STONE, Circuit Judge.

April 22, 1926, appellant filed its complaint charging infringement of patents: Steimer No.1,564,909 and Peiler No. 1,-573,742. The patents covered glass feeding machines. The accused device was a glass feeding machine of the so-called mechanical “plunger” type. A decree adjudging infringement of certain claims of these patents, enjoining infringement and ordering an accounting was entered December 17, 1928. That decree was affirmed in this court. Hartford-Empire Co. v. Obear-Nester Glass Co., 8 Cir., 39 F.2d 769. The report of the master on the accounting was filed August 31, 1935. Certain exceptions were sustained and re-reference ordered November 7, 1936°.

January 22, 1935, appellant filed a petition “For Supplemental Injunction or For Leave to File Supplemental Bill of Complaint.” April 8, 1936, an order was entered denying this petition because “not timely filed and presented to the Court.” From that order this appeal is brought.

Appellee raises two preliminary matters, either, of which, if well founded, would prevent consideration of the merits, of the objections to the order from which this appeal is taken. These contentions of appellee are not trivial but we think it unnecessary to discuss them as our determination upon the merits favors appellee. We will dispose of the appeal upon the merits of the order from which the appeal is taken.

The petition contemplates alternative matters. Primarily, a supplemental in june *416 tiou in the original case filed on April 22, 1926. If that relief be denied, leave to file a supplemental bill in the above case. These matters require separate consideration.

I. Supplemental Injunction.

First as to the rule of law governing supplemental injunctions in patent infringement suits where the purpose is to enjoin use of something differing from the accused device found to have been an infringement.

Not infrequently it occurs that, after a bill charging infringement of a patent has been filed, the defendant changes to a thing different from that which is accused in the bill. If the plaintiff desires protection from the new thing, as an infringement, he has a choice of remedies or methods depending upon the stage of the case, the situation in the litigation, and the difference between the accused and the new thing. If the new thing is discovered before trial, it can be brought in by amendment (Panoualias v. National Equipment Co., 2 Cir., 269 F. 630, certiorari denied 256 U.S. 691, 41 S.Ct. 534, 65 L.Ed. 1174) or by supplemental bill (Equity Rule 34, 28 U.S.C.A. following section 723) or it may be the subject of a separate action. If the discovery is made after interlocutory decree determining infringement by the accused thing and granting injunction and accounting, the new thing may be brought into the action by attachment for contempt of the injunction decree, by motion for supplemental injunction, by introduction into the accounting proceeding before the master or by supplemental bill; or there may be a new action. If the discovery is made after final decree, the new thing may be brought in by attachment for contempt, by motion for supplemental injunction or, possibly, by supplemental bill; or there may be a new action. The above enumeration of possible remedies is merely a statement of such as have been entertained by various federal courts. 1 It is not an acceptance of all of them as the procedural *417 policy of this court. Naturally, the stage of the litigation exerts some influence upon the propriety of use of certain of the above methods. For instance — amendment of the bill would vanish after decree. There are, however, two principles concerning which the courts seem in agreement: First, that the plaintiff may always, as a matter of right, bring a new action upon the new infringement; second, that use of any of the other remedies is subject to the judicial discretion of the court. Both of the alternative remedies here sought are such as fall within the second class.

The “discretion” here intended is a judicial discretion governed by the situation and circumstances affecting the exercise thereof. Even where an appellate court has power to review the exercise of such discretion, the inquiry is confined to whether such situation and circumstances clearly show an abuse of discretion, that is, arbitrary action not justifiable in view of such situation and circumstances.

We now consider the situation and circumstances bearing upon the exercise of discretion of the trial court in denying a supplemental injunction. Naturally, every investigation of this sort is governed by the situation and circumstances of the particular case. However, it sometimes happens that the same or closely similar situations and circumstances have reoccurred in enough cases to result in a measure or rule of decision. An instance of this kind exists in situations of proceeding against a new device where an infringement injunction is in force. That rule is that where the difference between the new and the accused device appears to be substantial or doubtful — not merely colorable — the plaintiff will be denied relief by attachment for contempt and relegated to other remedies. This rule is here applicable because a petition for supplemental injunction, while not the same, is little more than a milder proceeding intended to reach a situation where contempt might be proper, that is, of purely colorable differences. John W. Gottschalk Mfg. Co. v. Springfield Wire & Tinsel Co., 1 Cir., 90 F.2d 468, 472; Singer Mfg. Co. v. Seinfeld, 2 Cir., 89 F.2d 35, 37; Better Packages v. L. Link & Co., 2 Cir., 68 F.2d 904, 905; Wachs v. Balsam, 2 Cir., 38 F.2d 50, 52; Armstrong v. De Forest Radio T. & T. Co., 2 Cir., 10 F.2d 727, 728, 729, certiorari denied 270 U. S. 663, 46 S.Ct. 471, 70 L.Ed. 787; Charles Green Co. v. Henry P. Adams Co., 2 Cir., 247 F. 485, 486; National Metal Molding Co. v. Tubular Woven Fabric Co., 1 Cir., 239 F. 907, 908; Crown Cork & Seal Co. v. American Cork Specialty Co., 2 Cir., 211 F. 650, 653. Both in a petition for a supplemental injunction and in an attachment for contempt, the relief depends upon a construction of the existing injunction order (as to contempt, Flat Slab Patents Co. v. Turner, 8 Cir., 285 F. 257, 272, certiorari denied 262 U.S. 752, 43 S.Ct. 700, 67 L.Ed. 1215, and as to supplemental injunction, Armstrong v. De Forest Radio T. & T. Co., 2 Cir., 10 F.2d 727, 728, certiorari denied 270 U.S. 663, 46 S.Ct. 471, 70 L.Ed. 787). The inquiry in each is whether the new thing is within the injunction. The difference between the two proceedings lies in the relief. In petition for supplemental injunction, the relief is an adjudication that the new thing is within the injunction (resulting in a liability to contempt for future violations) and, possibly, inclusion in the accounting. In contempt proceeding, the relief is punishment for past violation of the original injunction with attendant damages. This difference in relief should not affect the application of the above rule.

With this rule in mind, we consider the situation and circumstances here to ascertain whether the new devices — designated as M-l and M-2 — are more than color-ably different from the accused device found to he an infringement. This comparison must, of course, be made in the light of the patent involved as construed by this court. Since the comparison is so to be made, it is necessary to examine the patent and the construction thereof by this court. Such examination is contemplated for the sole purpose of furnishing the background for a comparison of the new devices (M-l and M-2) with the accused device in order to determine whether the differences are only colorable or are substantial enough to arouse doubt of similarity.

The patents involved in this infringe *418 ment suit were Steimer, No. 1,564,909, and Peiler, No. 1,573,742. Both patents were in the art of mechanical devices to feed molten glass from furnaces into parison molds to form articles, such as bottles or jars. In another case (Hartford-Empire Co. v. Obear-Nester Glass Co., 8 Cir., 71 F.2d 539, 544-548, certiorari denied 293 U.S. 625, 55 S.Ct. 345, 79 L.Ed. 712), which involved the same devices sought to be brought in by this application for supplemental injunction but different patents from those here, though in the same art, this court endeavored to define that “art” and to set forth, generally, some of the stages of its development, including some of the types of apparatus. 2 The Steimer patent is for an apparatus, of the mechanical plunger *419 type. The Peiler patent was designed to combine in one patent the generic principles of. four earlier separate applications. It contained both method and apparatus claims —the latter being expressly related to the specifications as to forms of apparatus of the “paddle — [mechanical] plunger” (or needle) type and of the mechanical “plunger” (needle) type. The other types, “shown * * * for the purpose of illustrating the generic character of certain claims made herein” (patent p. 11, 1.1.107-110) were of *420 the “paddle” type and the “punty” type. All of the above types involved purely mechanical means to affect gravity flow in the determination of the weight (amount of glass) and the form of the gob fed into the mold.

On appeal from the interlocutory decree in this case (Hartford-Empire Co. v. Obear-Nester Glass Co., 8 Cir., 39 F.2d 769), this court determined that the ten method claims involved were invalid as covering only func *421 tions — “the problem to be solved.” 39 F.2d 769, at page 774. We sustained various of the apparatus claims in each patent, We sustained these claims on the basis of “combination” patents, each of which was an improvement over the existing art. 39 F. *422 2d 769, at page 774. 3 In the later case above, 8 Cir., 71 F.2d 539, at pages 546-548, 566, 567, we developed the somewhat crowded state of the art.

The accused device (found to infringe) in this case was a mechanical “plunger” type.

The essence of what has been said above is that the interlocutory decree herein enjoined a mechanical plunger type of apparatus as infringing the mechanical plunger apparatus claims of each of two combination patents which came into a somewhat crowded art.

With this background we take up a comparison of the new devices with the infringing device. Both of the new devices are feeders wherein no mechanical apparatus- — such as plungers, paddles or pun-ties — are present. In these two devices, the weight and shape of the gob fed to the mold is brought about by the exercise of air pressures upon the flowing glass. This pneumatic type of feed was the kind involved in the later case here. 71 F.2d 539. In the course of describing the art, we said (in that case, 71 F.2d 539, at page 548) : “The automatic means for exercising force to affect gravity in the glass movement are of two kinds: Mechanical and pneumatic. The character of each of these has affected the methods and apparatus for control. For example, Owens and Severin principles are based upon pneumatic force, and would be impossible mechanically, while Peiler’s paddle-dam feed method is just as strongly based upon mechanical means with impossibility of pneumatic accomplishment. On the other hand, certain feeding principles are susceptible either to mechanical or to pneumatic control, and some principles (such as the suspended gob feed) may employ either at the same stages and for the same purposes.”

From what has been said in the two paragraphs just preceding, it ought to be clear that the differences between the two new devices and the infringing device are not merely colorable but are such as arouse a decided doubt as to the former being within the scope of the injunction order in the interlocutory decree. We do not intimate any view as' to whether the new devices infringe the Steimer and Peiler patents involved in this main action. What we do decide is that these new devices are so different from the infringing device that infringement as to them should not be determined on a petition for supplemental injunction'but that plaintiff should be relegated to a new action or to a supplemental bill where a full hearing may be had on issues framed in pleadings. 4 It would have been an abuse of discretion to have com *423 pelled appellee to litigate this matter on a petition for supplemental injunction because such a remedy is not applicable in the situation and under the circumstances here. American Foundry & Mfg. Co. v. Josam Mfg. Co., 8 Cir., 79 F.2d 116, 118. In Charles Green Co. v. Henry P. Adams Co., 2 Cir., 247 F. 485, 486, 487, the court said: “If an asserted new infringement does not plainly render a new action and another trial an expensive futility, no supplementary injunction should issue. * * * If testimony and cross-examination are required to justify the issuing of supplementary injunction, there is small practical difference between such procedure and the institution of a new suit.”

II. Supplemental Bill of Complaint.

The alternate matter sought by appellant was leave to file a “Supplemental Bill of Complaint.” The purpose of the proposed supplemental bill was to bring into the litigation two new constructions, known in this record as M-l and M-2, which had been made and used subsequent to filing of the original bill and which were claimed to infringe Peiler No. 1,573,742. Leave was denied by the trial court for the reason that “said petition [for leave] was not timely filed and presented to the Court.”

Filing of supplemental pleadings (including supplemental bills of complaint) is governed by Equity Rule 34, 28 U.S.C.A. following section 723. The here pertinent provisions of the rule are: “Upon application of either party the court or judge, may, upon * * * such terms as are just, permit him to file and serve a supplemental pleading, alleging material facts occurring after his former pleading, or of which he was ignorant when it was made.” The Supreme Court has defined the function of a supplemental bill under rule 34 to be “to introduce matters occurring after the filing of the original bill, or not then known to the plaintifif.” General Inv. Co. v. Lake Shore Railway, 260 U.S. 261, 281, 43 S.Ct. 106, 115, 67 L.Ed. 244. Also, that court has declared that “An application for leave to file a supplemental bill is addressed to the discretion of the court, and the ruling thereon will not be disturbed on appeal unless *424 the discretion has been abused” (same case, 260 U.S. 261, at page 281, 43 S.Ct. 106, 115, 67 L.Ed. 244; and for other judicial expressions and applications see Selden Co. v. General Chemical Co., 3 Cir., 73 F.2d 195, 196; Continental Oil Co. v. Osage Oil & Refining Co., 10 Cir., 57 F.2d 527, 529, certiorari denied 287 U.S. 616, 53 S.Ct. 17, 77 L.Ed. 535; General Electric Co. v. Alexander, 2 Cir., 280 F. 852, 855; Rosemary Mfg. Co. v. Halifax Cotton Mills, 4 Cir., 266 F. 363; Berliner Gramaphone Co. v. Seaman, 4 Cir., 113 F. 750, 754; Sheffield & B., etc., Co. v. Newman, 5 Cir., 77 F. 787, 791; Hicklin v. Marco, 9 Cir., 56 F. 549, 552). Under the above statement of the applicable law, the issue here is whether there was an abuse in the exercise of discretion by the trial court. This issue is not whether this court, had it been the trial court, would have exercised this discretion differently, but it is whether the action of the trial court was so clearly wrong that its decision cannot be justified by the situation and circumstances which must, as matter of law, control the exercise of this discretion.

This situation and these circumstances are as follows.

April 22, 1926, appellant 5 filed a suit for infringement of patents, Steimer No. 1,-578,742 and Peiler No. 1,573,742, covering machines for automatically feeding molten glass to parison molds. The nature of the patented apparatus and of the accused device are set forth earlier in this opinion.

January 10, 1928 (six days before this cause was for trial), counsel for appellee (defendant) wrote a letter to counsel for appellant wherein it was set forth that the accused device had been discontinued about four months after the complaint was filed and appellee had no intention of using it further. Also that the accused device had been replaced by two other constructions of which blueprints were inclosed. These new constructions were M-l and M-2. The two constructions differed both from the apparatus covered by the two patents and from the accused device in that these two were air feeders while all of them were mechanical plunger, needle or paddle types.

The trial began January 16, 1928. During the trial (on January 18, 1928) and, apparently, near the beginning of appellee’s evidence, appellee sought to introduce the blueprints showing the M-l and M-2 construction. An objection 6 that they were outside the issues of the case, which were confined to the accused device, was sustained.

December 17, 1928, decree was entered holding Steimer patent valid and infringed, certain claims of Peiler patent invalid, other claims valid but not infringed^and still others valid and infringed. An injunction and accounting were ordered. Cross-appeals were taken from this decree.

February 26, 1929, appellant filed a separate suit charging infringement by M-2 of certain claims of three Peiler patents, Nos. *425 1,405,936, 1,662,436 and 1,662,437 and one claim of Ferngren No. 1,677,436 — the Peiler and Stei'mer patents in the original suit not being included.

February 24, 1930, this court affirmed the interlocutory decree of December 17, 1928.

During the accounting before a master, appellant attempted to include M-l and M-2 in the accounting. Appellee resisted this inclusion as not within the decree and was sustained by an order of December 20, 1930, instructing the master to exclude M-l and M-2 from the accounting.

February 19, 1931, appellant applied to this court for leave to file a petition for mandamus and prohibition requiring vacation of the order of instruction.

February 24, 1931, a decree was entered in the second suit (filed February 26, 1929) finding the claim of the Ferngren patent and seven claims of Peiler No. 1,-662,436 invalid, certain other claims of the Peiler patents valid but not infringed and dismissing the bill, D.C., 51 F.2d 85, 98.

March 26, 1931, this leave was denied “without prejudice to any rights the petitioner may be entitled to assert in this or any future litigation.”

May 25, 1931, appellant filed petition for rehearing of the order instructing the master.

June 1, 1931, this petition was denied.

June 25, 1931, appellant filed a petition, in this court, for a rehearing of its order denying leave to file a petition for mandamus and prohibition.

October 28, 1931, this petition was denied.

June 1, 1934, this court determined an appeal (by this appellant) in the second suit (filed February 26, 1929). That decision, 8 Cir., 71 F.2d 539, found some claims of the Peiler patents there involved invalid and some others not infringed and that the claim of the Ferngren patent involved was invalid.

July 30, 1934, rehearing was denied.

January 7, 1935, certiorari was denied by the Supreme Court. 293 U.S. 625, 55 S.Ct. 345, 79 L.Ed. 712.

January 22, 1935, this petition was filed. At that time, the testimony before the master on the accounting had been closed and appellant had filed its brief with the master.

The pertinent things ,to note in the above chronological outline of the litigation between these parties are the following.

The original bill herein was for infringement, by a mechanical plunger type of feeder, of Steimer and of Peiler No. 1,573,742. A few days before trial, appellant (plaintiff) was formally notified by appellee (defendant) that use of the accused mechanical plunger device had been permanently abandoned some months before and that defendant had substituted therefor a pneumatic type of feed which it was using and construction of which was shown by two attached blueprints.

In this situation, the plaintiff had a choice of several courses of action. It could bring a new action for infringement of the same two patents by the new devices. It could proceed upon the issues as then framed (including only the then accused mechanical plunger) to an interlocutory decree or it could bring the two new devices into that case — possibly, by amendment to the bill (Panoualias v. National Equipment Co., 2 Cir., 269 F. 630, 632) or, certainly, by a supplemental bill (Equity Rule 34). If it desired not to bring the new devices into that litigation, it had the further right, after securing an interlocutory decree, to include the new devices in the accounting ordered by the interlocutory decree (Flat Slab Patents Co. v. Turner, 8 Cir., 285 F. 257) or/and to protect its injunctive relief by an attachment for contempt or (if such proceeding is to be recognized) by a supplemental injunction — but the exercise of such rights would depend upon the new devices being merely “colorable” departures from the accused device (Flat Slab Patents Co. v. Turner, 8 Cir., 285 F. 257).

Plaintiff elected neither to bring the new devices into the litigation nor to file a new action. It did more by preventing the introduction of any evidence as to the new devices at the trial for the interlocutory decree.

Eleven months later, plaintiff secured an interlocutory decree adjudging infringement with injunction and accounting. From this decree, both parties-appealed — plaintiff, because certain claims of Peiler No. 1,573,-742 had been decreed invalid therein.

Two months after the above decree, plaintiff brought a new action charging infringement, by the two new devices, of four patents not including either of the two patents in the suit under decree.

*426 A year thereafter, this court affirmed the interlocutory decree.

About ten months thereafter, plaintiff sought, unsuccessfully, to have .the accounting (under the affirmed decree) include the two new devices. No move to apply the injunction to the new devices was made, although these devices were within the injunction if they were within the accounting.

About two months thereafter, a decree was entered dismissing the bill in the suit against the two new devices.

By eight months thereafter, plaintiff knew (in the mandamus-prohibition proceeding) this court would not interfere with the order ,of the trial court excluding the two new devices from the accounting.

During the next two years and eight months, the accounting went forward and was closed in the first suit and, also, the appeal proceedings in the second suit (until denial of certiorari by the Supreme Court on January 7, 1935).

On January 22, 1935, two weeks later, this petition was filed for leave to file a supplemental bill bringing the two new devices into the first suit. That suit had been tried seven years before resulting in an interlocutory decree more than six years before, which decree had been affirmed by this court almost five years before.

A matter of some importance is that while the new devices were in the same art — automatic feeding of glass into parison molds — as the patents in the first suit and the device accused therein, yet they were in different lines of development (within the art) which “affected the methods and apparatus for control” (71 F.2d 539, 548, 553, note 6), of the glass being fed. Another consideration is that this was a crowded art. 71 F.2d 539, 544-548, 566. Yet another matter is that, because of the state of the art and because of its development along different lines or principles, the relation of the new devices to the patents in the first suit and to the accused device therein was such as, naturally, to involve different comparisons with the art and with other patents. 7

Our inquiry thus comes down to this: Did the trial court abuse its discretion in the above outlined situation and circumstances, in denying leave to file a supplemental bill and thus relegating appellant to a new suit?

We think not. Appellant had had knowledge of the new devices and had a clear opportunity to bring them into this suit before trial on the interlocutory decision. It preferred first to move against them by a different suit based on other patents and by an attempt to bring them within the accounting. It never sought the present relief until two years and eight months after it knew it could not get them into the accounting and until it was denied certiorari by the Supreme Court from our affirmance of an adverse decree, entered by the trial court in the second suit almost four years before.

The differences between the two new devices and the accused device (first suit) are such that, obviously, a determination of infringement by the new devices will require different lines of evidence and a full hearing. It is entirely doubtful as to how far, or whether at all, the line of investigation of validity and, particularly, of infringement by the accused (mechanical plunger) device would be helpful in determining the same issues as to the new devices. Yet, after actively keeping the new devices out of the trial on the interlocutory decree, appellant now seeks to get the maximum of benefit from that decree by bringing these devices into the same suit long after entry of such decree and after trying other attacks upon them.

We cannot think the trial .court was so wholly wrong in its action as to amount to an abuse of discretion. We think the long delay, under the circumstances here, would be sufficient to prevent our reversing the trial court. Without such delay, had the trial court based its action upon the differ *427 ence in the devices (in view of the art) being such as to make a trial by new action rather than supplemental hill more just to the parties, we could not declare such action an abuse of discretion. In determining this appeal we expressly disavow any views which would affect the bringing of a new action or the issües which may be framed therein — such matters must be determined if and when they arise.

The order of the court is, with costs of this appeal against appellant, affirmed.

Notes

(a) The remedies allowable are not the same iu all jurisdictions. For example — after decree, the Second Circuit approves" the method of supplemental injunction or supplemental bill as preferable to attachment for contempt in most situations (Gordon v. Turco-Halvah Co., 247 F. 487, 490; Charles Green Co. v. Henry P. Adams Co., 247 F. 485, 486, 487) while the Third Circuit recognizes no such practice (Minerals Separation v. Miami Copper Co., 269 F. 265, 268).

(b) The following are some of the cases in the Supreme Court and the Courts of Appeals which have set forth the views as to the use of various of these remedies. This list is not intended to be complete but rather to be illustrative. There are many more cases in the District Courts (see lists of cases U.S.C.A., title 35, § 69, p. 487, note 55). Leman v. Krentler-Arnold Hinge Last Co., 284 U.S. 448, 52 S.Ct. 238, 76 L.Ed. 389, reversing Krentler-Arnold Hinge Last Co. v. Leman, 1 Cir., 50 F.2d 699, 707; General Inv. Co. v. Lake Shore Ry., 260 U.S. 261, 43 S.Ct. 106, 67 L.Ed. 244; Union Tool Co. v. Wilson, 259 U.S. 107, 42 S.Ct. 427, 66 L.Ed. 848, Worden v. Searls, 121 U.S. 14, 7 S.Ct. 814, 30 L.Ed. 853; Blanchard v. Putnam, 8 Wall. 420, 19 L.Ed. 433; John W. Gottschalk Mfg. Co. v. Springfield Wire & Tinsel Co., 1 Cir., 90 F.2d 468; Singer Mfg. Co. v. Seinfeld, 2 Cir., 89 F.2d 35; American Foundry & Mfg. Co. v. Josam Mfg. Co., 8 Cir., 79 F.2d 116; Selden Co. v. General Chemical Co., 3 Cir., 73 F.2d 195; Wadsworth Elec. Mfg. Co. v. Westinghouse Elec. & Mfg. Co., 6 Cir., 71 F. 2d 850, certiorari denied 293 U.S. 604, 55 S.Ct. 122, 79 L.Ed. 696, and Wadsworth Electric Mfg. Co. v. Sachs, 294 U.S. 724, 55 S.Ct. 552, 79 L.Ed. 1255; Better Packages, Inc., v. L. Link & Co., 2 Cir., 68 F.2d 904; Wachs v. Balsam, 2 Cir., 38 F.2d 50; Denaro v. McLaren Consol. Cone Corp., 1 Cir., 23 F.2d 384; Monroe Body Co. v. Herzog, 6 Cir., 13 F.2d 705; Field Body Corp. v. Highland Body Mfg. Co., 6 Cir., 13 F.2d 626; Armstrong v. De Forest Radio T. & T. Co., 2 Cir., 10 F.2d 727, certiorari denied 270 U.S. 663, 46 S.Ct. 471, 70 L.Ed. 787; United States Rubber Co. v. I. T. S. Rubber Co., 7 Cir., 288 F. 786, certiorari denied 262 U.S. 748, 43 S.Ct. 523, 67 L.Ed. 1213; Schey v. Giovanna, 2 Cir., 273 F. 515; Minerals Separation v. Miami Copper Co., 3 Cir., 269 F. 265; Eureka Tool Co. v. Wire Rope Appliance Co., 8 Cir., 265 F. 673; Individual Drinking Cup Co. v. Public Service Cup Co., 2 Cir., 250 F. 620; Gordon v. Turco-Halvah Co., 2 Cir., 247 F. 487; Charles Green Co. v. Henry P. Adams Co., 2 Cir., 247 F. 485; Frank F. Smith Metal Window Hardware Co. v. Tates, 2 Cir., 244 F. 793; Sundh Elec. Co. v. Cutler-Hammer Mfg. Co., 2 Cir., 244 F. 163, certiorari denied 245 U. S. 661, 38 S.Ct. 61, 62 L.Ed. 536; National Metal Molding Co. v. Tubular Woven Fabric Co., 1 Cir., 239 F. 907; Crown Cork & Seal Co. v. American Cork Specialty Co., 2 Cir., 211 F. 650; L. S. Starrett Co. v. Brown & Sharpe *417 Mfg. Co., 1 Cir., 208 F. 887; Rajah Auto Supply Co. v. Grossman, 2 Cir., 207 F. 84; D’Arcy v. Staples & Hanford Co., 6 Cir., 161 F. 733; Bullock Elec. & Mfg. Co. v. Westinghouse Elec. & Mfg. Co., 6 Cir., 129 F. 105, certiorari, denied 194 U.S. 636, 24 S.Ct. 859, 48 L.Ed. 1160; Cary Mfg. Co. v. Acme Flexible Clasp Co., 2 Cir., 108 F. 873, writ of error dismissed 187 U.S. 427, 23. S.Ct. 211, 47 L.Ed: 244.

This portion of that opinion is as follows:

“It has been well said that: ‘It is necessary for one who wishes to grasp successfully the principles underlying the process for the production of glass on a commercial scale to possess some knowledge of the sciences upon which glass technology is founded. The three most important are chemistry, physics, and engineering.’ ' A Text-Book of Glass Technology by Hodkin and Oouscn, Constable & Co., London, 1925. Our present purposes do not require an extended incursion into the very large field of the commercial production of all kinds of glass, but it is necessary briefly to scan this field in order properly to understand the art of our immediate concern and to appreciate the contributions of various cited devices to that art.

“Two matters should be noted at the beginning. The first is that the problems of chemistry, physics, and engineering — all of which are involved in any process of glass making — arise partly from the character (size, shape, and uses) of the article to be produced and from the conditions governing such production. The second is that there is a like sequence of general operations or steps in all processes. The first of these implies dissimilarity while the latter implies similarity. The natural result of the combined influences is that there are differences within the elements or steps of the sequence without disturbing the sequence. Broadly speaking, this sequence in mechanical (also called automatic) processes consists of four general operations: (1) Selection of the proper chemical substances, mixing thereof in correct proportions, and putting the mixture in a fusing receptacle; (2) fusing of the mixture into an harmonious fluid mass and conditioning (mainly ‘fining’) for working; (3) transfer (called ‘feeding’) from the fused mass to the forming apparatus, of a portion suitable for the particular working; and (4) ‘working’ into the desired article. In each of these steps an ‘art’ exists. It is within each of these steps that there occur those divergences .or differences caused by the character (size, shape, and use) of the article to be produced, or worked, and by the conditions governing such production, or working.

“Our interest is in the third step in the sequence — that of ‘feeding.’ But our particular concern is more concentrated than upon the art of feeding glass for making glass articles generally. The patents here in suit have to do with feeding glass to be blown in molds to make glass articles — mainly bottles and narrow mouthed ware. The size and shape of such articles and the peculiar conditions (blowing in molds) governing their manufacture cause a divergence or difference from feeding methods for many other glass articles. For example, a considerable quantity of glass may be taken from a fusing furnace by ladle, pot, or pouring and placed on a largo forming table when plate glass is to be rolled; a window glass cylinder may be blown from a fair-sized open pot into which the molten glass has been ladled or has flowed from the melting furnace; for large ware (formed by pressing in molds) the desired quantity of glass for a piece of ware may be taken in various ways from the furnace and placed in the wide open mold; while for bottles or similar ware blown in small-mouthed molds a particularity of quantity and shape of feed charge is required which is not of interest in the making- of many other articles. These differences in requirements in the feed produce differences in problems of feeding which have led to distinct branches of art within the general art of feeding glass. Not only this, but wide differences in principle of methods for feeding to molds for this kind of blown glass articles have revealed very different problems and required thought for solution along such different lines that there has arisen a further branching of art within the art of feeding to such molds. In the last analysis our special concern is with a particular principle or kind of feeding to molds for blowing glass bottles and similar ware; that is, the art of ‘suspended gob’ feeding.

“Having indicated the related arts and the particular art with which we have to do, it is necessary now to state certain matters which influence, with some *419 what varying force, all glass feeding and those peculiar to or more potent in feeding to molds for blowing bottles.

“Glass is an amorphous chemical compound created by fusing in a furnace. The chemical constituents and problems of glass, while always very important, are outside our concern, but the compound has certain physical characteristics important here. With some few exceptions (mainly optical lenses) not of consequence here, glass is ‘worked’ into the finished article while in a molten state. Its physical characteristics while in that state naturally influence or control the methods in which it can be worked and fed. The extent of such influence or control varies with the kind of article to be made. For the kind of articles which here concern us, the characteristics we need specially to notice are two. The first is extreme viscosity (increasing rapidly with loss of temperature) with a pronounced tendency to adhere to anything it touches — particularly if the thing be hot. The second is the rapid formation of an enveloping ‘skin’ when exposed to temperatures cooler than itself. While it is probable that all liquids (due to internal molecular attraction) form this surface tension or skin, yet the decided viscosity and the extreme susceptibility of molten glass to temperatures cooler than itself cause the rapid formation of a relatively tough skin of a nature to affect the working and feeding of it at atmospheric or near atmospheric temperatures. These characteristics deeply affect most methods of feeding and working glass into comjiaratively small mold-blown articles, such as bottles.

“Another matter is that ‘feeding’ is an intermediate step in a continuous and fairly rapid sequence beginning with melting of the compound and ending with working into the product. Being the connecting link (in a continuous process) between the molten mass in the furnace and the subsequent working, the engineering problems of methods and means of feeding have inevitably been strongly influenced by the kind of furnace at one end and by the kind of working at the other. For example, it would be impractical, if not impossible, to have any kind of feed from a pot furnace except some form of dripping therefrom— such as by punty, ladle, or suction. Again, working methods are controlled by the character of article made — thus plate glass is rolled, window glass is blown or drawn, flat ware is pressed in molds, bottles and hollow ware are blown in molds. Necessarily, each of these methods of working has vitally conditioned the methods and means of feeding for its particular purpose. While the influence of the furnace has been nowhere near as great as the working methods (as will hereinafter appear), yet both are to be kept in mind, for furnaces and working methods had mechanically been improved and were highly developed before mechanical feeding attracted much serious attention; therefore it was into such a situation that most mechanical feeders were designed to fit. In fact, the purpose of the patents now in suit is to supply, mechanically, the link between developed furnaces and developed working mechanisms.

“In connection with what has just been said, it will be helpful, briefly, to state the development of this situation in so far as bottle forming processes are concerned. Such history will throw light on the feeding needs and requirements to moot which most mechanical feeders were devised and upon some of the conceptions and devices (cited in this record) which arose to supply these needs and requirements. Slaking bottles and similar hollow glass containers by blowing is a very old process. Before mechanical “means for so doing appeared, this was done by men with few and crude tools. This process required the closely consecutive action of several skilled workmen. It was slow, expensive, and produced bottles lacking in uniformity of size, shape, and workmanship. In this hand process, the glass was taken from the furnace (through a furnace hole above the glass level) on a slightly knobbed and hot end of a small metal pipe which was twirled to accumulate and to separate from the furnace mass a portion of glass, called a ‘gather.’ When the amount of glass estimated by the workman as proper to make the bottle had been thus gathered and removed from the furnace, other workmen took the pipe and shaped the bottle by blowing and manipulating the gather. Other workmen broke the bottle from the blowpipe and shaped and strengthened the bottle mouth. , The first development in this age-old process was along the line of tools to lighten the *420 physical strain on the workers or to produce better and more uniform bottles. One of the latter sort was the introduction of molds in which the workman blew the finished bottle to better and more uniform shape and size. A step in this use of molds was for the workman roughly to shape the gather before it was placed in the mold for blowing. This shaping was done by manipulation of the gather on a slanting stone or metal slab called a ‘marver.’ When the gather had thus been shaped for the blowing mold, it was called a ‘parison.’ The next progressive step was to do away with this initial shaping on the marver by introducing a separate initial mold wherein the gather could be directly placed and would quickly form the ‘parison’ deemed desirable for use in the blowing mold and from which the parison was transferred on the pipe by the workman to the blowing mold. This initial mold was called the ‘parison mold.’ This term should be borne in mind for subsequent mechanical development preserved these two molds — parison and finishing (or blowing) and mechanically placing a charge by a feeding device into a mold means always into a parison mold. Shortly afterwards, came the first mechanical improvement through a machine to blow the bottle (in the finishing mold) by compressed air. At this stage, it was possible to dispense with the skilled human blower, and the blowpipe gave place to the ‘punty,’ which was a rod with a particularly shaped knob at the gathering end. But the gather from the furnace, the placing thereof into the parison mold, and the transfer of the parison therefrom to the finishing mold still required manual labor. Following developments resulted in multiplication of pairs of parison and finishing molds into mechanical units with synchronized mechanical transfer of the parison from its mold to the finishing mold. Such machines were called ‘semi-automatic,’ since they still required human agency in the process, but this agency was now reduced to gathering the charge from the furnace and placing it in the parison mold. As these machines were perfected and became more rapid in operation, the defects in this manual feed (waste through inaccuracy in the quantity of the charge, slowness of operation, and expense) were more and more acutely felt. This was the developed situation which called for relief from this manual feed and resulted in much effort to devise mechanical substitutes therefor. Some devices (illustrated by citations in this record) were not confined to feeding apparatus, but covered molding machinery and some even annealing, but the majority of the citations here are for feeding devices intended to meet the above situation and need.

“Another matter (mentioned above in outlining our particular art and also set out in the concluding sentence of footnote) is an important and a somewhat surprising thing which occurred in the efforts to devise mechanical feeding machines for bottle making molds. This is the difference not only in devices but in ‘principles’ of action thereof. Of course, all principles had in view to remedy, mechanically, those defects in the hand gather which made that method ineffective in the economical and rapid commercial production of bottles by machinery. However, the conceptions of the method of doing this differed widely. In part this may be explained by lack of knowledge of some of the problems engendered by substituting machinery for the intelligently directed action of a skilled' workman and in part by lack of knowledge of the reaction of molten glass to mechanical control. There are numerous examples of the former and some of the latter. As to the first, the necessities of an accurately measured or quantity of charge and rapid delivery of charges were generally appreciated, but the necessity of shaping the charge so that it would cleanly enter and accommodate itself to the parison mold was entirely outside the knowledge or endeavor of quite a few of the inventions cited herein. A failure early to recognize this necessity of shaping was not unnatural because shaping was a problem first brought to light in mechanical feeds. In the hand gather, the skilled attention, judgment, and action of the workman had taken care of the proper introduction of the charge into the mold. In quite a few of the patents cited here, the sole design was to deliver accurate quantities of glass to the molds rapidly, and this is true even where the de-vices of some such had, within harrow limits, an entirely unintended shaping effect; that is, the action of the apparatus changed the natural form of the charge to some other which was faix-ly constant. Nat *421 urally, this accidental shaping might or might not be an advantage according to the circumstances — such as mold mouth, bottle size, etc. Another matter not always appreciated was the necessity of timing or synchronizing the delivery of the charge with the movements of the molds into feeding position. An example of ignorance of molten glass reaction under control is shown by Cleveland, No. 901,881, which is a stream feed where the charges are separated by stopping the outflow orifice with the shears between the end of one charge and the beginning of another. Such action would inevitably and speedily result in stopping operations through smearing of the shears and orifice. But, whatever the causes therefor, the important things are the facts that such feeders were devised along different principles, and that such principles, in some instances, developed basic problems of control peculiar to themselves. Therefore, some of these principles developed subsidiary arts of their own because of these differences.

“Recognition of and acquaintance with the differences in these various principles are quite important in estimating the contribution of a given device to the art of bottle machine feeding. This is true because each principle gave birth to problems often peculiar to it; invention was inevitably controlled by these problems; and apparatus must be adapted to solution thereof. Therefore, while a device may be a decided contribution to discovery along its given principle, it may mean little or nothing to an investigator along some other. Although this may be self-evident it may be useful to emphasize and illustrate this now by several examples, as follows: Some form of shearing or definite separation of a charge from the furnace mass would seem a necessity, but Severin provides none, and the cup-shear of Brooke could suggest nothing useful to an inventor working along any principle other than the ‘stream’ feed. Again,- a forehearth, or furnace extension, is a necessity to most principles, but Severin uses none, and neither Owens’ revolving pot nor Hitchcock’s extended passage carry help as to forehearths to any other inventors in this record. Again, the main problems in most principles — to secure a measured or a measured and shaped charge to be transferred into the parison mold — are not encountered by Owens and Severin, who introduce a furnace mass portion directly into the parison mold.

“This difference in principles and resulting methods has, naturally, led to endeavors to classify feeders to bottle blowing molds. This classification has been attempted both outside this record (see ‘A Text-Book of Glass Technology,’ supra, p. 426) and within it. Such classifications are useful, but must be used with caution because they differ as to bases, because some devices do not readily fit into any one of such classes, and because considerable significant subelassincation is demanded if any one of them be adhered to. We will use such only in so far as seems advisable to make our meaning clear.

.“From what has been said it is evident that, broadly speaking, the practical engineering problem involved in all mechanical glass feeding is that of control of the glass for the purposes of the feed; that the control requirements differ according to the article to be made and the methods of making such article. Also it is clear that these requirements as to bottles (which are made by blowing in molds) are accuracy of quantity of charge, shaping of the charge to enter and fit the mold, rapid succession of charges, and synchronizing the time of charging to the movement of the parison molds into charging position. Also that attempts to devise feeds for bottle -machines have been along different lines presenting different problems meaning different solutions and solutions meaning methods (or conceptions) as to the way of solution, and also apparatus to carry such methods into effect. We are now to state (in so far as here helpful) the general considerations and means of control. It is with such that this case is vitally concerned.

“Broadly, control has to do with movement of the glass from the furnace mass into the parison mold. It is obvious that the fluid or semifluid condition of molten glass in the furnace is always affected by gravity, and, therefore automatic transfer (movement) of a charge from the furnace to the parison must reckon with this omnipresent natural force. This automatic control must involve forces which will result in some one or more of several effects: Overcoming, using, accelerating, or retarding gravity. In respect to such effects, the requirements in this movement from furnace to mold may differ or may differ at various stages in the progress of such movement — depending upon the principle of feeding. Thus, the Owens device is concerned only with *422 overcoming gravity; Bridges, No. 1,121,-008, utilizes gravity (twice) and overcomes gravity; Wilzin (Fig. 20) accelerates gravity; McCauley, No. 1,281,083, overcomes and accelerates gravity; Mc-Cauley, No. 1,322,318, ovorcomes, retards, and accelerates gravity; Lott (Fig. 29) utilizes gravity and (Fig. 30) overcomes and utilizes gravity; Miller utilizes and accelerates gravity. The differences in the above examples (there are many others) suggest the wide possibilities of divergence in the use of these four effects and in the sequence of use where more than one is employed.

“The automatic means for exercising force to affect gravity in the glass movement are of two kinds: Mechanical and pneumatic. The character of each of these has affected the methods and apparatus for control. For example, Owens and Severin principles are based upon pneumatic force, and would be impossible mechanically, while Peiler’s paddle-dam feed method is just as strongly based upon mechanical means with impossibility of pneumatic accomplishment. On the other hand, certain feeding principles are susceptible either to mech'anical or to pneumatic control, and some principles (such as the suspended gob feed) may employ either at the same stages and for the same purposes.”

In that case, 39 F.2d 769, at page 774, we said:

“Besides, I think there are other reasons bottomed on the fact that many other patents in the prior art disclose the notion of phase-changing, although they did not appropriate the designation, or attempt to do so. I have no doubt, however, that Peiler improved upon all such.

“So, I think both patents in suit are combination patents. Protracting and retracting mechanical shears, the plunger to force a definite mass of glass out of a vertical outlet, and synchronous co-action between the shears’ action, and the plunger movement, are found, crudely, at least, present, or plainly forecast, in many patents of the prior art. But I think it is clear from this record, that Steimer largely improved upon all of them, and that Peiler improved upon Steimer.”

In a case (Individual Drinking Cup Co. v. Public Service Cup Co., D.C.E.D. N.Y., 234 F. 653, at page 656), wherein *423 tlie situation was somewhat similar to the one here, Judge Chatfield said:

“It appears that the new device which the plaintiff now seeks to bring in as an infringement was upon the market and was actually in the plaintiff’s hands before the- decision of this case. It is a different device as a whole than the patent in suit. It appears, as has been stated, that it may involve in certain ways the other patent taken out by the plaintiff, and which the defendant alleges contradicts the broad scope now claimed by the plaintiff for the patent in suit. At any time when the plaintiff sought a finding by the court upon the charge of infringement by this particular device, the defendant was entitled to a day in court upon such issues as could be properly presented for the court’s consideration under the plaintiff’s charge of infringement.

“If a new action, charging infringement by an entirely different style of cup distributor, is brought under the broad claims of the patent in suit, and if the plaintiff claims that as between it and the defendant, the issue has been adjudicated by the decree with respect to these broad claims, then the court will be compelled to decide upon a full hearing how far the validity of .the broad claims of the patent has been settled by consideration of those broad claims in their application to the cup dispenser involved in the previous action.

“The court cannot take back, modify, or explain a previous finding of validity, which was correct upon the prior record; nor can the court, as between the same parties and in the same action, and in the absence of proper newly discovered evidence, retry the same issue, as to an exhibit alleged to be an infringement (which was in the possession of the plaintiff prior to the entry of the interlocutory decree, but which was in effect disregarded by the plaintiff until the independent action was started and discontinued), and thus allow the plaintiff to retrace its steps so as to throw upon the court the responsibility of deciding what the court’s decree would have been, in case the present structure had been on trial, in case the defendant had pleaded all the matters which it now seeks to allege, or in case it had defaulted, as to proof of part, in so doing.

“The only sure way to dispose of these matters on the merits is to deny the present application, which it is believed is within the discretion of the court, without thereby deciding in any way any of the issues that might be presented in a new action based upon the present form of structure. If the plaintiff thereby loses anything from the effect of the plea of res adjudicata, it will at least be after a full hearing, and if the plaintiff’s position should be sustained upon that trial, then the defendant will have had its day in court.”

Appellant and plaintiff, appellee and defendant are interchangeably used hereinafter.

The objection was as follows:

“Mr. Paul: I object, your Honor, to the introduction of any letter, or any drawing. In the answer to interrogatories we were furnished with the drawing which we have offered in evidence, and upon which we have made our case of infringement.

“The Court: Those drawings are not this drawing (indicating)?

“Mr. Paul: Those are not the drawings.

“Mr. Bakewell: This is (indicating).

“Mr. Paul: Yes, but the blue prints are not. One of these blue prints discloses, as I understand the statement, a structure that the defendant began manufacturing and using after the filing of the bill.

“Your Honor has ruled on the evidence. The question of infringement is a question as to the structure of what the defendant was doing prior to the filing of the bilí, and we are not litigating here any structure that this defendant has manufactured or used since the filing of the bill.

“I object to any testimony in reference to what the defendant has done, or is doing, since the filing of the bill. I object to any drawing different than the drawing that the defendant furnished us in answer to the interrogatories.

“The Court: As I understand it, you are basing infringement upon the device shown by the photograph I have in my hand, and which machine was being used at and before the time of the filing of the bill—

“Mr. Paul: The drawing of which was furnished us in answer to interrogatories.

“The Court: You are not basing it upon any machine, or upon any blue print depicting a machine, which the defendant has seen fit to manufacture and begin the use of since this—

“Mr. Paul: That is correct, your Hon- or.”

This is evidenced by the difference in the cited anticipatory patents in the two cases. While the difference in the patents involved in 'these two cases would account for some — probably much — of this difference in citations, yet that does not weaken the inference of a definite divergence or difference in principles within the art.

The citations discussed in the first case

(39 F.2d 769, 776, 777) are Brookfield No. 883,779; Cleveland No. 901,881; Hitchcock No. 805,067 and 805,068; Mansfield No. 854,511; Morrison No. 810,167; Proeger No. 1,059,634; Schulze-Berge No. 421,620; Severin No. 892,013 and Wilzin No. 439,150 (French). Compare the above list with those cited and discussed in the second suit. 71 F.2d 539, 553-557, 564, 567.