ALSCHULER, Circuit Judge (after stating the facts as above).

Appellants' contention of want of invention over the prior art is based on British patents to Phipps, 1844, No. 10234; to Tomlinson, 1867, No. 628; to Newton, 1871, No. 1825; to Maguire, 1885, No. 11755; to Marks, 1897, No. 18740; and to Annand, 1898, No. 9721; United States patents to Crowell, 1887, No. 365051; and to Moody, 1898, No. 598107. It is apparent, from examination of these patents and the detailed evidence of the patent experts concerning them, that no one of them embodies all of the elements which inventor Stone combined in his device. To fully analyze each of these patents, and compare it with that in issue, would enlarge this discussion far beyond beneficial scope.

Concededly, mounting a plurality of rolls of paper or other material upon a rotatable reel to bring the rolls successively into position for feeding the paper into a printing press, was not new with Stone. Neither was there novelty in a belt so mounted that it would movably rest upon the periphery of a roll of paper or other material which was being fed into a press. Some of the patents cited showed one or the other of such, but none showed both; and surely none showed them so assembled with reference to each other that they co-operated to make what is known as the "flying paster," which is Stone's achieved result over the prior art, by means of the several elements which he first brought together.

Tomlinson, which is perhaps the most persuasive citation, shows the plurality of rolls upon a reel, but without the belt. The fact that Tomlinson's machine was designed for printing calico might not of itself prevent its successful reference as prior art. Instead of making a "flying paster," it points out the operation of storing some of the material in a box, from which it would be taken for attachment to the material of the next roll.

It is pointed out that, with some modification, such as reversing one of the rollers, a flying paster might be successfully made by the Tomlinson device. We do not need to speculate whether such modification, enabling, if it did, the successful making of a flying paster, would in itself involve invention over Tomlinson's patent. Suffice to say, it is not the device of the Stone patent and, in our judgment, would not, and evidently did not, suggest it.

And in general we might distinguish each of the cited patents, all of which, except Annand, were cited and considered in the Patent Office during the extraordinarily protracted pendency there of Stone's application. It is of course true, as appellants contend, that Stone's patent does not cover a flying paster. It is a machine patent, but it is a patent upon an instrumentality whereby the flying paster is effected, and it secured to him the monopoly for the specific combination which brought it about. A structure combining wheels, springs, screws, and what not, however ingenious and interesting, unless it effects some useful purpose, is not properly the subject of a patent. As stated in Electric Signal Co. v. Hall Signal Co., 114 U.S. 87, 5 S. Ct. 1069, 29 L. Ed. 96: "The thing patented is the particular means devised by the inventor by which that result is attained, leaving it open to any other inventor to accomplish the same result by other means."

It is not material that in the claims the term "flying paster" is not employed. One can scarcely read the patent and file wrapper without reaching the conclusion that the improved result at which Stone was aiming, and which his combination was designed to bring about, was this flying paster, whereby without the unreeling and storage of the web, as in Tomlinson, and without the making of a loose loop of the web, as in Marks, or otherwise, the paper from the expiring roll might be joined to that of a new roll upon the same reel, and the printing operation continuously carried on without stoppage of the press.

It is manifest that in newspaper printing, where every saving of a moment's time is a "consummation devoutly to be wished," its achievement, either by means entirely new, or by a new and advantageous combination of old elements, strongly points toward the quality of invention. Prior to 1919 apparently few of the Stone installations had been made. But about the time of the installation in question, and afterwards, it was rapidly adopted by large newspaper printing offices, and by 1925 installations of the machines had been made by appellee in the aggregate price of $1,300,000, and, so far as appears, no installation, save that in question, was made in defiance of Stone's patent.

From our careful examination and comparison of the prior art patents, as well as the action of the Patent Office in granting the patent after the long pendency and its citation of nearly all the alleged prior art, also the very substantial public recognition of the patent, we would not be warranted in disturbing the conclusion of the District Court that the claims are not invalid upon the prior art.

The question on the contended commercial use by the Chicago Daily News more than two years before application date is not free from doubt. Stone had long been the mechanical superintendent of the Chicago Daily News office. New presses were installed there early in 1897, and about that time Stone supplied for each of two of the new presses a multi-reel feeding device with four arms, each carrying a roll of paper. An article of March 16, 1897, published in the Chicago Record (a daily newspaper of same ownership as the Daily News, and using the same equipment), gives a glowing account of the new installation of presses, headlining and stressing the use for the first time of electrical appliances for controlling the machinery. It refers to Irving Stone (the inventor) as the mechanical superintendent in charge of all the presses, and states that he had devised other novelties, one of which is a magazine reel, which can be loaded with four rolls of paper to feed the press, the reels standing far below in the basement. It describes the operation of the reel, but makes no mention of the belt feature, nor of the flying paster, or anything of that nature; the general practice then being to stop the press long enough to change to another roll. It emphasized the value of the reel device, in that it enabled the expired roll to be removed from the reel and replaced by another roll while the press was moving, thus effecting a saving in time. But the failure in that article to refer to the further time saving which the flying paster effects is quite persuasive that the means to that end were not then and there present — if, indeed, then conceived.

There was in evidence also a publication of the Electrical Engineer, of February 24, 1898, and of the Western Electrician, of April 16, 1898, and of Electricity, of February 22, 1899. These publications are, in the main, devoted to a description of the electrical equipment for governing these presses. None of them refers to the belt arrangement for mechanically making the flying paster. It appears from the record that at some time in this period the operatives became sufficiently expert, so they could cut out the expiring roll and manually paste the severed web to the new roll, thus continuing the operation of the press by very materially slowing it down, but not stopping it, and thereby effecting some further saving of time. This was, in a sense, a flying paster manually made, and it is reasonable to conclude that the second article, which refers to the printing of 60,000 papers an hour without stopping the press, was descriptive of this method. The same may be said of the last publication, which was about eight months within the two-year period; that is to say, that at the time the article was prepared (which the evidence does not show) the mechanical pasting, as effected by the belt, was not yet in use.

The other evidence upon this prior use was that of three witnesses, one of them a brother of the inventor and long employed in this pressroom, and of two others employed there for a time. While they say, with varying degrees of certitude, that the belt was part of the installation some time in 1897, their evidence in some respects is quite contradictory in details. While it is often true that contradictions in details between witnesses may tend to indicate their good faith and lack of their conscious co-operation, yet in proving a date or time, upon the establishment of which the issue depends, contradictions upon that very subject do not assist in fixing it.

It is quite elementary that, to strike down a patent upon the ground that the patented article was in public use more than two years before the application was filed, this fact must be established by evidence so clear as to satisfy beyond reasonable doubt; and the evidence must so satisfy, not alone as to the time, but also that the use was of the very thing which was patented. James Heddon's Sons, Inc., v. South Bend Bait Co. (C.C.A.) 14 F.2d 805.

Where the evidence is that of persons testifying from memory upon incidents remote in time, without corroboration by contemporaneous records, or writings, or definitely proved circumstances, the proof quite generally fails to establish with sufficient certainty the alleged prior use. Deering v. Winona Harvester Works, 155 U.S. 286, 15 S. Ct. 118, 39 L. Ed. 153; T.H. Symington Co. v. Natl. Malleable, etc., Co., 250 U.S. 383, 39 S. Ct. 542, 63 L. Ed. 1045; Eibel Process Co. v. Minnesota Ontario Paper Co., 261 U.S. 45, 43 S. Ct. 322, 67 L. Ed. 523.

The publications tending as they do rather to negative than to establish the alleged prior use, the proof here rests wholly on these witnesses. Testifying, as they did, solely from their recollection of things occurring about 25 years before, without corroboration by contemporaneous writings or records or articles, or other definite fortifying data or things, we would not be justified in substituting our opinion upon the weight of the evidence for that of the master, who had the added advantage of seeing and hearing the witnesses, and whose finding against the fact of prior use was sustained by the District Court.

Furthermore, appellee maintains that, even if it were concluded from the evidence that the Daily News original installation of 1897 embodied substantially the elements of the Stone patent, that installation was under the supervision of the inventor himself, and was experimental only, until within the two year period preceding application.

For appellants it is contended that the burden is on the inventor of proving that a prior installation was experimental, and that, under the evidence, appellee has not sustained this burden.

But, if it can be made to appear that such installation and use was for experimental purposes only — to perfect the device, or to demonstrate its utility — the two-year period does not begin to run until the experimentation ceases, and the device is perfected or its utility so demonstrated. Elizabeth v. American Nicholson Pavement Co., 97 U.S. 126, 24 L. Ed. 1000; Smith Griggs Mfg. Co. v. Sprague, 123 U.S. 249, 8 S. Ct. 122, 31 L. Ed. 141.

There is abundance of evidence in the record tending to show experiment and change in the device for a considerable time following the Daily News installation. For just how long, it is difficult to say. While it is upon the inventor to prove, if he so claims, that a prior use, if shown, was experimental, we cannot say that the master who heard the contradictory testimony thereon was wrong in his conclusion of fact that the here alleged prior use was experimental.

But we do not reach this contingency, if, as we have above indicated, we do not find that appellants have sustained their initial burden of establishing belond reasonable doubt that the Daily News installation of 1897 was in fact the embodiment of the patent which was subsequently applied for.

The defense of invalidity urged in the District Court, based on delay in prosecution of the application for the patent, is not seriously urged here. That the prosecution was most extraordinarily procrastinated is manifest from inspection of the file wrapper, which shows pendency of 14½ years, without as much as an interference proceeding to consume the time. Applicant made 12 amendments, in each instance delaying within a week of a full year following the next preceding action of the Patent Office. But, this being within the law as it then was, the validity of the patent is not thereby affected. Overland Motor Co. v. Packard Motor Co., 274 U.S. 417, 47 S. Ct. 672, 71 L. Ed. 1131.

Appellants contend that, if validity be conceded, the Tribune machines do not infringe, because, while to all appearance they may embody the elements of the patent, their contacting belts do not in fact move quite so fast as the roll revolves, and so the belts have the effect of retarding rather than of feeding the web to the press, and do not constitute a feeding device at all. Appellants' expert, who, ex parte during the hearing, made measurements, stated that the Tribune belts all moved slightly slower than the web on the roll, and their entire machine was not a feeding device, since it was the pull of the press alone that drew the paper from off the rolls and into the press.

It is true that in the Patent Office, in distinguishing from Tomlinson, the patentee stated that in Tomlinson the printing press itself pulled the calico from the rolls, while in the patent the contacting belt caused the roll of paper to revolve. We do not regard this distinction as controlling.

"Feeding device," or "feed," should not be regarded so narrowly as to indicate only the means or process whereby the web is brought to and delivered into the press, without influence of the press itself. These devices are no part of the press. They are conveniently located — in the various installations referred to generally in the basement, where the rolls are unloaded, below the level of the press floor. It would be impossible for the device to push up the paper from the roll and literally deliver it into the press. The "apparatus for feeding paper to printing presses" is, in each instance, a "feeding" device, in the sense that it holds the web or paper which is fed into the press, quite regardless of whether the roll is revolved by the drawing of the web into the press, or by means apart from the press, or by the co-operation of both.

If the revolution of the roll were controlled wholly by the contacting belt, and the latter moved the roll continuously and materially faster than was required in order to feed the press, there would soon be an accumulation of loose web somewhere between the roll and the press, which would seriously interfere with the intake of paper by the press. But, if this independently revolving belt would move the roll materially less rapidly than required by the press, there would, without compensation by way of slippage or otherwise, be an increase of tension upon the web, which would cause breakage, and render the device useless.

Compensation between the belt and the roll, because of the dimishing periphery of the latter through drawing off the web, by means of the changing position of the belt, which is pivotally suspended from its upper and driving pulley, was testified to. But it is evident that the experienced pressman, in the operation of the mechanism, sees to it that the relative movement is such as to secure and maintain the requisite degree of tension of the web to insure proper feeding into the press.

Whatever, in practice, may be the degree which the drawing of the press and the movement of the belt may contribute to the delivery of the paper to the press and maintenance of the proper tension of the web, the combination of the multiple reel, means for revolving it, and the contacting belt, as an instrumentality for bringing about, mechanically and without stopping the press, the flying paster or union between the web of the expiring roll and that of its immediate successor, is the essence of the patent, and is present in appellants' device.

Upon this record we would not be justified in disturbing the decree of the District Court, which is accordingly affirmed.