In the court below the Bemis Car Box Company, owner of patent No. 239,702, issued April, 1881, to Sumner A. Bemis, for a car axle box, brought an action at law against the J. G. Brill Company to recover the alleged damages suffered by it through infringement of said patent. The case having proceeded to issue, it was by consent referred to a referee, who by stipulation was to—
“hear and determine the case and report his conclusions thereon to the court. His rulings on the admission or rejection of testimony and his findings of fact and .conclusions of law shall be reviewable on exception by the United States Circuit Court, and his findings of fact shall have the same force and effect as the verdict of a jury.”
After taking proofs, the referee found the plaintiff had sustained damages in $186,642.24. The parties consented to a pro forma judgment in favor of plaintiff, the Brill Company sued out this writ, and the court now hears the case as if at first instance.
[1] From the proofs in the case it appears that in June, 1890,‘the Bemis Company brought a suit in equity against a street railway company in the Circuit Court of the United States for the District of Massachusetts, charging it with infringing said patent in the use of certain axle boxes made by the Brill Company. Thereafter the latter defended the suit and became the real defendant therein. The case was so proceeded with that in August, 1896, a decree was entered adjudging said patent valid and that the boxes in question infringed the same. Bemis Car Box Co. v. Boston & R. Electric St. Ry. Co., 75 Fed. 403. From such decree the Brill Company caused-an appeal to-be entered to the Circuit Court of Appeals of the First Circuit, which court in April, 1897, affirmed the decree below. 80 Fed. 287, 25 C. C. A. 420. Thereafter the latter court (98 Fed. 121, 38 C. C. A. 661) denied a petition of- the Brill Company for leave to file a bill of re*751View and permit the defendant in said cause to set up and prove the defense, now sought to be set up in the case now before us, of two years’ public use and abandonment to the public of the device covered by said patent. Subsequently the present suit was brought, and resulted in the judgment above recited.
An examination of the record of the recited litigation in the First circuit shows that the validity of the first claim of the Bemis patent was then in issue, and that such litigation resulted in a decree adjudging the title of the patent was in the complainant, that said patent, so far as the first claim thereof was concerned, was valid, and that it was infringed by the particular device of the Brill Company involved in that cause. On well-established principles it is clear that all the questions involved in that issue were, as between the parties to such litigation, merged and concluded in the final decree therein entered. The validity of the claim therefore became res adjudicata, and could not be called in question in any subsequent litigation between the parties. When, therefore, the Brill Company, in the present suit, sought to raise the questions o-f prior use and abandonment — questions which went to the validity of the patent — the referee properly held that as between these parties that question was concluded by the Massachusetts decree. The defendant herein there had an opportunity to litigate that issue, and in point of fact by its petition for a bill of review had contested the very question it now seeks to relitigate. The referee, therefore, was right in holding that the prior litigation had, as between those litigants, finally settled the validity of the claim. But whether a particular form of device, which was not passed upon in the preceding suit, was covered by that claim, was, of course, an unadjudicated question, and therefore open in the present suit. This •question of infringement we think the referee rightly decided against the defendant.
[2] The car box of the Bemis patent is so fully described in the opinions of Judge Carpenter (75 Fed. 403) and of Judge Putnam (80 Fed 287, 25 C. C. A. 420) that we avoid repetition by reference thereto. To us it is clear the patent disclosed, in its field, a generic device, in that it first gave to the art a washer which by its elasticity, its contact with the flange on the box, and a tapered periphery surface, in connection with the flange on the wall, so operated that, in the language of the specification:
‘•There is always contact between the end of the flange and the wall and the side of the washer, and also contact between the inner rim of the washer and the outer surface of the sleeve and the housing.”
Indeed, the generic character of Bemis’ device and the differentiation of the infringing device as a species are to our mind impliedly conceded in Brill’s patent No. 418,439, granted December 31, 1899, under which the defendant sought to justify its infringement, and wherein the patentee, in disclaiming the Bemis patent, said:
T am aware that a ear axle box shield, in which the axle box has in its rear end an annular chamber, with open end toward the wheel, and said chamber having an inclined step or ledge, against which a ring or shield is pressed by a flanged, collar, is old; but the same differs from my invention *752in tliat the flanged collar is either secured to the wheel or forms an integral part thereof, and as this location or securement of the collar flange is open to many disadvantages, which are fully set forth in said patent, I distinctly disclaim the same.”
Moreover, it will be noted that the Brill patent refers to its second ring as a mere duplication of the first or .alternative construction, thus:
“In some eases, where only one ring or shield is necessary or required, the ring or shield Q may alone be used, as shown in Fig. 2, or the ring or shield E may correspondingly be employed, as indicated in Fig. 3.”
In-the Massachusetts case the court had before it the Bemis device, embodying as its distinguishing feature a single elastic washer. The Brill patent, as we have seen, had two elastic washers. In his opinion Judge Putnam describes that element of Bemis’ claim as “(4) the location of the washer so as to be confined by contact,” and says:
“The word ‘crowd,’ found in the specification, relied on by the appellant, is at the most only descriptive, and is sufficiently apt to indicate crowding against the taper of the sleeve merely to make a close joint, exactly as the appellant crowds the washer against its ‘abutment’ for the same purpose. If it be conceded, however, that the appellant does not use the ‘tapered sleeve,’ or the flange projecting out from the side of the wheel, in precisely the forms as described in the appellee’s patent, the case becomes a question of equivalents. * * * This invention did not relate to a matter of mere simplicity of form, or of mere convenience, or to cheapening the cost. It involved a new and useful function, although, perhaps, in view of what the record shows of the art, in a limited field of operation. It is therefore entitled to some aid from the doctrine of equivalents; and we cannot conceive of any ease where it could he' so entitled, unless it is in the present one, where the departures are only in matters of form,, and of such character as to suggest that they are studied evasions of those described, in the claim in issue.”
This estimate of the Bemis device and construction of the claim in question commend themselves to us. Applying them to the device the Brill Company subsequently adopted, having a single washer, the referee was justified in holding it still infringed. The Brill patent, with its two alternative washers was before the Massachusetts court, and a device made with two was enjoined by that court. That a defendant in such case should relieve itself by using one, and especially when that one was by its own specification-admission an optional construction, is a proposition to which we cannot assent.
It is also contended that, inasmuch as the claim in question embodies “a washer placed upon said tapered sleeve and the box and there confined by contact with- the end of the ñcmge and the wall, substantially as described,” and as defendant’s point of contact is not at the end of the flange, but intermediate, it thereby avoids infringement. But it is clear this location is wholly one of form to escape infringement, while it utilizes the infringing function to appropriate the substance of the claim. The point of contact in defendant’s device is, when viewed from result, the functional end of the flange; for, as conceded at bar, the part of the flange beyond had no functional part in the operation of the device. To say that a shoulder or abutment which by its flange relation efficiently effects the elastic connection covered by the patent can be shorn of its infringing character by *753adding functionless further. length to the flange is to make form without substance the equivalent of substantial form. Such a contention is answered by this court, speaking bv Judge Gray, in Carnegie v. Brislin, 124 Fed. 221, 59 C. C. A. 659:
“Again, the claim of tlio patent in suit, as well as the description in tlie specification, requires a feed roller table pivoted at its outer end. Interpreted reasonably, this, of course, does not mean the very tip of the outer end, but the functional outer end for the purpose of pivoting.”
So in the present case the outer end of the flange, for the functional purpose of utilizing washer elasticity, was the point where connection was made in the Brill device. The referee being justified, therefore, in finding the defendant infringed, we turn to a consideration of the amount of the judgment-
At this point it may be as well to state what we understand to be the effect of the findings of fact that are stated by the referee in support of his award. The parties stipulated (as heretofore quoted in part) that the referee should—
“hear and determine the ease and report his conclusion to the court. His rulings on the admission or rejection of testimony, and his findings of Ia.ee and conclusions of law, shall be reviewable on exception by the-United States Circuit Court, and his findings of fact shall have the same force and effect as the verdict of a jury. Judgment shall be entered on such report in accordance with the decision of the United State's Circuit Court, and on appeal or writ of error from said judgment all questions raised by said exceptions or by the filial decision of the United States Circuit Court shall be reviewable by the United States Circuit Court of Appeals.”
It is not specifically stated under what authority, federal or state, the parties were referring- the case, and they differ in opinion now upon this subject; the difference being important in its bearing upon the power of the court to review the findings of fact. In behalf of the Bemis Company it is contended that we are confined to the inquiry whether there was submissible evidence to support a particular finding of fact, and that the finding is conclusive if such evidence shall be found to exist. This is the ordinary rule by which courts are governed in dealing with the effect of a jury’s verdict, and (as we understand the present situation) the plaintiff in error does not deny the proposition that, unless the present reference is controlled by the Pennsylvania practice, the findings of fact do bind us if there was submissible evidence to support them. Under the Pennsylvania practice the rule is apparently different. The act of 1874 (F. L,. 166), and the decisions of the Supreme Court of Pennsylvania thereon, gave to a referee’s findings of fact so controlling an effect that a change in the law was thought to be desirable. Merchants’ Bank v. Kern, 193 Pa. 86, 44 Atl. 334. Accordingly, in 1889 (P. L. 80) the legislature permitted the parties to file exceptions to the referee’s report, required him to re-examine the report in the light of such exceptions, and to amend it if they should appear to be well founded, and, if exceptions were so filed, made it the duty of the appropriate court to consider the report and the exceptions with the referee’s action thereon, and gave the court power “to confirm the report of the referee, or alter, *754amend, or reverse it, er send it back to the referee for further proceedings before him.” It enacted, further, that:
“A writ of error or appeal from tlie final judgment of the court may be taken by either party in like manner as in other cases of a similar kind, provided exceptions were duly filed with the referee.”
.' This statute gives the court of common pleas a large power over the whole controversy in spite of a referee’s finding (Bank v. Kern, supra); and if the contention of the plaintiff in error is correct, that we are to follow the Pennsylvania practice, the correctness of the findings of fact in the pending case is now open to attack, and this court (which is acting in effect as a court of first instance) is not confined to the inquiry whether the referee had before him enough evidence to carry the questions to a jury. For present purposes-we will assume that the plaintiff in error’s contention on this point is correct, and accordingly will examine the report on the subject of damages on that assumption. A summary would not do it full justice, and therefore we quote it at length as follows:
• “Since tbis suit is an action at law for infringement, and not an action in equity for an accounting, tbe plaintiff is entitled to recover compensation for tbe actual damage sustained by it because of defendant’s wrongdoing. This damage must be proved with a reasonable degree of probability. The action is in the nature of tort for a statutory wrong, and not of contract; there being no ’ agreement between the parties with regard to a license or royalty to be paid by the defendant for the use of the plaintiff’s invention. I find the facts as to the previous course of the plaintiff’s business to be as follows:
“The plaintiff company was engaged in the business of manufacturing and selling car axle boxes, pedestals, springs, and other parts, not including wheels, which together comprised the running gear for cars drawn by horses upon street railroads, from a period beginning in 1881, or earlier. It sold this apparatus (including the boxes), which was known in tbe trade as ‘gears,’ throughout a large territory, comprising the greater portion of the United States, and also sold large numbers of said boxes (the device covered by the patent in suit) separately; many such sales being for the purpose of replacing boxes previously sold by it as a part of said gears, and many being intended for use as a part of gears manufactured by others. The business proved to be profitable. There was a broad, general market, not only for the gears, but for the boxes separately. The box proved to be dust proof in practice, and sufficiently tight to prevent leakage of the lubricating oil placed, within it. It was, therefore, possible to operate cars for long periods of time, approximately in many instances a year, without replacing the oil; whereas boxes without the dust proof feature had to be packed with waste and cleaned every few days, and required additional lubricating material to be placed in them at frequent intervals. The exclusion of dust also prevented wearing of the bearings occasioned in other boxes by dust and grit finding its way into the bearing.
“These advantages were due to the mechanical features of the Bemis box, which were covered! by the patent in suit. Other parts of the box were covered by other patents; but all such other patents concerned petty details of construction, which in no substantial degree created a demand for the box, or contributed to its selling value. Plaintiff was able to establish a broad ■market for its box, by reason of its dust proof quality, and the market continued to expand until after 1893, when defendant had become an active competitor in the business with the aid of its infringing device. Prom that time on, defendant’s business gradually increased, and the plaintiff’s decreased, until, in 1898, plaintiff’s sales were only a few hundred boxes. The plaintiff during-the period from 1884 to 1892 had introduced its boxes on 288 rail*755roads. (Pee Plaintiff's Exhibit No. SO, Plaintiff’s Record, yol. 5, page 20.) In 1884 it had US railroads as customers. In 1885 it added thereto 66; in 1886 and 1887, 72; in 1888, 40; in 1889, 38; in 1890, 34; in 1891, 25; and its sales of boxes increased, each year. Among the car builders to whom it sold was the defendant, and from 1884 to 1891 the list of customers who were car builders increased steadily.
“The expansion and subsequent falling off in plaintiff’s business is shown by the testimony of Iloadley (Plaintiff’s Record, pages o24, 525), who furnished a list in tabulated form; the figures being given from September 1st to September 1st, the business year of the plaintiff company. The total number of boxes sold with or without gears or trucks in 1885 was 1,223; in 1886, 2,188; in 1887, 4,239; in 1888, 5,422; in 1889, 6,797; in 1890, 9,051; in 1891 (the year of the Baring panic), 6,761; in 1892, 9,390; in 1893, 4,858; in, .1894, 4,005; in 1805, 5,573; in 3896, 2,532; in 1897, 714; and in 1898, 488— the last item being for the period between September 1, 1897, and April 5, 1898, the date of the expiration of plaintiff’s letters patent. This list was examined by defendant’s expert accountant, Brown, who testified that his estimate varied only slightly from that of Mr. Iloadley. (See Defendant’s Record, page 467.)
“I find as a fact that during this period, and subsequently until the expiration of the patent, plaintiff possessed a plant and working force, and had available sufficient capital with which to continue to supply its goods and meet the demands of a broad market. The defendant argues that by reason of constant distribution of profits during the years from 1890, and earlier, to 1898, plaintiff’ company became financially unable to handle the business. It nowhere appears, however, that the plaintiff’s credit was impaired, nor can it he pi'esumed that the profits would have been so distributed if needed as working capital. The business was commenced and carried on with a small cash capital through its period of greatest expansion, and through the Baring panic of 1891, and there is no ground for inference that it could not have so continued.
“During the same period the defendant was engaged in the business of' manufacturing cars and trucks, and various component parts thereof. Prior to 1889, when it obtained its patent under which its infringing devices were manufactured, it had been a customer of the plaintiff company, and had manufactured many trucks in which jffnintiff’s boxes were used. Prom year to year its sales of trucks, including boxes, largely increased, and from June 28, 1892, until April 5, 1898, when the patent expired, I find as a fact that its sales of infringing boxes aggregated 37,867, of which 9,598 contained the large washer heretofore adjudicated to be an infringement, and 28,269 contained the small or axle washer only.
“This finding is based upon the testimony of the plaintiff’s witness Eugene Eble, who examined defendant’s sales books Nos. 3, 4, 5, 6, 7, and 8 l’or this period; that is, from June 27, 1892, to April 5, 1898. (P. R., pages 277 and 109 and 111.) It was made up from the entries contained in defendant’s l ooks and from defendant’s shop cards showing the sales of trucks, running gears equipped with boxes," and also sales of boxes alone, and also from examination of officers and employes of ihe defendant, John Brill, Ourwen, the defendant’s chief draughtsman, Herbert J. Dycett, and others. (P. R., pages 209-236 and 243-263.) Eble’s list will be found as Plaintiff’s Exhibit No. 38. (Plaintiff’s Record, vol. 5, page 7.) He testified that the total number of all infringing boxes made and sold by the Brill Company from 1892 to 1898 was 38,228. (P. R., page 285.) Certain boxes were eliminated therefrom, and allowed as a reduction by the plaintiff, leaving the total of infringing boxes 37,867.
“By the defendant’s witnesses Adams and Heulings it was clearly proved that the sales of the defendant increased from year to year from and after 1892, while the plaintiff’s sales were decreasing; ihe market absorbing approximately the same number as in 1890 and 1892 (when the plaintiff’s sales were largest), until 1897 and 1898, when there was a slight reduction of the total sales by both parties. The plaintiff granted no licenses to sell its boxes. Plaintiff’s and defendant’s markets and customers wore to a large extent. *756the same. They were In constant competition with each other during the period in question. During that period the prices realized by plaintiff were gradually reduced from about $12 per box to a little less than $7 per box. Its output was reduced, as already shown.
“Having ascertained the number of devices in question which were manufactured by each of the parties to the suit, the next step is to examine the evidence to show the cost of production, and the evidence with regard to the selling price. Much testimony was taken as to both plaintiff’s and defendant’s cost of construction. Plaintiff’s average cost of construction for the entire period was $5.14 per box. See the testimony of plaintiff’s witness, Hoadley (Plaintiff’s Record, pages 513, 514, 531 to 533, 56S to 581, 645 to 646, 682 to 686), who testified that the total cost of each box on the average, and taking one of the heavier and therefore more expensive boxes as a sample, was $4.67, including raw material, labor and proportion of general expense. To this he added 10 per cent. — that is, 47 cents — as an ordinary manufacturer's profit, fixing the total cost at $5.14, the figure above mentioned. The referee is satisfied that these figures are substantially accurate, and represent the maximum cost for the entire period, the books having been thoroughly examined by both parties.
“Defendant, persistently, through its employés who were upon the stand, declined to attribute any proportion of the total selling price of its trucks to the boxes. It was conceded that prices in the defendant’s establishment were fixed by a method of calculating the total cost of all component parts of trucks or cars, and by adding thereto a percentage of profits. Witnesses, however, stated that this percentage differed with different component parts, and that it was impossible to give even an approximate figure to show the profit made by the defendant at any time on any box. The referee finds it difficult to conclude that information could not have been given on this subject which would have been approximately accurate. If, therefore, any uncertainty arises from lack of information upon this point, the doubt should be resolved against the defendant.
“With regard to the selling prices, plaintiff’s testimony, made up from its sales books with reasonable accuracy, shows that the average annual selling prices obtained by the plaintiff were as follows (see Plaintiff’s Record, volume 5, page 58, Exhibit No. 64):
September 1, 1889, to September 1, 1890. $12.13
September 1, 1890, to September 1, 1891. 12.05
September 1, 1891, to September 1, 1892. 11.05
September 1, 1892, to September 1, 1S93. 11.14
September 1, 1893, to September 1, 1894. 10.79
September 1, 1894, to September 1, 1895 . 9.17
September 1, 1895, to September 1, 1S90. 8.SS
September 1, 1896 to September 1, 1897. 8.52
September 1, 1897 to September 1, 1S9S. 6.S9
or an average for the nine years of $10.06s/9 per box.
“These nine years included the six years covered by the suit, in accordance with the statute of limitations, beginning with June 27, 1892, and a preceding period of little less than three years from September 1, 1889; the total period being a little'less than nine years. It should be noted that defendant’s patent, under which it claimed the right to manufacture the infringing boxes, is dated December 31, 1889, and that the prior suit brought by the plaintiff, and defended by the defendant in this case, was commenced in June, 1890. The average selling price, therefore, covers a period of a little less than three years prior to the period covered by the suit, which three years I have included in the calculation, in order to take into consideration to a reasonable degree the prices obtained by the plaintiff before the competition of the defendant became severe. The average annual profit per box upon the plaintiff’s sales is, therefore, $10.068/9, less $5.14, or $4.92s/9.
“The question then arises whether the testimony as above summarized constitutes a sufficient basis for the computation of actual damages. Defendant has introduced much testimony to show that there was no market for boxes *757separate from trueles; that niter its competition with'the plaintiff became severe the device covered by plaintiffs patent had, of itself and separate from trucks, no substantial market, and that other satisfactory boxes were in general use. It appears that about 1890 or 1891, and during a few years immediately thereafter, conditions in the trade were altered by the introduction of electricity as motive power for street railroads. The cars in general use became longer and heavier from year to year and were operated at higher speed. This Occasioned the development of improvements in trucks, and within a few years the demand for the old-fashioned street car gears dwindled away, and their places were taken to a greater part by the four-wheeled, instead of two-wheeled, gears or trucks, which were developed and improved with marked success by the defendant, and to a less extent by the plaintiff. In 1894 and 1895 a process of consolidation of street.railroads commenced which, though it doubtless did not lessen the demand for trucks and gears of the new type, nevertheless largely decreased the number of customers in the market. The defendant company became one of the largest manufacturers of rolling stock with its various parts, though it had powerful competitors. The plaintiff could not hold its own in competition with defendant. It did. however, continue to sell old-fashioned, gears, also its own type of electric four-wheel .trucks, and its boxes.
“The defendant called many of its own employes, who testified that it sold trucks or cars, and not boxes: that it was impossible to designate any portion of the profit from such sales as attributable to boxes as such: that it built up a market for its trucks by reason of their various valuable qualities, other than the box used; and that the boxes which constituted the infringement, contributed but slightly, if at all, to the sales of defendant’s appliances. A number of witnesses were also called by the defendant, who liad been officers of street railroad companies during the period in question, and who testified in effect that the boxes were always satisfactory, but no more satisfactory than those furnished with trucks made by other manufacturers, and principally known as the ‘McGuire’ and ‘Peckhani’ trucks. I do not consider that this testimony carries much weight. It is at best an expression of opinion, and. so far as the defendant's employes are concerned, hearsay as well.
“The important facts are that the defendant after becoming familiar as a customer with the qualities of the plaintiff’s invention, imitated it by adopting and, so far as the evidence shows, using as a general practice in its truck and car business a box embodying the novelty of plaintiff’s invention, which differed from it only in trifling details of mechanical form, and which the defendant, when now called upon to defend its conduct, does not seriously attempt to distinguish from the plaintiff's device, except upon the ground that claim 1 of plaintiff's patent is too narrow to cover the substantial principle of the invention. Doubtless the Brill truck had many valuable qualities, and those qualities helped to sell it. I decline,' however, to find that the infringing features of its box did not contribute to the creation of a market for it, but find, on the contrary, in view' of the defendant’s own course of business, that the use of the infringing boxes was of substantial assistance to it in building up its truck and ear business against all competitors, that, its infringing box wras universally satisfactory to its customers, and in all probability somewhat more satisfactory than the other boxes in general use. It is true that the testimony appears to show' that at the end of the period, viz., in 1897 and 1898, some railroads operated by electricity began to operate at such high speed that this type of box became unsuitable, and they were obliged to adopt axle boxes like those used on steam railroads. It does not appear, however, that the defendant ceased to any substantial extent to use the dust proof box, or that it ever ceased to give satisfaction in city, as distinguished from interurban, operation, during the period covered by this suit.
“I find as a fact, therefore, that there was, during the period covered by the suit, a constant market for the plaintiff’s box, due to the device covered by the patent in suit, and that during that period the defendant, instead of purchasing the use of the invention, wrongfully appropriated it to itself and incorporated it in its own manufactured appliances in an attempt to evade the plaintiffs patent. Plaintiff’s actual damage, therefore, is the loss of prof*758its on the 37,S67 infringing boxes sold by the defendant within the period covered by the suit.
“It has been earnestly contended before me by counsel for the plaintiff that it is my duty in estimating damages to treat the conduct of defendant as a wanton infringement, and to recommend judgment for the plaintiff for a sum in excess of the actual damages. Such a recommendation, if within the province of a referee in any case, would be proper in this suit upon all the evidence, where it not for a single circumstance — the existence of defendant’s patent. There is an element of turpitude involved in conduct properly described as ‘wanton’ which is negatived by the existence of that.patent and claim of right thereunder. See Boyd v. Janesville Co., 158 U. S. 260 [15 Sup. Ct. 837, 39 L. Ed. 973]. Defendant’s conduct, however, while not technically wanton, appears to have been in the highest degree reckless, both of consequences to itself and of injury to the property rights of the plaintiff. Its conduct in this respect has been such as to justify a court in finding any doubtful questions of fact against it, rather than in its favor-.
“Defendants contend that there is no evidence indicating rvith reasonable probability that plaintiff would have made as many sales as defendant if there had been no infringement, or how many sales plaintiff would have made, if less than defendant, and that plaintiff is therefore limited to nominal damages. It seems to the referee that the plaintiff has furnished much reliable data. Its own sales before and after the infringement, its own cost and the prices it received from year to year, its prolonged hold upon the m,arket and the breadth of its trade among railroads and truck and car builders, have all been proved with most convincing detail. There is not the slightest doubt that the device was valuable, and that the demand for it existed and continued before and after the infringement. It is evident that in defendant’s business judgment no other box would have answered its purposes as well. Of course, there could be no absolute certainty upon such a subject; but probability based upon substantial data is sufficient, and every presumption is in favor of plaintiff’s ability to maintain the Output previously established by it as proved in the case. It should be observed that, in most of the cases where a demand for profits has been rejected as speculative, it has been a demand for future profits with reference to the date of the litigation, or for profits based upon a hoped-for establishment of a business, or incredse beyond the business already established, subject to innumerable contingencies with regard to commercial conditions which could only be tested by experience. No such elements of uncertainty exist in this case.
“On the other hand, plaintiff asks me to find, not only that it could have made additional sales equal in amount to the infringing sales made by defendant, but that it could-have obtained therefor the same prices which it was receiving at the date when the infringement began. It asks for a large sum for loss of profits on its own sales during the infringing period. There is much to be said in favor of this claim, especially in view of the defendant’s failure to furnish any useful data on the subject from its own experience. I consider, however, that on the subject of selling prices from year to year the evidence as a whole does not warrant a finding that the earlier rate .of profit could have been maintained. The general competition existing .might well lead railroad- companies which desired to use Bernis boxes to decline to pay the old prices for them. Other large ear builders like McGuire and Peckham were competitors in the field, and although there is nothing in the evidence to indicate, that their boxes were good enough to have driven the Bemis boxes out of the market, I am not satisfied that they would not have forced a reduction in prices. I decline, therefore, to adopt a measure of damages higher than the prices actually obtained by the plaintiff in the face of all the competition, including that of the infringer. I consider, however, that the data appearing in the testimony amply justifies me in fixing the profits lost upon defendant’s sales at the same figure as the average profits made by the plaintiff during the period of nine years upon its own sales. I find that the plaintiff’s damage is as follows: Doss of profits on sales of 37,867 boxes, at an average selling price per box of $10.06Vü, less the maximum cost of production (including 10 per cent, manufacturers’ profit) of $5.14, or an average profit per box of $4.92s/9, making a total damage of $186,642.24.’’’
*759But the referee also computed the damages on the basis of royalty as an infringing user, making the findings upon that subject noted as follows:
“If, however, there should be any doubt whether the plaintiff could have continued to sell these boxes to its customers other than the defendant, had the defendant not distributed them throughout the same market as a part of its trucks or cars, the referee linds that the figures above stated are a sufficiently accurate estimate of a reasonable royalty value for the patented device. The defendant in a certain sense might well be held to have competed as a user rather than a seller, for it appropriated to its own use the plaintiff’s patented invention in its business of manufacturing and selling trucks and cars. By its own witnesses the defendant insisted that the selling value of the boxes could not be separated from the selling value of the trucks or cars. As previously stated, the referee does not consider this testimony con. vincing. Defendant’s officers and employe's showed a marked disinclination to furnish any information contained upon their books and records. Even as to the amount of their sales the facts were not forthcoming until the witnesses called by the plaintiff for cross-examination were examined at length as to the manner in which defendant’s books were kept.
“Assuming, however, that it is true that the boxes were not sold as boxes, but were used in the business of manufacturing and selling trucks and cars, the referee finds that the record furnishes ample data •from which to estimate the amount of a reasonable royalty or license fee which plaintiff should have received from the defendant for the latter’s unlawful use of the plaintiff’s invention without compensation. No royalty having been agreed upon, it is a well-established rule that such royalty may be calculated upon evidence showing the usefulness of the invention, the cost of manufacturing the device', and its market value. The defendant’s underfaking to manufacture and sell the infringing articles and its continuing to do so recklessly in the face of the pending litigation in the Massachusetts court, with full knowledge that it was from day to day injuring the business of the plaintiff, shows how highly the defendant valued the invention. T find, therefore, that the devices were so satisfactory to the defendant and so well suited to the requirements of the market that the defendant would probably have continued to purchase them from the plaintiff, as it did originally, had it not chosen to manufacture the boxes itself at the risk of liability for infringement.
“The general rule for cases of this sort, where there is no established license fee, is stated in the leading case of Suffolk Co. v. Hayden, 70 U. S. (3 Wall.) 3X5, 18 L. Ed. 76. Tn that case the plaintiff had made no sales, it had given no licenses, and there was no evidence before the court of a market or a market price established by the plaintiff before the infringement. The court below’ nevertheless admitted evidence as to the uses and advantages of the device over devices previously used. There was also evidence of tile extent to which the defendant had used it. The court charged the jury that they were to ascertain from the evidence the actual damages sustained. On appeal to the Supreme Court, Mr. Justice Nelson, who wrote the opinion, said: ‘This question of damages under the rule given in the statute is always attended with difficulty and embarrassment, both to the court and jury. There being no established patent or license fee in the case, in order to get at the fair measure of damages, or even an approximation to it, general evidence must necessarily be resorted to. And what evidence could be more appropriate and pertinent than that of the utility and advantage of the invention over the old modes or devices that had been used for working out similar results. With the knowledge of these benefits to the persons who have used the invention, and the extent of the use by the in-fringer, a jury will be in possession of material and controlling facts that may enable them, in the exercise of a sound .judgment, to ascertain the damages, or, in other words, the loss to the patentee or owner, by the piracy, instead of the purchase of the use, of the invention.’
“In the present case the usefulness and advantage of the invention over *760the prior methods are established. The exclusion of dust and retention of oil, the saving of labor in inspecting and refilling the box with waste and lubricating material, and the avoidance of wear upon the bearing are shown by the persistent appropriation of the device by the defendant. We have, in addition, proof of the exact number of infringing articles used by the defendant in its car or truck manufacturing business, and (which was lacking in the Suffolk Case) the cost of manufacture and the average price which the plaintiff was able to obtain by selling the article itself. If an estimate of a reasonable royalty or license value for an invention is ever an appropriate and proper measure of damages, it is so in this case.
“In Rose v. Hirsh, 94 Fed. 177, 36 C. C. A. 132, 51 L. R. A. 801, the plaintiff owned a patent for an improved umbrella stick. The defendant, formerly a customer, ceased to purchase, and deliberately manufactured and used rods in its manufacture of umbrellas until enjoined. The court held that it was treasonable to conclude’ that if the appellees had not deliberately and wantonly become infringers, and wrongfully trespassed on appellant’s patent rights, they would have purchased from appellant the rods they used. ‘These facts united afford substantial, not mere conjectural, grounds on which to base the conclusion that the appellant by appellee’s wrongful acts lost the sale of these particular rods, and to that extent assuredly was damaged. * ... * The master, as we have seen, thought the evidence presented no definite basis on which damages could be assessed. When the facts^ and circumstances attending this case are considered, it seems to us the master was in this regard in error.’ The cost of manufacturing was proved, as well as the selling price previously existing between the parties. The court, therefore, found the damages suffered by the trespass to be the profit on each infringing article, based upon plaintiff’s figures of the cost and the selling price. The only element in the above-cited case which is not present in the case before the master is that there the defendant had no patent. The case would hardly have been decided otherwise, had the defendant succeeded in obtaining the apparent protection of a patent.
“In McCune v. Baltimore & Ohio Railroad Co., 154 Fed. 63, 83 C. C. A. 175, the plaintiff sued the defendant at law for damages for the infringement of a patent for locomotive ash pan. The plaintiff had never granted a license, established a royalty, or manufactured or sold the patented device. There was testimony tending to prove the defendant infringed the patent.' Thereupon the plaintiff offered to prove the utility and advantages of his patented device, that it had been appropriated and used by á number of other railroads, and that this use, being without his consent, prevented the establishment of a market value therefor. This offer the court rejected, and directed a verdict for the plaintiff for nominal damages. Upon error the court, in an opinion delivered by Judge Buffington, held upon the authority of Suffolk v. Hayden, above quoted, that general evidence must be resorted to in similar cases and directed a new trial.
“In Creamer v. Bowers (C. C.) 35 Fed. 206, Judge Wales, who delivered the opinion, says: ‘It is further objected that there was no satisfactory proof that, if the defendants had not infringed, they would have purchased from the complainant patented registers equal in number to the infringing registers made and used by them. That the defendants would have made such purchases is a reasonable conclusion from the evidence. It is certain that they made large purchases at first, and that they did not, during the time they were infringing, wholly cease buying from the complainant, and their conduct convinced the master that in all probability they would have continued to buy at the same rate as formerly if they had not infringed. This cannot be considered as a conjectural or speculative inference from the facts proved. It is true there may be a possible doubt on the subject, but in eases of wanton infringement every doubt and difficulty should be resolved against the infringer. Rubber Co. v. Goodyear, 9 Wall. 803, 19 L. Ed. 566. All that is necessary, in order to prove actual damages, is to furnish some reasonable basis or data on which to calculate them, and, whenever the evidence is sufficiently definite to show the pecuniary loss, suffered by the complainant, he is *761entitled to be reimbursed, no matter whether ihe infringer gained anything by the infringement or not.’
“In thus estimating a reasonable royalty value for the invention, I do not wish it to be understood that I decline to find that there is proof of loss of profits. On the contrary, my finding is, as stated above, that plaintiff did suffer loss of profits. This is based upon the proved cost of production and proved usefulness of the patented article and the strong circumstantial evidence that the market continued to absorb approximately a fixed number of said articles after the infringement began, though at lower prices. But whether the defendant sold the infringing boxes as boxes or not, it unquestionably used Ihom in its business as a car builder. For this use it is liable to the plaintiff in damages to an amount equal to a reasonable royalty. The amount of such a reasonable royalty (allowing a 10 per cent, manufacturer’s profit) is established by the proof in precisely the same manner as was adopted above in estimating the amount of profits lost by the plaintiff. I therefore find that the plaintiff's actual damage, whether on'the basis of a reasonable royalty value, or on the basis of a loss of profits on said 37,867 boxes is a total of $186,612.21.”
[3] It will thus be seen that the referee reached the same conclusion by either method, and the question for us is whether substantial error has been shown either in method or result. The size of the award, if nothing else, calls upon us to be cautious in decision, and we have therefore scrutinized his findings with the care demanded by the situation. But it is not to be supposed (and the Pennsylvania decisions support this statement) that the act of 1889 intended that the mere filing of exceptions should deprive such findings of all value. It contains no language that shows so unusual a purpose, and we have no doubt that, even when a referee’s findings of fact under that statute are questioned, they are still entitled to great weight, and should not be set aside unless they appear to be clearly erroneous. Scott v. Scptt, 196 Pa. 132, 46 Atl. 379; Hotchkiss v. Roehm, 181 Pa. 65, 37 Atl. 119; Bruch v. Philadelphia, 181 Pa. 588, 37 Atl. 818. And the burden of showing them to be erroneous is properly on the party excepting. If the court doubts concerning their correctness, the findings stand as the conclusions of a tribunal to whom the decision of a dispute has in the first instance been committed by the parties — as conclusions that should be accepted until overthrown by a clear conviction of error.
Applying these well-settled rules to the report now before us, we are unable to pronounce the findings of fact to be erroneous in any substantial respect. No doubt the referee was required to decide some difficult questions of fact. He had to choose between opposing witnesses on some matters; he had to decide which of two possible inferences should be drawn from testimony that would support either inference; and he had to weigh the value of evidence in the mass as well as in its component parts. But all this wa£ precisely his appropriate function, and having performed it conscientiously, with great care and marked ability, his conclusions are not to be set aside merely because we ourselves might not haye reached them, if we had been doing his work. In a word, we have not been convinced that the referee’s findings of fact are erroneous upon any vital point, and we therefore accept them as sound.
[4] If his findings are not to be disturbed, the damages were prop*762erly measured either according to the rule of profits or to the rule of reasonable royalty. As to the first rule, the recent decision by the Supreme Court in Westinghouse, etc., Co. v. Wagner, etc., Co., 225 U. S. 604, 32 Sup. Ct. 691, 56 L. Ed. 1222, renders it superfluous to do more than refer to the discussion contained in the opinion; but we may add, in passing, that it was considered and applied. Roth v. Harris (D. C.) 197 Fed. 929. If, however, the rule of reasonable royalty is the rule that should govern under such facts as are here disclosed, we may refer to the cases cited in Rose’s Notes to Suffolk Co. v. Hayden, 70 U. S. (3 Wall.) 315, 18 L. Ed. 76 (6 Rose, 504), in addition to the authorities discussed by the referee.
The referee finds, as noted, that:
“Defendant’s officers and employés showed a marked disinclination to furnish any information contained upon their books and records.”
He also reported that an award of punitive damages would have been justified in the case, had not the defendant’s conduct been relieved of wantonness by the patent under which it claimed to justify. He nevertheless found:
“Defendant’s conduct, however, while not technically wanton, appears to have been in the highest degree reckless, both of consequences to itself and of injury to the property rights of the plaintiff. Its conduct in this respect has been such as to justify a court in finding any doubtful questions of ’fact against it, rather than in its favor.”
It would seem likely, therefore, that the finding of the referee on the proofs before him did not bear as hard on the defendant as would have been the case, had all the evidence in their control been produced. But a study of the proofs given satisfies us that the referee had before him sufficient testimony which, if believed, warranted his conclusion. In the first place, we have the fact that the defendant was a large buyer for a number of years from plaintiff of the box of the patent. In that respect the referee found:
“Prior to 1899, when it [the defendant] obtained its patent under which its infringing devices were manufactured, it bad been a customer of the plaintiff company, and bad manufactured many trucks in wbicb plaintiff’s boxes were used.”
While thereafter defendant ceased purchasing plaintiff’s boxes, it did not stop using boxes of that type, for the infringing box was but the plaintiff’s box made by the defendant. The desire of the defendant to use that type of box would, therefore, had it not become an in-fringer, have probably caused it to continue to purchase from the plaintiff. Its conduct, both before and during the period of infringement, is wholly and only consistent with the conclusion that the defendant would have continued a customer of the plaintiff, had it not become an infringer. In Creamer v. Bowers (C. C.) 35 Fed. 206, a case in this circuit, it was said:
“It is further objected that there was no satisfactory proof that, if the defendants had not infringed, they would have purchased from the complainant patented registers equal in number to the infringing registers made and used by them. That the defendants would have made such purchases is a reasonable conclusion from the evidence. It is certain that they made large pur*763•chases at first, and that they did not, during the time they were infringing, wholly cease buying from the complainant, and their conduct convinced the master that in all probability they would have continued to buy at the same rate as formerly, if they had not infringed. This cannot be considered as a conjectural or speculating inference from the facts proved. It is true there may be a possible doubt on the subject, but in cases of wanton infringement every doubt and difficulty should be resolved against the infringer. Rubber Co. v. Goodyear, 9 Wall. 803 [19 L. Ed. 566].”
Indeed, the character of the patent the defendant took out, which, as we have seen, was a specific form of the generic device of the plaintiff, shows the defendant’s desire to continue the use of a box of that character. To say they could as well have purchased other devices is met by the fact that they did not. Indeed, their present contention that they could have used devices provided with springs is negatived by their admission in their own patent of the worthlessness of such devices:
‘‘They are objectionable, for the reason that the springs either lose their elasticity in time, or are apt to break under pressure of end thrust of the axle, necessitating replacement, besides which said spring plates are more or less expensive.”
Moreover, we think the evidence tended to show the serious character of damage to the plaintiff caused by defendant’s infringement. As found by the referee, the plaintiff’s business—
“proved to he profitable. There was a broad general market, not only for the gears [which included the boxes], but for the boxes separately. The box proved to be dust proof in practice, and sufficiently tight to prevent leakage of ihe lubricating oil placed within it. It was therefore possible to operate cars for long periods of time, approximating in many instances a year, without replacing the oil; whereas boxes without the dust proof feature had to be packed with waste and cleaned every few days, and required additional lubricating material to be placed in them at frequent intervals. The exclusion of dust also prevented wearing' of the hearings occasioned in other boxes by dust and grit finding its way into the bearings. * * * The plaintiff during the period from 1884 to 1892, had introduced its boxes on 288 railroads * * * and its sales of boxes increased each year. Among the car builders to whom it sold was the defendant, and from 1884 to 1891 the list of customers who were car builders increased steadily.”
The proofs show the volume and increase in the sale of boxes by the plaintiff was as follows: In 1885, 1,223; in 1886, 2,488; in 1887, 4,239; in 1888, 5,422; in 1889, 6,797; in 1890, 9,051; in 1891, the year of the Baring panic, 6,761; in 1892, 9,390. They further show that plaintiff sold such boxes in 1889 for $12.13, in 1890 for $12.05, in 1891 for $11.05, and in 1892 for $11.14, and that the cost to the plaintiff of manufacture, including material, labor, and a due proportion of general expense, was $4.67. It will thus be seen that during the six years ending with 1892 the plaintiff had sold 44,118 at a profit, averaging the selling price of those years at $11.59, of $6.92 per box, or approximately $304,296.56. From these figures the substantial character of plaintiff’s business and the value of its patent monopoly are apparent.
*764As to its ability to continue to handle that business, the referee found — ■
“as a 1’act that during this period and subsequently, until the expiration of the patent, plaintiff possessed a plant and working force and had available sufficient capital with which to continue to supply its goods and meet the demands of a broad market.”
The referee also finds, and, indeed, the facts proven by the defendant’s own witnesses and books show, that during the infringing period covered by this suit, and not barred by the statute of limitations, viz., from 1892 to 1898, the defendant sold 37,867 boxes. Of these 9,568 contained the large washer adjudicated in the Massachusetts case to be an infringement. The remaining 28,269 contained the small or axle washer only, which the referee found also infringed. The referee further' found:
“By the defendant’s witnesses Adams and Heulings it was clearly proved that the sales of the defendant increased from year to year from and after 1892, while the plaintiff’s sales were decreasing; the market absorbing approximately the same number as in 1890 and 1892 (When the plaintiff’s sales were largest) until 1897 and 1S98, when there was a slight reduction of the total sales by both parties. The plaintiff granted no licenses to sell its boxes. Plaintiff’s and defendant’s markets and customers were to a large extent the same. They were in constant competition with each other during the period in question. During that period the prices realized by the plaintiff were gradually reduced from about $12 per box to a little less than $7 per box.”
We think there was evidence tending to support these several findings by the referee. Had this case been tried by a jury, the evidence was such that a trial judge would not have been justified in giving a jury binding instructions that it could only find nominal damages for the plaintiff. On the contrary, the evidence was such that different men might reasonably draw different inferences and conclusions as to the quantum of the plaintiff’s damages, and consequently that question was for the jury. It follows, therefore, the question was equally one for the referee, and in the quantum of damages found by him we see nothing to constrain us, assuming we have that right, to set aside his findings and substitute some other method of assessing the damages sustained by the plaintiff. As to the cost of manufacturing the boxes, the referee was justified in his' finding by the proofs before him, and the only possible criticism as to his figures is one in favor of the plaintiff, in that it has been charged and the defendant credited with an added 10 per cent, of manufacturer’s profits, which would possibly have accrued to the plaintiff, had it not lost the opportunity of realizing such profit through defendant’s infringement.
[5] Complaint is made of the basis used by the referee in arriving at an average selling price. His method was to strike an average selling price, based on the six infringing years involved in this suit and the three years jpreceding. It is contended that the noninfringing years should be excluded. On the other hand, there is strong reason for basing the selling price on the noninfringing years and excluding the infringing years, for the former represent the normal legal status during which the plaintiff was reaping the legitimate profits from its *765patent monopoly, while the latter represent that price depreciated by the wrongful competition of the defendant. To base the price wholly on the infringing years is to enable the defendant to profit by its own wrong. Under these circumstances this court cannot say as matter of law that the referee erred in the basis he adopted in arriving at a selling price based on six infringing and three noninfringing years.
[6] Without entering into a detailed discussion of the many alleged errors of the referee, we may say that after careful consideration of them all, we find no substantial ground for reversing the report, which is manifestly most careful, thorough, and able. Indeed, to the case generally may be applied what was said by Judge McKennan in Bigelow Carpet Co. v. Dobson (C. C.) 10 Fed. 385:
“All tliat is left for presumption is that the infringing carpets displaced in the market the complainants' carpets, and hence that the profits which would have accrued to them upon the quantity of carpets put upon the market is the measure of their damages. This presumption, as against a wrongdoer, is not unreasonable, and it has the sanction of numerous decisions. Putnam v. Lomax [C. C.| 9 Fed. 418; American Saw Co. v. Emerson [C. C.] 8 Fed. 806; McComb v. Brodie. 2 O. G. 117 [Fed. Cas. No. 8,708]; Westlake v. Cartter, 4 C. C. 636 [Fed. Cas. No. 17,451].”
To these may be added Blake v. Greenwood (C. C.) 16 Fed. 676; Cincinnati Co. v. Western Co., 152 U. S. 200, 14 Sup. Ct. 523, 38 L. Ed. 411; Corbin v. Taussig (C. C.) 137 Fed. 151; Lazier Gas Engine Co. v. Du Bois, 130 Fed. 834, 65 C. C. A. 172; Callaghan v. Myers, 128 U. S. 617, 9 Sup. Ct. 177, 32 L. Ed. 547.
Bv these cases the general principle is established that the owners of a patent whose business is established, successful,*and well equipped may recover from an infringer in an action at law the profits he would probably have made on sales of the infringing device, and those profits are determined by the average profits made by plaintiff during a reasonable period. The financial outcome of this suit is serious to the defendant, but it must be borne in mind that this is but a tardy reparation for damages sustained by the plaintiff, whose established business has been undermined by the course of the defendant. Too often in the case of such unwarranted infringements the reparation made; by the law has not been reparation in fact. The value of a patent lies in the substantial protection the law affords in the way of injunction or substantial damages. Moved by these considerations the Supreme Court has lately .said (Westinghouse v. Wagner, supra):
“It is to be borne in mind that Congress has legislated (Itev. St. § 4921 [U. S. Comp. St. 1901, p. 3395]) with a view of affording the patentee ampie redress against the infringer, it not only makes the latter liable for damages ■ — sometimes three-fold damages — but for all profits derived from tile use or sale of plaintiff's invention. The rule as to the burden of proof has, however, been so applied that this statutory right has been often nullified by those in-fringers who had enough to smother the patent with improvements belonging to themselves or to third persons. In such eases the greater the wrong the greater the immunity; the greater the number of improvements the greater the difficulty of separating the profits. And if Umt difficulty could only be converted into an impossibility, the defendant retained all the gains, because tiie injured patentees could not separate what the guilty infringer has made impossible of separation. Manifestly such consequences demonstrate that either the rule or its application is wrong. The rule is sound, for it but *766announces the general proposition that the plaintiff must prove its 'case and carry the burden imposed by law upon every person seeking to recover money •or property from another. But the principle must not be pressed, so far as to override others equally important in the administration.”
Applying these general principles, we reach the ends of justice by approving the referee’s report and affirming the judgment entered thereon in the court below.