LYON v. BOH, (S.D.N.Y. 1924)

LEARNED HAND, District Judge.

As the case seems to me to turn on the question of infringement, I need not consider the validity of the patent. It will first be best to read the patent as it was granted, next to see what limitations, if any, were placed upon it in the proceedings in the Patent Office, and, finally, how far it may be extended under the doctrine of equivalents.

The specifications state that the invention "relates particularly to a motor vehicle buffer comprising opposed resilient parts or springs, each of which may comprise a transverse member and a longitudinal member, the transverse members overlapping and being secured together by detachable clips." Page 1, lines 14-20. At the outset I note the use of the word "may," which recurs throughout the specifications substantially to the entire exclusion of the indicative mood. Now a patent must show how the invention is to be practiced, and is valid only to the extent that it does. It is idle for a patentee later to insist that he occupied more room in the art than he disclosed, because his specifications were cast in the optional. Such equivocation will not avail to amplify what was described only as a possibility.

The only figures show two separate pieces of flat steel which overlap in front — i.e., "opposed resilient parts" — and are held together by clips. These are bent into loops to form the ends of the buffer, proper, and from the loops turn in a curve through 90 degrees, ending where they are fastened to the frame of the motor car. The specifications speak of the different sections of these pieces as separate members, being in that respect somewhat confusing. That part which engages the obstacle, and takes up the shock, is called the transverse member; that part which is at the end is the loop; that part which is set at 90 degrees to the transverse member is called the longitudinal or attaching member.

The description proceeds then to say: "The transverse members, 1, of the opposite parts of the buffer may overlap one another, * * * and they are preferably detachably, but rigidly held together by suitable connecting means." Page 1, lines 80-84. Again: "This detachable connecting means thus facilitates the relative transverse adjustment of the parts of the buffer, so as to make it readily applicable to vehicles having side frames, * * * which are located at different distances apart." Page 1, lines 91-96. Again: "The tightening of the clip as by the bolt 16 clamps the transverse members 1 in both a vertical and horizontal direction." Page 1, lines 105, 106; page 2, lines 1, 2. Again. "The overlapping transverse members which constitute the impact receiving portions in the front of the vehicle may be brought into alignment with each other and rigidly connected * * * and correspondingly stiffen and reinforce this part of the buffer." Page 2, lines 44-55. Again: The "laterally extending loop always extends outward a substantially fixed distance beyond the vehicle frame which is advantageous in many cases, and also, since the two elements of the buffer are symmetrical, the ends of its overlapping transverse members always project the same distance on opposite sides of the center of the buffer." Page 2, lines 58-66. Again: "The front portion of the buffer, because of the reinforcement secured by the overlapping of the connected transverse members, is very strong." Page 2, lines 85-88. Finally: "By making up this buffer of two tempered steel strips, its manufacture is greatly facilitated because of the smaller size of these parts and the greater ease of bending them, and particularly because these smaller units can be much more effectively and regularly hardened and tempered, with much less danger of warping a distortion, than if the buffer was made of one piece of spring steel. Page 2, lines 127-130; page 3, lines 1-6.

I have quoted so much of the specifications because it appears to me beyond any fair dispute that every one reading the description would understand that from top to bottom it was meant to cover only a buffer made of two pieces of steel, bent to form the right and left members of a jointed buffer, and that the adjustments were possible only because the buffer was split. What Lyon apparently supposed to be his invention was to secure adjustment while retaining a constant distance between the end of the buffer and the outside of the wheel, by using longitudinal bends and a split buffer, and also to strengthen his front by overlapping the ends of the transverse members.

I now turn to the claims. The case is the common one of needless elaboration of distinctions which are verbal and details which are trivial. It is absurd to apply in all such cases the rule that claims must at any cost be treated as patentably differentiated. Courts have descanted upon the abuse again and again, but the antlike persistency of solicitors has overcome, and I suppose will continue to overcome, the patience of examiners, and there is apparently always but one outcome.

The first six claims are for two separate steel spring strips, and in that regard follow the express disclosure. Claims 7 and 8 are for overlapping members, or a pair of transverse members. Claims 10, 11, 12, and 13 are for spring members, each with a transverse impact, and a longitudinal attaching section. Claim 15 requires the attaching means to be integral with the loops, and to extend rearwardly. Claims 16 and 17 call for two springs which overlap. If, therefore, the claims in suit, 9, 14, and 18, are for a single steel spring reinforced in front by a separate member, and without any longitudinal attaching section, they are egregious in this galaxy. I shall take them up in detail to try to see what they fairly mean.

Claim 9 is for a buffer comprising "transversely extending impact members and open ended lateral loops, connected attaching members, * * * connecting means connecting said impact receiving members, * * * and means providing lateral adjustment of said attaching members." The plaintiffs argue that in Grotenhuis Pancoast's patent, the alleged infringement, the means providing for lateral adjustment of the attaching members are the brackets, 5, which hold the rear bar to the frame. But the claim says that the attaching members must be themselves laterally adjustable. It is not an adjustment of these members to slip the brackets longitudinally along the bar, which is the only adjustment possible in the infringement. The phrase "attaching member" refers, not to the clamps, but to the longitudinal section of the continuous strips. Page 1, lines 26, 27, 75, 76; page 2, lines 16, 20, 31, 32, 36, 37, 74, 78, 79, 83. They can only be laterally adjusted if they are two. The straight rear bar is never laterally adjusted to different widths of frame. Thus, even if read literally, the claim is not infringed. Read on the specifications, as it should be, it is even less so.

Claim 14 is for a "spring having a transversely extending member and a rearwardly extending attaching member." So far, the use of the singular shows that the draughtsman had in mind only one-half of the buffer. Then it proceeds: "Said transversely extending member being arranged adjacent another transversely extending spring member," to be secured to it. If the second transverse member was not the same as the first, if it had no attaching member as one section, the whole buffer would hang in the air on one side. The first transverse member has but one perpendicular or longitudinal leg; the second must have one as well, just as disclosed.

Claim 18 does not mention a divided spring, and is particularly vague, but at its end it literally contradicts the possibility of single spring as much as claim 9, because it specifies that the "attaching members" must be "relatively adjustable" to fit frames or different width. "Relative" adjustment is even clearer than the language of claim 9. Rearwardly extending legs cannot be relatively adjustable, if the spring is continuous.

I have been over the language of these claims literally, because I think that literalism is the best chance of the plaintiffs. The conclusion is that both specifications and claims read upon the diagrams, and upon nothing else. Lyon supposed that he had an invention made up of two connected springs. They were to be adjustable to different frames, because they were split. Some advantages resulted from this: First, that they were more reliable in manufacture; second, that the distance between the outer side of each wheel and the end of the guard was constant, no matter what the width of the frame. If he had any idea that a single reinforced spring buffer was an invention or a form of his invention, he gave no indication of it from one end of his patent to another.

Before considering what latitude he is entitled to under the doctrine of equivalents, it is necessary to consider how far the proceedings in the Patent Office estop him from any such expansion. A patent is peculiar among solemn written unilateral contracts in this: That there are cases when although the promise of the promisor — i.e., the claims — do not, read them how we will, cover another machine, yet if they do so in function and means, the court will disregard the language in which the promise is couched, and apply to the case a kind of cy pres doctrine in aid of justice. Such a doctrine — i.e., the doctrine of equivalents — is unknown elsewhere, and is altogether anomalous, though it is, of course, in the primer of the law of patents.

Along with it, and equally anomalous, judged by other legal analogies, is the doctrine of estoppel by abandonment. Ordinarily the final writing, which incorporates a solemn agreement, is taken by the courts as the parties want it taken; that is, it is the sole resort for ascertaining their intentions, for the excellent reason that the parties meant it to be such. All prior negotiations are disregarded, since otherwise the chief purpose of reducing the contract to writing would be frustrated. However, there is the exception, well settled and very recently confirmed by the Supreme Court (Weber Electric Co. v. Freeman Electric Co., 256 U.S. 668, 41 Sup. Ct. 600, 65 L. Ed. 1162), that, if a patentee submits to the rejection of a claim while his application is in the Patent Office, he may not later insist that other claims which he does get are to be regarded as equivalent to that which was rejected.

All this is, of course, trite enough; but it seems to be supposed that the doctrine established for long in this circuit (A.G. Spalding Bros. v. Wanamaker [C.C.A. 2] 256 Fed. Rep. 530, 167 C.C.A. 602), that the courts will not look generally to the file wrapper, contradicts the theory of estoppel. It does not, and the two have always been understood to stand side by side. The defendants have, it is true, in their argument ignored the distinction, and ask me to look at arguments made by the patentee and replies by the Examiner, all of which are absolutely a closed book to me. Therefore, before deciding how far I may depart under the doctrine of equivalents from the plain writing of the contract, I have the right to, and indeed I must, see whether the patentee submitted to the rejection of claims which covered a one-piece spring. Not only did he do so, but he did so again and again. By that I mean that, with a degree of obduracy which is exceptional even in the Patent Office, the claim to such a buffer was put forward after rejection, disguised in one way or another, three or more times, either deliberately to hoodwink the Examiner, or because the solicitor had become enmeshed, like a silkworm, in his own emanations.

At the very outset the Examiner declared an interference with Hoover's application, and according to the practice of the Patent Office drew up a claim for "a continuous spring" supported by Figure 8 as it then was in the drawings, and by lines 8-13 of page 5 of the specifications. Claims 2, 3, and 4, as originally drawn, were upon this form of the invention. This was in effect Hoover's disclosure, Figures 2 and 4, except for the absence of the reinforcing strip. Lyon did not accept the claim, and it became Hoover's claim 1. Lyon lost it for all purposes as part of his invention, whichever of the two was in fact prior in time. His claims cannot cover such a structure; as patentee it is as to him a part of the prior art, though not, of course, in any suit in which he does not sue on the patent.

Thereupon Lyon erased claim 2 and Figure 8, along with the corresponding text. Claims 3 and 4 were soon rejected, and the application stood without any claims for Hoover's disclosure. On November 10, 1913, the applicant tried again, with five new claims, to get a single spring buffer, and these were all rejected, because of his withdrawal of the Hoover figure, and his failure to contest the interference. He did the same thing again on January 29, 1914 (claims 8-10), on July 23, 1914 (claims 7 and 8), and on June 23, 1915 (claims 8 and 9), and each time he was unsuccessful. When the whole application was renewed, he was bound by his former estoppels.

In speaking of Hoover's Figures 2 and 4, I shall always mean to exclude the reinforcing strip.

It is clear at a glance that Lyon's invention, as it was allowed, is a species of the genus so rejected; he himself calls it a "broader embodiment" of it. The form is the same — a continuous spring bent into loops and turned at right angles to the front till it reaches the frame to which the ends are fastened. What then did he do? How did he improve it? By making it adjustable, and by reinforcing the front; nothing more. These things he accomplished by the simple and admirable expedient of splitting the spring in two and allowing enough extra length to each half for an overlap. It remains the same open buffer as before, but it was more adaptable, easier to make, and stronger in front. In function it was a single strip buffer, with attaching legs at right angles; it was an improvement on Hoover, but it was of the same type. I do not, of course, suggest that Hoover was its origin, but only that, for the purposes of this suit, by his surrender in the Patent Office, Lyon put himself in the position of an improver upon Hoover. If he meant to claim larger rights, then was the time to assert them; I take him now at his own estimate then.

Thus the patent cannot claim to be a pioneer in the art of making steel strip buffers, with or without a rear bar. Fageol did not have a flat strip anywhere, nor did he have a spring in the front. Yet the form was the same as Lyon's, and as Hoover's, and there was means of adjustment in the front bar, as in Lyon, though reached by different means. But Fageol was not part of the prior art. Davis-Bournonville Co. v. Alex. Milburn Co., 1 F.2d 227 (C.C.A. 2). Nothing else in the art appears to me significant, while we are looking at buffers without rear bars, and I do not mean to throw any doubt upon Lyon's patent, if his claims are read naturally. So limited, this record does not change the result reached in the first suit.

Coming, now, to the question of equivalents, I have to decide whether the Grotenhuis Pancoast patent is an infringement. For the reasons I have given, I assume that the language of the specifications and claims do not suggest it, and that in any event the invention does not include a single strip, nonadjustable buffer of Hoover's Figures 2 and 4. The argument is, or must be, that it produces substantially Lyon's result in substantially his way.

The first result is adjustability. Lyon's buffer was adjustable, in the sense that it fitted varying frames, always maintaining a constant distance between the end of the buffer and the outside of the wheels. Grotenhuis Pancoast do not have this; their buffer is constant in width, a single strip of steel bent into a continuous perimeter, which cannot vary. It is not adjustable at all, except in the sense that Welton and Harris are adjustable, and its adaptability is limited by the amount to which it is tolerable to have the end of the fender project beyond the wheel; that is to say, since with the greatest width of frame the wheel must be protected, and since with every decrease in width the buffer will protrude unnecessarily, the buffer is limited in its use by such protrusions as are not obnoxious in practice. In short, the buffer is adaptable only to the extent that its inadaptability does not matter. It seems to me fair to say that this result of Lyon the Grotenhuis Pancoast patent does not substantially realize.

The other result is the reinforcement of the front by the overlap. The question is whether the reinforcing strip, 8, of Grotenhuis Pancoast, produces the same result as the overlap. The plaintiff has offered proof to show that the two buffers show similar results when tested, and that the Grotenhuis Pancoast reinforcing strip "creeps" over the main strip as the overlap creeps in Lyon. Perhaps, in the condition of the proof, I ought to assume that both act similarly when pressure tests are applied to both strips simultaneously. It does not, however, appear that, if only the upper or lower strip of Grotenhuis Pancoast's buffer strikes an obstacle, the result will be the same. Passing that possibility, there is certainly one respect in which they are clearly different in result. The reinforced section in Grotenhuis Pancoast is always of the same length; in Lyon, its length depends upon the width of the frame, which may vary by fourteen inches. The narrower the frame, the longer the section, and vice versa. Dyer did not know how great a difference in resistance the length of the reinforced section would make, but acknowledged that it would make some difference. It is possible that it might make a good deal, and so I understand the witness, in spite of his first answer.

Coming next to the differences in means by which the results are accomplished:

First, as to adjustability. This Lyon accomplished by splitting the spring into two parts, which might be separated as need were. Grotenhuis Pancoast, maintaining always a buffer of uniform width, had a straight bar at the back upon which the sleeves of the brackets, 5, slide, a method which is in substance disclosed in Harris and in Welton. These are quite different ways of reaching different results. Disregarding the difference in result the difference of means is substantially so distinct that it seems idle to elaborate upon it.

Second, as to the means of reinforcing the resistance of a single front strip. The reinforcing strip of Grotenhuis Pancoast may be regarded as a means equivalent to the overlap in view of the proofs, but the resistance of the two buffers is in general quite different in operation. As pressure is put on the Lyon bumper, the loops begin to close, and the ends of the loops to move forward past the original line of the front, thus putting a strain on their inner sides. When they have quite closed, the length of the carry is reduced, though the strain on the inside of the loops continues. So far we may say that the two buffers operate in the same way.

However, at about 1,200 pounds pressure the front bar of Grotenhuis Pancoast touches the rear, and thereafter both bear the strain until the back bar breaks. This occurs when the clamp loses its grip upon one end, which is what happened in the experiments, and what obviously must happen. Therefore, though we assume that the operation of the two is alike in its earlier phase, after the bars have met Grotenhuis Pancoast acts differently. It does not appear how much the front bar is bent at this time, but certainly it must be enough to allow the two to touch only at one point. Thereafter the front bar, regarded alone, may continue to act as Lyon's does; but it is obvious that the back bar during the succeeding phase co-operates with it, and that the resistance is built up by the two together. I have no means of knowing how far the back bar co-operates by its own elasticity and how far the strain upon it is tensile. Clearly when it breaks it is because the tensile strain is too much for the grip of the clamp.

It is apparent, therefore, that the two buffers do not operate in the same way in this later phase, which is the important one, since 1,200 pounds is a low pressure. Lyon depends upon the elasticity of his spring alone, and fastens its ends beyond rupture to the frame. Grotenhuis Pancoast merely float the whole bumper, as it were, against the frame ends, relying upon its internal structure as a whole to resist the shocks, as though it were a cross section of an elliptical cylinder. They do not depend upon the connection with the frame at all, but let the buffer do its work as a self-contained elastic body. Certainly the means so disclosed by the two parties are not substantially the same.

Such being the facts, I think that Grotenhuis Pancoast is not the equivalent of Lyon, and that it would be illegitimate to expand his claims, so as to cover that disclosure. The truth is that the two have an entirely different provenience. One is mechanically an evolution from the general conception, which is seen in the Abresch-Cramer truck buffer, in Fageol, and in Hoover's Figures 2 and 4. It was that of an open, flat strip of metal, the ends of which were firmly fixed to the frame, and which depended upon the resistance of front alone. The other is an evolution from the idea, as I have said, of the inherent resistance of the section of an extremely elliptical cylinder. It probably comes from Figures 1 and 3 of Hoover, and is dependent in part upon the added, and I think tensile, strength of the rear bar. The block was omitted and the front reinforced. It seems to me to be an entire misconception of the history of the art to try to make one include the other. The Patent Office did, indeed, treat all of Hoover's figures as one invention. To me it seems as though there were two; but the point is no more than whether the patent should have been divided. In any event, I think it clear that there is no intimation in word or in function which should give to Lyon's invention so wide a scope. It has its place, and a very important place, as any one may see who walks the streets; but, while he was prosecuting it in the Patent Office, he did not suppose that it was even for an open spring buffer generaliter. He is now trying to make it occupy even more, the place of a spring buffer with continuous perimeter, fixed in width, and operating as a unit.

The bill will be dismissed for noninfringement, with costs.