LEARNED HAND, District Judge.
I shall in the main consider only the “Silvertown cord” practice of the Diamond Company up to April 29, 1911. Though originally brought from England, yet, if “known or used by others in this country” (R. S. § 4886 [Comp. St. § 9430]), it will serve to anticipate McLeod’s invention, and he is not “the original and first inventor” (R. S- §' 4920, subd. 4 [Comp. St. § 9466]). The facts are these:
In November, 1909, the Diamond Rubber Company, which was absorbed in 1912 into the defendant, having heard of the success of the English “Palmer Cord Tyre,” concluded a contract with its owner, by which, at a cost of £60,000 and a royalty, they acquired the right to manufacture that tire in the United States. As allowed in the contract, they sent over an expert, Tew, to learn the process and to bring back plans and drawings with which to commence the manufacture at their own plant at Akron, Ohio. Tew stayed in England from January, 1910, until September of that year, bringing back with him, not only a complete set of drawings, but molds made in accordance with them, which were at once set up, and with which work was begun. The drawings for the molds and cores were put in evidence. Another Diamond emplo}ree, Converse, was also sent over while Tew was there, and he as well was taught the art as it was then practiced. Both these men were called and testified; the drawings were produced with apt dates upon them, and legends showing that they were made for the United States, as declared. Edgar, a former employee of the English company, was sent over to assist in April, 1910, and stayed for two years. The plaintiff called him, and he bore out the testimony of Converse and Tew in all respects, except as to the degree of stretch, a matter to be later considered. Newall, a present employee of the English company, was also examined, and swore to seeing Tew and Converse in England, to the making' of the drawings and molds, and to certain other records.
[902] The defendant also proved by very detailed “construction sheets” that at the very end of 1910 and in the early months of 1911 they had made and tested to destruction six tires, designed and built after the Silvertown specifications, and by “production sheets” that before April 29, 1911, they had sold upon the market over 400 tires identified orally as of the same make. The character of these tires did not, however, appear from the “production sheets”; the construction sheets for the classes to which the tires belonged, though once in existence, not being produced. These were said to have been lost in several movings of the defendant’s officers. It was, however, stipulated that the-Diamond Company began paying royalties under its agreement with the English company in January, and continued through February, March, and April, showing that they supposed themselves by the first of 1911 to be employing the process.
All this puts it beyond reasonable doubt, unless answered, that the Diamond Rubber Company began early in 1911 manufacturing under the Silvertown process, and that they tried so far as possible to follow that process as closely as they could. Indeed, no other supposition is in the least probable. They had just invested a very large sum of money in the process, certainly because they believed in it. They had sent over men to learn it; they had received Edgar to help them start; they meant to get the worth of their money. Whatever changes they later made, it is scarcely possible to suppose that at the start they should not have imitated the English process step by step. It follows that,, if that process was in fact an anticipation of the patented process it was “known or used” in this country before April 29, 1911. Before considering the process, however, I shall consider the plaintiff’s evidence to show that they did not practice it, or that it was not what the defendant’s witnesses say that it was.
First, as to the “construction sheets” of the six tires made in January and February, 1911, it is argued from their internal evidence that they were merely experiments, and should not be taken as anticipation. If they stood alone, I should agree; for the fact is so. While they are detailed in the last degree, and show the exact manufacture of the tires, it is very properly urged that the tires were run to destruction upon testing machines, and were obviously intended only as experiments. The defendant’s answer is that the chemist of the Diamond Rubber Company was not satisfied with the composition of the rubber used in the English process, and insisted upon a test of that element. The purpose of the tests appears to have been more general in character than Tew admitted, as will appear by examining the comments upon them; but the last one, of February 6th, was eminently satisfactory, and Tew commented at the bottom that it was “far superior to a canvas tire.” The “production sheets” speak from January, and considering the fact that the royalties were paid from that time, and that price lists were issued, there can be no doubt that they refer to tires' which the Diamond Company was trying to make after the Silvertown process. That they knew how to follow that process the test sheets show, and at the time they had no other molds and draw[903] ings than those brought over from England. It must be owned that the loss of the construction sheets for the commercial types was unfortunate, but not enough, I believe, to raise any fair doubt. It is incredible that the Diamond Company should have gone to the expense and trouble which it did, and have found the experiments in the United States as satisfactory as they found them, without making some effort to use the designs and molds which they got.
Finally, there is the proof of Denman, who says that early in 1913, while a draftsman for the defendant — the consolidation of the defendant and the Diamond Rubber Company having occurred in 1912-— he made drawings under Converse’s directions which provided for a greater stretch than that already employed. The implication, as I understand it, is that the defendant was changing its practice in the light of McLeod’s invention, of which they had already got wind. I think it patent that the drawings produced by the defendant, signed by Denman, must be those to which he referred as providing for an increased stretch. They answer his recollection exactly in showing that the metal of the mold was taken away at the tread, and a little on each side, and they provide for about the stretches he recalls. Moreover, there are no others which in any way answer, unless a set made in 1913 has been lost or suppressed. The difficulty is only that Denman supposed that his work increased the stretch, while in fact it is shown that it merely maintained the same stretch as had existed before. In the first place, he is in error by about six months, a most natural mistake, as to the date of these drawings. All but one were made in June and July, 1912, the earliest only a few days after McLeod’s patent issued, and all but that one refer back to molds made at the end of 1911 or very early in 1912. Theoretically I suppose it may be imagined that in November and December, 1911, the Diamond Company might have heard of McLeod’s invention, the application being filed on November 13, 1911; but such a suspicion seems to me rather fantastic. I shall assume, therefore, that the originals were made in innocence of what McLeod had done. In each of Denman’s drawings occurs the legend, “The mold increases the thickness of the tread over that shown” in the earlier drawing at the three places indicated. Unless this be a later addition, and therefore pro tanto a forgery, it accounts for the change and bears out Converse’s testimony beyond any question. Denman was either mistaken at the time, or now erroneously recalls that the stretch lie provided for was an increase.
So far as I know this is the only evidence which conflicts or may be thought to conflict with the defendant’s position that they were practicing the Silvertown process from the early part of 1911, and, if so, the anticipation turns upon what that practice was. Before considering that, however, it will be necessary to state what the patented process is. McLeod was speaking of a fabric tire, and of the weakness which it betrayed, owing to the fact that, if vulcanized without pressure, the fabric at the sides, especially near the beads, would be under less stretch than over the rim. This resulted from the manner in which the fabric [904] was laid over the core, which left a slack at the sides, that must be taken up by “stitching” or “milling,” or, as he better describes it, by “condensing.” His method was to allow room in the mold for expansion, and so to stretch the fabric to an even tension throughout, and this he was to accomplish by the direct injection of water through holes in the core.
At the time when McLeod filed his first application, and also when he substituted his second, his thought was to expand the tire to about service conditions; at least, that was the minimum that was necessary. To do so he set the cords in the vulcanized rubber at a uniform tension, and when inflated the pressures would not tend to be taken up by the loose or “pondensed” sides of the tire, but would be evenly distributed. He encountered difficulties with his original claims in the Patent Office, because of references which it is not necessary here to describe, and over which all his claims were allowed only by the interjection of what is at best an ambiguous phrase referring to the stretch of the threads; i. e., “to approximately the limit of their elasticity.” Apparently thislwas the result of some oral discussion between McLeod’s solicitor and the -Examiner, and the phrase is not elucidated in the specification at all. The plaintiff asks me to read it as no more than explanatory of the earlier language that the stretch should be at least equal to sendee conditions, and I must own that, if the specifications are to be taken as one piece, I should strongly incline to take that view.
There is, - however, some doubt whether the phrase can be treated in this way, owing to the fact that it was introduced in order to avoid the examiner’s rejection. Under the well-settled canon it must be held to have differentiated the claims from what they covered before, and the final answer depends upon whether the phrase “under tension,” which they all had originally, must have been understood as meaning “under a tension equal to service conditions.” If it must have been so understood, then the added phrase, “to approximately the limit of their elasticity,” must itself in turn be understood to mean something more. If, on the other hand, it meant “under some tension,” then the added phrase may possibly refer only to the stretch prescribed in the specifications, though the language is c'ertainly extremely inept. Were it necessary to a decision, I should perhaps think that the phrase was added merely for certainty, and accept the plaintiff’s theory. Indeed, it appears to me open to very grave doubt whether the disclosure is otherwise sufficient at all. The phrase is, strictly speaking, altogether inapplicable to textiles, and the specification gives no intimation whatever of how one is to stretch and set the cords at the limit of their elasticity, unless the phrase be referred to service conditions. ’ In any event, for the purposes of this case, where it is unnecessary to decide the point, I shall assume that it means no more than that the cords shall be stretched to the limit of their elasticity under service conditions.
The main features of the Silvertown practice, both here and in England, are not in dispute. The tires were built in the following way: A core of usual construction was covered with a thin layer of rubber, [905] and over this was laid diagonally across the top and around a set of pins at the bead a series of strands of a continuous cord. At the end of the first operation the cords had covered the whole surface of the core. The difference in circumference between the tread and the bead, accommodated usually by “condensing” the fabric, was, under the Silvertown practice, arranged for by having the cords elliptical in cross-section. They passed the pins with the minor axis parallel to the plane of the core; but they passed the tread with the major axis nearly parallel, somewhat tilted. After the first cords were laid, another layer of rubber was added, and upon that a second layer of cords crossing the first at an angle of about 76 degrees. The usual “breaker strip” and tread were then cemented on, and the raw tire taken off the core. It was next cured by the usual process of vulcanization between a mold and a water bag, under such pressure as to drive the tread with great force against the inner face of the mold. The crux of the dispute in the case at bar rests upon the difference in size between the mold, and the core, through which the tire was stretched in cross-section during the process, or, more particularly, how much the cords themselves were stretched. As these ran diagonally over the core, the percentage of their stretch was quite different (55 per cent.) from the stretch measured in cross-section — a consideration not relevant, however, if the claims be construed as I have construed them.
The stretches of the tires were measured in cross-section, and we have contemporaneous records to prove what they were. Thus the production cards of.1,000 tires made in England show stretches of from about 6Y2 per cent, to over 11 per cent. Next we have the stretch of the 6 experimental tires run to destruction iu the. Diamond shops, already referred to. These run from 4% per cent, in the case of a 3y¿-inch tire to over 9 per cent, in the case of a 5-inch. Finally, the blueprints of sections of tires made in England and sent over here to be used at Akron show various stretches from about 4% Per cent, to a little over 9 per cent. The drawings of molds and cores which accompanied these do not show similar figures, but I think I should assume that they were made in accordance with the tire sections. The defendant’s witnesses so swear, and the plaintiff has made no effort to contradict it. .Indeed, there is no ground whatever to contradict it, except the testimony of Edgar that the stretch was from 3% per cent, to 4 per cent. This he gives as the practice in England, which was followed in Akron, and the reasons for any stretch at all he says were convenience in handling and the existence of excrescences on the tread. The witness only professed to give these figures as nearly as he could remember, and it is not altogether surprising that he should have forgotten the amount; but I must confess it is strange, if the stretch was deliberately used to put the cords under tension during vulcanization, that he should have recalled it as merely a provision against too close a fit. Still, the documentary evidence is so strong, and the oral testimony on the other side so unanimous, that I do not hesitate to accept it. Moreover, it must be remembered that some of the stretches were little more than 4 per cent., [906] and it is possible that these alone remained in Edgar’s mind, and that the reason he never knew, since his employment was primarily to adjust the machines.
I conclude, therefore, that in the Silvertown practice there were varying stretches between 4*4 and 9 per cent. In its brief the plaintiff does not so much question this as it disputes the inferences to be drawn from it, relying upon the testimony of Mr. Bemis to show that the cross-sectional stretch did not put the cords under any tension at all, but merely straightened them out. The proof is unfortunately somewhat detailed, but its discussion cannot be avoided. Mr. Bemis’ arguments proceeded upon his deduction from the data given" of the width of the cord, the distance between the pins, and the necessity, as he assumed, that the cord leave the pin without bend. In all these respects his argument was fallacious. He proceeded upon the supposition that the United States “construction sheet” of the 5-inch experimental tire was correct, and that the longer axis of the flattened cord was .09". He then by experiment flattened a round piece of wire till its longer axis was .09" and found its shorter axis was .073". This he assumed to be the width of the cords as they passed the pins; their longer axis being concededly normal to the plane of the wheel at that point. On these data he by trial ascertained that they could leave the pins at most at an angle of 29 degrees to the radius, and then he developed a curve which the cords must take up on the side of the core, and which made them cross the tread at an angle of 49 degrees, instead of 38 degrees, which is the correct angle, if the cord is drawn straight.
At the outset it may be observed that his data were quite wrong, assuming that his underlying assumptions were correct. The actual thickness of the cord on its minor axis was .085", not .073", and two cords wholly filled the space between the pins which was .17" (the pins themselves are .08" in diameter and set at .25" between centers). His argument applied to the corrected data would therefore show the cords leaving the pins radially, and the curve necessarily much greater than he indicated, how great does not appear. If sound at all, his argument would therefore prove too much. But it is sheer assumption in any event, because the cords can and do bend very soon after rising their width above the pin, being pulled straight under their laying load of 3 or 5 pounds. Moreover, I cannot find the least justification for the curve which he actually, plotted. On his own assumption the cord was bent, and obviously it could be. What data he used for the bend he actually allowed I have been unable to find in the record. So far as appears, he took quite arbitrarily what appeared to him to be a reasonable bend, something near to a parabola in vertical projection. Perhaps I am wrong; but so far as I can see this was wholly fanciful, and in the absence of some evidence as to the resistance of the cords to distortion from a straight line must necessarily have been so. Therefore, in the absence of any evidence to the contrary, I do not see that his conclusions were reliable. However, there is such evidence in a photograph of the cords as actually laid in England before Tew left (Exhibit Q *), and in Figure 13 of the “Palmer Cord Tyre” circu[907] lar. Both these show the cords as laid by the machine, and they run directly from above the pin over the tread; they take a marked bend just above the pin, running necessarily radially for a short distance to clear.
As I have said, the stretch along the cord is less than 60 per cent, and more than SO per cent, (say 55 per cent.) of the sectional stretch; under Silvertown practice the 9 per cent, stretch was about 5 per cent, along the cord, and the 4y% per cent, stretch about 21/2 per cent. Therefore it appears to me that the defendant has proved beyond reasonable doubt that it was stretching the cords under uniform tension to that amount before McLeod made his invention. The plaintiff does.not assert, and indeed has "no way of asserting, that the defendant’s infringing process stretches the cords more than these percentages, and the case comes down, therefore, to the question whether the Silvertown process, so understood, was an anticipation. As to that, I may at the outset observe that it need not anticipate the process, if that be applied to a woven fabric, because the case involves only what the defendant chooses still to call a “cord” tire. In fact, these tires are made by a series of layers of so-called “Palmer fabric,” which was old in the art before McLeod, and which consists of cords laid side by side and buried in crude rubber. In some cases there is a cross-weave at relatively very wide intervals, but I think it is not denied that this is not a strain-resisting member.
Taking for the moment the product patent, the single claim requires “a fabric body and a coating of vulcanized rubber, the said fabric body having its threads uniformly tensioned in the casing structure to approximately the limit of their elasticity, and substantially held at such degree of tension by the vulcanized rubber." If the present “Palmer fabric” is a fabric within the meaning of the patent, as perhaps it is, then the old Silvertown manufacture was also a fabric. Each was made up of parallel cords buried in rubber. If the present tires are stretched to the limit of their elasticity, so were the Silvertown tires, because the only change has been in reduction of the stretch.
The difficulty arises rather over the process patent. The first four claims are substantially the same; they call for “the making” of “a casing structure of fabric and rubber.” At present the defendant makes this fabric before it is put upon the core; under the old Silver-town practice it made up the same thing by laying the cords between layers of raw rubber. As to claims 1, 2, 3, 4, 8, 9, and 10, the question of whether the Silvertown process anticipated may, of course, be tested by whether it would infringe, and that in turn depends upon whether under that process “the casing” was made up of “fabric and raw rubber.” Now a “fabric” is usually conceived as a piece of woven stuff, and claims 8 and 9 are so especially limited as to make it extremely questionable whether under any aspect the “Palmer fabric” could infringe them. Passing that, however, and assuming that the patent included such, was not the Silvertown casing, also made of a fabric, if the defendant’s present practice uses a fabric? It is true that the “fabric” was then made, even “woven,” if one chooses, on the [908] core; but the claims say nothing about how it shall be made. When the laying machine had done its work, the cords lay side by side across the core, just as they do now, and while the rubber was pressed between them at a later stage, instead of before they were laid on, there would be no defense to infringement in that, unless nothing is fabric which will not hold together when off the core.
I think, however, that the Silvertown process could hardly be said to infringe claims 5, 6, 7, and 11. In them the step of “condensing” the sides of the “fabric” is an element of the combination, and it would be a very doubtful interpretation of that word to include within it that part of the Silvertown process by which the cord was turned through 90 degrees, so that its minor axis was parallel'to the surface of the core at the bead, and its major at the tread. The phrase referred to a well-known step in the art, which did not appear in the old process at all, and which the defendant now uses. If the process would not infringe those claims, are they an invention over it? If so, it is because of the substitution of the old method of “condensing” the fabric, instead of laying it on the core over tire pins. In short, given a knowledge of the method of stretching the cords to service tension or more and the method of making the tire with layers of fabric, was it invention to combine the two into one process? As I have already shown, McLeod conceded by his amendments in the Patent Office that some tension in the cords was old, and he got his claims only by defining the tension specifically; i. e., service tension, as he claims, and as I am assuming for argument’s sake. If so, stretch and condensed fabric simpliciter are not invention; only condensed fabric and service tension stretch. And if service tension stretch, together with a fabric laid by a machine, were old, would it not be strong doctrine to say that service tension stretch and condensed fabric would be invention? Or, stated otherwise, if it was old to stretch condensed fabrics somewhat, and old to stretch uncondensed fabrics to service tension, both of which are established, was it not an obvious permutation of the old processes to stretch the condensed fabric to the same extent the old had formerly been stretched ? That is all that the defendants did at the most, and in so doing they did not apparently resort to McLeod’s invention. That invention they had heard of early in 1914, and the change to the “condensed” fabric they did not make till 1917. Of course, they may have needed those three years to see its value; but that seems unlikely.' If they had not known of the stretch, and if Denman’s drawings really indicated that they had changed their molds for that purpose as soon as they heard of the invention, a very different situation would have been disclosed. Little inference is,, however, permissible from the change made in 1917 to a style of fabric and a method used long before.
It may be urged that, granting that the Silvertown process stretched the cords, it does not follow that it stretched them to service pressures, and that this is what the patent covers. The answer is this: The defendant concedes a present stretch of 7% per cent, at times, and I understand that the plaintiff stands on Bemis’ testimony of a stretch of 8.7 per cent. Whether or not this is such a stretch as still allows [909] no expansion under service pressures, and the best proof is that it does not, at least whatever the defendant does now was done before April 29, 1911, and that is a complete answer. Nowhere in the patent does it appear how one is to get the stretch during vulcanization which will keep the tire from any expansion under service pressure, and we must proceed pro tanto in the dark. But 1 cannot see how the conclusion can be avoided that, if the defendant does follow the patent, then the Silvertown stretches anticipated it. Hence that prior use seems to me a sufficient answer.
The prior patents and the Ajax use I do not find it necessary to consider.
Finally, I come to the Michelin process, as described by Mitton and supported by drawings made in 1907. Although it is true this evidence is not large in volume, its character seems to me such as to make the conclusion irresistible that the process involved as great a stretch sec-tionally of the old woven fabric tire as the defendant uses. These drawings are beautifully and accurately made, as in the case of all skilled French mechanical work, and they show that there was intended a stretch measured by the difference between the interior face of the inflated tire (développement intéríeur de pneu gonflé) and the outside of the core (développement de la forme en bois), which in the case of 25-F was about 6% per cent. There can be no question in my mind that the phrase, “allongement des toiles,” is contrasted with “allongement cireonférentiel.” In Figure 43-F, two percentages are given— one, the stretch without anything laid over the core to make it larger; the other, with layers of 1.5 millimeters in thickness. The relative stretches are 8.866 per cent, and 6.6 minus per cent. In the case of 42-P, these percentages are 13 plus per cent, and 8 minus per cent.; in that of 41-F., 16 plus per cent, and 11% per cent. In all these cases the molds had a “bull ring” at the bottom, which held the fabric around the bead quite as McLeod’s disclosure. It is, of course, possible that the fabric pulled out around the bead, just as it did in the International use or the Ajax; but nothing prevents the same result in the patented process. That is a defect to be met by proper shop practice, and nowhere in the patent is one taught how to avoid it.
The critical point is that these drawings seem so unconditionally to corroborate Mitton’s testimony of the existence of a stretch as to leave nothing patentable in the invention that I can perceive. Even assuming, with the plaintiff, that the original cores were built up to avoid the original stretch, they were not built so far as to avoid a stretch greater than the defendant now uses. The Michelin company meant to get the stretch, and it of course makes no difference what were'their purposes in getting it. ‘ One cannot patent a new use, or a use never before observed. Therefore I think that, in spite of the paucity of the proof, its probative value is sufficient. These were fabric tires, so that the question does not arise, as in the case of the Silvertown process, of any change from cord proper to Palmer fabric.
Bill dismissed for lack of invention, with costs.