This suit for infringement of a 1957 patent was brought in 1968 by Scully Signal Co. (Scully), the assignee of the patent and its licensor. Electronics Corporation of America (ECA), the defendant, alleged both noninfringement and invalidity. The patent expired before trial, leaving only damages at issue. The case was tried in December, 1975 and January, 1976, and at the end of the presentation of evidence ECA moved to amend its pleadings to allege fraud against the Patent Office by Scully because of a failure to reveal allegedly anticipatory patents, which in turn would entitle ECA to damages. The district court held that ECA had infringed the disputed patent, and went on to hold that the patent had not been anticipated within the meaning of 35 U.S.C. § 1021 but was invalid for obviousness under 35 U.S.C. § 103. Denying ECA's motion to amend the pleadings, the court awarded attorneys fees to the plaintiff because of "exceptional" conduct on the part of ECA.

On appeal, Scully vigorously challenges the district court's determination of obviousness, accusing the court of substituting hindsight for a proper assessment of the level of ordinary skill in the pertinent art at the time of the supposed invention. ECA in a cross appeal seeks to overturn the district court's denial of its motion to amend the pleadings, although it does not appeal the award of attorneys fees to Scully.

The patent in question, No. 2,798,214, W. G. Rowell, Checking Technique and System ("Rowell '214"), describes a technique designed to incorporate " fail-safe" features into machines or systems whose unsafe failure would present dangerous consequences. The technique combines a monitoring system, a failure simulator, and a self-checking circuit that will activate an alarm and take corrective measures whenever either the unsafe condition appears or the checking system itself breaks down. As the word "fail-safe" implies, the system is designed to shut off the machine it regulates whenever anything goes wrong, even if the machine itself is operating as intended.

Rowell assigned the patent to his employer, Scully, which in turn sought licensees to manufacture devices applying the patented system. In particular Scully offered nonexclusive licenses to ECA and Minneapolis-Honeywell Regulator Co. (Honeywell), the principal manufacturers of burner control devices. While the patent does not show a burner monitor application, the district court found that a use "would be obvious to anyone minimally skilled in the art," and this is not disputed. After satisfying itself as to the validity of the patent, Honeywell took a license in 1960. The license was limited to

"(t)he field of flame detection in which a flame sensing means is arranged to detect the presence or absence of flame, provided the flame sensing means is connected to the input of an electrical amplifier having a feedback in the form of a relay controlling a chopper switch means or other chopper member disposed at or before the input of the amplifier for controlling the feedback so that the relay normally is caused to repetitively cycle upon the flame sensing means detecting a flame or detecting the absence of flame, as the case may be, there being a further switch means controlled by the relay to alternately and repetitively connect a capacitor to a source of energy to charge the capacitor and then to connect the charged capacitor to an electrical device (load) normally to maintain the electrical device (load) continuously energized only so long as the relay continues to cycle."

A diagram used by the district court, which we attach as Appendix A, illustrates this description more clearly. When the detector (5) picks up the light, an amplifier (6) transmits the signal to relay coil (7). When so charged, the relay coil holds the relay arm (8) in place with contact (9), which completes a circuit between the battery terminals (Bk) (B-), a storage capacitor (11), and a resistor (12) that regulates the current. When the interrupter (3) blocks the light, relay (7) receives no charge, the first circuit is broken as arm (8) drops to contact (10), and a new circuit is formed between the capacitor (11), the resistor (12), and the load relay (14). A small capacitor (13) draws off some of the current from this circuit. As long as current flows through it, the load relay (14) holds the arm (15) to contact (16), which may be a ground or some other circuit, signalling all is well. If current were to stop passing through the load relay (14), however, the arm (15) would drop to contact (17), setting off the alarm (18) and cutting off oil to the burner.

Current passes through the load relay (14), holding off the alarm, as long as a proper cycle between the two circuits is maintained. The continual charging occurs because the capacitors (11) and (13) each have the property of storing and dispensing current, depending on whether a stronger power source is attached to the circuit. When the light is on, capacitor (11) is storing energy from the battery (Bk) (B-), and capacitor (13) is giving off current to the load relay (14). When the light is off, capacitor (11) is giving off current to charge the load relay (14) and associated capacitor (13). Because each capacitor has only a limited storage capacity, however, each must be recharged continually by alternate completion of the two circuits. The choice of the components determines the proper rate for the cycle. Although the diagram does not show it, the solenoid (1) that operates the shutter (3), which in turn controls the alternating periods of light and darkness that trigger the respective circuits, can itself be hooked into one of the circuits so that it may respond to the cycle it controls. This "feedback" feature was mentioned in the patent, although the invention was meant to be used with or without this modification, and incorporated into the Honeywell license.

ECA refused Scully's offer of a license, citing the added cost of installing the self-checking system in burner monitors already on the market. In 1967, however, ECA brought on to the market its own self-checking burner monitor, the Fireye UVP-4S. The ECA device differed in material respects from that sold by Honeywell only in that it relied on an independent timer for the flame-interrupting shutter rather than on feedback.

Scully's licensing arrangement with Honeywell continued until its expiration in 1975, Honeywell's payments totalling over $400,000 during the fifteen year period. It is notable that in 1954 Honeywell itself drew Scully's attention to the two patents which ECA alleges Scully fraudulently concealed from the Patent Office, and thereafter accepted a license notwithstanding its awareness of them.

At trial the district court considered several patents which were alleged to anticipate Rowell '214. These included No. 2,659,880, A.E. Dodd, Apparatus for Detecting Recurrent Circuit Operation (Dodd); No. 2,605,334, C.H. Hines, Circuit Integrity Indicating System (Hines); German Patent No. 898,564, Ludwig, Photoelectric Security Installation (Ludwig); German Patent No. 696,166, Werner, Circuit for Signal Devices (Werner); No. 1,631,021, J.J. Dowling, Thermionic Indicating Means Responsive to Light Variations (Dowling II); No. 1,561,837, J.J. Dowling, Thermionic Indicating Means Responsive to Light Variations (Dowling I). The last four were not cited to the Patent Office during the prosecution of Rowell '214, although Honeywell had informed Scully of the two Dowling patents in 1954.

Dodd and Hines, both of which were cited to the Patent Office, referred to a code-following circuit2 as prior art. A code-following circuit described by a witness to have existed in the late 1940's is diagrammed in Appendix B. Relay CTR, analogous to relay (7) in the Honeywell device, alternately receives and does not receive signals from some external device. When charged, CTR switches the attached arm so as to complete a circuit between Bk and B-, a battery or other power source, a capacitor C, and a resistor R. When not charged, CTR causes a circuit to be formed between capacitor C, resistor R, and relay TR, with resistor R 1 wired parallel to relay TR. The effect of wiring resistor R 1 across relay TR is to delay the release of the relay during the period capacitor C is being charged and is not charging relay TR. The substitution of the resistor R 1 for the capacitor (13), the only distinguishing feature between the two circuits, was held to be irrelevant, as expert testimony indicated the desired effect of a delayed release load relay could be achieved in a variety of ways, any of which would have been obvious to one of ordinary skill in 1954. As a result, the district court held that Rowell '214's self-checking circuit was not by itself inventive.3 It formulated the sole remaining question as "whether it was obvious to use such a circuit in a flame-out monitoring device in a manner that achieved precise simulation of the predetermined event that the monitor is to detect."

The other patents considered by the district court, while employing self-checking circuits of varying degrees of efficacy, were relevant mainly because of the monitoring and interruption means that generated the on-off cycle transmitted to the checking circuit. Werner and Ludwig both involved space intrusion detectors, such as burglary alarms, designed to set off an alarm if some object interfered with a beam of light being sent into a photoelectric cell. Werner reflected the beam with a mirror from the light source to the detector; the portion of the beam between the mirror and the detector was projected across the protected space. The mirror was regularly jerked out of position, creating a steady pulse of light that went into the detector. Ludwig achieved the same effect through a circuit that switched off the light source upon receipt of the beam at the detector. Both systems embodied a feedback principle. The two Dowling patents were designed to detect variations in the intensity of light, such as occlusion caused by fog. A pierced disk which rotated in front of the beam of light was used in Dowling I. Dowling II substituted a vibrating prong, something like a tuning fork, which oscillated in the path of the light beam. The stimuli to the prong were controlled by the signals generated by the pulses of light, thereby embodying yet another form of feedback.

At trial Scully emphasized that the Rowell device, in exercising the monitoring system, simulated precisely the event to be detected by the monitor, namely disappearance of the flame. All other self-checking systems, it was maintained, created some other kind of interference with the operation of the detecting circuit that, because of a lack of exact correspondence with the looked for event, failed to achieve the same degree of reliability. In particular, Ludwig and Werner rather than blocking the beam of light, as would the intruder sought to be detected, turned off the light signal completely. Further, the Dowling systems, which were meant to detect variations in light intensity, employed instead rhythmic but total blockage of the beam. The district court held, however, that the distinction was without a difference, as Scully had failed to indicate how Rowell's "precise" simulation of the flame-out in any way enhanced reliability in comparison to the other systems. The court further held that the combination of a light interruption device, already considered prior art, with a self-checking circuit, also considered prior art, did not amount to a patentable invention.

Although Scully knew about the Dowling patents during prosecution of the Rowell patent, this prior art was not disclosed to the patent office. ECA contended that the Dowling II patent, by employing feedback in its monitoring circuit, completely anticipated Rowell '214 and would have resulted in the latter patent's invalidation if seasonably presented to the Patent Office. Rowell's feedback feature was not, however, essential to the invention and in other respects the Dodd and Hines patents, which were cited, seem more closely to have anticipated the Rowell system. Both Dowling patents were in the public domain for more than a decade before Rowell applied for his patent. Rowell in 1954 wrote two analyses for Scully of the Dowling patents, each of which contended that his invention contained substantial safety features not found in the earlier devices. The second of these memoranda, of which ECA made use during trial, accepted for the sake of argument that the self-checking circuit in Dowling II was as safe as that in Rowell '214 but went on to indicate other features of the earlier system that made it less safe than his own invention. Honeywell was sufficiently convinced by the memoranda to accept the Scully license.

It does not appear that anyone thought much of the Dowling patents until Rowell became embroiled in an unsavory dispute with Scully in 1970. Impugning his own invention and prior statements, Rowell surprisingly asserted that one of the Dowling patents was entirely anticipatory of his own invention; and undertook on this basis to sabotage Scully's suit against ECA. The district court nonetheless found, supportably we think, that, "given that plaintiff did in fact cite to the Patent Office numerous patents far more relevant than Dowling, to either a broad or narrow reading (of the Rowell patent), I cannot imagine that citing Dowling would have affected the Patent Office proceedings."

While "the ultimate question of patent validity is one of law", Graham v. John Deere Co., 383 U.S. 1, 17, 86 S. Ct. 684, 694, 15 L. Ed. 2d 545 (1965), this court has emphasized the highly factual context of a determination of § 103 obviousness, and the strong deference due a district court's reasoned judgment on the issue:

"More often . . . obviousness as an ultimate question cannot meaningfully be separated from those factual determinations which are peculiarly within the trial court's province, such as the credibility of the experts. The district court's supported findings on obviousness will therefore normally stand unless manifesting a misconception of the correct legal standard."

Forbro Design Corp. v. Raytheon Co., 532 F.2d 758, 763 (1st Cir. 1976). Scully contends, however, that the district court, although reciting the proper legal standard for determining obviousness, in fact applied the wrong criteria, namely obviousness to the court itself. Scully goes so far, indeed, as to deny that the record itself contains any evidence that would support the finding of obviousness, arguing that the court simply ignored the "years of expertise in the nuances of these circuits" of the Patent Office, which also had Hines and Dodd before it. Further, the court is said to have overlooked the demonstration "that the best the skilled engineers in this art had been able to evolve, over the past twenty years, despite their attempts to provide against unsafe failures, still ran the risk of . . . failures, that simply cannot fail unsafe with the Rowell technique." The entire technical community is said to have recognized the novelty and importance of the Rowell system. The district court is said to have ruled by "fiat", piecing together a multitude of prior inventions and patents by hindsight, in violation both of the admonitions of jurists and the Constitution itself.4

If the district court were guilty of such misdirected thinking we would agree that error had indeed occurred. Scully, however, ignores the substantial evidence supporting the district court's finding that the relevant techniques were all known to the art in 1957 when the patent was obtained, and the lack of persuasive evidence that Rowell's assemblage of these bits and pieces reflected, in the instant application at least, a novel insight.5 According to Pascoe, a Westinghouse engineer, the same ingenious self-checking circuit forming the backbone of the patented system had been employed in railroad signalling devices in the 1940's; it is referred to in the Dodd patent and in the Hines patent. Scully does not seriously contest this,6 but argues that since no one "had thought of the application of this kind of technique, suitably modified, for burner control safety monitoring", there was invention.

In response, the district court inquired whether using the precise event to be detected, in this case the light from the burner flame, with a light interrupter and detector in combination with the non-inventive self-checking circuit, was inventive. It concluded not. It would not be inventive to adopt a self-checking circuit to monitor the presence or absence of light, nor "to effect the pulsing needed to utilize the self-checking circuit by use of a shield or similar light occlusion device to cause light periodically to strike the detector." The latter technology was sufficiently revealed in both Dowling patents and in Werner and Ludwig. Pascoe, moreover, testified to a contemporary use of a light source, interrupter, and a detector with a self-checking circuit to signal the presence of a train.

The court then turned to Scully's emphasis upon the patent's teaching "that the precise predetermined event which the device is to monitor should be repetitively simulated to produce the checking pulse." Scully presented this as, in effect, the synergism which could transform a combination of familiar elements into an invention. The court was unimpressed warrantably, we think. It could find little evidence that the concept of precise simulation was itself the key to some advance over the prior art in averting unsafe failures. To the extent blockage of light from the flame to the detector was a species of "precise simulation", it found it to be just another obvious way of employing light interrupters merely "the recognition of an attribute of an existing device". Hence "at least as adapted to a nonfeedback burner flame monitor, the patent is invalid."

Given the level of technology which the court was entitled to find existed, we believe it was warranted in concluding that utilization of the burner flame itself, the interrupter, the detector, and the self-checking circuit was in 1957 within the competence of engineers ordinarily skilled in the art. To be sure, this presupposes knowledge of self-checking circuits in the railway field and of systems in other industries with common problems such as burglar alarms, fog detectors, and so forth. Rowell's patent, however, encompasses such a range of applications: indeed it describes a railway application but does not specifically describe a burner flame use at all. We think the "art to which said subject matter pertains", as defined in § 103 would embrace such devices.

On appeal, Scully does little to meet the district court on these grounds. Rather it belittles the district judge as one who has, never in his life, upheld a patent,7 and urges courts to stay out of matters that they don't understand. Its most credible argument, but one we also find deficient, is that the district court paid no attention to the ready commercial acceptance of Honeywell's licensed device, and its evidence of enthusiastic trade comment.

We would agree that secondary factors especially were they to show "long felt but unsolved needs, failure of others", Graham v. John Deere Co., supra, 383 U.S. at 17, 86 S. Ct. at 694 could be important evidence in a case such as this, but we do not agree that Scully's evidence measures up to the claims of its counsel. In Hand's famous compendium of "signposts" in Reiner v. I. Leon Co., 285 F.2d 501, 504 (2d Cir. 1960), cert. denied, 366 U.S. 929, 81 S. Ct. 1649, 6 L. Ed. 2d 388 (1961), the questions, "how long did the need exist" and "how many tried to find the way", appear side by side with the question of success. That Scully and Honeywell were the first to adapt and market a self-checking system in the burner industry, and that the product was safer than previous devices, says little about the inventiveness of the system in a technological sense. Beyond indication that earlier burner monitors were less reliable, it was not brought out what sort of an effort had been mounted in the burner industry to develop a comparable system. The industry's failure earlier to develop a self-checking system could as well have been due to lack of interest or appreciation of such a system's potential or marketability, as to want of technical know-how. Indeed, there was evidence that ECA, a major producer, refused a license initially because of a belief (whether or not misguided is beside the point) that what it had sufficed.

Scully introduced a variety of news clippings, lab reports, and related items dating from the period of invention, all of which remarked on the advance in flame monitoring safety achieved by the Rowell invention. The majority of these items, however, were either promotional literature put out by Scully or press reports cribbed directly therefrom. The lab reports established only a fact which is not in dispute: that the Rowell patent was the first to apply the self-checking circuit to burner flame monitoring. None of these reports were decisive or even especially germane to the inventiveness of this application. As the Supreme Court said recently, in discussing a patent held simply to arrange "old elements with each performing the same function it had been known to perform, although perhaps producing a more striking result than in previous combinations,"

"Though doubtless a matter of great convenience, producing a desired result in a cheaper and faster way, and enjoying commercial success, Dairy Establishment 'did not produce a "new or different function" . . . within the test of validity of combination patents.'. Anderson's-Black Rock v. Pavement Co., supra, at 60 (396 U.S. 57, 90 S. Ct. 305, 24 L. Ed. 2d 258). These desirable benefits 'without invention will not make patentability'. Great A. & P. Tea Co. v. Supermarket Corp., 340 U.S., at 153. (71 S. Ct. 127, 95 L. Ed. 162.) See Dann v. Johnston, ante, at 230 n.4."

Sakraida v. AG PRO, Inc., 425 U.S. 273, 282-83, 96 S. Ct. 1532, 1537, 47 L. Ed. 2d 784 (1976).

The foregoing authority, and the cases it cites, also dispose of Scully's argument that the district court was duty bound to treat the fact of issuance of the patent as itself conclusive of non-obviousness. While weight must be given to the presumption of validity, and this circuit is quite prepared to sustain patents which meet the statutory criteria, the time has long since gone, if it ever existed, when district courts and courts of appeal could refuse to make an independent assessment of § 103 obviousness in light of all the evidence presented. To criticize a court for making an independent assessment is to criticize it for doing what the law presently requires. The process involves the ever-present risk of an overuse of hindsight, as well as the possibility of blunders by lay judges; but this court has no license, even if it wanted one, to adopt another approach. Finding nothing even marginally erroneous in the analysis employed by the district court, we sustain the finding of invalidity.

Turning to the cross-appeal by ECA, it must be determined whether the district court violated the mandate of Federal Rule of Civil Procedure 15(b) to amend the pleadings to conform to issues tried with the express or implied consent of the parties. Having determined that the Dowling patents did not anticipate Rowell '214 and would not have affected the prosecution of the patent in light of closer prior art that was cited, the court refused to consider whether Scully nonetheless violated its duty of candor and good faith by not disclosing the two patents. The court noted that further evidence would be necessary to resolve the issue, and that the failure of the record to contain sufficient evidence to try the issue was due entirely to ECA's own misconduct. The court found that ECA had known of the two Dowling patents at least since 1972, although its counsel were not told of their existence until midway through the trial. In addition, ECA in its post-trial briefing on the issue had attempted to mislead the court as to the extent of an inventor's duty of disclosure at the time the Rowell patent was prosecuted. These factors all persuaded the court to deny the motion to amend.

Although Rule 15(b) by its terms requires amendment of the pleadings whenever an issue has been tried by express or implied consent, courts have refused to grant such motions if amendment would prejudice one of the parties, such as by requiring the presentation of additional evidence. See American Hot Rod Association, Inc. v. Carrier, 500 F.2d 1269, 1277-78 (4th Cir. 1974); United States v. An Article of Drug, 320 F.2d 564 (3rd Cir.), cert. denied, 375 U.S. 953, 84 S. Ct. 444, 11 L. Ed. 2d 313 (1963); 3 Moore's Federal Practice P 15.13(2), at 997 & n. 34 (2d ed. 1974). Professor Moore explains this practice as an implied finding that the issue involved was not tried by the consent of the parties. Id. Whether the district court's ruling be interpreted either as finding the issue had not in fact been tried, or that Scully had not consented to trying the issue, the denial of ECA's motion to amend did not exceed the court's discretion. The Dowling patents were put in evidence primarily to attack the validity of the Rowell patent, not to prove bad faith on the part of Scully. As the district court noted, establishing fraud on the part of Scully would require evidence of state of mind, see Norton Co. v. Carborundum Co., 530 F.2d 435, 441-42 (1st Cir. 1976), which neither side produced to sufficient degree. Requiring Scully to introduce new evidence of its intent and actions during the prosecution of Rowell '214, when the failure of the case to embrace this issue can be attributed entirely to ECA's neglect, would be sufficiently prejudicial to warrant the action taken by the district court.

We must comment on the entirely unacceptable tenor of argument by Scully's counsel. The right of appeal includes the right vigorously to challenge the decision of a lower court and to describe in every proper way its alleged errors. But appellate counsel may not give vent to their frustrations by undignified or discourteous remarks directed against the person of the deciding judge. Never suppressing any fact or proper argument, counsel have a professional responsibility to refer to the tribunals from which an appeal is taken, as well as those before which they appear, with reasonable respect and courtesy. Perhaps an attorney would have greater leeway if provoked by some act of judicial misconduct, but clearly there was no misconduct here only a decision which counsel believes to be wrong. The court's decision manifested care and diligence. While it might be natural for a layman, embittered by a decision, to lash out at a judge, such conduct cannot and will not be tolerated from a member of the bar of this court. We only refrain from taking some action because of the curious history of this case which, beginning with defendant's egregious misconduct, seems to have spawned an unusual atmosphere that seems unlikely of repetition. We make it quite clear, however, that counsel's personal asides in Scully's brief raise serious questions in our mind. See Mass.Sup.Jud.Ct. Rule 3:22; DR 7-106(c)(4); DR 7-106(c)(6). Should we receive anything approaching this from counsel in the future, we shall not hesitate to act.

Affirmed.

APPENDIX 'A'

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLE

Appendix 'B' to follow.APPENDIX 'B'

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLE