Dow Chemical Co. v. Coe

STEPHENS, Associate Justice: This is an appeal from a judgment of the District Court of the United States for the District of Columbia dismissing the appellant’s action brought under Rev.Stat. § 4915 (1875), 35 U.S.C.A. § 63. The action sought to authorize the Commissioner of Patents to issue to the appellant a patent containing claims Nos. 22 to 34, inclusive, of application for patent Serial No. 44,252, filed October 9, 1935, by Edgar C. Britton and others, and thereafter assigned to the appellant. The application relates to the preparation of alkyl halides by the direct addition of hydrogen halides to olefines, particularly to methods for the preparation of alkyl chlorides from olefines and hydrogen chloride, and, more particularly, to methods for the preparation of ethyl chloride by combining ethylene and hydrogen chloride. The term “alkyl halides” as used in the application means alkyl halides which have at least two carbon atoms in the molecule. Claim 22, which reads as follows, is representative. “22. In a process for the production of ethyl chloride, the step which consists in passing ethylene and hydrogen chloride, under substantially anhydrous conditions, into a liquid reaction bath consisting sub — 578 — stantially of aluminum chloride and a chlorinated lower aliphatic hydrocarbon having a boiling point above that of ethyl chloride, while the bath is maintained at a temperature above 12° C. and below the boiling temperature thereof and under approximately atmospheric pressure.” It is to be noted that this claim does not explicitly describe the process as a continuous, as distinguished from a batch, process, but other claims, to-wit, Nos. 30, 32, 33, and 34, do in terms describe the process as continuous; and it was testified by an expert witness, Dr. Henry B. Hass, called in behalf of the appellant, that if the teaching of claim 22 is followed the process described must be operated as a continuous one in the sense that the reactants are being supplied and the product being removed from the bath at the same time; and it is not in dispute that the process described in claim 22 is continuous within the definition of continuous stated by this witness. The trial court held the process unpatentable over the patent to Curme of December 9, 1924, No. 1,518,182, and the patent to Nutting and others of July 2, 1935, No. 2,007,144. The court made findings of fact describing the process of the appellant and of Curme and Nutting, and concluded : “1. It was not invention to use a chlorinated lower aliphatic hydrocarbon as the reaction solvent, or to carry out the process of the patent to Curme 1,518,182, continuously, in view of the patent to Nutting 2,007,144. “2. It was not invention to use the process disclosed by the patent to Nutting, et al. to produce ethyl chloride in view of the patent to Curme 1,518,182.” The -court held that claims 22 to 34, inclusive, of the Britton application were, therefore, not patentable over the prior art, and that the appellant in consequence is not entitled to a patent on the Britton application. The Britton process, as described in the application and explained by the evidence, is carried out thus: Ethylene and hydrogen chloride are passed into a reaction vessel which contains a liquid bath consisting of a chlorinated lower aliphatic hydrocarbon such as tetrachloroethane or tetrachloroethylene (non-aqueous liquid solvents which are non-reactive under the conditions of the process), in which is suspended a catalyst, aluminum chloride. The reactants and the vessel are maintained under approximately atmospheric conditions as respects pressure, and at a temperature ranging from about 12° C. to about 40° C. The lower figure is the boiling point of ethyl chloride and is below the boiling point of the liquid medium chosen for the reaction bath.1 Since the temperature is kept at a point at or above the boiling point of ethyl chloride and below that of the solvent, the ethyl chloride, as it is formed by the reaction between ethylene and hydrogen chloride in the presence of the catalyst, vaporizes -and rises above the reaction bath and is drawn from the reaction vessel in vaporous state and captured through condensation. The solvent does not enter into the reaction. Polymerization products of the reaction have the fluidity of lubricating oil, and they may, therefore, be and are withdrawn as the reaction proceeds, and consequently do not clog the reaction vessel. The process may be continued indefinitely if new reactants and catalyst are added to the bath as the old are being expended. Uncontradicted evidence for the appellant shows that the process has been continuously operated on a semi-plant scale for months, and that it produced yields of the end product of from 90 to 95 per cent of the possible maximum to be derived from the raw materials passed into the reaction vessel. The trial court found the process to be continuous, simple, inexpensive, and efficient. The chemical reaction involved in the Britton process is explained as follows: An ethylene radical contains an unsaturated bond between its two carbon atoms. This bond may be made the basis of an additive reaction resulting in a saturated compound by severance into two component bonds which are then free to attach themselves to two additional atoms. In the Britton process such an additive reaction takes place thus: Aided by the catalytic action of the aluminum chloride, the hydrogen and chlorine atoms of a molecule of hydrogen chloride separate from each other and adhere to the two free bonds formed by the severance of the unsaturated bond between the two carbon — 579 — atoms of the ethylene radical. The equation for the reaction is as follows: It is not disputed that this reaction was well known prior to the disclosure of Brit-ton. It is the reaction of Curme. The Curme patent claims a “Process of making ethyl chloride which comprises causing anhydrous hydrochloric acid to react on ethylene, under superatmospheric pressure and in the presence of aluminum chloride.” The specification states that since ethylene does not combine with hydrochloric acid as readily as propylene, “high pressures or the presence of a catalyst are required if any substantial yield is to be obtained. Pressures up to 160 atmospheres, without a catalyst, and aluminum chloride as a catalyst, without pressure, have each been tried. In either case ethyl chloride is produced in substantial quantity. The best method for producing ethyl chloride employs both a catalyst and pressure. In the presence of aluminum chloride and at pressures no higher than 35 atmospheres, commercial yields are obtainable even at room temperature.” The specification discloses the use of gaseous materials, but adds that “The invention is not limited to a reaction between the vapor phases of the combining substance[s] as these may react in solution.” It states that the solvent employed may be the reaction product, or other suitable solvent, preferably anhydrous. The process is a batch process, as distinguished from a continuous one. The chemical equation for the Curme reaction is, of course, identical with that for the reaction of the Britton process. The Nutting patent teaches an improved method of making ethylidene chloride by reacting hydrogen chloride with vinyl chloride in the presence of a catalyst. The specification states: “Broadly, our method consists in dissolving or suspending the catalyst, aluminum or ferric chloride, in a non-aqueous liquid medium and thereafter reacting the hydrogen chloride and vinyl chloride in the resultant mixture at between about 10° and 135° C., preferably between 15° and 60° C. The liquefied reactant, vinyl chloride, or the product, ethylidene chloride, is preferably employed as the reaction medium, since the reaction occurs smoothly and rapidly in the presence of either of said compounds and their use simplifies the step of purifying the product. However, other non-reactive solvents, e. g. trichloroethane, acetylene, tetrachloride, etc., may be employed, if desired.” The specification states also that the “reaction may be carried out at atmospheric pressure, but is more rapid and complete at a higher pressure.” All of the claims of the Nutting application describe a batch process, except one, claim 6. That claim describes a continuous process, which is the one pertinent here. It is referred to in the specification as follows: “A convenient mode of operation is to mix the aluminum or ferric chloride catalyst, with ethylidene chloride, pass hydrogen chloride and vinyl chloride simultaneously into the mixture, and distill the ethylidene chloride product directly from the reaction mixture either periodically or as it is formed.” Claim 6 describes this process thus: “In a method of making ethylidene chloride, the steps which consist in passing hydrogen chloride and vinyl chloride, simultaneously and at approximately atmospheric pressure, into a heated mixture of ethylidene chloride and a catalyst selected from the class consisting cf aluminum chloride and ferric chloride, distilling ethylidene chloride from the mixture as it is formed, condensing ethylidene chloride from the distilled vapors, and returning any uncondenced vapors to the reaction.” The equation for the reaction of the Nutting process is: — 580 — The trial court found that the Nutting patent disclosed “an operative continuous process” for the production of ethylidene chloride. — 579 — — 580 — • It will he noted that the Britton, Curme, and Nutting processes have certain similarities. The same catalyst is used in each. The additive type of reaction is the basis of each. Each adds hydrogen chloride to a compound containing an olefinic double bond. So far as formulae are concerned, there is no difference between Britton and Curme. Comparing Britton and Nutting, ethylene, the beginning reactant of the former, differs from the vinyl chloride of the latter only in the substitution of one atom of hydrogen for one of chlorine; and the end products of Britton and Nutting differ only in the substitution of one atom of chlorine in the ethylidene of Nutting for one of hydrogen in the ethyl chloride of Britton. Hydrochloric acid is used as a reactant in all three processes. The solvents of Britton and Nutting are also similar, both using chlorinated lower aliphatic hydrocarbons. Curme’s specification, being largely in terms of the vapor phases of the reactants, does not mention solvents except to say that the invention is not limited to a reaction between vapor phases of the combining substances as they may react in solution, that the solvent employed may be the reaction product or other suitable solvent, preferably anhydrous. But it is not contended that the solvents are critical, in respect of the question of invention, in any of the processes. The contention of the Commissioner is that one skilled in the chemical art would realize that a process (Nutting’s process) which adds hydrogen chloride to vinyl chloride could be used for adding hydrogen chloride to ethylene; that it is apparent that both ethylene and vinyl chloride have a reactive olefinic bond and that the fact that vinyl chloride, as compared with ethylene, has an atom of chlorine in place of one of hydrogen would not be understood as interfering with the additive reaction. Putting it otherwise, the Commissioner contends that it did not require invention to carry out Curme’s batch process continuously in view of the continuous variant of Nutting, or to substitute in the Nutting continuous variant the ethylene of Curme, to produce ethyl chloride. Stating it more broadly, the Commissioner asserts that the substances and processes of the prior art are so analogous to the Britton process that it cannot be said to have required more than the skill of the art, as distinguished from inventive genuis, to achieve it The appellant contends that it required invention to achieve the Britton process and bases the contention upon the testimony of two expert witnesses, Dr. Henry B. Hass, above mentioned, head of the Department of Chemistry at Purdue University, and Mr. James L. Amos, a chemist engaged in research for the appellant, and upon certain contentions of law and fact made in connection with this testimony. The testimony of Dr. Hass, which we state in part in terms and in part in substance and effect, was as follows: Brit-ton’s process consists essentially in four things: first, the utilization of an aluminum chloride catalyst; second, the use of a reaction bath boiling at a point higher than the boiling point of ethyl chloride, the product, and therefore capable of being maintained in the reactor as the ethyl chloride boils away; third, the use of a temperature above the boiling point of ethyl chloride and below that of the reaction bath; fourth, the use of atmospheric pressure. These four things combine to render possible the continuous operation of the process. In respect of the comparison between the Britton and Nutting processes, and particularly concerning the seeming identity of the two primary reactants, vinyl chloride and ethylene, Dr. Hass said in terms: "... To a chemist somewhat unfamiliar with ethylene and vinyl chloride, it might seem that there is an extremely close analogy here. The only difference between ethylene and vinyl chloride as far as structure and composition is concerned, is that vinyl chloride has a chlorine atom where ethylene has a hydrogen atom. The chemical and physical properties of vinyl chloride and ethylene are, however, widely dissimilar. In the first place, ethylene is uniquely unreactive among the simple olefines. It reacts more sluggishly toward practically every reagent which attaches itself to this double bond than is true of almost any other simple olefine. Vinyl chloride, on the other hand, is just the opposite. The introduction of that chlorine atom into the ethylene molecule, converting it to vinyl chloride, changes the nature of the molecule to the extent that vinyl chloride is an extremely reactive substance. For example, vinyl — 581 — chloride is used very widely for the manufacture o’f synthetic resins. . . . “. . .So the first essential difference between these two processes, as I view it, is that in the case of ethylene we have an extremely sluggish double bond, one which reacts with the greatest of difficulty; whereas in the case of vinyl chloride we have an extremely reactive double bond. Secondly, there is a considerable difference in boiling point between the ethyl chloride boiling at plus 12° C. which is produced when ethylene adds hydrogen chloride, and the ethylidene chloride boiling at 58° C, which is the compound formed when vinyl chloride adds HC1 or hydrogen chloride. This higher temperature of boiling means that if the process is to be carried out continuously at atmospheric pressure, as the claim specifies— "... —The Nutting claim specifies, Claim 6 ... it is necessary to heat the reaction bath to a temperature above the boiling point of the ethylidene chloride, the boiling point being 58° C., which is a much higher temperature than is necessary to distill ethyl chloride from its reaction bath, ethyl chloride boiling at 12° C., 46° C. lower than the boiling point of ethylidene chloride. “Now, the importance of this difference in boiling points consists in the reversibility of these hydrogen chloride additions. The addition of hydrogen chloride to ethylene or to vinyl chloride comes under the heading of what an organic or any chemist calls a reversible action. That means it is possible to start with the reaction products and produce from them the reagents used, and because of the nature of this reaction at higher temperatures, that reversibility becomes more and more important, so that if one goes to fairly high temperatures, not excessively high, it is possible to start with ethyl chloride, for example, and produce from it ethylene and hydrogen chloride by an exact reversal of the process in issue. “On the other hand, the same thing can happen with ethylidene chloride. It is possible to heat that compound and obtain from it hydrogen chloride and vinyl chloride. And because these reactions become more reversible as the temperature increases, the higher temperature necessitated, if we are to boil out the ethylidene chloride, becomes a matter of considerable importance. The most important difference in my judgment, however, between these two reactions which formerly seemed so analogous, is in the nature of the polymer produced. It is possible to polymerize ethylene in the presence of aluminum chloride. That has been done. The ethylene polymer when produced, however, (and it is produced in very small quantities as this reaction proceeds) it is of a character resembling lubricating oil; it drains easily from the reaction vessel. The polymer from the vinyl chloride, on the other hand, is a tarry, gummy, resinous material which does not drain out but which forms a pitchy, gummy material which plugs the reactor and prevents continuous operation from taking place. “To recapitulate then, there are, it seems to me, three essential differences here between these two apparently analogous reagents— “Firstly, the sluggish nature of ethylene towards polymerization compared to the very great ease of polymerization of vinyl chloride ; "Secondly, the higher temperature necessitated by the higher boiling point of ethylidene chloride as compared to ethylene chloride; and “Thirdly, the nature of the polymer which is produced, which in the case of ethylene drains easily from the reaction vessel, whereas in the case of the vinyl chloride it is still gummy.” In respect of the Curme patent, with which he said he was familiar prior to being consulted about the instant case, Dr. Hass testified in substance and effect as follows: Curme is one of the most outstanding chemists in this field in the country. He states that for commercial yields both high pressure and catalysts are necessary. Curme’s patent states that pressures up to 160 atmospheres, without a catalyst, and aluminum chloride as a catalyst, without pressure, have each been tried, and in either case ethyl chloride is produced in substantial quantity. Thus far in the patent it might be understood that one runs the process without pressure, but when the entire patent is examined, including the claims, every claim calls for both catalyst and pressure. Curme states that the best method for producing ethyl chloride employs both a catalyst and pressure and that in the presence of aluminum chloride and at pressures no higher than 35 atmospheres commercial yields are obtain — 582 — able, even at room temperature. That indicates that Curme was trying to teach that the desirable operating conditions and the only ones worth covering in the claim, the only practical ones, were those using super-atmospheric pressure. While it is a theoretically sound conclusion that any reaction which occurs at super-atmospheric pressure will occur to some extent at atmospheric pressure, nevertheless the amount of reaction may be so slight as to be wholly impractical at atmospheric pressure, while at super-atmospheric pressure the amount of the desired reaction may be very great and commercially feasible. With that law of chemistry in mind, a skilled chemist reading what Curme states would assume that he had concluded that his process, when operated at atmospheric pressure, does not give commercially desirable results. Asked what impression was made on him when he discovered by conference with representatives of the Dow Chemical Company that they were actually making ethyl chloride at atmospheric pressure (apparently the Britton process is here referred to), the witness said: “. . .1 was distinctly astonished to realize that Curme had, in the vernacular, ‘missed the boat’. When one deals with such a substance as hydrogen chloride, which under many conditions is corrosive, it is undesirable to work at super-atmospheric pressure if that can be avoided. Curme is one of our very best chemists. To think that he had been the pioneer in this field, had worked out a process which seemed very logical, and then to learn that someone coming later had found much more desirable conditions for carrying out this reaction, gave me a very distinct sensation of surprise.” On the question whether the disclosure of Nutting would have led to the Britton process, Dr. Hass’ testimony was this: “Q. Now, if you take the disclosure of Nutting, is there anything there that would lead you as a chemist to try the thing on the ethylene reaction to make ethyl chloride? A. Just to the contrary. Ethylene being the most sluggishly reacting of all the simple olefines, vinyl choride having as its outstanding characteristic its ease of adding to other molecules, I would conclude that a set of conditions suitable for vinyl chloride would in all probability be unsuitable for reaction with ethylene. “Q. The one wouldn’t lead you to try the other? A. The one would lead me directly away from the other.” Finally, as to whether a knowledge of Nutting in view of Curme would lead to Britton without experimentation, Dr. Hass testified, in substance and effect: Neither Nutting alone nor Curme alone nor Nutting and Curme together would lead to Britton without the need of experimentation. Because of the different reactive qualities (which the witness had already referred to) of ethylene and vinyl chloride, Nutting would not have suggested Britton; because of its teaching the use of super-atmospheric pressure, Curme would not have suggested Britton. The witness Amos testified concerning the success of -the Britton process when carried out in a semi-plant form, concerning the Nutting continuous variant when run in the laboratory, and concerning the stability of ethylidene chloride, Nutting’s end product, compared with ethyl chloride, that of Britton. In respect of the success of the Britton process, Mr. Amos stated, in substance and effect: In connection with work with the appellant he had carried out both in the laboratory and in a semi-commercial plant development the continuous ethyl chloride process of Britton. In the semi-commercial plant development a continuous unit was operated for a period of months without any trouble due to clogging or stoppage of the reactor by means of tars, gums, or other products produced by the reaction. Three hundred pounds a day of ethyl chloride were continuously produced. These tests followed the steps of claim 22 of Britton. In respect of the Nutting continuous variant (Nutting’s claim 6), the witness testified : This was run in glass, i. e., in the laboratory, with the reactants described by Nutting and with both tetrachloroethane and ethylidene chloride, the reaction product itself, as the bath, and within the temperature ranges described by Nutting and at approximately atmospheric pressure. After four hours of operation in a run designated Run 7 “The percent of total vinyl chloride which was passed through the reactor and converted to ethylidene chloride was only 33.8% as compared to the continuous ethyl chloride process of 90 to 95%. The percent of the vinyl chloride used in the reaction — that is, used up and transferred to ethylidene chloride, was only 53.5% as compared to the ethyl chlo — 583 — ride process of 90 to 95%. The percent vinyl chloride not recovered as vinyl chloride from the products coming from the reactor or as ethylidene chloride, was 29.5%. In other words, 29%% of the vinyl chloride did not go to either ethylidene chloride or was recovered as unreacted vinyl chloride.”2 The majority of the 29.5 per cent “polymerized to the black, thick residue which deposited on the rings of the reactor,” so that in the process only about 30 per cent of the vinyl chloride was made into the product desired, and almost as much was made into black tar, and at the end of the four-hour operation the apparatus was not in fit condition to continue for the reason that it plugged at the entrance of the gaseous materials into the reactor. Run 8, which was carried on with tetrachloroethane as a solvent in place of ethylidene chloride, ended in four and a half hours — again because of the development of tarry material. The results reached in Run 8 were more favorable, so far as the production of ethylidene chloride is concerned, than those of Run 7, but the maximum figure was only 42.7 per cent. In view of the experience with these two runs, the Nutting continuous variant was not to be considered a continuous process; the best runs obtainable were a very few hours during which time practically as much gum was made as product; the Nutting yields were not to be regarded as commercially substantial. In respect of the comparative stabilities of ethylidene chloride and ethyl chloride, Mr. Amos ran a test in the presence of the trial court. This, he testified, demonstrated the instability of ethylidene chloride as compared with ethyl chloride at the operating temperatures necessary for these compounds, ethyl chloride boiling at approximately 12° C. and ethylidene chloride at 58° C. The point of his testimony was that in the processes in question, that is, the Britton and Nutting processes, the instability of ethylidene chloride is such that a reversed reaction occurs, i. e., when this compound is produced it decomposes and in connection with the solvent reforms vinyl chloride and hydrogen chloride; whereas the ethyl chloride as used in the Britton process remains stable when once formed. Dr. Hass did not take part in the tests of the Nutting variant carried out by Mr. Amos to determine whether or not it was operative, but he testified in substance and effect that he was of the opinion that the process disclosed by claim 6 of Nutting would not work satisfactorily even before he was shown the results of the various tests made of it. His opinion in this respect was based upon his knowledge of the polymerizing characteristics of vinyl chloride and upon his knowledge of the tendency of ethylidene chloride to a reversed reaction. ' We shall discuss separately the several contentions of the appellant. 1. The first contention is that, although in the field of mechanics a claimed invention may be defeated because of the analogies of the prior art, in the field of chemistry analogical reasoning cannot be relied upon to defeat a claimed invention for the reason that chemistry is an experimental science in which prediction is limited, if not impossible. Chemistry is so empirical, asserts the appellant, that each new result is an independent discovery, and it is not permissible to reason from the prior art by analogy to the conclusion that a particular applicant’s result might have been predicted. Hence the appellant asserts that the basis of rejection of the Britton application by the Commissioner and the trial court is unsound. But in contending that analogical reasoning cannot be relied upon at all in determining whether or not an applicant — 584 — has achieved invention in chemistry, the appellant has overstated the law. A similar contention was made in Naamlooze Venootschafs v. Coe, 76 U.S.App.D.C. -, 132 F.2d 573, decided this day. In rejecting that contention, we there said: . . The appellant urges that analogical reasoning cannot be relied upon in the field of chemistry but only experiment. It is true that analogical reasoning is more restricted in chemistry than in the field of mechanics. Toledo Rex Spray Co. v. California Spray Chemical Co., 6 Cir., 1920, 268 F. 201; Naylor v. Alsop Process Co., 8 Cir., 1909, 168 F. 911. This is because chemistry is essentially an experimental science in which predictions cannot be made with the same certainty that they can be in mechanics. General Electric Co. v. Laco-Philips Co., 2 Cir., 1916, 233 F. 96; cf. Corona Co. v. Dovan Corp., 1928, 276 U.S. 358, 368, 48 S.Ct. 380, 72 L.Ed. 610; Tyler v. Boston, 1868, 7 Wall. 327, 330, 19 L.Ed. 93. It is therefore true to say that since Ecldand did not name hypochlorite solution as a bleaching agent but only ozone, it does not necessarily follow that hypochlorite solution would be a suitable substitute. Experimentation may well have been necessary to determine this. But from the facts that prevision is not certain in chemistry, that progress in the chemical art is reached largely through experiment, and that patents are often upheld where the inventor stumbles upon a discovery (Badische Anilin & Soda Fabrik v. Kalle, C.C.S.D.N.Y1899, 94 F. 163, it does not follow that every new and useful result accomplished by experiment is patentable. Hamilton Laboratories v. Massengill, 6 Cir., 1940, 111 F.2d 584, certiorari denied, 1940, 311 U.S. 688, 61 S.Ct. 65, 85 L.Ed. 444; L. Sonneborn Sons, Inc., v. Coe, 1939, 70 App.D.C. 97, 104 F.2d 230; Minnesota Mining & Mfg. Co. v. Coe, 1938, 69 App.D.C. 217, 99 F.2d 986; Ruben Condenser Co. v. Aerovox Corp., 2 Cir., 1935, 77 F.2d 266, certiorari denied 296 U.S. 623, 56 S.Ct. 145, 80 L.Ed. 443; cf. Zotos Corp. v. Rader, 2 Cir., 1937, 91 F.2d 935; Ruben Condenser Co. v. Copeland Refrigeration Corp., 2 Cir., 1936, 85 F.2d 537, 541; De Lore v. St. Louis Lithopone Co., 8 Cir., 1928, 26 F.2d 864, 868; Western Willite Co. v. Trinidad Asphalt Mfg. Co., 8 Cir., 1926, 16 F.2d 446, 450, certiorari denied, 1927, 274 U.S. 737, 47 S.Ct. 575, 71 L.Ed. 1317; Rodman Chemical Co. v. Steel Treating Equipment Co., 6 Cir., 1923, 288 F. 471, 474 .. . 2. Upon the basis of the testimony of Dr. Hass and Mr. Amos, the appellant contends that the Nutting variant described in claim 6 is not continuous and that, therefore, the contention of the Commissioner and the finding of the trial court that Nutting discloses a continuous process of making ethylidene chloride by the use of vinyl .chloride and hydrogen chloride is not supportable. With this contention we cannot agree. It is true that the testimony shows that the Nutting variant is not commercially operable as a continuous process because in about four hours it gums up with polymerization products, and shows that it is much less efficient than Britton in that it produces much less of its end product, ethylidene, out of a given quantity of vinyl chloride than the Brit-ton process produces of ethyl chloride out of a given quantity of ethylene. But the difference between a batch process and a continuous process is that in the latter the reactants are supplied and the end product is removed at the same time, whereas in the former a given charge of reactants is put into a reactor and the reaction allowed to proceed to completion, at which point only, the reactants having been exhausted, the end product is taken out. In other words, in a batch process the end product is not removed during the reaction but only at the end of it, whereas in a continuous process the end product is removed throughout the reaction — of whatever duration the same may be. With this distinction in mind we think it obvious that the Nutting variant was continuous, since throughout the entire period of its approximately four-hour operation, as described by the witness Amos, there was a continual feeding into the reactor of vinyl chloride and a continual drawing off of the end product ethylidene chloride. Specifically, in describing the process Mr. Amos testified: “The passage of the hydrochloric acid and vinyl chloride mixture up through the chlorinated solvent bath and aluminum chloride mixture resulted in the reaction, which consisted in the production of some ethylidene chloride, which boiled off at a temperature above the boiling point of ethylidene chloride (that is 58° C. thereabouts) and came off at the top of the reactor at substantially atmospheric pressure and entered a water-cooled condenser — 585 — from which flowed the condensed ethylene chloride into a receiver . . . Under the testimony the Nutting variant was not a commercially efficient continuous process. But it was certainly a continuous process in kind, as distinguished from a batch process. 3. The appellant urges that the argument -of the Commissioner, and the view of the trial court as expressed in the findings — -that in view of Curme’s use of ethylene in a batch process the Nutting continuous variant would lead to Britton —is not supportable for the reason that although superficial similarities between ethylene and vinyl chloride might impress a chemist not well informed in respect of these two compounds, a skilled chemist would understand that the properties of these reactants differ in such manner that Nutting would lead away from, rather than toward, Britton, i. e., lead away from the substitution of ethylene for vinyl chloride as a starting reactant. This contention of the appellant’s is based upon the testimony of Dr. Hass. But we think that in this aspect of the case Dr. Hass’ testimony proves too much. It is true that it shows that a skilled chemist would know that vinyl chloride is extremely reactive, that it tends to polymerize readily, and that in so doing it produces huge molecules of tarry or resinous character and would therefore not lend itself well to a continuous process; and the testimony shows that a skilled chemist would know also that ethylidene, the end product of Nutting, boils at 58° C. and at such temperature readily enters into a reversed reaction. But Dr. Hass’ testimony also proves that the same skilled chemist would know that ethylene, Britton’s starting reactant, is sluggish in its double bond reaction, does not polymerize readily, and produces, to the extent that it does polymerize, a fluid like lubricating oil which can easily be drained out of the reaction vessel in small quantities as produced; and Dr. Hass’ testimony shows that ethyl chloride, Britton’s end product, boils at approximately 12° C. and lends itself much less readily than ethylidene to a reversed reaction. Thus the knowledge of a skilled chemist would lead one seeking to achieve a continuous process of making ethyl chloride to use the reactants of Curme with the process of Nutting. The appellant contends -that in order to reach Britton’s continuous process from Nutting and Curme experiment would have been necessary. That may be conceded. We conclude as follows: In view of the unsoundness of the appellant’s contentions that the patent law permits no analogical reasoning whatever in the field of chemistry, and that Nutting is not continuous, and that Nutting necessarily leads away from rather than toward Britton, the appellant’s case for invention is reduced to the proposition that in the field of chemistry a new and useful process reached by experiment is necessarily inventive. Britton is technically new in that while Nutting is, as above explained, in a proper sense continuous, he did not use ethylene as a starting reactant; and while Curme did, ' his process was batch rather than continuous. Britton’s process is also concededly useful. But in the field of chemistry as elsewhere the patent law requires more than conception, reduction to practice, novelty and utility for invention. Something beyond the skill of the art is required — either to lead toward the experiments which result in the improvement or in carrying them out — if invention is to be said to exist. The appellant has failed to demonstrate that it required more than the skill of the art either to see that experiment would be necessary to reach Britton from Nutting in view of Curme or to-carry out the experiments necessary to achieve Britton. Affirmed. Footnotes [1] More specifically, the application states that the temperature range of the hath is from about' 12° C. to about 40° O. and that the boiling point of the various solvents which may be used varies from 119° O. to 146° C. [2] The quoted statement is confusing because in the second sentence the witness referred to the vinyl chloride used in the reaction and the vinyl chloride transferred to ethylidene chloride as identical. What the witness meant appears from later testimony to be the following: By “percent conversion” (amounting in Bun 7 to 33.8%) is meant the percentage of the total vinyl chloride passed into the reactor which went into the end product ethylidene chloride. By “percent yield” (amounting in Bud 7 to 53.5%) is meant the percentage of the total vinyl chloride passed into the reactor which reacted in some manner, i. e., either to produce ethylidene chloride or tarry residue, and which thus was not recovered as vinyl chloride from the products taken out of the reactor. The meaning of the witness’ testimony is that 33.8% of the vinyl chloride passed into the reactor went into the end product ethylidene chloride; that approximately 29.5% of the vinyl chloride passed into the reactor went into tarry residue; and that approximately 46.5% was recovered as vinyl chloride at the end of the reaction.