Appeal from the District Court of the United States for the District of Connecticut.
Before L. HAND, SWAN, and CHASE, Circuit Judges.
The very full and careful opinion of the District Judge makes it unnecessary for us to state the nature of the inventions, the scope of the patents and their claims, or the prior art. We refer to it and proceed at once to our own disposition. As to Sundback's patent the question of the validity of all the claims in suit except 18, depends upon the new feature of holding together the wings of the "slider" by a rivet to which the pull is attached. We should have to be well convinced that so slight a change would be invention, even if it were new; but it was not. Sundback's earlier patent, No. 1,302,606, had a "yoke member," 10, which held the wings against spreading. True, it was not a rivet, if that the important, which we doubt; but if it be, we cannot see how the claims can escape Aronson, 1,060,412, where an indubitable rivet held together the wings. It is true that the pull was not attached to it, any more than it was to Exhibit F, which apparently followed Aronson; but we cannot take very seriously such a detail, especially in the absence of any proof that it was important to the art. So far as we can see, the pull fastened to the rivet became desirable only after the cam lock came in, to which it was an incident. The art had gone away from a fixed, and towards a sliding, pull, which had obviously advantages. It is the merest question of convenience where to fix the pull; it required no inventor to select the rivet. Claim 18 is a little different; it includes the "tabs" on each side of the neck, bent over so as to form the inside of the divergent channels. We agree with the judge's analysis of Rocke's patent, No. 1,511,777; Rocke's "tabs" were not "integral" with the neck which held the wings together. The member, 17, in Sundback's patent, No. 1,302,606, was confessedly not "integral" with them, though it functioned as though it were. In Exhibit U, made in accordance with this Sundback patent, the change has been made, but it is not clear that this was produced before the application for the patent in suit. Again, however, it was a very trivial change in design to incorporate the "tabs" into the yoke. We think the claims in suit invalid.
While the Whitney patent is rather close to the line of invention we are disposed to give it the benefit of the doubt. It is quite true that the "cam lock" was very old, and had been applied to all sorts of other things. Moreover, there was nothing troublesome in adapting it to the "sliders" of this art, once one thought of using it. Nevertheless the evidence shows that its choice was not at once apparent. Aronson in his French patent, 442,132 of 1912, provided a lock for a "zipper" fastener; he chose something wholly different from the cam lock, though that had been well known long before. Binns in 1922, No. 1,506,286, tried again, and with a wholly different result. Prentice himself says that he reached the infringement independently of the patent in suit only after repeated failures; one of them, Exhibit J, is very far from the cam lock.
In all such cases where the mere idea of combining old elements is the whole of the invention, we have often said that we look rather to the history of the art than to our own divinations of what was easy and apparent. That it was desirable and desired to lock these fasteners appears from Aronson alone. It was not indeed much of a "problem," but it had been tried, and it was not satisfactorily answered for eleven years. Perhaps during that time such locks were not in much demand; we cannot say; they were certainly in some demand. Perhaps the suceess of the cam fastener is due only to its reversible action, which is attributable to the defendant's own lock; again we cannot say; it seems most unlikely. But it does appear that from Whitney dates the most feasible means of doing what people had been trying to do for some time. That seems to us enough to answer any doubts we might otherwise have.
The defendant insists that there can be no invention in using for a new purpose a well known device; it particularly relies upon Stearns & Co. v. Russell, 85 F. 218 (C.C.A. 6). The dispute is one of the oldest in the patent law and the decisions so numerous, on one side and the other, that discussion of them would not be very fruitful. We conceive the rule to be that if the invention be merely of a new use for an old machine, it is never patentable; the statute does not authorize patents for uses, though processes come close aboard at times. But if the patent be for a new machine, there is no such doctrine, and indeed could not be, because substantially every machine is sure to be composed of old elements. The real difficulty is, as it usually is, in fixing the marches where these conflicting doctrines meet. When may one speak of a combination of elements as a "machine"; when is such an effort abortive, in spite of an inducement such as appears in many claims; "in a combination," etc.; or "in a machine for doing so and so"? Certainly such verbiage will not always avoid the objection that the patent is merely a new use; one could not get a claim for unbreakable glass in a Pullman sleeping car by beginning, "in a communal vehicle for the dormitory accommodation of nocturnal viators," etc. We conceive that the question in the end depends upon whether the putative unity is factitious; whether the elements can fairly be said as a whole to constitute a "machine, manufacture, or composition of matter." Rev. St. § 4886 (35 USCA § 31). There can be no doubt in this case that, so tested, the lock is only one part of a true machine, which people would think of, and use, as such. Such a canon alone can in this instance avoid the barren dialectic which pervades so much of the patent law.
We think that the judge was right in allowing the plaintiff to withdraw its bill as to the Statham patent, without disposing of it on the merits.
Decree affirmed, except as to the Sundback patent, as to which the bill is dismissed for lack of invention. No costs.