(dissenting). I am so deeply convinced that the foregoing opinion is erroneous as regards both rules of law applied and facts assumed to exist, and that it will give much trouble in future patent accountings, besides working a grave injustice to the defendant, that I feel constrained to express my dissent at more than customary length. Ordinarily the expression of differences among judges of an appellate court over the facts of a case should be carefully avoided, and it is with reluctance that I discuss them here. It appears to me, however, that, aside from the punishment of the defendant to an unmerited degree, the conclusions of fact announced are not only contrary to prior holdings in this case, but that, even if they were true, they afford an insufficient basis for the rales of law applied. As to their departure from the prior holdings of this court I particularly feel that I should protest.
A brief review of the case will explain the situation. Claim 4 of plaintiff’s patent is as follows:
“Tie combination substantially as described, of an electric converter (transformer) constructed with open spaces in its core, and inclosing case, and a nonconducting fluid or gas in said case adapted to circulate through said spaces and about the converter.”
The validity of the claim and the fact of infringement were established in the Second Circuit in what is known as the Carbide Case against a customer of defendant. That suit was begun May 10, 1900, was decided November 11, 1901, by the court of first instance (Westinghouse Electric & Mfg. Co. v. Union Carbide Co. [C. C.] 112 Fed. 417), and its decree (so far as it concerned claim 4) was affirmed by the Court of Appeals of that circuit May 29, 1902 (117 Fed. 495, 55 C. C. A. 230). The present suit in this circuit for an injunction and an accounting was begun June 24, 1902, and a preliminary injunction was promptly awarded with consent of defendant. About a year afterwards the plaintiff applied to the trial court for an extension of the scope of the injunction and for an attachment of defendant for contempt in *468making and selling transforméis of a certain construction. The court, Judge Amidon presiding, denied the application May 22, 1903. He said:
“It appears from the evidence in the present ease upon this hearing that the defendant, Wagner Electric Manufacturing Company, had been engaged for some five years in the manufacture of the device which is now complained of. It further appears that the defendant, in order to obviate the infringement which was established in the suit in New York, transformed all its converters which had been manufactured in conformity with the device there in evidence to the form of device now complained of. It has gone forward in the manufacture and sale of these devices. In so doing I am satisfied that it has acted in the utmost good faith and under the advice of counsel. I am the more inclined to accept as entirely true the representations of the defendant in this regard from the fact that the defendant had been engaged for some five years in the manufacture of the same device, had advertised it extensively to the trade, and had sold it throughout the country without any complaint on the part of the complainant. It is quite manifest that no showing has been made here which would justify the punishment of the defendant for contempt for a violation of the preliminary injunction.”
The court further held that, aside from the question of contemptuous conduct, the defendant’s device, against which plaintiff’s motion was directed, did not infringe claim 4. The evidence upon which Judge Amidon acted is not shown in the records before us. The case again arose February 13, 1904, in the trial court, on a hearing for an interlocutory decree, Judge Adams, afterwa-rds a member of this court, presiding. (C. C.) 129 Fed. 604. The same questions that were before Judge Amidon were presented to Judge Adams. Judge Adams said, February 19, 1904:
“Again, it clearly appears that the defendant was engaged in making its transformer for some years prior to the hearing of the Carbide Case, and that the witnesses in that case knew of this fact. They certainly knew that defendant was manufacturing a transformer with open space between the coils and a rectangular opening into the core for the insertion of the coils. But no attempt was there made to hold these features to be an infringement of claim 4 of the patent. Complainant there contented itself by claiming the parallel open spaces throughout the core itself to be an infringement. After judgment was rendered in that case, the complainant, with full knowledge that the defendant was manufacturing transformers with the open spaces and openings now claimed to be an infringement of claim 4, never sought to hold the defendant guilty of contempt for violation of the injunctive order in that case. After that injunctive order became final, the defendant conformed thereto by changing its transformer, so as to close up all the parallel open spaces throughout the core itself, leaving the core one solid mass of iron plates, with no open spaces in it, but retaining in its structure the open spaces between the coils and between the core and coils as before, and has continued to manufacture such transformers, so modified, from that day to this. The proof shows no claim that the use of this modified structure constituted a violation of the injunctive order in that case. * * * To now enjoin the defendant, after it has been manufacturing, selling, and advertising the transformer complained of for a period of about eight years, as shown by the testimony, and under circumstances disclosed by this record, would in my opinion be grossly inequitable.”
The first reference was then made to a master. It may be said here that the accounting period was from June 24, 1896, to November 14, 1901, and that the accounting was for profits only, not damages. At the hearing before the master the Fayman schedule, so call*469ed, was introduced and became a part of the record. The master held that claim 4 was an entirety, also that $132,000, round figures, were the profits realized from all the transformers in the Dayman schedule, and, there being no evidence to apportion them between the infringing and noninfringing elements of the transformers, the plaintiff was entitled to recover the entire sum, and he recommended a decree accordingly. On exceptions the trial court (Judge Dyer) disapproved of the decision of the master. It held that the noninfringing features of the transformers contributed to the profits, that the burden was on the plaintiff to separate them, and, not having done so, the plaintiff was entitled to a nominal recovery only. As afterwards said by the Supreme Court:
“The master held that the entire commercial value of the transformer was due to the invention covered by claim 4, and that therefore all the profits belonged to the Westinghouse Company. The court, on the other hand, found that the defendant’s additions were not infringements, and had contributed to the profits, and that because of the failure to make a separation the plaintiff was entitled only to nominal damages.”
Then followed the first appeal to this court. On that appeal the decree of the trial court was affirmed. 173 Fed. 361, 97 C. C. A. 621. In the opinion of this court the character of claim 4 and the defendant’s construction, which plaintiff still insisted infringed, was exhaustively discussed. This court said:
“The appellee [defendant] manufactured and sold two types of transformers ■—the device involved in the Carbide suit, which was there held to be an infringement of claim 4 of appellant’s patent, and the device referred to in the record as Type M. Both are claimed by appellant to be infringements of claim 4 of its patent.”
And we held that claim A was a limited detailed claim, that open spaces in the coils and between the coils and the core constituted a material and substantial improvement over claim A, that type M, not having open spaces in the core, did not infringe; and that the burden of apportioning the profits between infringing and noninfringing elements of the transformers was on the plaintiff. The case then went to the Supreme Court by writ of certiorari, and that court held that the rule stated as to the burden of proof in the matter of apportionment was erroneous. It was expressly said, however, that the writ was not granted for the purpose of re-examining the decision below on the question of infringement. The Supreme Court remanded the cause for a new hearing—
“on all the questions involved in the original reference, and, on evidence already submitted and such additional testimony as may be offered, for further proceedings not inconsistent with this opinion.” 225 U. S. 604, 32 Sup. Ct. 691, 56 L. Ed. 1222, 41 L. R. A. (N. S.) 658.
When the cause reached the trial court it recommitted the accounting to the same master. In the meantime the defendant claimed to have discovered that by mistake and oversight large numbers of transformers of various types, sizes, and kinds which did not infringe claim 4 had been erroneously included in the Dayman schedule, and on that subject and on the questions of profits and apportionment, as *470to which the Supreme Court had laid down various rules, it introduced about 300 pages of additional evidence and more than 200 pages of schedules, tabulations, etc. The plaintiff introduced no evidence. Notwithstanding the above, the master again recommended a decree for plaintiff in precisely the same amount as before. This conclusion plainly resulted from two positions taken by the master, both so obviously erroneous as scarcely to merit discussion. He held', first, that the remand of the cause by the Supreme Court for “a new hearing on all the questions involved in the original reference, and on evidence already submitted and such additional testimony as may be offered, for further proceedings,” etc., gave him no power to impair the evidence tlmt was in the record before the Supreme. Court, but that the defendant was estopped to contest its force and effect (which force and effect he had previously decided); and, second, that the Layman schedule in connection with descriptive recitals in a stipulation to avoid the expense of printing, constituted an estoppel against defendant and that defendant, therefore, could not show that it contained other than infringing transformers. The trial court again vacated the report of the master. It held that he did not give the hearing contemplated by the reference and also that defendant’s contention that the Layman schedule was incorrect “seems to the court to have heen fully sustained by the evidence offered by the defendant at the present hearing.” The court then, .upon its own consideration of the evidence, found that the profits to which the plaintiff was entitled did not in its judgment exceed $10,000, and it entered a decree for that amount. Then followed the second appeal to this court. 233 Fed. 752, 147 C. C. A. 518. We fully sustained the trial court in its view regarding the character of the rehearing before the master and the master’s manifest error, and as to the Layman schedule we said: “At the second hearing the defendant offered substantial evidence tending quite strongly to show a mistake had been made, and that the Layman schedule was materially inaccurate in several respects, but principally in that it included large numbers of transformers of a size and type in which the patented invention had never been used ” etc.
We stated this carefully and temperately, so as not to prejudice the plaintiff upon another hearing of the cause, when with additional evidence,’ if offered, the actual number of infringing transformers in the Layman schedule should be considered. We held, however, that the trial court should have again referred the cause to a master instead of feeling obliged to institute an investigation of its own, following Chicago, etc., Ry. Co. v. Tompkins, 176 U. S. 167, 20 Sup. Ct. 336, 44 L. Ed. 417, and we therefore reversed the cause and' remanded it to the trial court for a new hearing upon the same and additional evidence —using the same language in that respect as was employed by the Supreme Court. A third reference was then made to a different master, and was heard by him on the record as it existed at the second reference. In other words, no witnesses appeared and no additional evidence was received from either party. This master recommended a decree for $182,000, round figures, or $50,000 more than the first master on the same record, and more than the entire amount of profits *471of defendant from all the departments of its business in the accounting period. The report of this master is unjudicial in tone, intemperate in its reflections upon witnesses who had not appeared' before him, and inaccurate in its statements of important facts. No one familiar with the case can read his report without being impressed by the absence of that dispassionate, impartial attitude to which every litigant in a court of justice is entitled.
It is important to note here that the award of $182,000 recommended by the master was based, not on the Layman schedule of transformers, but upon a part of the Williams schedule produced on behalf of defendant at the second reference to the first master, and purporting to be a list of all transformers, infringing and noninfringing, sold by defendant during the accounting period. Taking a part of the Williams schedule as a basis the master put upon defendant the burden of proving affirmatively which of the transformers listed in it did not infringe ; and he applied the burden of proof with so heavy a hand that in practical effect he ignored or dissipated for one reason or another, and not always logically or consistently, every other part of the voluminous evidence introduced by defendant at the reference to the first master after the case came back from the Supreme Court. He ruled that the part of the Williams schedule which he took as a basis was a declaration by defendant agoÁnst interest. The other part of the Williams schedule and all the other evidence, oral and documentary, bearing upon the number of infringing transformers, the relative commercial value of plaintiff’s construction, the value of defendant’s noninfringing openings, and the matter of apportionment were held for naught. In view of the character and volume of this evidence, admitted under the mandate of the Supreme Court, and the opinions of the trial court and this court as to its probative effect, the conclusion of the master is so extraordinary that its gross error is quite manifest. He also held claim 4 was unitary. When the report of the second master came before the trial court, instead of setting it aside wholly and referring the accounting to some one who would give the inquiry a judicial treatment, it sustained exceptions to vital parts and again reduced the award to plaintiff, this time to $9,174.12. The grounds for this appear in the court’s opinion. The plaintiff’s appeal from this last decree is now before us.
From the course of the case to this point the following conclusions appear, which would ordinarily be the law of the case unless this court, now differently constituted than before, decides not to give them effect: (1) Claim 4 is a limited detailed claim; open spaces in the core being a distinctive feature. (2) Oil-cooled transformers—that is, tranformers in a container with oil, with open spaces in the coils, and also between the coils and the core, but without such spaces in the core itself—do not infringe, and defendant had a right to make and sell them. (3) The Layman schedule did not operate as an estoppel, but, on the contrary, the character of the transformers listed in it, infringing or noninfringing, should be determined from all the evidence on that subject. In this particular we said there was substantial evidence before the first master on the second reference tending quite *472strongly to show that the Layman schedule “included large numbers of transformers of a size and type in which the patented invention had never been used.” (4) Defendant’s type M transformers, the core of which was without spaces, did not infringe.
There are also some matters of fact which I think ought reasonably be regarded as settled in this case, or at least be given more weight than has been accorded:
1. Open spaces in the coils of an oil-cooled transformer, and between the coils and the core, constituted “a material and substantial improvement” over claim 4; they “performed a very substantial office” even, in connection with core spaces. See opinion of this court on the first appeal. (The serious effort at apportionment by proof ought to be seriously considered.)
2. The trial court (Judge Amidon), in May, 1903, held (a) that defendant had been making for some five years a transformer with open spaces for the circulation of the cooling liquid, but without such spaces in the core, and that they did not infringe because of that construction ; and (b) that “the defendant in order to obviate the infringement which was established in the suit in New York (Carbide Case) transformed all its converters (transformers) which had been manufactured in'conformity with the device there in evidence (with core spaces) to the form of the device now complained of” (that is, without core spaces). Two things are apparent here: The making by defendant during the accounting period of oil-cooled transformers which plaintiff claimed before Judge Amidon to infringe but which lie held did not; and, next, upon the decision in the Carbide Case that those having core spaces infringed, the defendant obviated that infringement by closing the spaces. Again, the trial court (Judge Adams) held substantially the same in February, 1904, and the same observations may be made about his decision as about that of Judge Amidon. In other words, defendant had been making before the hearing in the Carbide Case oil-cooled transformers without core spaces (of which type M was an example), as well as transformers with such spaces in the core, and also spaces between the coils and between the coils and the core, and that upon the decree in that case, holding the latter to infringe, it obviated the infringement in that part of its output by closing the spaces, thereby making them in that particular like others of its manufacture. That this is true conclusively appears from the evidence adduced before Judge Adams. (It will be observed that he said in 1904 that spaces between the coils and between the coils and the core had been employed by defendant for eight years.) (This evidence will presently be adverted to in detail. When the case afterwards came to this court on the first appeal it was said:
“The appellee [defendant] manufactured and sold two types of transformers—the device involved in the Carbide suit, which was there held to be an infringement of claim 4 of appellant’s patent, and the device referred to in the record as ‘type M.’ Both are claimed by appellant to be infringements of claim 4 of its patent.”
Type M was taken as a notable illustration of defendant’s transformers made during and after the accounting period which did not *473contain core spaces. It was not the only type of that particular construction, nor the only type with the noninfringing spaces. As will appear hereafter, as many as 1,206 type M transformers Were sold during the accounting period, the first sale being January 9, 1901, most of them being 15 kilowatt or less in size. The first master evolved the theory that there were two type M’s, one with infringing core spaces made during the accounting period and the other without core spaces, to which all the infringing transformers were adjusted after the decree in the Carbide Case. He therefore held that all type M’s in the accounting period infringed. The only substantial foundation for this, aside from the conclusive effect erroneously accorded the Layman schedule, is a recital (225 U. S. 608, 32 Sup. Ct. 693, 56 L. Ed. 1222, 41 L. R. A. [N. S.] 653), in the “Statement of the Case” preceding the opinion of the Supreme Court as follows:
“It appeared that after the decree in the Carbide Case the Wagner Company had instructed its experts to build a transformer that would not infringe the Westinghouse patent. They thereupon devised one, referred to herein as type M, which omitted the (a) open spaces in the core, but substituted (b) spaces between the coil, and (c) spaces between the coil mié the core.”
This recital was employed historically or as a preface to the opening sentence of the opinion itself that—
“The statute makes the decision of the Circuit Court of Appeals final in patent cases, and the plaintiff’s petition for the writ of certiorari herein was not granted for the purpose of re-examining the court’s ruling that defendant’s type M transformer was not an infringement of claim 4 of the Westinghouse patent.”
■I am quite sure the Supreme Court did not intend this as a decision upon an important question of fact. It was so taken to be by the second master in his report on the third reference. He said that the later type M machines (without core spaces) resulted from instructions by defendant after the decree in the Carbide Case, and he added:
“ I so hold, as did the Supreme Court.’ 225 U. S. 608” (32 Sup. Ct. 691, 86 L. Ed. 1222, 41 L R. A. [N. S.] 653).
The evidence in the record that was before the Supreme Court shows without contradiction that defendant had been making transformers with open spaces in the coils and between the coils and the core for years before the decision of the Carbide Case, also that some of them had spaces in the core, some did not, and that type M, originating early in 1901, were of the latter kind. The photographic illustrations in the pamphlet of 1901 issued in June of that year show this last structure. The second master invokes another item of testimony for the conclusion that there were two type M’s, and that all of that type in the accounting period had open spaces in the core. At the first reference to the first master after the Layman schedule had been produced the plaintiff asked Mr. Layman to explain a dozen descriptive terms applied to transformers in the schedule. He did so by saying that they were factory terms applying to the number, type of mechanism, or electrical construction, different forms of design, particular characteristics of winding or of the boxes, or some other *474feature. Plaintiff then asked this question about seven of the types previously referred to:
“Constructional differences between G, F, Spec. F, K, L, and M have nothing to do with the issues in this matter, I presume, so far as you understand them?”
Mr. Layman answered: “No, sir; they have not.” Much is made of this, because some of the types other than M infringed, and the conclusion that type M infringed also was enforced by the last master against the positive testimony of three witnesses, besides Mr. Layman, at the second reference, that it had no core spaces. This indicates quite a curious condition of mind. If the question quoted had been framed as the witness might reasonably have understood it, and doubtless did in view of those just preceding and his answers to them, it would- have read:
“Have the constructional differences you have just mentioned, which led your company to apply the different factory terms, anything to do with the issues in this case?”
Almost anything can be proved by taking questions and' answers away from their context.
The evidence before Judge Adams shows clearly there were not two type M’s differing as to core spaces. At that hearing, which was for the interlocutory decree and before reference to a master, the plaintiff introduced defendant’s “Catalogue No. 102, Bulletin No. 59.” The bulletin was a printed pamphlet, dated “1903,” and entitled “Types M and M H Transformers.” It says type M is “our well-known high-efficiency type.” The illustrations in this pamphlet show these transformers without open spaces in the core. It was stipulated by the parties for the hearing mentioned that defendant made and sold transformers described and illustrated in this pamphlet “prior to the filing of the bill of complaint” in the trial court. The bill of complaint was filed in the court below on June 24, 1902. On the other hand, defendant introduced in evidence its Bulletin No. 43 of the same catalogue. This was a printed pamphlet, dated “1901,” was entitled “Type M Transformers,” and the text of it began as follows: “We continue to advocate the shell construction for transformers.” The illustrations in this pamphlet also showed the cores without spaces. A witness for defendant (Wells), who had been in its service from September, 1898, to July, 1901, said this pamphlet was issued just before he left; that, as near as he could recall, it was issued in June, 1901, and was for public distribution. Mr. Layman testified that it was issued' “early in the year 1901,” and that not less than 5,000 of them were distributed throughout the United States. An illustrated cross-section of a 15-kilowatt type M in this pamphlet shows open spaces between the coils and between the coils and the core, and the text discusses the provisions made for the circulation of the oil. Other illustrations, one of a 10-kilowatt type M, show the core without spaces. Mr. Layman also testified that defendant had 'been continuously making transformers with open spaces between the coils and between the coils and the core for 8 or 10 years (note the opinion of Judge Adams), *475and that its pamphlet issued and circulated in December, 1897, a copy of which was received in evidence, disclosed that construction. It was not meant by this that type M’s had been made that long. On cross-examination of Mr. Dayman, he said he thought defendant had been making type M’s since June, 1900; but, after seeing the pamphlet of 1901 (put out, as Wells testified, in June of that year), he said that to the best, of his recollection it was building type M’s about six months before that publication was issued. That would make the time about January, 1901. There was no testimony before Judge Adams on that subject less favorable to defendant than that above recited; and the conclusion that defendant had for years during the accounting period been making and selling transformers with the open spaces other than in the core, and that after about January, 1901, it had been making those known as type M without spaces in the core, was undisputed. There was no evidence to the contrary. It was this state of evidence to which Judge Adams and this court referred, and presumably Judge Amidon also. The Williams schedule, received in evidence at the second reference to the first master, discloses that the first sale of a type M transformer occurred January 9, 1901, thus corroborating the above evidence before Judge Adams.'
The Dayman schedule 'came into the record at the first reference to the first master. The importance hitherto attached to this item of evidence has been lessened, because a part of the subsequent Williams schedule (second reference), covering all transformers, infringing and noninfringing, made and sold during the accounting period, has been taken (third reference) as the basis of the decree against defendant. The Williams schedule embraces more transformers than Dayman’s for reasons hereafter mentioned. The Dayman schedule has been erroneously referred to a number of times as listing only about 6,000 transformers. The exact number is probably not very important, excepting for comparison with Williams’, and upon the contention of the defendant that the master’s stenographer and defendant’s clerks, in compiling Dayman’s, had apparently taken a list of all transformers (except the very smallest ones) appearing on the records they consulted, regardless of infringement. But any one who takes the trouble to compute the number will find that the Dayman schedule lists at least 16,300 transformers. There is no doubt about this. The first master at the second reference reported the exact number as 16,366. Mr. Dayman’s official positions with defendant and his connection with the mechanical construction of its transformers bear on the evidential value of that schedule as one exclusively of infringing machines. From 1892, when he was about 23 years old, to 1898, 2 years after the beginning of the accounting period (1896-1901), he was actively employed in tire transformer department, and participated in supervising the construction, testing, and installing of transformers. From 1898 to 1902 he was assistant manager and treasurer, and from 1902 to 1903, in which latter year he first testified for the hearing before Judge Adams on the interlocutory decree before the accounting was referred to a master, he was manager and treasurer. After 1898 his duties were principally commercial; he had no active supervision of *476the transformer department. It should be remembered that defendant was extensively engaged in making other things than transformers, such as motors, switchboards, etc. Transformers were but a department of its business. Witnesses for both sides agreed that Mr., Schwedtman, who was one of the founders of defendant’s business, but who had severed his connection with it before he testified at the first reference in 1905, had an intimate connection with the making of transformers; and plaintiff’s principal witness at the first reference (Johannesen) said no one knew more than Mr. Schwedtman did about it. Mr. Schwedtman did not aid in preparing the Layman schedule. He did not testify about it at the first reference, nor does it appear that he knew anything of it, until the controversy about it arose at the second reference, after the case came back from the Supreme Court.
The Layman schedule came into the case in this way: At the first reference Mr. Layman was put upon the stand by the plaintiff. • Defendant’s records having been produced or held subject to inspection under the master’s ruling, plaintiff proceeded to question him about each sale of transformers from the beginning of the accounting period. That course appearing tedious and interminable, plaintiff suggested that the witness produce at a later session a complete transcript of all transformers “hereinbefore inquired about” from the records of defendant produced before the master and other records of its business. It is fair to say, however, that there was always enough in plaintiff’s questions on this subject to indicate that infringing transformers only were meant, although at times, in following the questions in that aspect, rather close attention was required. Thereafter the Layman schedule was produced, and was assumed to contain only infringing transformers. It gave the following information as to each transformer listed: Defendant’s order number; the name of the customer; the residence of the customer; the date of the order; the number of transformers embraced in each order and sale; their size; their kind or type; the price; the total price to the customer; the agent’s commission; and finally the net receipt of defendant, being total price less commission. There is no doubt that at the time the schedule was produced as and taken to be a list of infringing transformers. It was so referred to a number of times by Mr. Layman. It is equally plain that this was to a considerable extent the result of assumption on his part, instead of a conclusion from his own knowledge derived from personal investigation or inspection. At times he referred to the schedule as “covering” or “incorporating” all infringing transformers. He made it clear at the outset that, while he believed the schedule accurate for the purpose sought, it was not checked by him, but was prepared under his instructions from the records by a representative of the master and the clerks of defendant. Again, at that reference the attention of defendant was directed more to the relative value of the core spaces in infringing transformers, and the belief that the burden of separation was on the plaintiff. The Supreme Court afterwards referred to this in declining to direct a decree on the record before it, but remanding the case for a new hearing. 225 U. S. 622, 32 Sup. Ct. 691, 56 L. Ed. 1222, 41 L. R. A. (N. S.) 653. At that same first *477reference Mr. Eayman testified to facts quite inconsistent with the conclusion that the schedule contained none but infringements. For example, he said the core spaces were made by driving plugs between the sheets of metal composing the core, which served to compress the sheets into separated groups; that this was a mechanical expedient, in lieu of external compression of all the core sheets into a solid mass; that it was employed in some transformers of various types, and not in others; and that it was not employed in small transformers. Yet the schedule contains large numbers of the very smallest, also types which with little question have been excluded from the recovery, notwithstanding the supposed estoppels, burdens of proof, etc., which have been so liberally imposed on defendant at various stages of this case.
It may be said at this point that defendant’s literature nowhere discloses that core spaces were an advertised feature. Moreover, the infringement by the very large transformers involved in the Carbide Case, used at Niagara Falls, was removed by simply withdrawing the plugs and compressing the sheets of the core together. Thereafter they continued in-use without invasion of plaintiff’s rights. I have said that Mr. Eayman testified, at the first reference at which the schedule was produced, that core spaces were not used in very small transformers (thousands of which were nevertheless listed in the schedule). Having a confirmatory bearing on this was the business practice of the General Electric Company, one of the very large manufacturers of transformers in this country. It was a licensee of the Westinghouse patent in which claim 4 appears. An engineer in the service of the General Electric Company, testifying for the plaintiff in the Carbide Case, said that a regular line of oil-cooled transformers up to 50 kilowatts in size were made by his company without the internal openings. In other words, that extensive concern, with legal right to employ the device of claim 4, omitted it from a standard line of oil-cooled transformers of less than 50 k. w. size. The great majority of those listed in the Eayman schedule were less than the size mentioned. I do not mean by this that defendant’s practice should be judged by that of another company, or that' it was not right at the first reference to take the Eayman schedule as it was evidently intended; but I do say that there were numerous indications, even at that time, of which I have mentioned but a few, that Mr. Eayman was mistaken as to the character of his schedule. There was proof at the second reference of the inaccuracy of the Eayman schedule in this particular, which was voluminous, direct, and positive. There is too much of it to recite in detail in this opinion. Part of it was given by men who were in a position personally to know whereof they testified. There was proof that in certain types and certain sizes of transformers core spaces were never employed, and that in other types and other sizes core spaces were employed, but not exclusively. Estimates and tabulations were also, given as to the extent of infringement in the latter class. It is this volume of evidence which the trial court said “fully sustained” defendant’s contention, and to which this court referred when it said that it was substantial and tended “quite strongly *478to show” the Layman schedule materially inaccurate in several respects, but principally in that it included transformers of a size and type in which the patented invention had never been used. To escape the effect of all this about the Layman schedule, the second master on the third reference shifted the ground. He took a part of the Williams schedule of all transformers of every kind as a basis, held that the burden was on defendant to show which of them did not infringe, and then decided that, except to a comparatively small extent, the burden was not discharged. This resulted in what I conceive to be a miscarriage of justice.
The Williams schedule was produced by defendant at the second reference to the first master, when the character of the Layman schedule was under investigation. Mr. Williams was a public accountant, not a patent expert. He was employed by defendant to check the Layman schedule with its records, and to make from the records a list of all transformers of every kind, size, and description sold by it during the accounting period, regardless of the question of infringement. He found that the Layman schedule had evidently been taken from the order records, instead of the sales records of defendant, and that, besides many errors appearing from a comparison of it with the records consulted by its compilers, it was incomplete and defective in three important particulars: First, it omitted orders from customers booked prior to June 24, 1896, the beginning of the accounting period, but filled within that period; second, it omitted transformers consigned to agents whose sales did not get on the order record; and, third, it took no account of customers’ returns of transformers and allowances to them. Williams made up his schedule in two parts, which, when introduced in evidence, were designated A and B. The first was a list of all transformers sold, with information like that in Layman’s, but not so complete; and the second was a list of returns and allowances. As I have said, the Layman schedule, according to the first master, listed 16,366 transformers. Upon the same authority the Williams schedule listed for the same period 3,381 more than Layman’s. A comparison of the two schedules, Layman’s and Williams’, together with the explanation of the major differences between them above noted, and the fact that Layman’s confessedly listed a large number of noninfringing transformers, effectively destroys in any reasonable, impartial mind the assumption that the Layman schedule was in truth a list of infringing transformers. But, after all, that is no longer of much moment. The exhaustive effort of defendant to disclose the character of that schedule was turned to its disadvantage at the last reference by the second master, who took Williams’ schedule A as his foundation of defendant’s liability as prima facie, and then practically disregarded its evidence.
At this point arose one of the curious things in the course of the third reference. The master’s report as first filed ignored Williams’ schedule B, and at defendant’s instance the cause was recommitted to him for a report upon it. In a supplemental report he disposed of Williams’ schedule B as follows: First, it was not proven to be (a) “from books of original entry,” or (b) “such books being those used in the course of defendant’s business,” or (c) “from entries made at or about *479the time of the transaction.” (The Williams A schedule, however, was accepted as a declaration against interest; Williams’ B was held to be self-serving.) To the above it should be said that the order of the court .referred the cause to this master for a hearing “on the evidence already submitted and such additional evidence as may be offered by either party.” All the evidence had already been submitted to the first master at the second reference. No additional evidence was received by the second master. When the Williams schedule was offered at the previous hearing, B was offered first, and immediately afterwards A was offered. There was a general objection to each as “incompetent and immaterialas there was too much other evidence that tended to impugn the Layman schedule; but there was no suggestion whatever that the schedule, or either part of it, was objectionable for the reasons stated by the last master. Both A and B were received in evidence subject to the objections. B was drawn from defendant’s books and records, which were already in evidence. Moreover, Williams’ testimony as to his work showed what he consulted to be inherently of an admissible character, even were the objections timely and sufficient in all other respects. This ruling of the master is so frivolous as hardly to deserve a serious consideration. Just before A and B were offered in evidence, Williams had testified that the actual number of transformers sold by defendant would be shown by deducting those returned in B from those listed in A. The master said that, if he were in error in die above ruling', then, second, “in testing the probative weight of Williams’ B schedule, we must first note that we have no oral proof to show that the returns and allowances may not all relate to transformers or other merchandise sold before the accounting period begins or to what they do refer. Without oral explanation the items are meaningless.” The record before the master showed this not true. Mr. Williams testified before the two parts of his schedule were formally offered in evidence as follows:
“There were some transformers returned in the early part of this accounting that were sold before the accounting period began. I did not list them. I simply listed those that were sold during the period and returned.”
Third, the master next criticized the Williams B schedule for insufficiency of data, but this was no more substantial than his other objections. The essential data were there, as other parts of the master’s report show. Fourth, he finally pointed out a number of discrepancies which he said appeared from a comparison of it with the Layman schedule, and because of them condemned the entire document upon the assumption that the Layman schedule was correct. He gave defendant no opportunity to explain them, to show which was correct, or whether there was in fact any real discrepancy, but he ruled as follows: “The master disallows each and every credit claimed in it.” If there were in fact any discrepancies between the two schedules, the undisputed proof as to the manner in which the schedules were compiled, who compiled them, and the pains taken with them, would seem to require, more consideration of Williams’ than of Layman’s in the matter in question. But, even if this were not so, the discrepancies indicated went to but *480a comparatively small part of the returns and allowances found in Williams’. Again, the discrepancies, if any, may well have been in defendant’s records, instead of originating in Williams’ schedule, or the Layman schedule may have been at fault. But, passing all the above, it occurs to me that most men exercising such functions would have been content under the circumstances with the exclusion of the discrepancies, leaving the remainder in effect.
Another feature of the report of the last master may be referred-to. It concerns fuse plugs. Mr. Williams testified, and his evidence was undisputed, that defendant sometimes sold its transformers without fuse plugs, but in most cases with them; that in some cases the transformers and the fuse plugs were invoiced together, without separation of the price, and in others they were invoiced separately; that in the latter part of the accounting period the price of the plugs was generally shown separately on the invoices; that the prices of the plugs specified ranged from about $1.75 to $4 per pair, the most frequent price being $2.50 per pair. Where the plugs were invoiced separately, he did not extend the prices with those of the transformers on his schedule A, ,but simply gave the prices of the transformers; and where the transformers and the plugs were invoiced together without separation, he applied the price of $2.50 to the plugs and extended them separately in his schedule. The master treated this subject as follows:
First. “In the absence of precedent, the master sees no reason for allowing the credit claimed.” It seems to me one would hardly expect a fuse plug precedent. There is an abundance of precedent, if any is needed, for the exclusion of things sold that were not a part of the infringed invention.
Second. “The fuse plug is a safety device, and performed the same function in the defendant’s transformers which it performs in electric house lighting or in trolley cars. It is as much an integral part of a transformer as an oil cup is of an engine or other machine.” That their function is the same as in house lighting or traction service is unimportant. They are not as intimately a part of a transformer as an oil cup is of an engine. They are separate and distinct devices, hung outside the metal container of the transformer, to protect it from an overcharge of electricity, as by lightning. A fuse plug is a complete mechanism in itself, and is not illustrated or remotely referred to in plaintiff’s patent. Indeed, if any inferences are to be drawn from the words of the patent, fuse plugs would be excluded by the specification, immediately preceding the claims, that “the entire converter may be sealed into an inclosing case,” etc.
Third. “It was not sold separately, save in the few cases specified in the Layman schedule, and, of course, was not invoiced separately.” This is not true. Fuse plugs were sold and invoiced separately in several thousand instances, as shown by Williams’ testimony and his schedule A. Nor can I find that the Layman schedule refers to fuse plugs at all.
Fourth. “No record was kept by defendant to warrant or justify Mr. Williams’ conjecture as to what their selling price might have been, if sold separately.” I have already stated the substance of Mr. Wil*481Hams’ testimony, and he was testifying to what defendant’s books and records showed.
Fifth. Then, as if the above objection might not hold, the master applies his general one that, “aside from what has just been said, the claim for this credit is supported by evidence so vague and indefinite and the amount is-so clearly excessive,” etc. The evidence was not vague or -indefinite. Except as to the price to be attributed to those fuse plugs invoiced with that of the transformers, it was definite and precise. Every sale involving the two prices in a single invoice was specified, with descriptive data by which it could be checked from the books; there were no generalizations and no estimates. When it came to the price to be allowed for the fuse plugs not separately invoiced, Mr. Williams took $2.50 per pair, because that was the most frequent price of those specified. It was less than the average of $1.75 and $4, the two extremes. Where the master found warrant for his statement that “the amount is so clearly excessive” does not appear in the record. If it was regarded as excessive, it was his duty to find a reasonable amount. These aspects of his report are not the most important ones. They do not affect the recovery he recommended as much as some others, but they illustrate -clearly the way he handled the matters referred to him. It was a case of all signboards pointing to the destination when one is on his way.
I now come to the opinion to which this dissent is addressed. The opinion does not indicate the amount of the decree rendered or directed, nor does it show whether the basis for the decree shall be the Layman schedule or that of Williams; there being between them a difference of about $50,000 in the result. It does not confirm or reject, or direct the confirmation or rejection of, the report of the last master upon the third reference. I am at a loss to know just what has been or will be done. Some things, however, are said to which I must record my objection. It is said in the foregoing opinion that there are two questions in this case:
(1) “How many infringing transformers were manufactured and sold by the defendant? and (2) wbat was the profit of the defendant upon tbe infringing transformers sold?"
I had supposed there was a third and very important question, to which the two above mentioned were in a degree preliminary, and to which the Supreme Court devoted considerable attention. Having found the number of infringing transformers and the profit on them, then (3) what part of that profit was attributable to the patented feature, and what part to the features not patented? I will not discuss this last matter, save to say that upon it, and also upon the questions of the number of infringing transformers and the profits on them, defendant introduced, under the warrant of the opinion of the Supreme Court, about 300'pages of additional evidence and more than 200 pages of schedules, tabulations, etc. The plaintiff introduced no additional evidence whatever. It seems to me incredible that all of this effort was wholly futile and fruitless, as it is held to be. I cannot here go into the details of that evidence, but will allow the proposition to rest upon the inherent moral probabilities of the situation.
*482I wish, however, to make clear the following: In the foregoing opinion the first two questions, namely, the number of infringing transformers sold and the profit on them, are stated and considered together as.subject to the same tests of fact and law, and as though both were subject to the rule of apportionment that the Supreme Court applied only to profits on an infringing transformer containing noninfringing elements. Recitals of evidence are applied to both questions indiscriminately, and the conclusion is stated that the burden of proof is upon the defendant as to both. In other words,. defendant’s records having shown the total number of transformers of all sizes, types, and kinds sold by defendant during the accounting period (Layman’s schedule, 16,366; or Williams’ schedule, 19,747), it is held that the burden of proof was on defendant to prove which of them did not embody plaintiff’s invention and that, carrying that burden, all that defendant could not definitely identify as free from spaces in the core went into the class of infringements. Cases of inadvertent (but legally wrongful) infringements might easily be imagined, in which such a rule would bankrupt a manufacturing enterprise and work a manifest miscarriage of justice.
There is a clause in the opinion from which an inference might naturally be drawn that the Supreme Court put on defendant the burden of proving which of its transformers did not infringe. It is said:
“Upon the second hearing, the Supreme Court having placed the burden of proof upon the defendant, the Wagner Company proceeded to attack the Layman schedule and stipulation aforesaid, and the principal witness produced was one John R. Williams, an accountant, who was permitted by the master to present a transcript from the card index and record of the defendant, containing a total list of all the transformers manufactured and sold by the defendant, infringing and noninfringing.”
The Supreme Court did not put this burden of proof upon the defendant. It is an extension in the foregoing opinion of the rule announced by the Supreme Court applicable to the apportionment of profits between infringing and noiiinfringing elements in infringing transformers. Furthermore, the testimony and schedule of Williams, who, it is said, “was permitted by the master to present a transcript,” went much deeper into the business of the defendant than the above quotation would seem to imply. His schedules were an exhaustive exposition of defendant’s sales of transformers during the accounting period as disclosed by its records. Again, it is said in the foregoing opinion of the second reference to the first master:
“But defendant had the burden of proof as to the number of transformers before any evidence was introduced. In addition to this it had the burden of overcoming the weight of the [Layman] schedule and stipulation; so that, when it came to the second hearing, it must be confessed that only reasonably clear and- convincing evidence contradicting the Layman schedule would be sufficient to sustain the burden resting upon the defendant.”
In this way this court, now differently constituted than before, effectively reverses what upon full consideration we once said regarding this matter. It not only puts the burden of proof upon the defendant, but erects the Layman schedule as a barrier, to be surmounted only by clear and convincing evidence.
*483Then, again, the opinion refers to the findings of fact upon three references to masters and to the rule as to the credence to be given in such cases. The rule is not applicable. The first report of the first master was held by the Supreme Court insufficient for a decree under the circumstances of the hearing before him. After the case came back, it was again referred to him, and the additional evidence, was adduced. This court set aside that master’s second report because he misconceived his duties; and, upon the assumption that the Layman schedule was an estoppel, and that he could not allow the additional evidence to impair the force and effect of the evidence upon the first hearing, had in effect given no rehearing at all. I have already referred to the third reference to the last master with considerable detail. But, so far as concerns the rule as to findings of masters, it is sufficient to say of the findings of the last master that not a particle of oral or other evidence was introduced before him. He simply took the printed and written pages of the old record just as now before us. In the foregoing opinion it is said:
“Furthermore, the court cannot shut its eyes to the well-established rule that, ‘so far as it depends upon conflicting testimony, or upon the credibility of witnesses, or so far as there is any testimony consistent with the finding, it [the master’s report] must be treated as unassailable.’ ”
The rule in the opinion above quoted was taken from Davis v. Schwartz, 155 U. S. 631, 15 Sup. Ct. 237, 39 L. Ed. 289. The reference to the master in that case was by consent of the parties, instead of being made by the court under the ordinary chancery practice. Because of that fact the Supreme Court applied the doctrine of Kimberly v. Arms, 129 U. S. 512, 9 Sup. Ct. 355, 32 L. Ed. 764. That this is so appears from the next succeeding sentence in the report from which the quotation was taken. The Supreme Court said further:
“As the reference in this case was by consent to find the facts, we think the rule in Kimberly v. Arms applies, and as there is nothing to show that the findings of fact were unsupported by the evidence, we think they must be treated as conclusive.”
Of course, that is not the case here; the reference was under the general practice, to which a far different rule of presumption applies. The doctrine of Kimberly v. Arms is well known in federal practice. As stated by Mr. Justice Field, it is:
“A master in chancery is an officer appointed by the court to assist it in various proceedings incidental to the progress of a cause before it, and is usually employed to take and state accounts, to take and report testimony, and to perform such duties as require computation of interest, the value of annuities, the amount of damages in particular cases, the auditing and ascertaining-of liens upon property involved, and similar services. The information which he may communicate by his findings in such cases, upon the evidence presented to him, is merely advisory to the court, which it may accept and act upon, or disregard, in whole or in part, according to its own judgment as to the weight of the evidence, s * * But when the parties consent to the reference of a case to a master or other officer to hear and decide all the issues therein, and report his findings, both of fact and of law, and such reference is entered as a rule of the court, the master is clothed with very different powers from those which he exercises upon ordinary references, without such consent, and his determinations are not subject to be set aside and disregarded at the mere discretion of the court.”
*484This court has had occasion to páss upon the doctrine of Kimberly v. Arms and Davis v. Schwartz a number of times, and to point out the conditions to which it does or does not apply, and I think we should observe them here. To surrender the trial of a case to a master and to treat his report as unassailable approaches an abdication of judicial functions. The results, so far as a review is concerned, are similar to those of the trial of an action at law by a District Court without a jury. Neither can be done without the expressed consent of the parties. In Mastin v. Noble, 157 Fed. 506, 85 C. C. A. 98, Judge Adams, speaking for this court, said:
“The reference to the special master to find the facts and conclusions of law was not made by consent of the parties. Accordingly his report is not clothed with that presumption in. its favor which attends reports made on reference by consent. The trial court could not of its own motion abdicate its proper function or delegate it to any 'other person. The report, therefore, could be treated, as the record clearly shows the Circuit Court did treat it, as advisory only.”
In Guarantee Gold Bond, etc., Co. v. Edwards, 164 Fed. 809, 90 C. C. A. 585, Judge Sanborn said for the court:
“The next contention is that, since the case was referred to the master by consent, and since his finding that the deed was not procured fraudulently, and was not intended as a mortgage, was founded upon conflicting evidence, it was unassailable, and the courts below erred in reversing it under the decision in Kimberly v. Arms, 129 U. S. 512, 516, 524, 9 Sup. Ct. 355, 32 L. Ed. 764, and Davis v. Schwartz, 155 U. S. 631, 633, 637, 15 Sup. Ct. 237, 39 L. Ed. 289. But in those and similar cases, where the rule here invoked prevailed the parties consented to the references to the masters, while in this case the reference was made by a general order before the suit was commenced, without the knowledge or consent of any one connected with it. The fact that the parties to this suit proceeded with the prosecution of it under this general order without objection was not a request for or a consent to the order; it was nothing but a compliance with it, and the case does not fall under the rule applicable to cases of consent.”
In International Harvester Co. v. Carlson, 217 Fed. 736, 133 C. C. A. 430 (a bankruptcy case), Judge Amidon said for the court:
“Such a reference, however, is not by consent, and the report of the referee can be treated as advisory only”—citing Kimberly v. Arms and Davis v. Schwartz.
See, also, Babcock v. De Mott, 160 Fed. 882, 88 C. C. A. 64, 67.
Were this not enough, reference might be made to Denver v. Denver Union Water Co., 246 U. S. 178, 38 Sup. Ct. 278, 62 L. Ed. 649. Justice Pitney said:
“In our opinion, the District Court erred in declining to pass upon the questions raised by the exceptions. Although no. opinion was filed, the ruling appears to have been based upon the theory that, because the order of reference was made by consent of parties, the conclusions of the master were not open to question. Kimberly v. Arms, 129 U. S. 512, 524, and Davis v. Schwartz, 155 U. S. 631, 633, 636, are cited in support, but they are distinguishable. In the former case, the reference, made by consent of the parties, authorized the master to hear the evidence and decide all the issues between them, and it was because of this that the court held the findings were not merely advisory, as in the ordinary case, but were to be taken as presumptively correct,” etc.
*485• He then added that the order of reference to which the parties consented showed that the master’s report was for the “advisement of the court.” In the case at bar the trial court has twice expressed in decided terms its dissatisfaction with tire reports of the masters (second and third references), and it would be somewhat extraordinary to find that, having sent the case to them for its advisement, it had constituted subordinate tribunals whose reports “must be treated as unassailable.” Generally speaking, such references, whether to special masters, auditors, examiners, or commissioners, are merely as aids to the judges who appoint them (Ex parte Peterson, 253 U. S. 300, 40 Sup. Ct. 543, 64 L. Ed. 919), and their actions are subject to their supervision and must be acceptable. (The new equity rules require such references to be the exception, not the custom.) It is true that a degree of verity is generally accorded the report of a master, but it is like any other status brought into being by a court, and subject to its confirmation or rejection. It is accorded when the master has heard the evidence himself, and particularly when the testimony is oral, but not when he has failed to observe the purpose of the reference, or appears to have ignored legal principles. When the confirmation of his report by the trial court is given, a very strong presumption arises on appeal.
Tilghman v. Proctor, 125 U. S. 136, 8 Sup. Ct. 894, 31 L. Ed. 664, cited in the foregoing opinion, is not an authority for the rule there quoted. The reference was an ordinary one, the evidence, including oral testimony, was heard by the master, and his report was confirmed by the trial court. Even so, the Supreme Court, upon an extended consideration of the evidence, rejected the master’s conclusion upon a branch of the case, saying:
“Much of the testimony on which he chiefly relies was in the record upon which the case had been previously heard before this court, * * 0 of which the court then said: ‘We have examined the evidence on this point, and are satisfied that it shows the objection to be unfounded.’ ”
This excerpt is rather pertinent to the case at bar. Part of the evidence before the last master was before this court oh the first appeal, when, the-character of plaintiff's patent claim and of defendant’s additions were discussed, and all of it in the record on the second appeal, when we decided that the hearing contemplated had not been given by the first master on the second reference and definitely expressed our view upon the evidence affecting the Eayman schedule.
Adamson v. Gilliland, 242 U. S. 350, 37 Sup. Ct. 169, 61 L. Ed. 356. also cited in the foregoing opinion, was a case of a finding of fact by a District Judge upon oral testimony of witnesses before him. It is true that Justice Holmes cites Davis v. Schwartz as authority for the doctrine of unassailability of a finding so made, but it was probably inadvertent, as Davis v. Schwartz was a reference by consent of parties as pointed out by Justice Pitney in the later case (246 U. S. 178, 38 Sup. Ct. 278, 62 L. Ed. 649), in which the doctrine in question is discussed and distinguished.
A number of reasons are recited for imposing the twofold burden of proof (as to the number of infringing transformers and profits) upon the defendant. But, as I have said, the reasons are not stated distrib*486utively, and some appear without relation to both subjects jointly. It is said first:
“The infringing element in the transformers was hidden from view by the metal casing,” etc. “The metal casing was part of the transformer, and was firmly and permanently fastened in place in the manufacturing process.”
It is true that the transformer proper was intended to be submerged in oil in a metal container with a top, but I can find no warrant in the record for the Statement that the casing was permanently fastened in place in the manufacturing process. On the contrary, the tops of the cases were easily removable, and it was the custom to sell hangers by means of which the transformers could be easily lifted from the oil for repairs to the windings, etc. It was important that this be so. The photographs produced by defendant of a number of transformers sold throughout the country during the accounting period to show that they were without core spaces, and the testimony regarding them, indicated ho difficulty in the particular mentioned. Moreover, I have never heard the reason relied on assigned as one for transferring from a plaintiff to a defendant the burden of proof as to the number of infringing machines sold. There were methods open to the plaintiff of determining which of defendant’s transformers infringed, and there is evidence in the record', much of it, direct and positive, that thousands of transformers included in the recovery indicated did not have the infringing device.
Second. It is said that—
“During the time of infringement, 1896 to 1902, the plaintiff had no access to the factory of the defendant or the books or records of the defendant, and after infringing and up to the time of taking testimony in this case, plaintiff had no knowledge of defendant’s business except as disclosed in the infringing litigation. So that the number of infringing transformers manufactured and sold, and the cost, selling price, and profits were known only to the defendant.”
As a ground for imposing upon the defendant the burden of proof, either as to the number of infringing transformers or as to the profits made from them (I am not now speaking of the apportionment rule announced by the Supreme Court), I will allow the above quotation to speak for itself.
Third. The books and records of defendant did not disclose how many infringing transformers were sold or the profits made from those sold, or from"any element included in them. As to this I will say there was much proof as to the number, sizes, and kinds of transformers which did not infringe, and there was proof as to the profits made.
Fourth. Destruction of records. It is true that some records were destroyed, probably in 1910; but it is not intimated that it was done in view of this litigation, and if it were so intimated there is no basis for it. It had been the custom in defendant’s engineering department to make advance estimates of the cost of labor and material going into transformers. Some of those estimates were in the shape of a chart with an ascending curve. In those days the estimates were used in making up periodical inventories of manufactured stock on hand. Both the advance estimates and the inventories are referred to in the foregoing opinion. The former were destroyed as being useless when a *487new and more accurate cost method was adopted. It is common knowledge that in the past 20 years or so there has been a great development in cost ascertainment in factories and the allocation of cost to items of product. It has been gradual, and as new methods were adopted old ones were naturally discarded.
It was claimed by defendant’s witnesses that the engineer’s estimates of costs were generally too low and were of little practical value. The evidence shows the inventories themselves were not destroyed, but neither side called for their production. But, however this may be, it is not seriously contended that the advance estimates of cost that were destroyed had any bearing upon the question whether the transformers to which they related infringed or did not infringe. Two of those estimates which survived, and which were produced in evidence by former employés of defendant, but then in the service of the plaintiff, threw no light upon the question of infringement. This matter ma.y have a bearing on the question of profits, but so far as that matter is concerned a measure was easily found in the evidence that was available, and it was applied. As to the finding of profits made it is certain that the defendant secured no advantage through the absence of. the engineer’s advanced estimates of cost or otherwise.
The above are the reasons, so far as I can gather them from the opinion, for putting upon the defendant the burden of proving which of all the transformers made by it during the accounting period did not infringe.
It is said that defendant had notice of the claim of infringement as early as 1897, and continued “from that time until July-2, 1902” (November 14, 1901 ?). Any inference from this that defendant knowingly and willfully persisted in the infringement of particular claim 4 of this particular patent is erroneous. It is true that on July 13, 1897, notice of infringement was served upon defendant; but the notice covered 26 different patents, which doubtless contained many times that number of claims. While such a notice might be sufficient for the purpose of the statute, it is not regarded as making the person to whom it is sent an intentional, willful infringer for punitive visitation.
In conclusion: I think that the plaintiff is entitled to a substantial recovery in this case. On the other hand, I think the defendant is entitled to have a real trial of its case and a real hearing before a master, who will accord it the just consideration that it has not so far received. No record is too voluminous or trouble too great for that end. I think the decree now before us should be vacated, and the cause sent back to the trial court for reference to such a master, so that when it again comes before court it will come with such presumptions as are proper to be' given. Also in the interest of consistency in the opinions and conclusions of this court, especially in the same case on different appeals, I earnestly protest against the foregoing' opinion. The observance of that consistency has hitherto been scrupulously given by the members of this court, but here the prior judgment of this court upon the quality of the evidence brought into the record at the second reference, expressed by judges who presumably gave it their careful, painstaking consideration, is not only waved aside for one *488reason or another, but also subordinated to the findings of masters which were successively disapproved by the trial court. And in doing this there has been no tempering of expression. What was once in this case regarded as “substantial evidence” is called “guesswork.” It is as though the case had never been here before.