CINCINNATI CAR CO.
v.
NEW YORK RAPID TRANSIT CORPORATION.

Toulmin & Toulmin, of Dayton, Ohio, and Curtin & Glynn, of New York City (Baker, Hostetler, Sidlo & Patterson and Arthur C. Denison, all of Cleveland, Ohio, and H. A. Toulmin, Jr., of Dayton, Ohio, of counsel), for appellant.

Alfred W. Kiddle, Harry E. Knight, Herbert H. Knight, and William J. Dowd, all of New York City, for appellee.

Before L. HAND, AUGUSTUS N. HAND, and CHASE, Circuit Judges.

L. HAND, Circuit Judge.

The plaintiff sued the defendant for infringement of patent No. 1,501,325, issued on July 15, 1924, to Thomas Elliott, and included two others in the bill. We held the patent just mentioned, and one of the others infringed; the third, not infringed. Cincinnati Car Co. v. New York Rapid Transit Co., 35 F.(2d) 679. Later we suspended the injunction to allow the defendant to substitute another device. 37 F.(2d) 100. It did so, and when the new structure came before us, we held that it escaped the claims of the patent. 52 F.(2d) 44. The cause had gone to a master to compute the plaintiff's damages and the defendant's profits, and he reported that neither had been proved, but that the plaintiff was entitled to a reasonable royalty of fifty dollars for each infringing "articulation," making $12,100 in all. The plaintiff appealed from the decree entered upon this report.

The chief question is whether the plaintiff is entitled to an accounting for profits. Elaborate calculations have been made, based upon the earnings of the road out of the infringing trains, as well as upon the savings from their substitution; but both assume that the profits are to be allocated to the invention alone. The supposititious increase in fares was clearly too speculative. Added fares, due to the larger accommodation for passengers of a train of three articulated cars over two unarticulated, presuppose that passengers, who could not have boarded unarticulated cars, would have left the subway, or not have entered because of the crowds. That would indeed have been a happy consummation, but there is no reason to assume it. That the road took in more fares than it would otherwise have done, must remain a conjecture on which we cannot base an award. More could be said for the initial saving in the cost of the articulated trains, if they had been patented as a whole. While it might indeed be illegitimate to include in the saving more than that upon the bare frame of the cars without fittings, on the other hand, it might be proper to add the decrease in cost of upkeep, had that been properly proved. Some fair measure of equivalency in accommodation was established *593 between the two systems, and we will not say that if the defendant was liable for profits on the trains as a whole, it was not possible to fix the savings.

The situation was not that, however, but one so common in patent accountings, in which the invention is not of the article as a whole, but of a small detail. The difficulty of allocating profits in such cases has plagued the courts from the outset, and will continue to do so, unless some formal and conventional rule is laid down, which is not likely. Properly, the question is in its nature unanswerable. It is of course possible to imagine an invention for a machine, or composition, or process, which is a complete innovation, emerging, full grown, like Athene, from its parent's head. It would then be easy to say that profits were to be attributed wholly to the invention. Such inventions are however mythological. All have a background in the past, and are additions to the existing stock of knowledge which infringing articles embody along with the invention. It is generally impossible to allocate quantitatively the shares of the old and the new, and the party on whom that duty falls, will usually lose. If the patentee is required to assess the contribution of his invention to the profits, he will find it impossible; vice versa, if this is demanded of the infringer. The burden of proof in such cases is the key to the result.

Before Westinghouse Co. v. Wagner Co., 225 U.S. 604, 32 S. Ct. 691, 56 L. Ed. 1222, 41 L. R. A. (N. S.) 653, it was generally assumed that the burden lay with the patentee, though there were exceptions even then. That case has at times been thought to lay down a different rule, treating the infringer in all cases as a trustee ex maleficio, and therefore subject to the severe standard imposed upon malversators. A rigid insistence upon this would cast him for full profits in all cases except those in which by artificial and unreal distinctions courts should come to satisfy themselves that they could dissect the contribution of the prior art from that of the invention. This would be as unsatisfactory a result as that which imposed upon the patentee the same duty, and, while it might be answered that he is a victim, and the infringer a tort-feasor, the character of the tort ought not really to have such sanguinary results. Patent infringement often involves nice and casuistical questions which it is mere artifice to treat as involving moral delinquency.

We do not think that Westinghouse Co. v. Wagner Co., supra, 225 U.S. 604, 32 S. Ct. 691, 56 L. Ed. 1222, 41 L. R. A. (N. S.) 653, involved any such radical shift. The infringing article in that case was completely covered by the claims; the invention so far permeated the whole device that it could be said, if one did not press the analysis too far, that the patent covered the article as a whole. But the infringer had improved it, and he insisted that the patentee should allocate the total profits between the patent and his advances upon it. This the court denied, holding that in such a situation he must unravel the knot. In Dowagiac Manufacturing Co. v. Minnesota Moline Plow Co., 235 U.S. 641, 35 S. Ct. 221, 59 L. Ed. 398, the situation was reversed, the invention being itself for an avowed improvement upon an earlier reaper, and the new parts structurally distinguishable. There the court refused to impose the burden upon the infringer, though he had indubitably created the confusion by his wrong, and might, if he was caput lupinum, be held for the whole consequences. The patentee was required to make the division, and because he could not, was relegated to a reasonable royalty, the only satisfactory solution; perhaps because it abandons the appearance of rationalizing the irrational. There is indeed a passage in the opinion at page 646 of 235 U. S., 35 S. Ct. 221, in which the court gives as a reason for the result that the patentee had not shown that apportionment was impossible, from which it may indeed be inferred that if he had, the result would have been otherwise. However, this has not been forced so far, either by the Sixth Circuit, or by us. Taylor v. Steuernagel, 19 F.(2d) 298, (C. C. A. 2); Egry Register Co. v. Standard Register Co., 23 F.(2d) 438 (C. C. A. 6); Rockwood v. General Fire Extinguisher Co., 37 F.(2d) 62 (C. C. A. 2). We adhere to that interpretation.

Nevertheless, there always remains the inquiry when the invention covers the whole article, and when it is only for an improvement. Once it be agreed that the mere text of the claims does not determine this, the answer must be in terms of degree and cannot be found à priori. As we have already said, in practice all inventions are for improvements; all involve the use of earlier knowledge; all stand upon accumulated stores of the past. Often, perhaps generally, the test will be structural, as for example, whether the addition cannot be isolated as a separate physical part; though that is by no means a final test. Be that as it may, the doctrine cannot be more particularly stated, and was in substance so declared, as we apprehend, in Westinghouse Co. v. Wagner Co., supra, 225 U.S. 604, 32 S. Ct. 691, 56 L. Ed. 1222, 41 *594 L. R. A. (N. S.) 653. Applied to the case at bar, we have no doubt that Dowagiac Manufacturing Co. v. Minnesota Moline Plow Co., supra, 235 U.S. 641, 35 S. Ct. 221, 59 L. Ed. 398, rules. Articulated trains were very old; Grealey's invention so nearly approached Elliott's that we held a minor change in the original trunnions and bolster to fall within it, rather than within Elliott. We generalized his disclosure far enough to include the original form, but the boundary was close at hand, and the changed form crossed it. The situation was one in which the patent could be only for an improvement.

In Van Kannel Revolving Door Co. v. Uhrich, 297 F. 363 (C. C. A. 8), Judge Sanborn applied the doctrine of Westinghouse Co. v. Wagner Co., supra, 225 U.S. 604, 32 S. Ct. 691, 56 L. Ed. 1222, 41 L. R. A. (N. S.) 653, to what was clearly an improvement patent, but the decision went primarily, not upon that, but upon the doctrine of Goulds Manufacturing Co. v. Cowing, 105 U.S. 253, 26 L. Ed. 987. In that case the market was a limited one, open only to drills with the patented improvement; and by using them the infringer had taken the patentee's customers away. The decision seems really to have been an adjudication of the patentee's damages, not of the infringer's profits at all, for the court said (page 256 of 105 U. S.), that the patentee should recover the infringer's profits as damages, apparently assuming that these could in that case be so measured. In Warren v. Keep, 155 U.S. 265, 268, 15 S. Ct. 83, 39 L. Ed. 144, this decision is spoken of as one where the invention was for a new article as a whole, ignoring the ratio decidendi; perhaps it is to be so classed. We cannot find that the Supreme Court has thrown further light upon it, though it was mentioned in Dobson v. Hartford Carpet Company, 114 U.S. 439, 445, 5 S. Ct. 945, 29 L. Ed. 177, rather vaguely, as an instance where the entire value of the machine was properly attributable to the invention, under the exception in Garretson v. Clark, 111 U.S. 120, 121, 4 S. Ct. 291, 28 L. Ed. 371. When it is to be so attributed was left at large; but we doubt whether it is enough that the article could not have been sold or used without the improvement. That is indeed material on the issue of damages, since it goes far to show that the patentee, who alone could have supplied the demand, would have got the sales. But it cannot really determine how much of the infringer's profits are due to the patent. True, the improvement in such a case is a condition, but so is everything else which the infringer uses and owes to the prior art. The question again arises which lies at the bottom of the problem in any form; quantitatively to allocate the joint product of two factors, each a condition to the result.

The burden of proof must fall on one side or the other; in cases of improvement we are not satisfied that it shifts to the infringer, merely by showing that the improvement was necessary to the use or sales. But we need not commit ourselves, because here at any rate the road could have continued to use unarticulated trains. That was not indeed quite the same use; it was a substitute, and a less desirable one. But by hypothesis an improvement always somewhat modifies the prior art, and is an improvement only because its use is different. If we were to press the supposed doctrine to an extreme, every improvement patent would fall within it, for the especial use would not be answered by anything but the improvement. Either it must be confined to the seizure of the plaintiff's market, in which case it is properly relevant only to damages; or it must be limited to cases where there is no close substitute. We hold that the burden was not upon the defendant, and that the plaintiff has failed to establish a case for the recovery of profits.

There was no proof of damages. Both sides agree that there was no established royalty, and the only possible position for the plaintiff is that it lost the contract. However, there is no reason to suppose that it would have got the job on its own terms. True, it was properly equipped to perform it, and the railroad's engineer went to its factories and studied its designs. He was certainly interested in, and made some use of, the patented device, but that is very far from saying that the defendant would have accepted the plaintiff as its manufacturer, had it observed the patent. The parties had never come even to the point of negotiating a contract. Several alternatives were open to the railroad. It might have got a license; indeed the only evidence is that before infringement one was offered on easy terms. The plaintiff's past treatment of infringers does not suggest that it would have held out for high tribute; it was scarcely in a position to do so. Or, the railroad might have set about devising an articulation which should avoid the claims. It succeeded, when forced by us to do so. We do not mean that such a structure may be used as a "standard of comparison" to limit profits. Perhaps not. Western Glass Co. v. Schmertz Wire Glass Co., 226 F. 730, 737 (C. C. A. 7); Philadelphia Rubber Works Co. v. United States Rubber Reclaiming Works, 277 *595 F. 171, 177 (C. C. A. 2). We need not say. But it is idle to deny, since the fact became manifest later, that the railroad was not inevitably limited to an infringing train, when we are considering whether the contract was sure to be awarded. In any event the case is quite other than Wallace v. Syracuse, 45 F. (2d) 693 (C. C. A. 2), where the City had specified the patented process, the patentee had bid, and would have been accepted, had he not been undercut. The case therefore appears to us to be one for a reasonable royalty under section 70 of title 35, U. S. Code (35 USCA § 70).

The plaintiff had made five settlements, four of them after infringement, and upon all of its three patents, one of which we declared uninfringed. These license fees ranged from $250 to $75. The whole notion of a reasonable royalty is a device in aid of justice, by which that which is really incalculable shall be approximated, rather than that the patentee, who has suffered an indubitable wrong, shall be dismissed with empty hands. It is no more impossible to estimate than the damages in many other torts, as for example, personal injuries with their accompanying pain and mutilation. Though the testimony of experts was recognized as competent in Dowagiac Mfg. Co. v. Minnesota Plow Co., supra, 235 U.S. 641, 35 S. Ct. 221, 59 L. Ed. 398, it is generally of small help. Here it is true that the witnesses agreed within limits as to the percentage upon cost which might be taken, though some spoke in terms of a percentage on profits. But the base used was widely different, the plaintiff's witnesses taking the whole cost of the cars; the defendant's the cost of the articulations alone. We are thrown back therefore upon very little that is tangible, and while any conclusion must inevitably be somewhat speculative, we must find some basis in the evidence; we cannot conjure figures from our own minds. Though the payments were not established royalties, we need not disregard them, any more than the master did. It is true that they were settlements for infringements, but both parties may have been influenced by a wish to be done with litigation; that consideration is a sword with two edges. The plaintiff had not yet got any decision on its three patents; one turned out not to be infringed; another was of slight moment. It was naturally concerned to get further recognition of them without suit. On the other hand the railroad was about to make a large investment, and the plaintiff saw the opportunity for an important exploitation of its invention. Perhaps it could have exacted the highest figure it had exacted before; possibly even more, though the last would be mere speculation. So far as we can find, Menden's affidavit is uncontradicted that for all three patents he had been offered a license of less than $100; how much less he does not say. This was not hearsay evidence as the plaintiff asserts; the parties stipulated that he would so testify, and presumptively the offer came from a responsible official, for the contrary was not shown, though it rested in the plaintiff's power to show it. The patent now before us was much the most important of the three; we doubt that any reduction would have been made, had the license been confined to it. But the plaintiff was playing on a very small margin, as the event proved, and might well have been content not to press the railroad's inventive powers too far. It seems to us that a royalty of $100 for each articulation is the utmost that we can fix upon this record, and we find that amount as reasonable. This should carry interest from the date of the first infringement, that being the period when presumptively it would have been paid. The cost of the appeal will be divided.

Decree modified in accordance with the foregoing.