Circuit Court of Appeals, Second Circuit.
*306 *307 Mayer, Warfield & Watson, of New York City (F. P. Warfield and Lawrence Bristol, of New York City, and W. T. Cashman, of Cleveland, Ohio, and C. P. Townsend, of New York City, of counsel), for appellant.
Dean S. Edmonds, Frank E. Barrows and Raymond F. Adams, all of New York City, for appellee.
Before MANTON, L. HAND, and SWAN, Circuit Judges.
L. HAND, Circuit Judge (after stating the facts as above).
We agree with the plaintiff that the patent in suit describes or rather attempts to describe the same invention as the German patent to Bayer & Company of December 31, 1913, and it cannot be properly understood without recourse to that document. This declares first that bases, other than those described in the earlier patents, are accelerators, regardless of their constitution, provided they have the necessary constant. (The first draft of this patent had indeed stated that while all bases were accelerators, only those above the limit had any "noticeable technical effect.") It then added that ammonia compounds, organic or inorganic, though not properly bases, had the same action if they were alkaline at vulcanization temperatures. All the other foreign patents were in the same form as the German.
It is not quite clear just what this language means. It may have meant that ammonia compounds which were not true bases, but, so to say, "quasi-bases," would be equally good accelerators as true "bases" if they had the requisite constant, or that all ammonia compounds if "quasi-bases" would be good regardless of their constants. Two of the examples given, aldehyde-ammonia and sodium amid, have either no definitely known constant or one which was too low; the naphthylenediamins and quaternary ammonium bases are in greater part below the limit. Unless the inventors mistook the constants of these substances, it would seem that the ammonia compounds intended need not have the specified constant. We incline to that view; that is, to read the original invention as being twofold; any bases of the required constant, any ammonia compounds which are bases or quasi-bases. In view of our doubts we shall, however, treat the invention on both assumptions.
On the assumption which we prefer, the patent in suit was an inadequate disclosure. It asserted, first that the accelerators in question were to be "bases," and also to have an alkaline reaction at vulcanization temperatures, i. e., "quasi-bases." Unless some true bases lose their alkalinity at those temperatures, the addition was surplusage. If they do, the disclosure was not complete because the original made no such limitation upon "bases." Added to these requisites was the possession of the prescribed constant. If so far the patent correctly described the first part of the invention, it failed to describe the second. It is true that it declared that among those bases which it covered were ammonia compounds, and it enumerated a number of these. But it nowhere even intimated that all such compounds were within the invention regardless of their constants, if only they were bases or quasi-bases. While most of those which it enumerated had either no constant or a constant below the limit, this did not disclose the invention, if it included all ammonia bases or quasi-bases. No one could gather this generalization from the instances. Hence we conclude that if the original invention was as we incline to understand it, there was not that "full, clear, concise and exact" disclosure which the law demands, Béné v. Jeantet, 129 U.S. 683, 9 S. Ct. 428, 32 L. Ed. 803.
If on the other hand the German invention is to be understood as merely adding to any bases of the required constants those ammonia compounds which have that constant and are quasi-bases, the patent is equally imperfect. The German patent begins by prescribing "bases" as we have already said; the addition was not of bases but of quasi-bases, and these do not appear in the patent in suit. It is quite true that among the enumerated substances are sodium amid and borate of ammonia and that the third example is of sodium amid. It is also true that after the enumeration the specifications declare that the enumerated compounds are either basic or have the alkaline reaction at vulcanization temperatures. This, however, leaves it uncertain whether we are to ignore the first element, i. e., that the substances must be bases. While we can understand how the confusion arose by a reference to the original, the patent, standing alone, is contradictory and speaks with an uncertain voice.
However, this is not the chief difficulty. When an inventor gives examples or instances of a generalization, he necessarily means that they contain all the determinants of *308 which the generalization is composed. If he is wrong, either it is because the generalization itself is wrong, or because he has not properly understood the examples. In either event the disclosure contradicts itself and furnishes no guidance to the art. The examples are certainly part of the evidence on which the generalization is founded. Not only do they not support, but they confute it. If it is right, it can only be by a chance; it cannot be founded upon genuine discovery and therefore will not support a patent, Matheson v. Campbell, 78 F. 910 (C. C. A. 2).
Therefore, we think that whichever construction of the original invention is taken, the specifications are valueless. It may be that, construed as we prefer, the original invention might stand, though we must confess that it has even then rather the appearance of a shot in the dark than a discovery based upon scientific inference. Certainly it is true that most commercial accelerators have been of less than the necessary constant, or have had an indeterminate constant. It can scarcely be an answer to say that all above have been proved to be accelerators, for the first draft of the original invention declares that all bases of every kind are accelerators. Possibly, this disparity in results can be accommodated if we include among commercial accelerators below the limit, ammonia compounds which are bases or quasi-bases. If so, it might be too much to say that no invention had ever been made. No more can we say that if the claims are not expanded, the specifications cannot be amended so as to state the invention whatever it was. But we may and do hold that, as it stands, the patent is invalid, and that until it is reissued, it does not support the claims.
This would indeed be enough to dispose of the case, were there no other difficulties in the path, but there are. We need not hold because some of the members of a class are in the prior art, that the class itself cannot be patented. The discovery of the class means that its determinants have been found, and the results attendant upon all its members so defined. It may not necessarily be a fatal objection to the novelty of that discovery that among them some have been already known, so long as they were not known as members of the class, that is, as possessing the class determinants. Ansonia Brass Co. v. Electric Supply Co., 144 U.S. 11, 12 S. Ct. 601, 36 L. Ed. 327, is not to the contrary; it merely holds that when an article has been in public use, no patent can be had upon it because of the discovery of one of its unknown qualities. It is of course always true that no members of the class already known may be monopolized, merely because they were later discovered to be within the class. The claim must exclude them, else it trenches on the public demesne; but if it does, there would seem to be left a genuine discovery which could be protected. At least we may assume arguendo that this is true.
In the case at bar we think that the claims should be read as excluding the patents and applications mentioned in the specifications. This includes the "methylene base" patents, among which the plaintiff insists that Spence's para-amino-dimethylaniline must be included. The argument is based upon the definition of "methylene bases" contained in the applications referred to in the patent in suit which later resulted in patents, 1,126,469 and 1,130,903. In the later of these patents that term is defined as including substances with open or closed carbon chain bases which are "saturated," and not "unsaturated such as anilin and pyridin." Strictly some, if not all, of the substances described in those patents are not methylene bases at all, for they contain only methyl radicals; but that point we pass, assuming with the plaintiff that the inventors might define their terms as they chose. However, if the words "saturated bases" refer to the carbon ring, the carbon rings of all those substances are saturated except one, benzaldipiperal, which like aniline is not. Hence it is argued that the word cannot refer to the carbon ring but must refer to some other part of the molecule. The experts are in disagreement as to what can be the meaning of the phrase, and we confess to being unable to decide between them. It seems to us unnecessary. It is clear that if the claims in suit are to depend for their validity upon the exclusion of the substances mentioned in these patents, these must be clear enough, when thus incorporated by reference, to advise the art of what is so excepted. That is to say, the definition given to "methylene bases" must itself be clear, and we think that the word "saturated" does not with certainty include para-amino-dimethylanilin, merely because, among many substances mentioned, one and one alone is "unsaturated," if that term were understood in its more ordinary sense. In short, if the plaintiff's argument is good at all, it does not sufficiently appear what was meant by "saturated bases."
Therefore we cannot agree that Spence's para-amino-dimethylanilin was disclaimed in the original specifications. The proof is ample *309 that it was used for vulcanization in large quantities before December 31, 1913, in the factory of the Diamond Rubber Company and that tires were made of the product and sold. This was the only kind of use possible and it was public. In Egbert v. Lippmann, 104 U.S. 333, 26 L. Ed. 755, and Hall v. McNeale, 107 U.S. 90, 2 S. Ct. 73, 27 L. Ed. 367, it was held that a use is public though the invention cannot be learned unless the embodiment of it is destroyed. If the accelerator could have been discovered by dissolving the product into its constituents, these cases would directly apply. However, this was not the case before 1914 and there was no way of learning from the product either its composition or the process by which it had been made. Such was the case also in Macbeth Evans Glass Co. v. General Electric Co. (D. C. Ohio) 231 F. 183, and it seems to us that the distinction cannot serve to prevent anticipation. It is quite true that the art is not enriched by such a use, but then it is not enriched even if the invention can be learned by destroying the embodiment. The fact that a curious person might disentangle it does not mean that any one will do so, and the discovery remains inaccessible for practical purposes. It might have been proper theoretically to exclude such uses, but we think that the statute has cut the knot. Once the invention has been embodied in goods which are put in public use it becomes impossible for a later inventor to secure a patent.
We are not sure indeed that the process was in any case secret in the sense which would prevent its being an anticipation. A number of employees knew it and, while it is true that it was kept in code and not generally disclosed, they felt no obligation of secrecy except while employed in the factory and were not at least expressly enjoined from using it after they left. This they certainly did in the case of thiocarbanilide, another accelerator kept equally secret, and so far as appears the Diamond Company and its successor made no complaint. We prefer, however, not to rest our decision on this feature of the evidence.
In any event, Spence's patent for devulcanization of June 27, 1914, was an adequate disclosure. In this he declared that the true solution of the problem of devulcanization lay in the use of a powerful vulcanizing accelerator, and that the more powerful the accelerator, the more marked its devulcanizing effect. His experiments had led him to employ the most powerful accelerators, though it was impossible to state all that he had prepared. Among the examples given of the process was one in which para-amino-dimethylanilin was used, and we think that this disclosed that substance as a powerful accelerator of vulcanization. Boggs discovered the use of para-phenylenediamin before December 31, 1913, and was a prior inventor even though that date be taken for the patent in suit, which it may not. In any case his naphthylamine patent applied for in the spring of 1914 disclosed para-phenylenediamin though it did not claim it, and he was a prior inventor, Milburn v. Davis-Bournonville Co., 270 U.S. 390, 46 S. Ct. 324, 70 L. Ed. 651. De Meeus also disclosed, and he claimed, para-phenylenediamin in an application filed before July 31, 1914.
The question arises as to these three applications whether the plaintiff can carry back the date of the invention to December 31, 1913. We have already discussed the relation between the German and the American patents and have shown that the patent in suit is an unsuccessful effort to disclose the first, from which it follows that the two are sufficiently identified to comply with the statute; the analogy of a reissue is apt. Obviously unless some differences are presupposed, identity would be impossible if there is any change in the applications.
There remains, however, the question whether the German application was filed by the same person who filed the application in suit. The formal applicant of the German patent was the Bayer Company and it may be argued that for this reason the statute is inapplicable, the applicants not being the same individual. While we cannot find that the point has ever been decided, it seems to us that the assignee of an invention ought to be entitled to take advantage of the earlier foreign filing date. To be sure the situation at bar is not quite that, but the plaintiff's contention is that the patentees, Hofmann and Gottlob, had assigned, or in some other way transferred, the German rights to the Bayer Company, which filed the German application by virtue of the transfer. This we think if proved would have been enough.
It was competently proved that they were chemists in the employ of the Bayer Company in December, 1913; Hofmann being still living and accessible though Gottlob was dead. It was also proved that there are corrections upon the two drafts of the German application in their hands and also in the hand of one Fertig, a lawyer of the company. Again that Hofmann and Gottlob had signed a statement upon the English and French applications that they had read them and found them correct. All this evidence, except that *310 they were chemists, was introduced over the defendant's objection that it was "secondary."
The proof of Hofmann's and Gottlob's hands was not competent evidence that they had read and approved the English and French applications, whatever that would have been worth as a basis of inference that they were the inventors. It was a declaration not under oath of that fact and was hearsay; it was properly objected to and should have been excluded. At most, therefore, the only competent proof was that they had corrected the German drafts, which we agree the proof of their hands showed. But the same is also true of Fertig, whom no one claims to have been an inventor. The evidence did not prove that they were inventors either, and in so far as it is sought to use it as a declaration to that effect, it would be incompetent. Thus the proof was far from carrying the severe burden imposed upon one who would carry back the date of an invention. We think that the plaintiff is confined to July 31, 1914, for the date of its invention.
It results that as respects para-aminodimethylanilin and para-phenylenediamin the claims cover what was in the public demesne. As to the naphthylamine we agree that it is an obvious misprint for naphthylenediamin, the context showing that the specifications are speaking of a diamin. There is no pretense that para-phenylenediamin is not within the claims and the plaintiff has not attempted to disclaim it. It has, however, disclaimed para-amino-dimethylanilin and the question arises of the validity of that disclaimer. This is a substance included in the general description of the invention and nowhere distinguished in the specifications. The question is whether a patentee may thus correct his generalization by successive retrenchment as he finds himself bound to retreat. If the part which the patentee disclaims need not be distinguishable on the face of the specification, there would never be need for a reissue on the second ground given in the statute, i. e., "by reason of the patentee claiming as his invention or discovery more than he had a right to claim as new." U. S. Code, tit. 35, § 64 (35 USCA § 64; Comp. St. § 9461). This language must include something which cannot be reached by disclaimer, else it was redundant. We adhere to our ruling in Strause Gas Iron Co. v. Crane Co. (C. C. A.) 235 F. 126, that the distinction is between disclaiming a part separated in the patent itself as opposed to something comprehended in its general language. Hailes v. Albany Stove Co., 123 U.S. 582, 8 S. Ct. 262, 31 L. Ed. 284; Albany Steam Trap Co. v. Worthington, 79 F. 966 (C. C. A. 2); Simplex Ry. Appliance Co. v. Pressed Steel Car Co., 189 F. 70 (C. C. A. 2). So far as the holding of the majority in Thompson v. N. T. Bushnell Co., 96 F. 238, 241, 242 (C. C. A. 2), is to the contrary, in view of our later rulings and of the other decisions cited, we cannot adhere to it.
The distinction is not merely formal. If the specifications themselves separate the part disclaimed from that retained, that part is exscinded without affecting the residue which must be of itself enough to support the claims. True, the patentee has been wrong in asserting that the whole disclosure was new, but he has been right as to a part which does not depend upon what he disclaims. But when the specifications do not so separate the parts, the disclaimer can succeed only by interpolating a limitation into the part which must remain, if the claims are to survive. Nothing is taken out of the specifications, though something is taken out from their meaning. What is changed was originally false and has now by the added limitation been made true. This however is to reframe the language, not to drop out a part. It is not the office of a disclaimer to do this, but of a reissue.
We therefore hold that claims one and four of the patent are invalid, first, because the specifications are insufficient, and, second, because the claims are too broad. We should not, nor indeed can we, say whether if reissued it would be valid; we cannot even know what it would be when it emerged. It is, however, not improper to observe that of the accelerators which have had substantial commercial use, only a small part are within the claims as they stand. The patentees' contribution may have been greater than that; their claims may have been unduly limited, but the dissociation constant taken alone has not apparently played a leading role in the development of this art.
Decree affirmed.
