L. HAND, Circuit Judge

(after stating the facts as above). The suit is so complicated, and such different considerations apply to the disposition of each patent, that it is necessary to treat them separately.

The Basie Patent.

If we disregard the first supposed infringement at Modem in 1910 and the experimental plant at Midvale, the defendant’s continuous exploitation of the “basic” patent begins in the spring of 1912, 11% years before the bill was filed. It is not, however, quite fair to ignore these earlier efforts, and so to cut off 2 'years from the plaintiff’s period of inaction. Dwight saw the Modern photograph some time in 1910, and, while it showed no details, it disclosed a plant which was making “sinter.” We cannot suppose that it did not excite his interest, or that he did not inquire into details, and it seems in every way probable that he then learned of the process then carried on. In September, 1910, when he paid the license fee to Greenawalt under his earlier patents, he did not suggest that he had any complaint. It is answered that at that time the Modem plant had shut down, which is true; but this gave him no *827 assurance that Greenawalt had abandoned the process. In fact, Greenawalt was active in 1910 and 1911 in getting his process installed, and Dwight was not altogether ignorant of his doings, because in 1911 and 1912 he threatened some of his licensees with suit. It is not as if Greenawalt were a recent or sporadic interloper in the field; he was in the business, a by no means infertile inventor, an active and driving competitor. We think that Dwight must be charged with knowledge of what Greenawalt had done and proposed to do for certainly 13 years before the suit was filed.

Meanwhile Greenawalt had built up a very large business, which he estimates to amount in all to about 300,000 tons a year. It is quite true that this has not been by building his own plante; he has exploited his invention only by granting licenses, but the assurance bom of such long inaction is for that reason no less. In the case of any accounting, the result of a decree is never more than to take from the infringer money that he has collected in the past, or to charge him with damages. The disuse of his plant follows only upon an injunction, and, as substantially all the decisions turn upon the right to an accounting, they are as relevant when the infringer has not manufactured as when he has.

It is often said that mere delay will not serve as a bar even to the accounting. Ide v. Trorlicht, etc., Co., 115 F. 137, 148 (C. C. A. 8); Columbia Graphophone Co. v. Searchlight Horn Co., 236 F. 135, 140 (C. C. A. 9); Drum v. Turner, 219 F. 188, 198 (C. C. A. 8); Baltzley v. Spengler Loomis Mfg. Co., 262 F. 423, 426 (C. C. A. 2); Smith Hardware Co. v. Pomeroy, 299 F. 544 (C. C. A. 2). These eases say that there must be some assurance of impunity, the disappointment of which amounts to an estoppel. We are so to understand the law, and yet it remains true that in most of those cases where relief has been denied no further assurance was given than that arising from the patentee’s long inaction. Each year as it passes inevitably builds up a belief, if nothing has been done, that the patentee does not suppose his rights invaded, and, if the time be long enough, the infringer gamers the harvest of even the earliest of the 6 years to which recovery is in any event limited, with just confidence that he will not be disturbed. To allow this reasonable inference to be upset, and his financial life perhaps to be gravely deranged, is thought to be incommensurate with the importance of the patentee’s stale claims. At any rate, whatever the explanation, there is abundant authority to deny an accounting when the patentee has let the infringer slowly build up a large business without protest. Mosler v. Lurie, 209 F. 364, 370 (C. C. A. 2); Wolf, Sayer & Heller v. U. S., etc, Co., 261 F. 195 (C. C. A. 7); Window Glass Mach. Co. v. Pittsburgh Plate Glass Co., 284 F. 645, 650 (C. C. A. 3); Richardson v. Osborne, 93 F. 828 (C. C. A. 2); Universal Arch Co. v. American Arch Co., 290 F. 647, 653 (C. C. A. 7); Cummings v. Wilson, etc., Co., 4 F.(2d) 453 (C. C. A. 9); Simpson v. Newport, etc., Co. (D. C.) 18 F.(2d) 318, affirmed 18 F.(2d) 325 (C. C. A. 2).

A delay shorter than that at bar has repeatedly been held enough, and indeed the plaintiff scarcely denies as much, relying for its excuse upon the pendency of the interference proceedings. Its position is that the real issue there was priority in the “basic” process, and that it was not bound to prosecute while this remained undetermined. It is impossible to see how the priority of the “basic” process could have been more than indirectly involved in the interference; all that was actually at stake was the priority between the parties in the discovery of the ignition devices claimed. Even to the limited degree that interference decisions are an estoppel at all, so far as we have found, their effect is limited to subsequent controversies between the parties upon the patents then involved. Besides, in such cases the decision, even of the Court of Appeals of the District of Columbia, is not res judicata (Butterworth v. United States ex rel. Hoe, 112 U. S. 50, 60, 5 S. Ct. 25, 26 L. Ed. 656; United, etc., Co. v. Muther, 288 F. 283 [C. C. A. 1]), whatever its prima facie validity until successfully overcome.

Dwight & Lloyd would have remained free, therefore, to sue the defendant even if priority had been awarded to him ; they had no excuse for waiting, because they gained little by the delay, nothing at any rate which would have precluded the defendant from contesting all the issues. Indeed, it cannot be doubted that, if the ignition claims had gone to Greenawalt, they would eagerly and properly have protested that the “basic” patent remained unaffected. To allow them to hold in reserve this patent appears to us to invite just those delays which are so oppressive an incident to patent suite. If we are to recognize such an excuse, we weave one more thread in the mesh which entangles those who must pass through their mazes. The delay of 13 years, not otherwise excused, therefore appears to us fatal to any accounting, and the decree as to the “basic” patent is affirmed.

*828 The Sintering Pan Patent.

As this was issued in July, 1914, the time is shorter, something over 9 years. Nevertheless, through all that period the infringement was open and known, and it cannot be even plausibly urged that the interference proceedings had any relation to this patent. So far as concerns the accounting, 9 years may be quite as good as 13; but as the patent has not expired, and the question would arise whether the bar extends to an injunction as well, we do not rest our decision upon the delay.

Although the disclosure is of an up draft pan, the patentee intended it to be operated by a down draft as well. Nevertheless, claim 4 seems to us limited by the feature, “means for positively restraining the ore particles from agitation.” In view of the specifications, this element comprises more than the earlier one in the same claim, “a holder * * * arranged to support a layer or stratum of the mass.” It refers to the top plate, which the defendant does not use. Claim 3 has not this feature, and the defendant infringes, except for the fact that what passes through the “support” is the product of combustion, and not its “supporter.” This is a narrow construction, which we should scarcely accept, were it not necessary to do so in order to save the claim.

We think it is necessary, however, if it is to escape anticipation by Eldred’s patent, 885,328. This was not, it is true, for “sintering” in the sense that Dwight used it, but it was for a closely analogous purpose, roasting a sulphide of lead, “galena,” or similar sulphide ores. The use for which the apparatus was intended is irrelevant, if it could be employed without change for the purposes of the patent; the statute authorizes the patenting of machines, not of their uses. So far as we can sée, the disclosed apparatus could be used for “sintering” without any change whatever, except to reverse the fans, a matter of operation. If the hood, 4, was lifted there would remain the “shallow converter,” 1, the exact equivalent of Dwight’s “receptacle,” 24, with a “perforate false bottom, 3,” again the equivalent of Dwight’s “perforated partition,” 25. ’Below the false bottom is a “chamber, 9,” corresponding to Dwight’s ‘Tower chamber, 25.” The “converter” or “receptacle” is mounted on trunnions, and the air enters, or the products of combustion leave, this chamber through a passage in one of the trunnions.

How there could be a more entire fac simile of the apparatus we cannot see. It is true that the valves which connect with the pipe, 11, must be closed, but, once that be done, if the fans are operated to draw air through the layer of the ore to be roasted or “sintered,” the apparatus would work precisely as Dwight’s, if the defendant is to infringe. This would not be true if the air was sucked1 into the chamber, 9, for then a top plate would be necessary. We do not, therefore, suggest that the patent in suit is invalid if the claims be confined to an up draft, though otherwise they would be. We prefer to say that claiip three must be so limited, and that it is not, therefore, infringed. The decree is affirmed as to this patent.

The Iron Patent.

The application for the “iron” patent was filed on June 17, 1911, and it remained nearly 7 years in the Patent Office. Of this time it was in interference for over a year, was finally allowed on June 3, 1915, and renewed on June 2,1917, to be reallowed on the 27th. Even then the fee was paid on the last day possible, December 26, 1917. The evidence of a purpose to keep it in the office to the last moment lawfully possible is very . strong.

The original specifications do not mention the possibility of silicates. The peg on which the amendments, made 4 years later, must hang, are certain passages and claims which prescribe the addition of enough coal to leave an unbumed residue after the “sinter” is complete. This and the statement that the ore is “partly” melted, and that the process is like smelting and a step in it, are thought to show that temperature control against silicification, which was later so much amplified. The argument is that, as one cannot preserve any coal without keeping the temperature too low for silicates of iron to form, some temperature control was implied. Even so, a follower of the specifications who cared nothing about keeping unbumed coal in his ‘^sinter” would receive no information that he would avoid silicates, if he did; nor would he get any intimation that, if he burned out the coal, they would appear. We decline to accept such obscure references as even premonitions of the matter added later. They are straws at which the patentee now clutches to extend his invention.

Therefore we think the matter so introduced a new invention, and it might be argued that claims 1 and 6 were invalid, irrespective of whether intervening rights had occurred or how much time had elapsed (Chicago & N. W. Railway Co. v. Sayles, 97 U. S. 554, 24 L. Ed. 1053). But we need not go so far; the delay was for over 2 years after *829 Greenawalt’s plant had been erected at Modern and was in operation, and, if his process involves temperature control at all, a disputed question, the amendment was too late. At least some satisfactory excuse must be given. Mahn v. Harwood, 112 U. S. 354, 5 S. Ct. 174, 6 S. Ct. 451, 28 L. Ed. 665; Webster Electric Co. v. Splitdorf Electrical Co., 264 U. S. 463, 44 S. Ct. 342, 68 L. Ed. 792; Westinghouse, etc., Co. v. Jeffrey-De Witt Co., 22 F.(2d) 277 (C. C. A. 2). Just what that excuse might be does not appear in the books, and we need not here inquire what might serve, since none is offered which deserves discussion. Claims 1 and 6, if infringed, were therefore invalid. That is enough to dispose of them here; it is not necessary to make an unconditional finding that they are invalid. As to them the bill may be dismissed.

Claims 4 and 8 do not expressly involve temperature control, and claim 8 certainly should not be so construed, being introduced a year before that theme entered the disclosure at all. Although claim 4 describes some of the chemical reactions thought to take place, we do not see-that it adds anything to claim 8. They occur whenever the “basic” process is practiced with iron oxides and coal, coke, or the like. These two claims differ only in that the first is confined to such metals, and the second is applicable to any “fine metal-bearing material.” So viewed, their validity depends upon whether it was invention to use the “basic” process for “sintering” iron oxides with the addition of coal. The specifications of that patent themselves contemplated the “sintering” of iron sulphides, for which no addition of coal or other carbon products was necessary. Upon this, as upon other matters relating to the arts, it is best to make no á priori assumptions; what seems simple may in fact be difficult.

However, omitting, as in considering these claims we must, temperature control and the avoidance of silicates of iron, the process appears to us completely anticipated by Heberlein’s patent, 844,355, which produced satisfactory “sinter,” save for silicates. We can see nothing left of the mere process of “sintering” iron oxides by the addition of coal, except that Heberlein says that his process may be conveniently carried on in his large holders. That might conceivably be enough, were it not for the “basic” patent itself; but that, too, was in the art. To say that, given that patent, it was invention to use it for nonsulphides in combination with Heberlein’s method, appears to us too much. The merit of the invention chiefly urged is its avoidance of silicates; it does not appear that merely to use carbonaceous material when the ores did riot contain their own combustible proved a great step in the art. We think these claims anticipated, and it is not necessary to consider the issue of infringement as to them, any more than as to claims 1 and 6.

The Ignition Patents.

On December 23, 1907, Dwight filed an application showing a machine to practice the “basic” process. This necessarily involved igniting the surface of the ore, and this was disclosed in various forms though none of them was claimed; the applicant at that time obviously supposing that they were not inventions. Figures 1 and 4 showed a pipé, which crossed the surface of the ore in advance of the air draft, and which had orifices through which the flame set fire to the surface. Figure 7 was for a brazier of coals, with a perforated screen within the suction area or pressure box. Figure 8 was for two opposed jets playing in a line across the ore, and figure 15 for a bottomless trunk or duet crossing the surface. All these were for direct ignition; the rest were for indirect.

Originally, as we have said, Dwight claimed none of these, but on February 21, 1910, he added three claims, which read, however, only upon the indirect heating igniters, all that he meant to reserve in that application. At the same time he amended his application to say that he excluded figures 7,14, and 15,-because they had been divided out on February 4,1910; that division finally resulting in patent 1,433,350. The three claims so introduced he canceled on September 22, 1910, with a reservation that he might put them into a divisional application filed later. Two divisions he did file in March, 1912, and some of the claims now in suit were the originals of those applications. A third application was later divided out of one of these two. Meanwhile the division of February 4, 1910, claimed the direct flame contact of the open duet, and an amendment to it of February 15th, the brazier. These features were continued till issue.

When Dwight abandoned the three indirect igniting claims on September 22, 1910, his reservation of their subject-matter for later divisional application was no broader than the canceled claims themselves. These had, of course, not included flame contact, which had been already divided out on February 4, 1910. True, the pipe with orifices remained, but that was not claimed, and was anticipated anyway. When, therefore, in *830 March, 1912, he for the first time introduced flame contact into those applications, which resulted in patents 1,433,349 and 1,433,351, it was new matter which he was no longer entitled to claim, since G-reenawalt had already more than 2 years earlier been using such a process to set fire to the ore.

In Dwight & Lloyd’s patents, 882,518 and 916,393, an igniter had been disclosed, operating by direct flame contact. These were pending at the time the original application was filed, and they were not eopending applications of the same inventor. Dwight & Lloyd jointly were different inventors from Dwight alone, and the joint invention of both is a good anticipation to the later invention of either. Bannerman v. Sanford, 99 F. 294 (C. C. A. 2). Hence, under Milburn Co. v. Davis-Bournonville, 270 U. S. 390, 46 S. Ct. 324, 70 L. Ed. 651, the disclosures in the two joint patents just cited are valid anticipations of direct flame contact, taken broadly. For two reasons, therefore, claims 1 and 2 of 1,433,349 are invalid.

Claim 3 of patent 1,433,349 includes as an element that the down draft shall draw the flame' into “intimate contact with” the ore. This was interjected by changes in the drawings and specifications, which were an afterthought. Nothing of the sort appeared in the original specifications, but rather the contrary. These read as follows: “The ore as it passes under the igniter L,, will be ignited and immediately come under the influence of the air currents entering the suction box.” This does not suggest any suction of the flame itself. Moreover, it is, if not theoretically, at least practically, impossible that, placed where they are, the flames will be sucked into the ore. While the pallets have no transverse ends, the air will follow the path of least resistance; it can scarcely 'move diagonally from the burner through the ore, when it has a direct path from above.

At most the draft might divert the flame laterally away from the ore, which is just what the claim does not admit. Claim 7 of patent' 1,433,351 apparently contemplates and provides against exactly this, though nothing in the specifications of the patent suggests the possibility, but rather that the flame will be drawn into the ore (page 2, lines 61-65), a passage copied from the specifications in patent 1,433,349. This element was, therefore, not only introduced into the claim too late, but it does not in fact exist in operation. We cannot resist the conclusion that it was added to cover Greenawalt’s flame hood, which had meanwhile gone into extensive use.

Claim 7 of patent 1,433,351 contains as an added element a flame of enough head not to be diverted by the draft. This element, taken alone, appears to us too insignificant to support a patent, when all the rest was old.

The claims in suit of patent 1,433,352 are for an igniter within the hood; at least claims 1 and 5 are expressly so, and in the light of the specifications perhaps also claims 8 and 9. It-is true that in the original application there was disclosed a brazier situated within the hood. This was not claimed originally, and, of course, the indirect ignition claims added on February 21, 1910, did not cover it; the brazier being already divided out. It is fairly plain that the only reason why the brazier'was ever placed within the hood was that it was supposed that the draft was neeessary to keep it active. At any rate, neither the three indirect ignition claims of February 21, 1910, nor any other claims before March, 1912, carried any intimation that Dwight regarded this feature as patentable, or meant to claim it. It was therefore too late on March 14, 1912, to introduce claims to a matter which had already been 2 years in the public demesne. Claims 1 and 5 are invalid, and, if claims 8 and 9 do not contain the element that the igniter shall be in the hood, they are equally bad, as the claims of 1,433,349.

Patent 1,433,350 does not, however, fall within any of these objections; certainly not claims 8 and 9. As we have said, the open flaming trunk or duet was continuously claimed from February 4, 1910, and we can see no laches to affect their validity. Moreover, the defendant’s hood appears to us to infringe these claims, because, while it is true that Dwight’s disclosure was of an open chamber connected with a flue, the defendant’s flame hood is still such a “chamber,” though it has no flue. Perhaps it is a better device, but it is nevertheless the device originally disclosed and claimed in season. Indeed, the four claims introduced on February 4, 191-0, would themselves have been enough. In claim 10 as issued the phrase, “through the material * * * during * * * the ignition,” cannot mean that the down draft draws the flame through the material, because the specifications disclose nothing of the sort, and the drawings forbid the conclusion. The claim might perhaps mean that the draft passes through the material in an earlier pallet, while a later pallet is being ignited. We need not, however, deeide how far that is possible, because the claims are invalid for another reason.

*831 It is now settled law, in spite of any doubts that may have been east upon it by certain observations in Chapman v. Wintroath, 252 U. S. 126, 40 S. Ct. 234, 64 L. Ed. 491, that a divisional application is no more than an amendment in the parent application, at least unless “new matter” be introduced in the specification, which, whatever latitude it gives, was not the case here. Webster v. Splitdorf, 264 U. S. 463, 44 S. Ct. 342, 68 L. Ed. 792; American Co. v. Prosperity Co., 295 F. 819 (C. C. A. 2); Westinghouse v. Jeffrey, 22 F.(2d) 277 (C. C. A. 2); Wagenhorst v. Hydraulic Co. (C. C. A. 6), 27 F.(2d) 27. Obviously this must be true, if the division is to take its date from the original, as it does; it could not be a proliferation of that original, unless the latter already adequately disclosed it. Until divided, the parent must therefore be understood to contain all those inventions,' no matter how many are later divided out of it. At least, until then its fate is theirs; it carries them in its bosom.

In the case at bar the original application was allowed on March 14, 1908, before any of the “ignition” patents in suit had been divided out. It was renewed on February 17, 1910, 13 days after patent 1,433,350 had on February 4th been divided out. The division was itself a renewal of the application, so far as concerns the inventions which it contained, and, as it was within 2 years after allowance, it was seasonable. The divided application was thereupon itself allowed on May 26, 1911, but was permitted to lapse, supposedly because of failure to pay the statutory fee. It was renewed on April 26,1912, 11 months after the allowance of May 26, 1911, but over 4 years after the allowance of March 14,1908.

If the division is -to be treated as a step in the one continuous application, then the case is directly within our ruling in Weston Electrical Co. v. Empire Electrical Co. (C. C. A.) 136 F. 599. There we held that the phrase in Revised Statutes, §' 4897 (35 USCA § 38; Comp. St. § 9443), “after the allowance of the original application,” meant its first allowance, and that an applicant might not indefinitely withhold his patent in terrorem, by tacking successive periods of 2 years at his pleasure, meanwhile making such amendments as would keep it alive. Considering, as we do, the division as not a new application and not the “original application,” that rule applies equally to this situation. We can see no conceivable reason why a division should, as ¡regards this matter, be treated in another way than any other amendment of which it is only an instance. We therefore think that the learned District Judge was right in dismissing the bill.

The Defendant’s Hood Patent.

The relevant part of this patent consists in the flame-confining hood, which is set over the pan when the charge is to be fired. This is disclosed in the specifications as “fitting closely” upon the flanged rim of the pan, so as to make “a seal or air and gas tight joint.” There are slits in the top of the hood, through which air is drawn during ignition so as to carry the flame into the ore body. The fuel is introduced by nozzles, which from the side spray it into the interior of the hood, where it mixes with the air and fills the whole interior with flame. For the purposes of this case, no more need be said of the plaintiff’s hood than that it corresponds generally with that of the patent, except that it does not fit closely upon the pan. This was held enough to avoid infringement by the Circuit Court of Appeals for the Ninth Circuit in Greenawalt v. American S. & R. Co., 10 F.(2d) 98, and by the learned District Judge in this suit, although verbally it imposes on the 'claims in suit a new limitation. The considerations which were thought to require this arise from what developed upon the interference proceedings between this patent and the four “ignition” patents.

Thirty claims were at issue in those proceedings, drawn from the claims proposed by both sides upon their applications, of which the defendant had 2 pending and Dwight 7. It would be impossible and bewildering to set out all these, with their combinations of elements, infected, like most proceedings in the Office, with the belief that safety lies only in logistic perfection. The main issue determined was of the priority between the inventors as respects ignition apparatus, and in this Dwight was generally successful. 'Indeed, the Examiner in a long opinion awarded all the claims to him, including those in which the hood fitted tightly on the pan. The Examiners in Chief affirmed the general finding of priority, but distinguished between a tight hood and one which was not. Finding that Dwight had not invented the first, they gave to Greenawalt all claims which they understood to cover that feature. A second appeal then went to the Commissioner, who affirmed the Examiners in Chief, except as to two immaterial claims, and finally the Court of Appeals affirmed the Commissioner, one justice dissenting, who agreed with the original Examiner.

*832 All this took nearly 7 years, and the question is whether anything was really settled. First, it is said that the Commissioner in his finding distinguished between a “closed” hood and a tightly fitting one. His language at times would not be wholly clear, were it not for his own definition of terms at the close of his opinion, where he says that the element, whieh for convenience he calls “f,” was a hood “closed (air tight).” There is no ambiguity in his language, when the word “closed” is so understood; he meant what the Examiners in Chief meant, and the Court of Appeals had no choice but so to understand him.

Thus it appears that the last three tribunals, whieh awarded to the defendant his claims, did so by reading them strictly on his specifications, whieh, indeed, suggested nothing else. We are asked in the face of all this, as well as of the five other judges who have passed upon the question, to give the claims another meaning. Quite aside from any question of law, we should refuse to do so, except in the plainest case; our diffidence ought to go so far, and we go further. After the claims have received an interpretation on whieh aldne they secured any allowance at all, we conceive ourselves bound by the meaning so put upon them. Anything else makes an absurdity of the proceeding itself. The defendant canceled the other claims, as he had to, and any express limitation whieh he was compelled to make would, of course, be conclusive upon him. Sutter v. Robinson, 119 U. S. 530, 541, 7 S. Ct. 376, 30 L. Ed. 492. His only relief would be a suit under Revised Statutes, § 4915 (35 ÚSCA § 63; Comp. St. § 9460). There can be no tenable distinction between limitations which are express and those whieh were presupposed by the authorities which allowed the claims and were the basis of their action.

But it is argued that this cannot be, since 2 of the very claims now in suit — i. e., Nos. 48 and,49 — were drawn by Dwight upon his own disclosure, whieh did not show an airtight hood. Whatever the meaning the claims had when Dwight used them, once they were authoritatively interpreted upon their allowance, the applicant, who accepted them, took them cum onere. Had that interpretation extended them beyond their proper import when properly construed, it might, indeed, be necessary to restrict them. The public cannot be charged with notice of a broad interpretation, drawn from matter dehors the specifications, and to put it to an examination of what took place in such an interference as this would be intolerable. No matter what the claims meant when Dwight drew them, in Greenawalt’s application they may not be extended beyond their reasonable meaning in that context. But this doctrine need not work conversely; Greenawalt’s monopoly may be limited, and the public demesne enlarged, by an interpretation whieh contracts- his claims to less than they might cover, if the specifications stood bare. And so we think they must be.

The same limitations apply to claims 13 and 21, which were added after the interference and a personal interview with the Examiner. The scope of the defendant’s invention had then been authoritatively determined by his superiors, whom he had no power to overrule. We do not suggest that this was his purpose; but, whatever it was, it was irrelevant. The claims were necessarily limited to the air-tight hood, because that was the only invention that had been awarded. Perhaps this was wrong; with that we have nothing to do. It is enough that Greenawalt may not now press substantially the same language as that already construed, to cover what was disallowed.'

There remains only the question of the American Ore Reclamation Company. This company intervened, and, while calling itself a defendant, was in substance a eoplaintiff. Greenawalt answered, and made the company a party defendant to his counterclaim. The only matter of consequence here is whether the company is estopped to plead noninfidngement of Greenawalt’s patent. For this Greenawalt relies on certain pleadings in a suit between the plaintiff and the ore company, and upon an agreement between himself and that company. The pleadings may be admissions, but they are nothing more; therefore we ignore them. The agreement gave the ore company the right to license certain other persons under 4 of Greenawalt’s issued patents, and “any further inventions in the same class which may be made or acquired by Greenawalt during the life of his aforesaid patents”; and the invention of the patent in suit was made within that period.

We fail to see how, if this agreement is to be held to be a license at all, it estops the ore company, or how, though licenses may be material on the question of validity, they can be upon infringement. When the patentee sues upon the license, the licensee may litigate the question whether what he has done eomes within the claims. Pressed, etc., Co. v. Union Pac. R. R. Co., 270 F. 518, 524 (C. C. A. 2). A fortiori, when the patentee sues, not upon the license, but repudiating it, and asserting that the defendant is merely an in *833 fringer. Chadeloid Chemical Co. v. McAdam, 298 F. 713 (C. C. A. 2). In neither event is the licensee estopped on the issue of infringement. On the other hand, if the agreement was not a license, it is immaterial for any purpose.

We agree with the learned District Judge on the disposition of the cross-bill.

Decree affirmed; no costs.