WOOLLEY, Circuit Judge.

Speaking of the parties as they stood in the trial court, the plaintiff, owner of Letters Patent No. 1,-281,884, issued in 1918 to M. J. Trumble, brought this action against the defendant (Skelly Oil Company) for infringement of claims 1, 2, 3 and 4. Against the defenses of want of invention, anticipation and non-infringement the trial court found the four claims in suit valid and process claim 2 and apparatus claim 4 infringed. (D. C.) 20 F.(2d) 995; (D. C.) 23 F.(2d) 111. The defendant took this appeal which is directed, and limited, to validity and infringement of the latter two claims under the same defenses.

The Trumble patent relates to the art of refining petroleum oils and purports to provide a process and apparatus by which petroleum oils of high specific gravity (heavy oils) may be converted into oils of low specific gravity (light oils) at a saving of cost. More particularly, the invention relates to that branch of the art of refining petroleum oils known as “cracking.” The art is old and as Trumble’s invention is an improvement upon one of its late developments it will be necessary to state the art in outline in order to understand what Trumble did in an inventive way and what he accomplished.

Crude petroleum consists of a complex mixture of hydrocarbons differing in specific gravity. Stated according to the range of their specific gravity from low to high, they *428 are ineondensible gases, gasoline, kerosene, gas oil and lubricating oil, leaving fuel oil, tar, pitch, asphalt or coke as a residue. These closely combined mixtures may be physically separated and recovered at atmospheric pressure by progressive vaporization at their respective boiling points through the familiar process of fractional distillation practiced in many arts. Air Reduction Co. v. Carbo-Oxygen Co. (D. C.) 17 F.(2d) 138; (C. C. A.) 19 F.(2d) 1014; Southern Electro-Chemical Co. v. DuPont Co. (C. C. A.) 20 F.(2d) 97,101. So, also, they may be recovered by cracking. But cracking is an operas tion wholly different from fractional distillation. It involves a chemical reaction and effects a partial conversion, or breaking up, of the complex hydrocarbon molecules of the heavier constituents of the mixture, by the application of heat and pressure, into the molecular structure of the desired lighter distillates, for instance, gasoline, the one now greatly demanded of the industry and chiefly concerned in this litigation. An unavoidable incident of the cracking process, however, is that the rearrangement of molecules yields free carbon and heavier oil and expels some ineondensible gases which are not desired because of little value and mainly because they add to the cost and danger of producing the valuable light distillates, particularly of producing gasoline, which requires more heat (about 750° P.) and higher pressure (about 75 pounds) than in producing kerosene, the heavier distillate, whereby these unwelcome by-products are increased correspondingly.

Tar, coke and foreign matter, inevitably produced in some degree in every cracking operation, are more than undesirable; they are highly objectionable for two reasons; one, they accumulate on the heating surface of the still and thus retard the transfer of heat to the oil and, what is more serious, they cause overheating of the still walls with consequent danger of leaks, blow-outs and fire; and the other, as the operation must from time to time be stopped and the still, thus idle, be cleaned out, they determine the length of the run or. time the still may remain “on stream,” that is, the time the still is in operation and earning money. To meet these problems the art progressed on lines and developed oil cracking devices and procedures differently stated by the opposing parties. These differences can, by some elaboration, be reconciled.

The defendant divides devices and procedures of the art before Trumble into two classes: (a) Bulk-oil cracking in boiler stills; (b) stream-oil cracking in coil stills; claiming that its practice is within the former and Trumble’s within the latter and, in view of the latter, lacks invention.

The plaintiff divides the devices and procedures of the art prior to Trumble into three classes: (a) The directly fired shell or batch still; (b) the once through still or system; (c) the cycle still or system; and claims that Trumble’s method and apparatus are improvements upon the last; that, in view of what had gone before, they involve inven-' tion; and that the defendant’s practice also comes within the last as improved by Trumble, and infringes.

The essential features of a directly fired, or batch, still are shown in the late Burton still (Patent No. 1,049,667 issued in 1913). It comprises a cylindrical shell or boiler, charged about one-half full with oil, the heating flames playing directly on the bottom sheets. The generated gases arise in the vapor space above the oil. The heavier vapors are condensed and return to the still while the lighter vapors pass to a coil condenser, whence, after condensation into liquid, they are conducted to a receiver and there recovered.

It is important to note that in the operation of a batch still no residue is taken off during the run; that a still of this type can-remain on stream only about thirty-six hours; that it must then be shut down, the tar, coke and foreign matter removed and the bottom scraped; that it takes about twenty-four hours to clean, out the still, and that this involves substantial cost and operative loss. Antecedent variations of the set-up of the Burton batch still may be found in'Letters Patent No. 419,931 to Dewar and Redwood, 1890; No. 28,246 to Atwood, 1860; and No. 52,151 to Pales, 1866.

Next, the art thought that sedimentation of foreign matter upon the walls of a still with its consequent physical and financial disadvantages could be avoided, or at least lessened, if the oil were moved over the heating surface so rapidly that there would not be time for the matter to settle. So there developed the “once through” still, which takes its name from what happens in the still. Of this type patents Nos. 342,564 and 342,565 to Benton in 1886 are fair disclosures. Briefly described, this set-up comprises a furnace with a pipe positioned above the flames and extending throughout and beyond the furnace length. Oil is fed into the pipe at one end, through an opening in the pipe intermediate the two ends lighter vapors flow into a releaser and then are conveyed into a condenser and what is left of the oil is dis *429 charged, from the other end of the pipe. It should be observed that oil passes through the heating medium once, hence the term “once through.” It is therefore heated only once, and then in transit. There were several infirmities in this system; one, the loss of heat attendant upon the single passage of the oil through the pipe; the other, incomplete cracking in one passage with consequent loss of the residue or cost of re-treatment.

Systems of this type are described in one of the set-ups in the Russian patent No. 4,782 to Sehuchow and Gavrilow, 1891; patent No. 27,842 to Stombs and Brace, 1860; showing a battery of directly fired shell batch stills with a continuous bulk-oil feed flow and discharge, circulation increased by patent No. 340,878 to Henderson, 1886; No. 477,153 to Pielsticker, 1892, as to one phase.

Whether or not the once through system was an advance step in the art, clearly it was a step away from the batch still, and it is to be credited with the kernel of a new and, later, very useful conception. As the feed and discharge of the once through system are continuous and vaporization likewise is continuous, it became one of two systems embodying the idea of continuous operation, the other being the cycle system.

The cycle type of still, like that of the once through type, derived its name from what happens inside, In the process of being cracked, oil moves in a cycle and the organization of the still is adapted to produce such a movement. The apparatus disclosed in connection with the method patent to Clark, No. 1,119,496, an improvement on Burton, issued in 1914, is a simple illustration. There, oil in an elevated reservoir located away from the furnace descends to a pump whence it is forced through a pipe to a coil within a furnace ; ascending, it returns heated to the reservoir from which it started and continues on and on in this cycle while the resultant vapors pass from the top of the reservoir through a pipe to a condenser and finally to a drum for receiving the distillate. No residuum is taken off during the on-stream operation. The Clark-Burton set-up seems to be a commercial adaptation of the Clark invention. Here, too, no residuum is drawn off when on stream. While an improvement on the batch and once through systems, the cycle system also must be shut down for cleaning certain of its parts, in this instance after a run of approximately forty-eight hours, an increase of twelve hours over the batch still run. -

The Russian patent, supra, will be discussed in some detail under another issue. It will be enough to say now that its second set-up discloses a cycle movement of oil with the usual drawing off of vapors but no drawing off of the residuum during a run.

It was just here that Trumble appeared and proposed an improvement upon the cycle system by recirculating the heavy uncracked oil and doing it in a way to draw off in large measure the objectionable free carbon, tar and other foreign matter. What he did, and apparently all he did, was to take a typical cycle system with furnace, heating coils, condenser, pump, etc., and after suitably arranging them, disclose a practice and suggest a means continuously to draw off a small portion of the revolving residuum (loaded with the objectionable matter) in each of its cycles through the dosed system and replenish the bulk with an equal amount of new oil stock and thereby maintain the bulk constant in volume. That was all. What was the result? An increase in a run from forty-eight hours to seventeen days. For this he was awarded the patent in suit, claims 2 and 4, in issue, reading as follows: “2. The process of converting a heavy petroleum oil into a light one which consists in circulating the heavy oil through a closed ring, continuously heating a portion of the ring, maintaining a pressure on the portion of the ring so heated, continuously taking off light vapors from the ring, continuously relieving the ring of a small amount of heavy oil, and continuously supplying fresh oil in sufficient quantities to maintain constant the volume of oil in the closed ring.”

“4. An apparatus for converting heavy petroleum mixtures into light petroleum oils, comprising a heating means; a vapor releaser; means for conducting the heated mixture from the primary heating means to the vapor releaser; means for withdrawing light vapors from the vapor releaser; means for continuously withdrawing a portion of the mixture; means for forcing the residuum from the vapor releaser through the primary heating means; and means for injecting sufficient fresh heavy oil into said residuum to maintain the volume of oil in the apparatus approximately constant.”

Trumble’s invention resides not in the cycle movement of oil in a closed system or structure, for that was old, but in drawing objectionable foreign matter out of the system by continuously discharging or drawing off a portion of the degraded oil as it circulates with the foreign matter in suspension and continuously feeding back an equal amount of fresh oil whereby the bulk of oil in the system, thus continuously changing in *430 quality, shall all the while be kept constant in quantity as it goes around and around in the closed system, which Trumble called “a closed ring.” Trumble’s closed ring as described in the patent comprises a feed pipe constantly charging fresh oil into a heated coil located in a furnace above the flames where the oil is heated to cracking temperature; a pipe leading from the coil and conveying the heated oil and vapors to a releaser from which the vapors are drawn off and through which the residue of heavy oil descends to another pipe having an outlet through which a portion of the heavy oil is drawn out of the system and the balance is returned to the heating coils where it joins the main body of oil refreshed by new oil from the feed pipe and then proceeds on another cycle; circulation of the always changing bulk being effected at a determined speed by a pump. If this is not a mere variation of what already existed in the art but is new in that it is something abruptly different from what existed before, then, utility being evident, the patent is not invalid for want of invention in view of the art. The plaintiff declares, again and again, that no process or apparatus in the prior art operating on the cycle system used or embodied these elements, and that Trumble was the first to realize that the period of a run of a cycle system with its attendant advantages could be materially lengthened by continuously withdrawing from the ring an appropriate amount of tar and coke bearing oil and providing a continuous supply of fresh charging stock. The defendant has not successfully met, nor have we found anything in the evidence that contradicts, these assertions. Therefore we find invention and, on the record before us, hold the claims of the patent which are in suit valid in view of the art — unless, as the defendant next urges, they are invalid because anticipated by the art.

The defendant says that the claims of the Trumble patent, as the District Court interpreted them to inelude its practice, are directly anticipated by United States patents No. 340,878 granted to Henderson in 1886; No. 1,170,884 granted to Edwards in 1916 after the date of Trumble’s application on an application filed May 8, 1915, which was before Trumble’s application on July 19 of the same year; No. 1,239,424 granted to Smith September 4, 1917, on an application filed April 28, 1915; No. 836,732 granted to Barbet in 1906, and a corresponding Drench patent granted to him in 1902; and finally the Russian patent (No. 175) granted to Schuehow and Gavrilow in 1891.

The Barbet patents disclose a directly, fired shell batch still having a continuous feed of charging stock and a continuous draw-off-. of residuum actuated by a “stirrer” mounted in the shells. They do not disclose oil of changing quality circulated through a closed ring or anything that told the art precisely or substantially what Trumble told it. These references do not meet the evidentiary rule of anticipation.

The Edwards and Smith patents are held not to anticipate on a finding that, if covering the same subject matter, their earliest dates of invention were later than Trumble’s.

Henderson added to the Stombs and Brace continuously operated battery of directly fired batch stills a means “to promote circulation” in the form of revolving blades, dishes, or buckets hinged to the side walls of the stills. Although the patent deals with “circulation” we fail to find a true cycle system disclosed. While the dishes or buckets may reduce sedimentation and aid circulation in the sense of promoting the oil movement, their operation more nearly resembles agitation than circulation.

The defendant places its main reliance for proof of anticipation on the Russian patent to Schuehow and Gavrilow. Primarily, and according to the patent title and disclosure, the system is a “continuous” one; but, as noted before, a continuous system may be either a cycle or a once through system. The patent discloses both. Each acts independently of the other and for different purposes according to the will of the operator.

However read, the patent discloses alternate systems. The oil comes in from a pump at B, passes through the pipe C, thence up through helical heating coils into a cylinder A, from which, after the light vapors have been drawn off through the pipe E, it descends as reflux condensate through the pipe D at whose bottom it may be drawn off at the valve P. That, falling a little short of a single cycle or revolution, is a once through system. Or valve P may be elosed and the oil shot back through valve R to the pipe B whence, continuing through pipe C, the oil may be re-circulated through the system indefinitely. That is a eyele system. As to these two systems the applicable words of the patent are: “In accordance with the speed of circulation of the oil in the pipes, and depending on the purpose of the distillation«, the liquid oil which has settled in the cylinder A is discharged by the pipe D and through the valve P, without being subjected to further ’ distillation, or by valve Q enters the pump once more, where it is mixed with a *431 new supply of oil running towards the pump, and resumes circulation through the coiled tubes.” If, in the second set-up, the “new supply of oil running towards the pump” be a “continuous feed,” we find no reference to a corresponding continuous bleed. There is of course a continuous diminution of the bulk by evaporation.

It would seem that the patent thus discloses only two methods of operation, alternate and optional with the operator. But the defendant says that the apparatus is capable of a third method, namely; the simultaneous opening of both valves P and Q a little bit whereby through the former a portion of the degraded oil will be drawn-off and through the latter the remainder (refreshed by new stock) will be carried back through the system and recirculated. Looking at the diagram submitted as an illustration, this would seem to be possible, but we find no such conception or use disclosed by the patentees. The defendant emphasizes these words of the patent: “In this manner, by means of the valves P and Q, a greater or less degree of a repeated process of circulation is obtained, as also of the distillation dr cracking of the oil.” The “manner” there referred to is the alternative manner of operation described in the sentence last quoted and immediately preceding this sentence. We think the two sentences, being placed together, should be read together and when so read they do not suggest a third system.

This third system proposed by the defendant is really a composite of the two systems described by the patent. It is only by combining these alternative systems with their alternative valves P and Q, that the defendant produces Trumble’s continuous bleed and feed.

As there is no evidence of the use of the invention of the Russian patent in the practical art, we are limited for information to the disclosures of the patent itself. Does the patent disclose this third practice? We find nothing to that effect. Does it teach that both valves P and Q should be opened simultaneously, one to draw off some of the oil loaded with impurities and the other to replenish the bulk, and together permit a continuous recirculation of old and new? We cannot find that it does. Nor does the patent state the practicability or desirability of the third practice? Apparently it never occurred to the patentees. The defendant’s conception of the third method is only an inference from the specification. Inferences as distinguished from disclosures, especially when drawn in the light of after events, cannot be accepted as a basis of anticipation.

A patent relied upon as an anticipation must itself speak. Its specification must give in substance the same knowledge and the same directions as the specification of the patent in suit. Otto v. Linford, 46 L. T. (N. S.) 35, 44. It is not enough to prove that a. method or apparatus described in an earlier specification can be made to produce this or that result. Flour Oxidizing Co. v. Carr & Co., 35 R. P. C. 457. A singularly sensible test of the rule of anticipation is given in British Thomson-Houston Co. v. Metropolitan Vickers Electrical Co., 45 R. P. C. 22, by asking the question — “Would a man who was grappling with the problem solved by the patent attacked, and having no knowledge of that patent, if he had had the alleged anticipation in his hand, have said: ‘That gives me what I wish?”’ The Pope Alliance Corporation v. The Spanish River Pulp & Paper Mills, Ltd. (Privy Council Appeals No. 33 of 1928). On the evidence in this case, read in the light of the rules of law applicable to anticipation, we do not find the invention of the patent in suit anticipated.

It is difficult to state the issue of infringement without,repeating the evidence and reproducing the diagrams of the contesting apparatus. Omitting details, the defendant’s set-up comprises the usual elements of furnace, heating means, dephlegmator, eonneeting pipes and pump. In place of either coiled tubes or shell still as a heating means the defendant uses a modem high pressure Babcock & Wilcox boiler with its familiar bank of inclined tubes in which is heated and cracked a very large volume of oil. The Trumble patent discloses a long coiled pipe as a heating means in which the volume of oil is not large. Since the form and size of the heating element of the invention are not claimed by the’ patent as novel and as they have nothing to do with the essence of the invention, the defendant’s substitution of one kind of heater for that preferably or incidentally disclosed by the patent falls out of the issue of infringement.

Oil is fed to the defendant’s system through a pipe leading into the dephlegmator through which, gathering heat in transit, it descends to the boiler, or rather to a drum on one side of the boiler, whence by force of a propeller pump it ascends through the battery of inclined tubes, subjected to flames from the furnace,, to another drum on the other side, then ascends to the open space in the upper part of the boiler and thereafter revolves around and around within the boiler *432 at a very rapid speed. The vapors arising from the surface of the bulk of oil in the upper part of the boiler, away from the flames, ascend through a pipe to the dephlegmator from which they are drawn off and the unsuccessfully cracked residuum is returned, as in other cycle systems, to the h oiler and there, joining the rapidly moving hulk of oil, it is again subjected to the oil cracking temperature. Except as to the type of heater, the defendant’s set-up thus far described discloses the main characteristics of the Burton batch still and Clark’s cycle improvement upon it. If the defendant had stopped there it would not have infringed Trumble. But to the operative essentials of Burton and Clark, or, speaking generally, to the essentials found in prior systems, the defendant made at least one very important addition. This was an outlet from the second drum through which it continuously drew off a quantity of degraded oil equal to fresh oil which it injected continuously through the inlet in another part of the system. This addition to the defendant’s otherwise unoffending set-up manifestly constituted a drawoff or bleed closely akin to that of Trumble. Even so, the defendant, while admitting a continuous circulation of oil coupled with continuous feed of fresh oil and continuous drawoff of heavy residuum, centers its position on the cycle within the boiler and maintains that its system retains the characteristics of bulk-oil process in boiler stills and is not a true cycle within the sense of Trumble’s closed ring because the circulation is through the battery of heated tubes and the vapor releaser part of the boiler where alone, it says, heating and cracking occur, and because its cycle is only a cycle of movement while in Trumble’s system the oil is subjected to a cycle of varied process steps. We have not been able to follow this line of reasoning for, as it seems to us, there are in the defendant’s set-up really two cycles, one within the other; each, if considered apart from the other, having a continuous drawoff and feed in a closed system. One cycle is within the boiler or heater itself, on which the defendant lays stress as a bulk-oil process in a boiler still, for it is there the bulk of the oil is circulated and cracked and there it receives all the heat it ever receives; but over and beyond the cycle movement of oil in bulk within the boiler there is another cycle movement of oil and vapors under different conditions of heat and undergoing different treatment. This cycle is the movement from the boiler to the dephlegmator, and, after the vapors have there been withdrawn, the movement of the residuum back to and through the boiler for re-treatment.

As the defendant’s system is not restricted to the boiler but extends to and embraces the other elements, it appears to be a true cycle system differing from other such systems in the character of boiler or heater employed and distinguished from others mainly by the larger bulk of charging stock, but resembling Trumble in the addition of a continuous feed and bleed in a measure that maintains the charge in constant volume as it revolves in a closed ring, of which the boiler is only a part. True, the defendant feeds and -withdraws a larger quantity of oil than Trumble suggests in his specification, but as the claims contain no limitation upon the quantity, that will not avoid infringement. We are constrained to find that by its practice and set-up the defendant has infringed process claim 2 and apparatus claim 4 of the patent.

The decree of the District Court is affirmed.