving Langmuir, assignor to the plaintiff, October 20, 1929, for “electrical discharge apparatus and process of preparing and using the same,” upon an application filed October 16, 1913, and renewed March 14, 1916, are in issue. It is known as the high-vacuum tube patent. Infringement is conceded, if the claims are valid. Validity is denied for want of invention and novelty, and because of prior knowledge and use, for double patenting, and upon the ground that Harold D. Arnold, not Langmuir, was the first inventor.

The specification, claims, and testimony are so interwoven with scientific theories, more or less abstruse, that it is somewhat difficult to isolate and keep separate from the principles of physics involved Langmuir’s conception and idea of means. Yet an analysis of the Langmuir device and its elements discloses that, apart from its functioning and results, it is nothing more nor less than a Fleming valve or Do Forest audion, for example, in which a higher vacuum, maintainable during operation, exists. How much higher the vacuum must bo to bring the tube within the patent is not definitely fixed in terms of pressure. The specification states merely that “the evacuation of the device should be preferably carried to a pressure as low as a few hundredths of a micron, or even lower, but no definite limits can be assigned.” A micron is a pressure equal to a millionth of a meter, or a thousandth of a millimeter, of mercury. It corresponds to a barometer reading of that value. The normal barometer reading in the open air is about 760 millimeters.

The justification for the failure to assign definite pressure limits is that the invention “comprises devices in which the electrical current is carried by negative charges called electrons, emanating from the cathode, independently of gaseous ionization” — a device in which the residual gases play no part in the operation. The results desired to be obtained are stability and reproducibility of operation and increase of power. The production of these results by the specified manner of functioning is dependent, not on the degree of vacuum alone, but upon four coordinating variables — geometry or relation of parts of the tube, temperature of cathode, voltage of anode, and degree of vacuum. This fact makes it apparently unavoidable that the claims be functional in form. But they must, of course, be construed to cover the tiling. Walker on Patents, § 183.

Claim 26 is comparatively simple. It is:

“A discharge device comprising a sealed-off envelope and electrodes therein, the envelope being shaped and the electrodes located so that the energy of the discharge is delivered mainly upon one or more of the electrodes, the inner wall of the envelope and electrodes being so free from occluded gas and the pressure in the tube being sufficiently low, with respect to the spacing of the electrodes and the energy of the discharge, so that the device is capable of passing a sustained discharge of one-tenth of a milliampere at forty volts, without developing any substantial positive ionization effects.”

The Fleming and Do Forest tubes are “discharge devices comprising a sealed-off envelope and electrodes therein, the envelope being shaped and the electrodes located so that the energy of the discharge is delivered mainly upon one or more of the electrodes.” The next clause of the claim — “the inner wall of the envelope and electrodes being so free from occluded gas and the pressure in the tube being sufficiently -low, with respect to the spacing of the electrodes and the energy of the discharge”- — has to do with the manner and degree of evacuation. Underlying this clause are these facts:

The glass wall and the electrodes harbor large volumes of occluded gases. Heat is necessary to expel these gases, so they may be removed by the pump. If not removed before the tube is sealed off, the heat occurring during the operation of the tube expels them, destroying in large degree the vacuum. “The spacing of the electrodes” — one illustration of the geometry of the tube — is a factor, because the permissible gas pressure in the tube increases with the proximity of the electrodes to each other. This varying relation is due to the fact that the opportunity for the electrons to collide with gas molecules varies with the length of the path the electrons have to travel.

“The energy of the discharge” may be interpreted’in terms of power, which is the product of the current values and the voltage. The unit of current flow is the ampere, or the current which flows at a pressure of one volt through a closed circuit having a resistance of one ohm. The power of a circuit having a pressure of one volt and a current of one ampere is one watt. The power output of the device of claim 26 is .004 of a watt, or 4 milliwatts. An increase of energy means an increase in the number and velocity of the electrons passing from eathpde to anode, and consequently an increase in the number and effect of collisions of electrons with gas molecules. It follows, of course, that, if the maximum voltage at which a given tube will function without ionization is increased, ionization will occur and the tube fail to function in the manner specified in the patent.

“Ionization,” by which is meant the conversion of atoms or molecules into ions, is explained by the “electron theory.” By this all atoms known to chemistry are deemed to be made up of a nucleus endowed with a permanent positive electrical charge, around which one or more electrons, each of which has mass or weight, and carries, or is, a natural unit of electricity, revolve at inconceivably great speeds in regular orbits in somewhat the same manner as that in which the planets of our solar system revolve about the sun. According to this hypothesis the chief difference between the chemical atoms lies in the number and arrangement of their electrons. Under normal conditions the electrons are maintained within their atoms and in their orbits by the attractive force of the positive nucleus, which, electrically, exactly balances and neutralizes the negative charge of the electrons of the atom. Consequently every atom having its normal complement of electrons exhibits no electrical effects..

If, however, by impact or otherwise, an atom loses an electron and becomes an ion, it is thrown out of electrical balance. The positively charged nucleus predominates, and exhibits a unit positive charge. In thermionic tubes, ions are created by the impact upon the gas atoms or molecules of some of' the myriads of electrons constituting the current flowing from the hot cathode to the cold anode — the Edison effect. See Westinghouse Elec. & Mfg. Co. v. De Forest Radio Tel. & Tel. Co. (C. C. A.) 21 F.(2d) 918. To dislodge an electron from its atom, the colliding electron must have sufficient momentum. The velocity required for this momentum is given to the electrons mainly by the pull or attractive force of the positively charged anode. The strength of the pull increases, other conditions being stable, with the anode voltage. For the gases with which we are here concerned 15 to 25 or 30 volts constitute the ionizing voltages.

When substantial numbers of atoms or molecules are being converted into positive ions by the electron stream, the “ionization effects” forbidden by claim 26 occur. One effect of ionization is an increase in the current across the tube over the value it would have at the same voltage in the particular tube, were the vacuum sufficiently high to prevent ionization. Such increase is due to two facts:

First, the electrons, being negative charges of electricity, tend, in the absence of a countervailing force, violently to repel one another, and thus to diminish the number, and so the value of the current, passing to the anode. Positive ions in the field constitute a countervailing force, and consequently expedite the flow of the electrons to the-anode. Second, the positively charged ions are drawn with great force to the cathode, the negative electrode, bombard it, and thus increase its temperature. As the electron emissivity of the cathode increases with its temperature, the ion bombardment of the cathode serves to increase the supply of electrons, and consequently the current passing to the anode. This bombardment is so severe that, if continued, it destroys the filamentary cathode.

Another effect to which substantial ionization usually gives rise is a blue glow in the .tube. This, as I understand it, is accounted for upon the hypothesis that, as the electrons are shifted in the atom, by impact of the electrons, from one orbit to another, at prodigious speeds, they produce waves or vibrations in the ether, which constitute light. The ultimate effect or result of ionization is an unstable, nonreprodueible, and, in a measure, noneontrollable current. As it is necessary, in order to prevent the destruction of the cathode and obtain a stable current, to operate gassy tubes below ionization, their power output — the product of current and voltage — was limited. These eonstitute’obvious defects.

These defects are not present in the tube of the patent. It operates without substantial ionization. When there is no substantial ionization, the current increases and decreases in a fixed relation to the voltage of the anode, unless and until the anode voltage is made sufficiently high to draw to the anode all the electrons emitted at the cathode. The electronic emissivity of a given cathode in-, creases with its temperature. This fact is known as “Richardson’s Law.” When a voltage high enough to carry across the tube all the electrons emitted is reached, a further increase of voltage, which has no power to increase the supply of electrons of which the current consists beyond the limit of einissivity imposed by the temperature of the cathode, produces, of course, substantially no change in current. The current that is thus substantially independent of the voltage is known as the saturation current.

Operation of the tube above and below this critical voltage is spoken of as in the saturation region and below saturation, respectively. If a tube is operating below saturation, an increase of the electron supply by an increase of cathode temperature, but with no increase in voltage, produces no marked change in the current. Decreasing the electrons emitted by the cathode, by decreasing the cathode temperature, is likewise without effect upon the current, provided, of course,that the decrease in the supply of electrons is limited to the electrons emitted in excess of those pulled across the tube by the given voltage. The fact that, if the anode voltage remains stable, the current cannot be increased by an increase of the cathode temperature, beyond that required to supply the electrons capable of being pulled across the tube at the given voltage, is illustrated by the “flat top temperature curve” on the following graph :

image

The reason that the current does not increase beyond a certain point, dependent upon anode voltage, with the increase of the temperature of the cathode, is that the electrons, all being negatively charged, exert powerful repelling forces on one another. When this force of repulsion equalizes the pull or effect of the voltage on the anode, the anode is without power to draw to itself a greater number of electrons. The current-limiting effect of the electric field of the electrons on one another is known as the “space charge effect.” It is space charge which produces the flat top of the temperature emission curve in the foregoing graph.

When the temperature of a cathode is fixed and constant, and the voltage is increased, the limitation upon the electrons limits the current. This is illustrated by the following graph, whereby it appears that the last few increases in voltage make- no increase in current-there being no more electrons to draw across the tube:

image

The curve of this graph, which, when plotted on logarithmic paper, appears, up to the point approaching saturation, as a straight line, discloses a relation existing between current and voltage. That relation is expressed in the specification thus:

“In most devices of simple construction which embody my invention, the discharge current passing through a given space with the cathode at a sufficiently high temperature with respect to the voltages employed, varies directly with the 3/2 power [square root of the cube] of the impressed voltages.”

A to D of the curve on the current-voltage chart, in which the temperature is constant, is the space charge part, or that in which the current is limited by anode voltage, and follows the 3/2 power law. The flat top represents the saturation part, or that in which the current is limited by the cathode and its temperature. The converse' is true in the first or current-temperature chart, in which the voltage is constant. In that the rising part is the part limited by cathode temperature, while the flat top is the part limited by space charge.

, Another characteristic peculiar to the tube of the patent, as stated in the specification, is that the device is “practically independent of the pressure when the pressure is below a certain value, which depends upon the size and shape of the device and upon the current and voltage of the discharge.” That statement is undoubtedly true, if it be interpreted to mean that if, in a given tube, the pressure is low enough to enable the tube to function at a given voltage without ionization — that is, as a pure electron discharge device — a further 'decrease in the pressure will not affect the discharge. It is likewise true that, if the pressure is less than that required for a given voltage in. a given tube, it may be increased up to the pressure immediately below that at which substantial ionization would occur at the given voltage without affecting the discharge.

The discharge, however, is in no sense independent of the pressure if the maximum voltage at which a given tube will function without ionization is increased, for ionization would instantly occur, and the tube fail to function in the manner specified in the patent. In such instance a decrease in pressure simultaneously with the increase in voltage would be necessary to avoid ionization. In its simplest aspect, this is due, of course, to the fact that, if ionization is to be avoided above ionization voltages, collisions must be avoided between electrons and gas molecules, and that the number of collisions in a given tube, dependent upon the relative density of the two, is increased, the pressure being stable, by an increase in the number of electrons pulled to the anode by the increased voltage. ,

These scientific explanations, arising mainly out of the 'electron theory, of the functioning of the high vacuum tube, change not at all the fact that it differs structurally from the Meming valve and De Forest audion only in the degree of vacuum maintainable during operation. This was expressly conceded during the pendency of the application in the Patent Office. The fundamental questions here in issue are not changed, and their solutions are, I think, but little affected, by the theories by which the functioning of the high Vacuum tube is accounted for. If it was not known to those skilled in the art how to produce, in a Fleming or De Forest structure, for example, a vacuum sufficiently high to enable the tube to function above ionization voltages without substantial ionization, and Langmuir solved the problem and made the necessary disclosure of its solution in his patent, manifestly he would be entitled to a patent for his disclosed method and device as a reward, quite apart from his knowledge or lack of -it with respect to scientific principles or laws of physics explaining the functioning brought about by the high vacuum.

Moreover, independently of whether or not invention was necessary to create a vacuum of the required degree, Langmuir would be entitled, I take it, to a patent for his device, regardless of his ability or inability to explain its functioning in principles of physics, if he was the first to appreciate that a high vacuum maintainable during operation in a Fleming or De Forest 'structure, for example, would constitute means -whereby the results produced by the high vacuum tube could be accomplished, provided, of course, he reduced to practice with reasonable diligence the device of his mind.

Upon the question whether the claims for the process of making the high vacuum tube or those for the tube itself, which, because of its stability, reproducibility, and power, has made possible radio broadcasting, modem radio reception, and long-distance telephony, involve novelty and invention, the parties differ widely. The plaintiff asserts, with apparent confidence, that the patent is for invention of a high order — equal to or greater than that of De Forest for the three electrode tube. The defendant, with equal confidence, finds not only all that Langmuir did, but all that he discloses with respect to the principles in accordance with which the device functions spread upon the records of the prior art. Since the Fleming valve and the De Forest audion were old, and the structure of the tube of the patent differs from these only in the degree of vacuum maintainable during operation, the first crucial question is whether claims 30, 31, and 32, directed to the process of making tubes having a vacuum of the required degree, are valid. Claim 31 is: ■ '

“In the process of producing a hot cathode discharge device, treating the interior surface of the envelope and the operating parts to liberate therefrom occluded gas, and removing from the envelope gas liberated by said treatment and free gas originally contained therein, continuing these operations until the occluded and original free gas is so thoroughly removed as to enable the finished sealed-off' device to operate below saturation in a stable and reproducible manner at a voltage above 50 volts and with a current greater than one milliampere and with the energy of the discharge delivered mainly at one or more anodes.”

To find wherein, if at all, this and the other process claims go beyond the discoveries and ideas of others, beyond the use of wise judgment in selecting and combining those discoveries and ideas, or beyond mechanical skill in applying them to practical results, the specification and the prior art must be examined. The specification states:

“For the evacuation of the device the glass walls of the tube are carefully heated to as high a temperature as the glass will stand without softening, and in general the most approved methods of incandescent lamp exhaust are used. The evacuation of the tube preferably while still heated is carried out by means of a suitable evacuating means, for example, a Gaede molecular pump, which removes vapors as well as gases. Chemical evacuating means, such as electrically vaporized calcium or magnesium, may also be used. Either before or during the evacuation the anodes may be heated, especially when the anodes are to be run at elevated temperature during the normal operation of the device. In this case the temperature is preferably carried close to brilliant incandescence. In the case of the structure shown in Fig. 1, the heating may take place by passing an electrical current through the wire 12. The heating, especially of solid anodes, such as shown in Figure 2, may take place in a suitable vacuum furnace; the temperature preferably being raised to 2,500° C, or even higher. Bombardment is a very effective means of removing occluded gas from anodes.”

The steps and precautions requisite to successfully evacuate by bombardment are then described, after which the specification adds:

“After the metal has been freed from occluded gas, reabsorption of gas will not readily take place, even though it is exposed to the air or other gases. For example, anodes thus treated may be removed to other apparatus, which then may be evacuated with less electron bombardment of the anode or anodes. * * * It is also true that, when the anode has been carefully freed from gas, residual free gas, even if present in a sufficient amount to cause some gas ionization when the apparatus is first started, does little harm, as it is quickly removed by the gas clean-up effect when the device is operated.”

The prior art, however, as I understand it, discloses that, long before the earliest date claimed for Langmuir, it was known, not only that large quantities of gas are occluded in the glass walls and electrodes of an electrical discharge device, and that unless removed before or during evacuation the gas will be driven out by the heat of operation, and thus detrimentally increase the pressure of the vacuum, but also that the occluded gases could be removed and the vacuum created and maintained by the very steps prescribed by Langmuir for removing the occluded and free gases of his device. In May, 1896, the American Electrician published an article by Thomas Duncan in which he said:

“If an incandescent lamp be exhausted until no air can be seen escaping through the valves of the pump, and the filament then incandesced, considerable quantities of air will be expelled at each operation of the pump, thereby showing that this air has been occluded in the pores of the carbon filament, and could not have been gotten rid of without heating the filament and driving the air out into the vacuous space surrounding it.

“This heating also assists in drawing off the air which clings or adheres against the glass upon the inside of the bulb. * * * If the electrodes are made from aluminum or other porous metal, the secondary terminals from a spark coil can be applied to them, and set in operation in a similar manner to that in which the tube is to be used after-wards. This cannot help but be a ready and satisfactory means of eliminating the air occluded within the electrodes.”

A resume of “Incandescent Electric Lamp Exhausting” by S. E. Doane, published in the Electrical World and Engineer of May 21, 1904, says:

“The exhausting of' incandescent electric lamps has developed into a highly specialized branch of the lamp manufacture. * * * The completed lamp must not only be free from gases and vapors, but must also have had those parts capable of giving off gases and vapors, thoroughly freed from them. The only way we knew to accomplish this was to highly heat everything we wished to free from gas. We had three things to heat: The glass, the filaments, and the joints between filaments and leading-in wires.

“The glass was heated by an asbestos cone-shaped hood painted black on the inside. The heat from the lighted lamp furnished all the heat necessary. The lamps could be heated hot enough to cause the bulb to collapse if the hoods were made too thick. The filaments were also heated sufficiently, by the current. * * * ”

Dwyer, in his patent No. 596,694, granted January 4, 1898, for a process of producing high vacuums in incandescent electric lamps or similar receptacles during their manufacture, recites that in the past such lamps have been exhausted by mechanical pumps, “by chemical processes involving the displacement or absorption of the gases contained in the lamp, or by combinations of these processes.” He says that a partial vaeuum is readily obtained by a good mechanical or mercurial pump, but that complete exhaustion by this means alone is not possible. He then adds:

“To carry the vaeuum still higher, it has been a common practice to heat the lamp bulb by either an external source of heat, such as a gas flame, or by an electric current passed through the lamp filament, or by both of these means combined, so as to expand the gases in the lamp, and also to drive off from the lamp filament itself, and its supports, moré particularly the joint between the leading-in wires and the filament, which joint is usually made of some form of hydrocarbon paste, gases occluded therein, while continuing the pumping operation.”

Soddy, in his patent No. 859,021, of July 2, 1907, for the employment of certain reagents in the process of producing high vaeuum, declares:

“In the production of high vacua, not only has the air to be removed, but also the gases occluded or condensed, on the walls of the vessel being exhausted and on the contained electrodes or filaments or other contents of the said vessel, which condensed gases are given up slowly after exhaustion, and if not removed during the exhaustion cause the vaeuum to deteriorate. * * *

“The heating and expelling of the condensed and occluded gases may be effected in any usual, or suitable, manner of expelling condensed and occluded gases from the walls and contents of vessels being exhausted. The usual manner of effecting such an operation is to heat the vessel while it and its contents are subjected to the same treatment as they are subjected to in after use. In incandescent electric lamps, for example, an eleetrie current is passed through the filament, and in X-ray bulbs an electric discharge is passed between the electrodes, during the continuation of the exhaustion.”

Many other references disclose the baking of the walls of the envelope of a discharge device, the heating of the electrodes, incandescing the filament, and electron bombardment of metallic electrodes to expel occluded gases. The gas clean-up effect was likewise old at the time Langmuir began his operations. One instance of its disclosure is in patent to Thatcher, No. 1,028,636, granted June 4,1912, application for which was filed March 30, 1910, for method of exhausting vessels, in which it is said in part:

“I institute the intensive incandescence of the filament proper;1 that is, I bring the current voltage up to a sufficient intensity to produce the blue haze phenomenon, and maintain the same until the blue haze disappears and the desired almost perfect vaeuum is attained. * * * ”

Dr. E. Leon Chaffee testified:

“It has long been known that the existence of this blue glow improves the vaeuum, probably by bombarding the walls with gas ions and consequently causing them to stick to the glass walls. * * * ”

These steps of the prior art are the identical1 steps disclosed by Langmuir. Nor are his disclosures with' respect to how these steps are to be employed any more ample or complete, with the possible exception of heating by bombardment. But bombardment has no saving power, for the claims are not restricted to heating by bombardment. They are broad enough to include any method of “treating the interior surface of the envelope and the operating parts to liberate therefrom occluded gas, and removing from the envelope gas liberated by said treatment and free gas originally contained therein.”

The plaintiff, however, places some reliance upon the clause in the claims directing continuance of the steps until the desired vacuum is attained, and cites Badische v. Halle (C. C. A.) 104 F. 802. But in that case those skilled in the art were of the opinion that a more prolonged or intense application of the old steps would not accomplish the desired result. Such is not the evidence here, as I understand it. Moreover, the tools of the evacuating art were constantly improving. In the early days of the.incandescent lamp art, manufacturers used the Sprengle pump to obtain their vacua. It gave way to the Geissler pump, which, in its turn, was superseded or supplemented by chemical evacuating processes. It is, I think, not denied that these old pumps and at least some of the chemical processes left in the tube mercury or other vapors having a pressure too, high for the tube of the patent. But, if Langmuir makes in his specification any disclosure of steps or means to avoid that result, it is by the use of the new, improved Gaode molecular pump. That pump is highly efficient, and, when properly used, leaves no free gas or vapor that will prevent the functioning of a thermionic tube above ionization voltages without substantial ionization. But Langmuir is entitled to no credit for that. I think the process claims are invalid.

This conclusion carries with it as a necessary corollary that the device itself, apart from its functioning and use, is lacking in patentable novelty. The inventive quality of the deviee claims, if any they have, must consequently reside in the discovery of a new principle that thermionic discharges above ionization voltages can be produced without substantial ionization, if the vacuum he sufficiently high, and the disclosure that the device of the claims constitutes suitable means for putting the principle into practice to obtain useful results, or, in the application of that principle, if known, lo a useful purpose —in the realization that the instability, nonreprodueibility, and lower power output of the gassy tube was due to inadequate vacuum and the conception of a tube with a higher vacuum as a means whereby stability, reproducibility, and high-power output in thermionic devices might be had.

That Langmuir was not the first to discover that a pure electron discharge, a discharge without substantial ionization, can be produced in a thermionic deviee, if the vacuum is sufficiently high, was indicated by Richardson’s law, evolved and formulated about 1903. If, perchance, that law was not universally accepted by physicists as being operative in high vacua, as was suggested by Soddy, Pring, Parker, and some others, the publications of Lilienfeld in 1910 were clear, complete, and specific upon this point. In fact, Langmuir admits that Lilienfeld had thermionic discharges above ionization voltages, “substantially free from the effects of positive ionization.”

Moreover, if a thermionic discharge is unaccompanied by substantial ionization, all the asserted “earmarks” or characteristics of plaintiffs deviee are present, for they are but incidents flowing from a sufficiently high vacuum for the voltages employed, or, otherwise expressed, from an absence of ionization. Those characteristics are, or include, (1) regularity and reproducibility of operation; (2) 3/2 power relation of current to voltage; (3) no blue glow or other visible evidence of discharge; (4) cathode not heated by the discharge; (5) current limited by space-charge effect; and (6) discharge independent of pressure in vacuum below certain pressure in the sense hereinbefore stated.

As I understand the evidence, these “earmarks” are likewise characteristics of the Fleming valve and original De Forest audion when operated at nonionizing voltages. It thus seems obvious that the merit of the invention, if any, must lie in the realization that tho known pure electron discharge to he had in a high vacuum would overcome the special difficulties and limitations of thermionic discharge devices. Whether this realization and the increase of vacuum in structures of the Fleming and De Forest type was a probable natural development of the art, or one requiring inventive skill, is, like the question of the presence or absence of invention in other cases, not without some difficulty.

One class of evidence, which often carries much weight, is the necessity for a series of experiments before the result can be arrived at. Such evidence raises a strong inference that the result was not obvious, and constitutes a real addition to public knowledge. But for such experiments to he of much value the patentee must have started his experiments with knowledge of the state of the art, or his experiments may have succeeded in enabling him to redisclose only what was already known. Langmuir made many experiments over a long period of time. Those that were first made were, however, for the purpose of discovering abstract principles, and not for the purpose of arriving at practical results. He made discoveries, but he did not immediately, or for some time, perceive that they had any practical use. Moreover, the discoveries to which his experiments led him, so far as we are here concerned, were those already made by Lilienfeld and others.

Great credit, bearing upon the issue of invention, is claimed by the plaintiff for Langmuir for discovering from his experiments that electrons, will boil out of a hot cathode in a high vacuum, that in a high vacuum the current varies with tho 3/2 power of the voltage, and other characteristics or incidents of a discharge in a high vacuum without substantial ionization. But disclosures of the prior art reveal that Langmuir, great and renowned physicist that he is justly conceded to be, was, in this instance, traveling, all unknown to him, over a field already well explored. Fleming, the inventor of the Fleming valve, published in the Scientific American Supplement for January 20, 1906, an article on the electric conductivity of a vacuum, disclosing, not only that electrons are emitted by a hot cathode in a high vacuum, but also “that a high vacuum may be a very good conductor, provided that the negative electrode is rendered incandescent. This shows that the obstruction which a high vacuum offers to the passage of electricity is not due to the mere reduction of the matter in the space (that is, to the rarefaction of the gas), but is, in some way or other, largely dependent upon the temperature of the electrodes; the resistance to the passage of electricity through a vacuum can be rendered very small by intensely heating the negative electrode”— thus revealing that a current could be had passing from cathode to anode.

In addition, Fleming stated: “If we take measurements of the current passing through the vacuum, when the negative electrode is an incandescent carbon filament and the other a cold metal plate, and, at the same time, measure the difference of potential between the electrodes, we find that the conductivity, as measured by the ratio of current to- voltage, is not constant, but varies with the electromotive force.” Furthermore, that Fleming knew the difference between a high and low vacuum is made manifest by his statement: “The vaéuum is said to be low when the original normal air pressure in it is reduced to one-thousandth atmosphere, or it may be only to one three-hundredth. It is said to- be high when it is reduced to one hundredth-millionth of an atmosphere.”

It is true, as shown by the Fleming curve sheet, that Fleming did not employ in his valve a high vacuum, but it is also true that Langmuir conceded: “There is no advantage in the device of the character of the Fleming valve to go to pressures of less than 50 microns.” Both C. D. Child, in an article in • the Physieal Review, vol. 32 (1911), and Lilienfeld, in an article entitled “Conduction of Electricity in Extreme Vacuum,” published in Annalen der Physik, vol. 32 (1910), disclosed that with the aid of certain simplifying assumptions, such as the ignoring of the initial velocity with which the electrons are boiled out of the cathode, the discharge current varies in .accordance with the three-half power of the voltage.

This is conclusive evidence, as I understand it, that in a high vacuum the current is under control, stable, and reproducible, and, since high voltages were employed, that high-power levels may be attained by the employment of a high vacuum. “Space charge effect” was, likewise, but a rediscovery. This is made-certain by an article by Lilienfeld in Physikalische Zeitsehrift, 1908, in reply to an article by Soddy dealing with the Wehnelt cathode in high vacuum, in which Lilienfeld said:

“ * * * It is not exclusively the electrons emitted from the oxide cathode which, "as was often assumed, determined the discharge in high vacua, but that the discharge depends essentially on a second factor. I pointed out more specifically in my second paper this view, that in consequence of the absence of the last traces of gas the number (volume density) of the negative electrons supplying the current becomes great in comparison with the number of gas molecules per unit volume, so that therefore the tendency exists for the formation of the greatest possible negative space charge — which opposes the current transport by the negative electrons.

“The higher the vacuum, the greater the current density, the more prominent this new-kind of discharge becomes * * ”

That the discharge in high vacuum devices is practically independent of the pressure when the pressure is below a certain value was discovered by Langmuir, after Lilienfeld had made the disclosure in his article of 1910, where he said:

“ * * * From a certain maximum value of the gas density downwards the number of positive charges formed per negative charge — i. e., per negative electron — is independent of the gas density. Since we usually regard positive charges as originating from molecules (or atoms), we can put this result in the following form: The number of positive charges per molecule increases continually with diminishing gas density in the region of validity of the function K, which is independent of the gas density, provided other conditions (the potential gradient) are kept constant.”

In view of these publications, it seems to me that the only permissible inference to be drawn from Langmuir’s experiments is not that they constitute evidence of invention, but is rather that Langmuir was not abreast of the developments in this particular art when he made his experiments.

Other evidence bearing upon the issue of invention is the great utility of the high vacuum tube. But in 1912 the radio art was young. Fleming did not need for his purposes, as Langmuir testified, a higher vacuum than that employed by him. The De Forest audion was not in wide use. Langmuir did not see one until 1913. Under such circumstances, it would be difficult to say that the high vacuum tube satisfied a long-felt want, or produced results long sought, or that the evidence of utility is here indicative of anything more than a proper natural development of the art.

Moreover, the fact, that falls but little short of a demonstration, that the employment of the known pure electron discharge above ionization voltages in tubes of the De Forest type was but the availment of those skilled in the art of the store of knowledge that had been accumulated and lay ready at hand, is that, when Arnold, who had had long experience with electrical discharges in high vacuum, was shown a De Forest audion for the first time on November 1, 1912, and saw it in operation, he immediately realized and said, in effect, that by increasing the vacuum the discharge would be sufficiently stable and at adequate power levels to enable the tube to bo employed as a relay device in transcontinental telephony. While not unmindful that invention may sometimes be the result of an instantaneous happy thought, yet the state of the art is strongly confirmative of Arnold’s own view a,t the time that the improvement suggested by him was obvious to one skilled in the art. The act of Arnold turns the scale so overwhelmingly that it lies athwart a finding of invention.

The patent is likewise invalid, in my opinion, because of prior knowledge and use by De Forest. The specification states: “Ordinarily the gas ionization in the audion begins to be important somewhere between 20 and 30 volts.” That, as I understand it, means that at this voltage the current becomes affected by positive ionization, and the tube ceases to function as a pure electron discharge device. Yet, prior to August, 1912, the Federal Telegraph Company of California, by which De Forest was then employed, used a B battery of 36 cells, another of 45 cells, or 54 and 67% volts, respectively, with the audion. When used as an amplifier, the amplification increases with the voltage.

The testimony of Langmuir, as I understand it, makes it clear that this voltage was employed for amplifying purposes, and so with a discharge current substantially unaffected by ionization. That De Forest, though not skilled in the art of evacuating his device, knew the effect thereon of an increased vacuum, though ho was very probably unaware of the extent to which it could be carried, is disclosed by his letters of July 26, 1912, and August 14, 1912, to the McCandless Company, the manufacturer of auctions for the Federal Telegraph Company. The first letter is:

“Out of the last lot of 24 audion bulbs which you manufactured for us, there are only 4 that came up to proper sensitiveness test. Apparently the vacuum in the other bulbs is too low, because in these latter bulbs we notice that the blue are or cathode beam starts suddenly at 15 or 20 volts, under which condition the bulb is not sensitive.”

The other letter reads:

“We took two of the poorest of the discarded 20 to a San Francisco glass-blower,1' unknown to fame, and had him exhaust same to as high a vacuum as his pumps would take. These re-exhausted bulbs now show a sensitiveness equal to those of the first four, which proves exclusively that our contention is correct. For the work which we demand from these bulbs they are not sufficiently sensitive, and we would bo glad, therefore, if you will re-exhaust them. You certainly should be able to put them in as good condition as the aforesaid San Francisco glass-blower.”

The testimony discloses that the glassblower referred to heated the tubes under an asbestos hood almost to the collapsing point of the tube while they were being exhausted. The references of the prior art, which reveal that a lamp may be made to collapse by heating under an asbestos hood, leave no room to doubt this testimony. Such heat was sufficient to heat to a certain degree the metal electrodes, as well as the walls of the tube, thereby driving off from the electrodes, as well as the walls, some of the gases occluded therein. The energy of the discharge of the De Forest tube was low. The bombardment was probably not adequate to heat the anode to a temperature even approximating that to which it was raised by the Bunsen burner under the asbestos hood. Consequently, little, if any, occluded gas would escape from the anode, and the gas clean-up effect probably tended to improve the vacuum more than it was caused to deteriorate by the molecules of gas that escaped from the anode. The tubes were thus enabled to function substantially above ionization voltages without substantial ionization.

Upon the issue of priority, too, I think the plaintiff must fail. In the contest in the Patent Office between Arnold and Langmuir upon this issue, the Examiner of Interferences awarded priority to Langmuir. The Examiners in Chief arrived at the same conclusion, but for reasons conflicting with those of the Examiner of Interferences. The Commissioner of Patents awarded priority to Arnold. The Court of Appeals of the District of Columbia decided ini favor of Langmuir, for the reasons stated by. the Examiner of Interferences and the Board of Examiners in Chief. I am in accord with neither of the divergent views expressed by the Examiner of Interferences and the Board of Examiners in Chief, but, generally, with the reasoning and conclusion of the Commissioner of Patents.

The defendant asserts that Arnold is entitled to the date of November 1,1912. Upon that date Arnold, who because of his training and experience in the matter of electrical discharges in vacuum and upon the recommendation of Dr. Robert A. Millikan, had been theretofore employed by Western Electric Company to aid in making possible 'transcontinental telephone service at the opening of the San Francisco Exposition, was shown by Drs. Jewett and Colpitts a De Forest tube that had been brought by De Forest to the Western Electric Company a day or so before. Arnold examined the audion and tested its performance as a repeater of speech currents. It functioned well with weak currents, but choked, distorted the speech, and blue-hazed when-increased plate voltage or loud speech currents were applied.

Though Arnold had been engaged in per-, feeting a different device of his own to be employed for telephone relay purposes, he instantly saw that the performance of the tube which made it too unstable in its existing state for a telephone repeater was due to ionization, and that this could be removed by increasing the vacuum. This he then, disclosed to Colpitts and Jewett. He also understood and then explained to Colpitts that the. hot cathode would emit electrons in the high vacuum, that they would be subject to the space charge effect, and that it would require high plate potentials to obtain the desired electron flow from eathode to anode.

Corroboration of the oral testimony of Arnold, Colpitts, and Jewett, if needed, is to be found in the faet that every act of the telephone company from that time on is consistent only with the intended >use of the audion made more rugged and with a higher vacuum in its transcontinental service. Arnold’s conception of the invention, consisting in the full performance of the mental part of the in-, ventive act, was complete. Nothing remained to be done, but to give the device practical embodiment with reasonable diligence. If, in view of the faet that those skilled in the evacuating art knew how to reduce to physical form Arnold’s mental conception, a physical reduction of his conception to practice was necessary, and I think it was not (McCormick Harvester Mach. Co. v. Minneapolis Harvester Works [C. C.] 42 F. 152, 155; Otto v. Linford, 46 Law T. [N. S.] 35), this requirement was performed by Arnold with reasonable diligence, thus giving to him an effective date of November 1, 1912. He obtained, knowingly, in that month with audion 7-A a current of 3 milliamperes at 80 volts without substantial ionization.

Again, Arnold met the requirements of reduction to practice in April and May of 1913. In the tubes of these experiments 250 volts were impressed upon the plate without ionization effect. The tubes used in Philadelphia in the fall of 1913 functioned well for many hours, one as long as 267 .hours, as commercial repeaters. The fact that the output energy of one of these tubes was less than that of the usual original De Forest tube does not negative its being a tube of the patent, for the output energy of claim 26 is likewise less than the output energy of the usual old De Forest tube. Moreover, it seems to me that other claims of the patent — for example, 11 — which was an original claim, differs from 26 in the same way, and only in the same way, that claim 26 differs from the original De Forest "device.

Langmuir began his experiments in August of 1912. He now asserts, and the defendant denies, that he then obtained a pure electron discharge, substantially unaffected by positive ionization, but concedes that, though he now knows that he did then obtain such a discharge, he did not know it then. This lack of contemporaneous knowledge denies to his August work any effective value upon the issue of priority. Robinson on Patents, p. 121, § 79. In November, Langmuir made other experiments; but these he concedes were for the purpose of ascertaining the accuracy and existence of certain laws of physics. I find no evidence that he then thought of applying the scientific principles in which he was interested to any practical result in the useful arts. Consequently there is no satisfactory evidence that Langmuir had a conception of the device of the patent before the end of November. “It is not while the creative energies are at work, shaping and sharpening the idea of means, that the conception of the invention becomes perfect. It is when these energies have ceased to act, when the idea stands before the mental vision of the inventor as clearly as a concrete invention before the eyes of an observer, that the conception is complete and the work of reduction is ready to begin.” Robinson on Patents, § 379.

Tested by this rule, even the statements of Langmuir to Coolidge in December, 1912, did not reveal a complete idea of means, and consequently did not disclose a complete conception, within the meaning of the patent law. It was not until March 21, 1913, as I read the record, that Langmuir’s conception or idea of means to a useful end began to take shape. He then made a record of it in these words: “Now, what I intend to do is to construct an audion which will operate entirely and exclusively on the electron discharge. I shall attempt to get the vacuum so good that positive ions are present in negligible amounts.” It is not necessary to inquire when the conception became complete, and his idea reduced to practice, for Arnold had had the conception long’ before, and had proceeded diligently to reduce it to practice.

The bill of complaint in this case must be dismissed.

In the second of the two suits between the same parties, tried together, three patents are involved. The first is No. 1,244,216, g-ranted to Irving Langmuir, assignor to the plaintiff, October 23, 1917, upon an application filed July 15, 1914. In the record this is called the second patent; that for the high vacuum tube being referred to as the first. It is likewise known as the thoriated tungsten cathode patent. The specification states:

“The present invention relates to electrical devices operating with a pure electron discharge and comprises a novel cathode material and the method of its preparation.”

It recites that various refractory metals —platinum, carbon, tungsten, molybdenum— have been used as cathode materials and then announces:

“I have discovered that the electron emissivity of pure thorium is of an entirely different order of magnitude than the emissivity of refractory materials heretofore used in electron discharge devices. It is not necessary that the electrode should consist entirely of thorium. When, for example, a thorium compound, such as the oxide, has been added during the process of production to one of the highly refractory metals, for example, tungsten, and the metal subjected to a preliminary heat treatment in a high vacuum, the electron emission in a high vacuum is enormously increased.”

The method of obtaining the high vacuum, enabling the device to operate with a pure electron discharge, is stated thus:

“The preliminary evacuation of the envelope is carried out by the usual methods of producing high vacuum, which includes baking out the envelope to remove water vapor. The final state of the evacuation is preferably, but not necessarily, carried out by a Gaede molecular pump to the highest possible vacuum obtainable by this means; that is, to about .001 micron. While the apparatus is still on the pump the filaments 2, 3, are heated to a temperature of about 2900° K. (absolute) for a short time, and the envelope 1 is baked out in an oven at a temperature of about 360 to 450° K. [C.]. The apparatus may then be sealed from the vacuum system at the contraction 10.”

Incidentally this statement is strongly confirmative, it seems to me, of the conclusion arrived at with respect to the process claims of the fix'st patent.

Directions for aging follow. The disclosure that the vacuum may be increased by vaporizing one of the filaments is added. The economic and practical advantage arising from increased electron emissivity at the same temperatui’e, or an equal emissivity at a lower temperature, is the lessened expense incident to the lessened energy required to heat the cathode.

The claims in issue are 1 to 7, inclusive, 10, and 11. Claim 3 is:

“A cathode for electron discharge apparatus, consisting largely of a highly refractory metal and a surface layer of material having an electron emissivity at a given temperature materially greater per unit surface than said refractory metal, independently of •and in the absence of positive ionization.”

The “surface layer” element of the claim is based upon the fact or theory that by incandescing the filament some of the thoria therein is converted into metallic thorium, which forms into a film or skin one atom thick on the surface of the cathode. The pith of invention, if it exists, lies in the employment of a thoriated tungsten filament for the cathode in a high vacuum tube.

That invention is to be found in any claim the defendant denies. This denial it rests primarily upon the ground that thoriated tungsten filaments wex'e in general use in incandescent lamps for some time before the dato of Langmuir’s application, and were gradually supplanting all other filaments. It asserts that, by reason of the analogy existing between the incandescent lamp and the radio tube, it was both natural and obvious to use in the tube a filament the same as that employed in the lamp. It shows by evidence that, since the origin of radio tubes, filaments the same as those used in incandescent lamps had been employed in the tube — in fact, that it was not unusual actually to take filaments from lamps' and put them in tubes. It points out that tbe preparation of filaments is tbe same for tube and lamp. It contends that, it being natural and obvious to use in tubes filaments identical with those employed in lamps, and it appearing that it was not unusual actually to transfer filaments from lamp to tube, the analogous user is not converted into a nonanalogous use and invention by the discovery that a thoriated tungsten filament possesses greater electronic emissivity than the earlier filaments. It rages that the public should not be barred from reaping all the advantages and benefits incident to doing that whieh was the natural and obvious thing to do.

It is conceded that lamps with thoriated tungsten filaments were on the market as early as the fall of 1912, and that they gradually superseded lamps with other filaments. It is not denied that filaments identical with those employed in incandescent lamps were used in radio tubes. In fact, the specification of the patent in suit states that platinum, tungsten, and molybdenum, as well as carbon —lamp filament materials — had been used as cathode materials. The evidence is clear that it was not unusual in the manufacture and repair of De Forest audions to obtain cathode filaments directly from incandescent lamps. Indeed, the evidence indicates a strong probability that thoriated tungsten filaments were used in radio tubes before Langmuir’s effective date, although it lacks, I think, the degree of certainty required to sustain a finding of prior knowledge and use.

Nor is the parallelism or identity of filament in lamp and .tube surprising. Fleming, in patent No. 803,684, issued in 1905, employed for the original radio tube “a carbon filament like the carbon filament of an incandescent lamp.” Each carries current, is heated in operation to a high temperature, and so must not be readily fusible. It is true that the ultimate purpose of the current in the lamp filament is light, and in the filament of the tube is electronic emission; yet these are not independent of each other, as witness the Edison effect discovered in the lamp and the light visible in the radio tube. To be sure, the different materials differ in their electronic emissivity. But this was pointed out at least as early as 1910, by Fleming, in his . patent No. 954,619.

When the thoriated tungsten filament went into general use in the lamp art, those skilled in the radio art knew with certainty that it had all the requirements for that art, if it possessed the property of electronic emissivity to a satisfactory degree. The ascertainment of that fact called, I think, only for the routine work of those skilled in the art, in the selection of the best among the mar terials known to be available. At most, it required only the judgment and skill of the engineer, and not the acumen and genius of the inventor. The question pertinent in all eases of this character — whether the new use lies so far out of the track of the former use as not naturally to suggest itself to a person skilled in the art, turning his mind to the subject — must, consequently, I think, be answered in the negative.

The third patent of this record, No. 1,-244,217, granted to Irving Langmuir, assignor to the plaintiff, October 23, 1917, upon an application filed October 28, 1915, relates to electrical devices having a thoriated cathode. The specification sets out that in the preparation of these devices “great care must be exercised to remove the last traces of oxygenous gases, particularly water vapor, and precautions must be observed not to evolve water vapor from the bulb walls during the operation of the device.” The object or purpose of the invention is to improve the device and render its preparation easier, “by providing <in the envelope a quantity of a vaporizable reagent of low vapor pressure, capable of preventing the oxidation of thorium.” An alkali metal — for example, potassium — is preferred for this purpose, but it is stated other reducing materials may be used. It is asserted in the specification that: “The presence of the reducing agent in the main receptacle maintains the thorium reduced in spite of traces of water vapor which 'may be liberated from the glass. In any event, the high electron emissivity of the .thoriated cathode is more easily secured and maintained in the presence of the reducing agent than without it.”

The claims in issue are:

“1. An electrical discharge device comprising electrodes, one of whieh contains thorium, an inclosing envelope the space within which is evacuated to a pressure so low that positive ionization is substantially absent, and a quantity of vaporizable reagent capable of preventing the oxidation of thorium in communication with the space within said envelope.”

“4. An electrical discharge device comprising electrodes, one of which consists of metallic tungsten and a thorium material in an oxidizable state, an inclosing envelope; and a quantity of a vaporizable reagent within said envelope capable of preventing the oxidation of thorium at an elevated temperature!”

Conceding infringement, if the claims are valid, the defendant denies validity upon the grounds of anticipation, noninvention, aggregation, and double patenting.

Many exhibits evince the earlier knowledge that certain vaporizable reagents constitute means of obtaining — “getting”—a vacuum of a high degree. As early as 1894 Malignani discovered, and disclosed in his patent, No. 537,693, granted in 1895, this process for evacuating incandescent lamps. The reagent was put in a tubular elongation of the bulb, and the bulb exhausted by pumping. The vacuum having been obtained, the lower end of the tube was soldered up and the carbon filament brought to intensive incandescence, to expel the gases contained therein. Simultaneously the part containing the vaporizable substance was heated by suitable means, such as a spirit lamp to vaporize the chemical reagent. The vapors so generated combine with the free gas and vapors of the tube and those expelled from the filament, and form a solid or liquid precipitation, “so that an almost perfect vacuum is obtained.” Arsenic and iodine were the reagents employed by Malignani.

Dwyer also, in patent No. 595,695, granted in 1898, made use of a chemical “having the property of evolving a condensable gas or vapor which, when the heating is discontinued, goes down as a solid upon the-cooler portions of the receptacle, and requiring quite a high temperature to be afterward again vaporized.” Dwyer limited the amount of the chemical, so as to have the least possible amount of foreign substance deposited as a film upon the inner surface of the lamp bulb. Doane describes the employment of phosphorous as a chemical exhausting means or .aid.

Merritt, in patent No. 909,811, granted in 1909, after reciting'the earlier use of suitable substance to act upon the deleterious gases, particularly oxygen, remaining in the bulb after pumping, and to convert them into harmless compounds, shows the advantages to be had from putting the chemical in the bulb, instead of in the tubulature. He restricted, however, the amount, so there would be no excess to discolor the bulb.

Soddy, patent No. 859,021, issued in 1907, described the process of bringing a heated alkaline earth metal — particularly calcium— capable of absorbing gas into the presence of the rarefied atmosphere whereby the remaining gases were absorbed. His article in Na/ture, of November 21, 1907, showed that by means of calcium a copious mirror deposit on the inner walls of the tube might be had. But he does not indicate that such a deposit is useful or desirable.

Clark, in patent No. 1,121,802, issued in 1914, made use of vaporized sodium or potassium to coat the inner wall of projector lamps with reflecting surfaces. He explained that, since sodium and potassium oxidize very readily when in the presence of oxygen, it is necessary to apply the coating with no trace of oxygen.

Van Recklinghausen, in patent No. 946,-068, applied for in 1904 and granted in 1910, made use of magnesium, in mercury vapor apparatus, for absorbing some of the gases which remain aftdr a certain degree of exhaustion by pumping has been attained. He claimed: “The combination, with a closed container, of a holder within the container, magnesium within the holder, and a source of electric current connected with the opposite ends of the magnesium, whereby the same may be rendered active for absorbing injurious gases in the container.”

An analysis of these and other disclosures of the prior art shows, however, that they all fall short of the underlying idea of the third patent. They reveal that chemical agents had become a most potent aid to the pump in creating — “getting”—vacua of the required degree. But nowhere is it indicated that the reagent has any practical value after the moment of volatilization — after it becomes cold — and nowhere is there a suggestion that by retaining the volatilized reagent within the sealed-off bulb the vacuum obtained may be maintained. In short, the pri- or art discloses the use of chemicals as “getters.” It does not disclose their use as “keepers.” Langmuir made a step forward. He grasped the fact that, by retaining the reagent within the tube, it would continue to function after the moment of volatilization, even though cold, and so combine with and form into a harmless solid compound any molecules of free gas, particularly oxygen, that might evade the evacuation process, and any molecules of occluded gas that might escape from walls or electrodes during operation of the tube. By this means the efficiency of the readily oxidizable thoriated cathode is preserved and its life prolonged. It is a true combination — not an aggregation.

Patent No. 1,159,307, applied for March 20, 1914, and granted November 2, 1915, to William C. White, assignor to the plaintiff herein, upon which defendant herein relies to show invalidity of the claims of the third Langmuir patent here in issue, on the ground of double patenting, is for a different purpose. The vaporizable substance retained in the tube by White is of a character to maininin the gas in the tube at the definite pressure at which the device possesses a critical region of sensitiveness. That retained by Langmuir, within the meaning of the claims, is of a character to capture and combine with the molecules of gas and maintain the vacuum. I think the claims in issue of the third patent are valid and infringed!

The fourth and last patent in suit is No. 1,529,597, applied for August 11, 1921, and granted March 10, 1925, to Irving Langmuir,, assignor to the plaintiff herein. Like the third patent, it has to‘do with “keepers.” The advance marked by it is in the employment for that purpose of an alkaline earth metal, preferably magnesium, instead of potassium or other alkaline metals suggested in the third patent. The claims in issue are 12, 13, 18, 19, 21, and 22. Claim 13 is:

“An electrical discharge device, comprising a sealed envelope, electrodes therein, one of which is of the thoriated type, and a film of magnesium deposited on the inner surface of said envelope.”

The claim to invention rests upon the discovery of special advantages incident to the use of magnesium. One advantage is set out in the specification thus:

“A thoriated cathode may be deleteriously affected by positive ion bombardment; that is, when voltages materially higher than the ionization voltages are impressed between a thoriated cathode and a positive electrode in the presence of ionizable gas, then the beneficial effect of the thorium may be largely or entirely destroyed by positive ion bombardment. Although the alkali metals are entirely satisfactory as agents to remove deleterious gases when the operating temperature of the bulb, is only moderately above room temperature, as in the ease of radio receiving tubes, these alkali metals have a sufficiently high vapor pressure, at the operating temperature of power devices in which cathodes of the thoriated type are used, to become ionized at higher voltages, and to cause removal of the active film of thorium from the cathode, by bombardment. I have found that it is advantageous in many cases to provide a device containing a cathode of the thoriated type with a material having an affinity for gases, such as magnesium, for example, which, at the operating temperature of - the device, has substantially no vapor pressure. * * * ”

Another advantage lies in the avoidance of the danger incident to the use of potassium, which is so oxidizable that it must be kept under oil, for it breaks into flame when exposed' to the air. Magnesium is handled without difficulty. Obtained in the form of wire or foil, it can be cut up in the open air and fed to the machines-without danger.

The broad question here presented is, of course, whether making the keeper element, of the old device, of magnesium, instead of some other material, particularly an alkali metal, involved invention. As I understand the law, there may be invention in selecting a particular material for a particular purpose, although other materials of the same class have been used before for the same purpose, and even though a patent has been pre- ' viously granted for the use of the whole class. C. & A. Potts & Co. v. Creager, 155 U. S. 609, 15 S. Ct. 194, 39 L. Ed. 275; Andrew’s Patent, 85 R. P. C. 477.

But courts which scrutinize such patents somewhat narrowly require, before invention may be found, that it must at least appear, not only that there is a special advantage attached to the use of the particular substance claimed, but also that its use is not the mere substitution of a superior material for an inferior, is not the result of judgment and skill in the selection of materials, nor a probable consequence of the natural development of the art. That magnesium has an affinity for gases, particularly oxygen, has long been well known. The availability of magnesium as a “getter," because of that property, was disclosed by Yon Recklinghausen, in patent No. 946,068, by British patent No. 15,788 of 1914, granted to the British Thomson-Houston Company, Limited, as a result of a communication from the plaintiff herein, and by MacKay, in patent No. 1,208,597. Soddy, in patent No. 859,021, made use of alkaline earth metals, preferably calcium, which is suggested by the patent in suit as an equivalent for magnesium, to absorb the residual gases of a vacuum not sufficiently high.

It is, of course, true that magnesium, though highly oxidizable, does not oxidize so readily as potassium; yet, in view of its known efficiency as a “getter,” I find nothing ,in the record to indicate that its use as a “keeper” involved more than judgment and skill in the selection of materials by those conversant with the art.

The language of claims 1 and 4 of the broad “keeper” patent and the known properties of the alkaline earth metals make it certain, I think, that those metals lie within the scope of those claims. Langmuir could not have been restricted in the exercise of the invention of those claims by a subsequent patent to another for the employment of the alkaline earth metals as keepers. Conversely, Langmuir can make no valid claim whose effect would be to extend the term of the monopoly that he obtained by the grant of the third patent. This makes inevitable a conclusion that the claims here in issue of the fourth patent are invalid.

Decrees in conformity herewith may be presented.