This appeal is from a decree of the District Court holding valid and infringed claims 1 and 5 of the complainant’s Letters Patent, No. 818,033, issued April 17, 1906, to Heinrich Koppers, for improvements in by-product coke-ovens, and dismissing a counterclaim of infringement of defendant’s Letters Patent, No. 673,928, issued to Schniewind.
The defenses pleaded respectively to the claim and counterclaim of infringement are: (1) Non-infringement; and (2) invalidity of the patent claims in suit, both for anticipation and lack of patentable invention.
Though the decision in this case will affect interests of considerable magnitude, the accompanying opinion probably will be read only by those who are engaged in the by-product coke-oven art and in kindred arts. As discussion in this opinion will be addressed to those who are conversant with the very extensive and complex subject matter of the patents in suit, we shall review the art, prior and present, in no greater detail than we deem necessary to make known the grounds of our decision.
The subject matter of the Koppers patent is by-product coke-ovens and the invention of the patent relates to improved means for heating them. The art of by-product coke-ovens was highly developed at the time of Koppers’ invention. It was in itself a broad art. Because of its dose relation, industrially and economically, to the great metallurgical arts, it had for many years attracted capital in ample measure and had invited the attention of scientists of the first order throughout the world. Notwithstanding the great advance which the art had *124made, the state of the art — that is, the point which the art had reached and at which it had stopped when Koppers made the invention of the patent in suit — disclosed a problem which was present at its beginning, which had persisted throughout its development, and which remained in a large measure unsolved. This was the problem of applying heat to the walls of a coking-chamber with an uniformity of temperature that would insure uniformity in the coking-operation and in the resultant products. Uniformity of heat distribution to coals in process of coking is essential to the production of coke of the structure and composition required for use in metallurgical industries as distinguished from coke suited only for general or domestic uses. Uniformity of heat distribution is essential also' to the saving of those distilled by-products — such as benzol and toluol — that are peculiarly subject to destruction during distillation.
Koppers freely availed himself of the art and cleverly combined in his oven many features of merit which are not embraced in his invention. Therefore,' in order correctly to estimate the issues of this case and properly to confine their decision to Koppers’ invention, it will be necessary to distinguish in Koppers’ oven the things which he invented from the things which already existed. This will require a short excursion into the prior art.
Coke is the mass of carbon which is left after the volatile matter of coal has been driven off by heat applied in such a manner as not to burn the carbon. Coking coals are converted into coke in the industrial arts in two principal ways, the gas-retort method and the coke-oven method.
The gas-retort method produces illuminating gas as a main product and yields as a by-product a soft, spongy and fragmentary coke known as “domestic coke.” The coke-oven method produces a hard, strong and structurally coherent coke known as “metallurgical coke” because of its utility in foundry and blast furnace processes. In the production of .coke of the latter grade, the gases of coal distillation may be altogether consumed as in the wasteful bee-hive coke-ovens, or they may be saved as highly valuable by-products as in by-product coke-ovens. The difference in coke products reflects the difference in methods of producing them and in the problems incident to their production. It is with by-product coke-ovens operated primarily to produce metallurgical coke, and secondarily, gaseous by-products, that we are concerned in this case.
The by-product coke-oven typical of the prior art, though appearing on its outside to be one huge, homogeneous structure, contained within itself two distinct organizations. One had to do with combustion and the other with supplying materials for combustion. These organizations, while inseparably related in their operation, occupied separate parts of the structure and were commonly referred to by reason of their location as the “upper story” and the “lower story” of the oven.
A by-product coke-oven of the prior art, while generally spoken of in the singular, comprised actually a plurality of ovens or coking-chambers, the size of the oven as a whole being determined by the number *125of coking-chambers desired. The upper story of the oven was made up of a battery of any number of long, high and narrow coking-chambers extending from side to side of the oven. Intermediate the coking-chambers and parallel with them was a like number of heating-chambers of somewhat similar dimensions. In this alternate arrangement of coking-chambers and heating-chambers, a wall of a heating-chamber was similarly a wall of its adjacent coking-cliamber, and so on throughout the battery. Heat was conveyed from a heating-chamber, in which combustion occurred, through the wall dividing it from a coking-chamber to the coking-coals there undergoing distillation.
To facilitate uniformity of heat distribution to its walls, the heating-chamber was divided into many parts by narrow vertical flues extending from near the bottom of the chamber to a channel or flue running horizontally across its top. Nearly midway the heating-chamber, a partition, extending from the bottom upward to this horizontal flue, divided the chamber into two parts. These batteries of coking-cham-' bers and beating-chambers, and certain hot air delivery flues presently to be mentioned, together with the supporting and separating masonry-all made of highly refractory and heat resisting bricks — comprised the combustion organization of the oven located in the upper story.
The lower story of the oven was largely made up of pillars and walls, likewise of refractory brick, to support the upper story. In the center of the lower story, or at its sides, or, indeed, on its Outside, was placed one or more sets of regenerators. A “set” of regenerators comprised two regenerators. Between the regenerators and the sides of the oven, when the regenerators were placed in the center, or between the regenerators and the center of the oven, when the re-generators were placed at the sides, was a system of cooling-flues and open arches used to prevent melting or fluxing of the bottom masonry of the upper story, incident to the sub-bottom combustion system employed.
A regenerator, described very generally, is an oven-like structure of firebrick which contains at the bottom an air flue or sole-channel and checkerwork of firebrick upward from the sole-channel nearly to the dome. The functions of a regenerator are, primarily, to heat air, and secondarily, to so heat it that it attains an approach to uniformity of temperature before it is delivered to the heating-chambers of the second story, where it is used with gas as combustion material. Re-generators of the prior art — in coke-ovens as distinguished from gas generators — were placed longitudinal of the battery of heating-chambers, that is, at right angles to each heating-chamber and to each coking-chamber throughout the length or depth of the battery. This was the organization for supplying materials for combustion.
The connection between the two organizations, situate respectively in the upper and lower stories of the oven, was made by at least two flues leading from the regenerator to each heating-chamber, the first extending vertically to the second, and the second extending horizontally under the vertical flame-flues of each heating-chamber. The horizontal connecting flue was termed “bus-flue,” and was in the very bot*126tom of the upper story. It had as many air openings as there were vertical flame-flues. There were as many connections of this kind with the one longitudinal regenerator as there were heating-chambers to be served with preheated air.
In the operation of a regenerative coke-oven, unheated air was drawn by smokestack draft through the sole-channel into the body of the re-generator. It was there heated on its passage upward through a check-erwork of bricks, then conveyed from the regenerator to the bus-flues and thence to one side of the divided heating-chambers, where, in its travel, it mingled with inflowing gas.and became ignited, thereby producing a gas jet for each flame-flue on that side of the division. Heat was in this way conveyed to the adjacent coking-chambers with some degree of uniformity. The hot, waste gases arising from combustion were then carried horizontally along the flues at the top of the heating-chambers, across the division, down through corresponding though inactive flame-flues in the other side of the heating-chambers, then into the bus-flues and thence on to the companion regenerator, where they served to reheat its checkerwork as they passed off into the outer air.
This operation was reversed at stated intervals, with the result that one regenerator of the set apd one side of the connected heating-chambers were actively heating a corresponding part of the coking-chambers, lying intermediate the heating-chambers, while the other side of the heating-chambers and the other regenerator of the set were inactive except in being reheated for use when the operation was again reversed.
To understand Koppers’ invention, it is necessary to have a better understanding of the prior art than can be had from this inadequate statement. Therefore, we insert at this point diagrams of Hoffman’s coke-oven (U. S. No. 492, 400—1893), an excellent illustration of the elements of the prior art in combination.
*127
An alternate arrangement of coking-chambers and heating-chambers in parallel alignment appears in Boult (British Patent No. 5,979 — 1883); Hoffman (U. S. Patent No. 492,400 — 1893); Otto (German Patent No. 88,200 — 1895); Hilgenstock (U. S. Patent No. 649,450— 1900); Schniewind (U. S. Patent No. 673,92S — 1901); Koppers (U. S. Patent No. 738,918 — 1903). Vertical flame-flues are seen as early as Herberz (German Patent No. 25,526 — 1883) in which air and gas for combustion are introduced and ignited by means for regulating their supply. Flue combustion is found also in Koppers (British Patent No. 3,026 — 1899), in Hoffman (U. S. Patent No. 492,400 — 1893) and in many other patents. Regenerators in sets, each set ordinarily containing two, and positioned longitudinally the heating-chambers battery, each regenerator serving the whole battery and operated under the system of periodical reversal of air flow, are shown in the cited patents of Hoffman, Schniewind, Otto, and Herberz, and in the early Koppers.
Vertical connections from regenerator to horizontal bus-flues, and vertical connections thence to flame-flues of heating-chambers, for delivery of preheated air for combustion, appear in the same patents.
Koppers invented none of these elements of a by-product coke-oven; neither did he invent their method of operation. All were old in the art, yet all are found in one relation or another in the byproduct coke-oven of his patent. As so much of the prior art is found in the organization of Koppers’ coke-oven, what is there in his oven that distinguishes it from others in the art? What is the invention for which he was granted a patent?
Koppers has answered these questions by the claims of his patent in suit. They are as follows:
“1. In a coke-oven, a series of heating-chambers, and coking-cliambers intermediate the heating chambers, combined with a series of regenerators *128below and parallel to the heating-chambers and communicating directly therewith, substantially as specified.”
• “5. A coke-oven provided with -coking-chambers* two sets of heating-chambers intermediate the coking-chambers, two sets of regenerators commu-nicatingf with the heating-chambers and arranged below the colving-chambers, and a partition between the regenerators, substantially as specified.”
Prior to Koppers’ patent, the efforts of all inventors, including Koppers himself, to produce uniformity of heat intensity over the walls of heating-chambers of by-product, coke-ovens, were directed along the line of delivering and distributing preheated air to heating-chambers from a set of two regenerators which served the whole battery of heating-chambers, whatever the number, and which were positioned longitudinally the battery and at right angles to each heating-chamber. With heating-chambers crosswise the oven and regen-erators lengthwise the oven and therefore at right angles to one another, the delivery of preheated air from regenerator to heating-chamber was necessarily roundabout or indirect. As the air, in its transit from regenerator to heating-chamber, traveled up the vertical flue and then along the horizontal channel or bus-flue to the flame-flues, it encountered inequality of air pressure in the flue, with consequent inequalities in air deliveries to the heating-chamber and in flame distribution over its walls.
This being the state of the art, how did Koppers meet its problems ? He took what was good in the art and discarded what was bad. He retained the upper story combustion organization substantially as it was, but he departed abruptly from the prior art organization of the lower story. He abandoned the one set of regenerators which had served, the whole battery of heating-chambers and provided in their stead a séparate regenerator to serve each heating-chamber. In doing this, he greatly multiplied the number of regenerators and filled up the arched spaces which theretofore had been left open to keep the masonry below the bottom of the coking-chambers from fluxing, and in doing this he overcame the chief defect of the sub-bottom combustion system. He then changed the position of regenerators from longitudinal of the battery to crosswise the battery, placed each individual regenerator below and parallel to each heating-chamber and made multiple connections between the two which were vertical and direct instead of horizontal and indirect, thereby causing the preheated air to pass from each regenerator evenly and directly to its companion heating-chamber, where, on combustion, heat intensity is distributed over its wajls with practical uniformity.
When Koppei's made these changes, he wrought a change in the art from a stagnant struggle for uniform distribution of heat over heating-chamber walls to the achievement of that result. The change was immediately reflected in economy of operation and increase of production. Coke production for a given coking time was substantially doubled. The coke product was of the best metallurgical grade, and the saving of distilled by-pi'odücts was increased, with a reduced waste of fuel gas used in the heating operation. The art promptly recognized these achievements and adopted Koppers’ arrangement to *129the extent that over eighty per cent, of the by-product coke-ovens built during the last ten years in the United States have been Kop-pers ovens, amounting to 4,700 and representing an outlay of between $50,000,000 and $100,000,000. They exceed in number all competing ovens combined.
Koppers’ changes from prior art ovens, involving changes in the number, size, and position of regenerators and in their relation to and connection with heating-chambers, may best be understood by comparing Koppers’ patent diagrams here inserted with the diagrams of the Hoffman oven previously inserted.
*130Measured by all standards, the things which Koppers did involved invention, if he was the first to do them.
But the defendants say, that in multiplying regenerators, in changing their position, and in connecting them directly with the heating-chambers, Koppers was not the first to do these things, citing as anticipations, first, the published report of the Fiftieth Congress of the Technical Society of the Gas Industry of France, held June 24, 1878, at Paris, on the subject of the Paris gas-retort furnaces; and, second, Dods U. S. Patent No. 571,558 — 1896, for a gas generator.
The fundamental difference between these references and the invention of the patent in suit — which we observe on the threshold and which we shall keep in mind throughout the discussion without burdening this opinion with repeated allusions to it — is the difference in the arts to which the references and the invention of the patent respectively belong. Paris gas-retort furnaces and the Dods gas generator belong to an art fairly described by their names, in which gas is the main product and coke the by-product. It may be termed the art of by-product gasrovens. The invention of the patent belongs to an art in which coke is the main product and gas the by-product. This is the art of by-product coke-ovens.
[1] Though the two arts, producing the same products from the same source, are concededly related, yet differences in them, if fundamental, may validly prevent inventions in one from operating as anticipations of inventions in the other. While distillation of coal is practiced in both arts, and, therefore, coke-ovens or coking-chambers are included in the organization of both, difference in the process of distillation, in the function of the ovens, and in the products sought and obtained, mark the difference in the arts. The issue of invalidity of an invention in one art because of anticipation by inventions in another art, when the arts are different, though related, must be determined with a cautious regard to the differences that distinguish the two.
The Paris gas-retort contains coking-chambers, and regenerators, or recuperators, for preheating air. The coking-chambers are in the form of semi-cylindrical or tubular retorts, which lie horizontally across the furnace structure, each holding several hundred pounds of coal. These tubular coking-chambers are stacked in rows in what is substantially a flame-filled heating-chamber, in which flames are diffused throughout the chamber, above, below and around all parts of the coking-retorts. The system employed to preheat air for combustion is without reversal in the direction of flow and without alternate change of position of the flames. While coke is a product of the Paris gas-retort, it is incidental, the main object of the Paris gas-retort being the production of gas. In this furnace, there is nothing that requires uniformity of heat distribution, and therefore there is no means to perform that function. In a furnace of this type, the problems of a by-product coke-oven are not present, and, as we regard it, there is nothing in it to suggest the means which Koppers 28 years later employed to solve such problems.
The Dods gas generator (U. S. No. 571,558 — 1896) comprises com*131bustion-chambers, heating-chambers, and regenerators. Heating-chambers and combustion-chambers are intermediate one another, as in coke-ovens, and regenerators (two in number) are placed below each heating-chamber. The prior art coke-oven system of reversing the flow of preheated air for combustion and of waste gases for reheating regenerators is used. On first view, Dods’ organization seems to resemble rather closely the invention of Koppers. On examination, however, it is seen, that the coking-chambers of the Dods gas gen-orator are tapered and in position are steeply inclined, the broader end of the tapered chamber being lower than the upper end; and that the intermediate heating-chambers are of similar dimensions and positions. These heating chambers are open in the sense of being without flame-flues to confine combustion to a particular direction or to control its distribution. As the furnace is a gas-retort, its primary product being gas, there is neither need nor means for the uniform distribution of heat.
[2] The pairs of regenerators under the heating-chamber are stepped one above the other and so are out of parallel with the inclined heating-chambers. Connections between regenerators and beating-chambers for the passage of preheated air are long and uueciual conduits. This inequality of delivery conduits, we are told, makes uniform delivery of preheated air impossible. Gas is delivered to the heating-chamber by gas conduits of unequal length parallel with the air conduits, and combustion takes place on the meeting of gas and air in the open heating-chamber. These limitations, we are shown by the testimony, conflict with any practice of coke-oven distillation. As Dods’ invention has never reached the industries — although it preceded Koppers by ten years — we do not know whether it is operative. We are inclined to believe, because of its lack of means for uniform heat distribution, that it will not produce metallurgical coke or those by-product gases which are destructively decomposed by overheating and are recovered only by uniform heating. We cannot believe that either the Paris gas-retort furnaces or the Dods gas generator did the things which Koppers did when he invented the coke-oven of his patent, or that they suggested to Koppers the conception which later he put into practice. Therefore, we agree with the learned trial judge, that the Paris gas-retort furnaces and the Dods gas generator do not anticipate Koppers’ invention, and find with him that claims 1 and 5 of the Koppers patent in suit are valid.
We shall next direct our discussion to the issue of infringement of the Koppers patent in suit.
The defendants’ alleged infringing by-product coke-ovens were built under patents to Wilputte (U. S. Nos. 1,212,865, 1,212,866 — 1917). Wilputte, like Koppers, made free use of what he found in the art — ■ with this difference: He found Koppers there. Wilputte took the combustion organization of the upper story of the prior art by-prodUct coke-oven and embodied it in his oven without change, except the elimination of the bus-flue. Koppers cannot complain of this, for he did the same thing. Wilputte’s changes, like Koppers’, were made in the organization of the lower story. • Unless these changes were *132his conceptions, the question is: Where did Wilputte get them? From Koppers or from the prior art ?
The elements of the Wilputte oven are elements of the prior art. Their organization, confessedly, is not the organization of the prior art. Then whose organization is it? The plaintiff admits that in certain structural details Wilputte differs from Koppers, but contends that, basically, it is the organization of the Koppers invention. The defendants admit similarity in certain structural essentials, but insist that the organization involves a fundamental difference — conceived by Wilputte — which makes infringement of Koppers impossible. In order to compare the Wilputte oven with Koppers’ and to distinguish the two, on the issue of infringement, it is necessary to show the principal characteristics of the construction and operation of the Wilputte oven. An examination of diagrams of the last Wilputte patent, here inserted, will, on comparison with diagrams of the Kop-pers patent, previously inserted, be of assistance.
Externally a'nd considered as a unit, the Wilputte oven is not distinguishable from Koppers. There is no infringement here, for in this respect Koppers is scarcely distinguishable from the prior art. Internally, and considered with reference to its duality of organization, the Wilputte oven is not different from Koppers. Here again there is no infringement, for in this Koppers copied the prior art. The am-bustión organization of. Wilputte, located in the upper stop? of the oven, is the same as that of Koppers, but as Koppers took this organization from the prior art, it was equally free to Wilputte. The issue of infringement begins at the point where Koppers left the art and began to malee changes. This is in the organization of the lower story
*133
and in the connections between the organizations of the two stories of the oven. Wilputte’s first change, like that of Koppers, was in the number of regenerators, the change being from a set of two to a number equal to the number of heating-chambers. His next change, like that of Koppers, was in the position of the regenerators from longitudinal of the oven to crosswise the oven and from right angles with the heating-chambers to a position below and parallel to them. Wil-putte’s last change, like Koppers’, was from an indirect connection between the regenerator and bus-flue of each heating-chamber to a multiple direct connection between regenerator and heating-chamber.
Assuming this comparison of the two ovens to be correct so far as it goes, claims 1 and 5 of the Koppers patent read literally on the Wil-putte structure. To avoid the legal consequence of such reading, Wil-putte distinguishes his structure from Koppers by specifying a difference in the position of the regenerators with reference to the heating-chambers and by indicating a difference in the types of regen-erators and in their mode of operation.
It is conceded by the plaintiff, that if the regenerators of the Wil-putte oven are longitudinal of the oven, Willputte does not infringe Koppers. This, the defendants maintain, is the case. The first phase of the issue of infringement, therefore, is the position of the Wilputte regenerators.
Viewed as a regenerator extending crosswise the oven, the Wilputte regenerator is the Koppers regenerator substantially in size and dimensions and actually in its location below and parallel to the crosswise heating-chamber. The claimed difference is this: Koppers’ regenerator is an unitary structure, while Wilputte’s regenerator is divided into twenty-eight compartments by partitions extending from the sole-*134channel to the top. Again like Koppers, the Wilputte regenerators are made into a battery of any desired number, extending from the front to thg back of the oven. In the battery of Koppers’ regenerators, each regenerator is divided from its neighboring regenerator by the supporting walls of the oven. The same is true in Wilputte, with the difference that in the latter there are spaces left open in' each supporting wall between all regenerators, the spaces being equal in area to that of two bricks. These spaces are termed “equalizing ports.” As these wall-slits or equalizing ports appear in each supporting wall, they leave openings longitudinal of the entire oven. The defendants maintain, first, that the frontal splitting up of the (Koppers) cross-re-generator into multiple compartments makes the partition of each compartment the side of a smaller regenerator, and, second, that the sequence of wall-slits or equalizing ports, being longitudinal of the battery, gives the Wilputte compartmental regenerator a position longitudinal of the battery and not crosswise, and that, in consequence, the regenerator so shaped and positioned is in effect a reversion to the prior art and avoids infringing Koppers.
Elaborate argument was made on this contention. We shall not discuss it in this opinion, because we are satisfied that the inconsiderable openings in the supporting walls between the Wilputte compartmental regenerators do not make the position of the regenerators longitudinal; nor do they make them functionally longitudinal, because under normal conditions, when air is driven into each regenerator compartment by fans and drawn through the regenerator by smokestack draft, the equalizing ports have nothing to equalize and perform no function. Air thus driven and drawn naturally moves upward in each regenerator without stopping to pass laterally through the equalizing ports. In fact, the whole theory of difference between Wilputte and Koppers in mode of operation is based on the fact that into each regenerator compartment of the Wilputte oven a dose of air is fed precisely equal to the capacity of that compartment. This equality of dosage makes the equalizing ports functionless, under normal conditions.
The sole-channel is as certainly a part of a regenerator as the-check-erwork of firebricks. The brick checkerwork cannot heat air until air is brought to it. It is the function of the sole-channel to bring unheated air to the regenerator and to feed it at multiple points to the checkerwork above, there to be heated. As the sole-channel is a part of a regenerator, the direction of the sole-channel is a fair indication of the position of a regenerator. The sole-channel of all re-generators of the prior art (also in Koppers) is lengthwise the regen-erator, whatever the position of the regenerator with reference to heating-chambers; and so in Wilputte, the sole-channel, or its equivalent, is crosswise the oven and lengthwise the regenerator, passing under and serving unheated air to its many compartments.
For' the reasons we have indicated, without elaborating them, we are clearly of opinion, that the Wilputte regenerators are not longitudinal of the battery, but are crosswise, and in this position they follow Koppers.
*135[3] The next defence is, that, even i£ the Wilputte regenerator be a cross-regenerator in position, it is different from Koppers’ regenerator in structure. This contention is based on the division of the re-generator into compartments. The division of whose regenerator? The division of the regenerator of the old by-product coke-oven art? No, because this art (as distinguished from the related art of gas generators) contained no such regenerator. The regenerator there found was a long regenerator, being longitudinal of the oven and serving all heating-chambers. Was it a division of Koppers’ regenerator into compartments? Manifestly so, because Koppers’ was the only regen-erator of that shape and in that position then in the art. Therefore, the substance of what Wilputte did structurally with reference to his regenerator, was to take Koppers’ regenerator and divide it by partitions into as many compartments as there were flame-flues in the heating-chamber above. This, Wilputte could not have done if Koppers or someone else had not invented the cross-regenerator. It follows then, that Wilputte built his regenerator on Koppers. In doing this, he took Koppers’ regenerator as it was, both in structure and position, took nothing from it, and added something to it. Being an addition, he did not change its basic structure; the most that he did was to improve upon it. Infringement of a patent device cannot be avoided by merely improving it.
Wilputte’s multiple 'connections between his compartmenial regen-erator and the flame-flues above, though limited to one for each regen-erator compartment, are, taken together, so precisely the vertical and directly communicating air ports of the Koppers invention, that discussion of infringement of this feature is unnecessary. Therefore, considered with reference to the structure of the two ovens, we are forced to find that, basically, they are the same.
But the defendants maintain further, that even if the two ovens are alike in position and structure, the Wilputte multiple compartmental regenerator is fundamentally different from the Koppers unitary regenerator in its mode of operation, and that this difference is so substantial that the coking operation of the oven of the Koppers patent is not, and, indeed, cannot be followed in the Wilputte oven. This contention of difference in mode of operation of the two ovens is based upon a claimed difference in the operation of their regenera-tors. This contention requires us to discuss somewhat in detail the structure and function of regenerators in by-product coke-ovens.
A regenerator, defined without any attempt to be technical, is a firebrick chamber of a coke-oven used to heat air before it flows into the heating-chamber for combustion. It has three operative parts. The first is the sole-channel. This is a firebrick conduit, which brings unheated air from outdoors and carries it along the bottom of the chamber. The sole-channel contains in its top or roof a number of openings through which the unheated air passes into a checkerwork of firebrick — the second part — which is heated for weeks before the coking operation begins. As the air passes up through the checker-work, two things happen: First, the air is heated, and second, any inequality in the volume of its initial delivery through the ports of the *136sole-channel are baffled into equality of temperature and volume as the air rises to the dome. The third part of the regenerator is the outlet through which the air after it has been heated passes to the heating-chamber. In the prior art regenerator, there was but one air outlet from regenerator to each heating-chamber. This was first vertical, then horizontal into the bus-flue, and then vertical again. There were in the bus-flue as many air outlets as there were vertical flame-flues to be served. Koppers’ regenerator is structurally and operatively the regenerator of the prior art differentiated in number, shape and position to serve one instead of many heating-chambers, and differentiated chiefly in having not one but- as many air outlets as there are vertical flame-flues to be served and in having these outlets vertical and leading directly to the flame-flues above.
Wilputte’s compartmental regenerator contains precisely the same elements, but it is organized to operate in a different way. Instead of the sole-channel of the prior art as an inlet for unheated air, Wilputte has a 6-inch pipe with as many air outlets or orifices as there are compartments to be supplied with air, each outlet being graduated in size to meet the capacity of each compartment. In each compartment there is the old checkerwork of firebricks and from each compartment there is an air outlet leading directly to the flame-flue above. When the operation is reversed, the outflow of waste gases is taken care of by another channel, and, it seems to us, by the 6-inch air inlet pipe also. Be that as it may, we regard the pipe and the exit-channel together as the equivalent of the sole-channel of the prior art, for taken together they perform the same functions, even if they perform them better. The main difference in the regenerative function between Wil-putte on the one hand and Koppers and the prior art on the other, is that in Wilputte’s regenerator compartments there is no distribution or equalization of air by the baffling operation of the checkerwork beyond the confined area of a compartment, while in Koppers such distribution may extend throughout the regenerator. The problem of equal air distribution in Wilputte is wholly taken care of when the air' is brought to the compartments. This is done by the initial injection into each compartment of a quantity of air precisely equal to the capacity of each compartment. This air allotment is termed “air dose.” Air distribution from one compartment to another is impossible because of the intervening partitions. Air distribution among the many compartments is regulated by the air doses measured and delivered to each compartment. After delivery to a compartment, the air is heated as it ascends through the checkerwork as before, and after it has been heated it passes from the top of the compartment through a port directly to its corresponding flame-flue above.
It must be conceded,' that in mode of operation, there is a difference between the contesting regenerators. On this difference, the defence of non-infringement must rest, for structurally, and, in all else, functionally, Wilputte has followed Koppers.
Notwithstanding the claims of the Koppers patent read literally on the Wilputte oven, the defendants maintain that the true invention of Koppers is not disclosed by the patent claims, but is found in the pat*137ent specification, and particularly in the specification of his original application as it appeared before amendment. In the latter specification, Koppers said:
“It is the object of the present invention, by interposing the regenerator in the path of the hot gases, to render it possible to utilize said regenerator as distributing means. The invention is based upon the principle that the descending hot gases become uniformly distributed, over the regenerator and that the ascending gases which are to be preliminarily heated also undergo such distribution. This parallel arrangement of heating flues and regenera-tors is attained by the transverse disposition of these latter, in such manner, that the combustible substances to be preliminarily heated are conducted to the part of the regenerator belonging to one-half of the furnace and distributed and burned throughout the length of the furnace, whereupon the burnt gases, descending through the other half of the regenerator reach the chimney.”
A similar idea is contained in Koppers’ description of his invention in his German patent. It is as follows:
“The leading idea of the present invention is this: so to introduce such a single-chamber air-heater (regenerator) in the path of the hot gases, air and off-heat that it acts simultaneously also as a distributor, and indeed in so far that while the descending hot gases -distribute, themselves uniformly over the regenerator, the ascending gases to be preheated, lileemse undergo such distribution. This parallel arrangement of heating-flues and heat magazines (regenerators) is attained in this wise, that the latter are located transversely, that is, in the longitudinal direction of the individual ovens, in such manner, that the combustion substances (air and gas) to he preheated are introduced into the part of the regenerator belonging to one-half of the oven and are distributed to the oven length, then burned throughout the oven and Anally descending as burnt gases through the other half of the regenerator, arrive at the exit-flue.”
These expressions, touching the operation of the invention, do not appear in the patent in suit, but as they are Koppers’ utterances they cannot be disregarded. From them, the defendants have construed Koppers’ invention to be primarily, if not entirely, the distribution of air by the checkerwork in the regenerator, and thus construed his invention to be limited to distribution of that kind. If such is the invention, then, manifestly, Wilputte does not infringe, for he distributes air in his regenerators in a different way.
The only expression in the specification of the patent, as granted, which bears upon this aspect of the interpretation of the claims, is the following :
“The regenerators above described have the advantage that the passing gases are evenly distributed. As the regenerators are placed directly in the way of tlie heating-gases, the latter are also evenly distributed in the healing-chambers so that the formation of any whirls and stagnation is avoided.”
On this and on the other specification expressions, the defendants’ argument of regenerator-distribution as the essence of Koppers’ invention is based. But as we read the specification of the patent and the specifications of Koppers’ other patents for the same invention, we do not gather that air distribution in the checkerwork of the re-generator is the essence of the invention, or that air distribution in the regenerator is the whole invention, for manifestly the invention *138includes air distribution to the heating-chambers. Repeating the specification as quoted, Koppers says:
“The regenerators above described have the advantage that the passing gases are evenly distributed."
“Evenly distributed” — in what ? In the regenerator, to be sure, not in the checkerwork of any particular regenerator. Air distribution takes place in the regenerators of both ovens; in Koppers, by the sole-channel and main body of checkerwork; in Wilputte, by graduated sole-channel air deliveries to checkerwork of confined area. “Evenly distributed” — to what? Manifestly, to the heating-chamber. This is true, because the patent specification says so — “As the regenerators are placed directly in the way of the heating-gases (that is, directly in their “path”), the latter (heating-gases) are also evenly distributed in the heating-chamber!’ It is also true because air is passed from re-generator to heating-chamber of both ovens, in a way that effects distribution in its passage, namely, by the distributing operation of multiple air ports.
. Distribution of preheated air with uniformity to the heating-chamber is what Koppers tried to achieve by his invention. Distribution of air to this chamber involves, first, distribution to the re-generator; second, distribution in the regenerator; and, third, distribution from the regenerator. No matter how unheated air is distributed to a regenerator, or how it is distributed, in a regenerator, it will not be distributed from the regenerator to the multiple flame-flues of the heating-chamber in uniform volume, there to produce on combustion uniformity in heat intensity, unless it’ is ported from the regenerator and distributed to the flame-flues uniformly. Eor such uniform distribution from regenerator to heating-chamber, Koppers provided multiple ports that distribute the air vertically and directly. This was the capital thing that Koppers did. It was wholly new. It was made possible only by the change in the number and position of regenerators, and it immediately overcame the factors of inequality of air distribution incident to the horizontal and indirect bus-flue delivery of the prior' art. Air distribution by the baffle of the checker-work of a cross-regenerator would be of no more value than similar distribution in a longitudinal regenerator, if not followed by multiple direct distribution from regenerator to heating-chamber. This the prior art clearly shows. And Wilputte’s internal air distribution by graduated air feed would be of no value, if not combined with Kop-pers’ multiple direct distribution to heating-chamber. If, instead of multiple air ports in the regenerators of Koppers and Wilputte, whose sole function is to distribute preheated air to the heating-chambers, there were in either regenerator a single air port, there would be a complete reversion to the prior art, and the infirmities of the bus-flue, now overcome, would be revived.
In his invention, Koppers uses the sole-channel and checkerwork of the prior art regenerator to perform the old functions of preheating and distributing air until it reaches the top of the regenerator, when the new thing occurs. This new thing is the direct and vertical de*139livery to each flame-flue of regenerator-equalized air as distinguished from the old thing of a horizontal and indirect delivery of air similarly equalized. Wilputte does the same thing. He brings air to each compartment of his regenerator by the equivalent of a sole-channel, gives each compartment a precise dose, thus equalizing the quantity of air by sole-channel deliveries, then by the checkerwork heats just the air the dose contains. This much is new with Wilputte. But the new thing which Koppers invented then occurs in Wilputte, namely, multiple port distribution of air from the top of the cross-regenprator vertically and directly to its corresponding flame-flues above,
if Koppers’ invention were a regenerator, then Wilputte clearly would not infringe. But Koppers’ invention is not a regenerator; it is an organization of which a regenerator is only a part. This organization comprises the elements of coking-chamber, heating^chamber, regenerator, and connections, and in this organization, it is intended, that each element should perform its ordinary function. In his patent claims, Koppers covers simply a series of heating-chambers without specifying the type: a series of coking-chambers intermediate the heating-chambers, without specifying the type; combined with a series of regenerators, without specifying the type; but specifying very particularly that the regenerators shall be “below and parallel to the heating-chambers and communicating directly therewith.” Invention is in this clause, for here it is that the change in regenerators is made from the longitudinal .position of the prior art to the cross position of the patent and from indirect air distribution of the prior art to the direct air distribution of the patent. As Koppers did not invent a regenerator, Wilputte is free to use his improved regenerators in a coke-oven or to use the regenerators of the type Koppers adopted, if he does not place them crosswise the oven and parallel to the heating-chambers and does not connect them directly with the heating-chambers in the manner disclosed by Koppers’ claims. But when Wilputte placed his eompartmental regenerator crosswise the oven and made as many connections between it and its corresponding heating-chamber as there are compartments in one and flame-flues in the other, and made the connections vertical and direct, he did the things which Koppers taught. He placed the regenerators “directly' in the way of the heating-gases” (as the specification says), obtained the normal “advantage” of regenerator “distribution” and also the new advantage of multiple port direct distribution to the heating-chambers, and got the air “evenly distributed in the heating-chambers.”
[4] The heart of Koppers’ invention is the number and position of regenerators in relation to heating-chambers and in the direct connections between the two. His invention does not consist in making a new kind of regenerator; it consists in placing a regenerator of whatever kind in a new position — that is, below and parallel to the heating-chamhers — with new connections that will insure an uniformity of air deliveries. The regenerators of Koppers and Wilputte, even though differing in mode of operation, are none the less regenerators which perform broadly the function familiar to the art, that is, they receive unheated air, preheat it, and deliver it to the top of the re-*140generator at comparatively one temperature, thence to be distributed elsewhere. Variations in the method of regenerator distribution whether entirely in the sole-channel as in Wiíputte, or partly in the sole-channel and partly in the checkerwork as in Koppers, do not make the chamber any the less a regenerator. Koppers -dealt with regenerators. He dealt with their number and with their position in relation to the heating-chambers, and with their connection to the heating-chambers, not with a kind of regenerator. Given a regenerator, infringement occurs or is avoided, not according to the type of regenerator, but according to its position with reference to the heating-chamber and according to the connections between the two. When Wilputte placed his regenerators below and parallel to the heating-chambers, and made connections between the two which were multiple and direct, he got the air distribution of the Koppers invention, and in following him, we think, the defendants’ infringed.
To the plaintiff’s charge of infringement of the Koppers patent, the defendants interposed a counterclaim of infringement of claim 1 of U. S. Patent No. 673,928, granted to Schniewind, and later assigned to Otto Coking Company, Inc., one of the defendants. This issue of infringement has been elaborately developed and vigorously contested by the parties, and has been studiously followed by us. The charge of infringement is of a claim somewhat limited by specific references to the patent drawings, and is based on a very careful comparison of the elements of the two inventions. On' the surface, these elements are markedly dissimilar,'yet are claimed to be in all respects identical or equivalent. We shall do nothing more in disposing of this issue than make a very brief comparison of the two ovens.
[5] Schniewind has a pair of regenerators built entirely independent of the oven structure, to avoid the effect of their expansion on the oven brickwork. Being longitudinal of the oven, the two regenerators serve a battery of any number of heating-chambers. Koppers’ re-generators correspond in number with the heating-chambers. They are built entirely within the oven structure, and, like the heating-chambers are positioned crosswise the oven, where each regenerator serves a separate heating-chamber. In Schniewind, connection between regenerator and heating-chamber is by one flue which is vertical, then by a connecting flue which is horizontal, running crosswise .the heating-chamber like the bus-flue of the prior art, as against Kop-pers’ multiple-vertical-direct flue connection. -In Schniewind combustion takes place in a number of horizontally extending combustion-chambers, which are without divisions at the bottom and are divided into four vertical flame-flues toward the top. In Koppers, there are no such chambers, and combustion takes place in the flame-flues themselves. In Schniewind, gas is ported to the combustion-chambers where it meets the air from the horizontal air flue and ignites into a flame that extends first over the undivided portion of the horizontally extending combustion-chambers and then is carried into the vertical flame-flues. In Koppers, the air meets gas in the direct flame-flue ports and on ignition the flame is carried directly into the flame-flues and nowhere else. These, we believe to be the substantial differences *141between the two ovens. Instead of regarding them as differences, the defendants argue that they are in substance the same things. We have not been persuaded tp this conclusion. As we are clear that Schniewind, being earlier, does not anticipate Koppers, it follows that Koppers does not infringe Schniewind.
The decree below is affirmed. '