Ferguson Beauregard/Logic Controls, Division of Dover Resources, Inc. ("Ferguson") and Delaware Capital Formation, Inc. ("Delaware Capital") appeal various aspects of a judgment of the United States District Court for the Eastern District of Texas following a bench trial concerning, inter alia, allegations of infringement of four patents and multiple versions of accused infringing devices. Ferguson Beauregard/Logic Controls, Div. of Dover Res., Inc. v. Mega Sys., LLC, 6:99CV437 (E.D.Tex. Feb. 6, 2002) ("Judgment"). Mega Systems, LLC ("Mega") and James Bartley ("Bartley"), president and majority owner of Mega, cross-appeal other aspects of the same judgment.

In particular, Ferguson and Delaware Capital appeal: (1) the district court's construction of certain claim limitations of U.S. Patent No. 5,146,991 ("the '991 patent"), owned by Delaware Capital and licensed to Ferguson; (2) the district court's determination that a certain version of a product made and sold by Mega—version 3 of the APC 1000 device — did not infringe the '991 patent; (3) the district court's conclusion that Bartley was not personally liable for inducing Mega to infringe the '991 patent and U.S. Patent No. 4,352,376 ("the '376 patent"), also owned by Delaware Capital and licensed to Ferguson; (4) the district court's denial of Ferguson's motion to amend the claims to assert willful infringement; and (5) the district court's conclusion that it lacked jurisdiction to consider whether U.S. Patent No. 4,921,048 ("the '048 patent"), assigned to Mega, had been improperly revived at the U.S. Patent and Trademark Office ("PTO") after it lapsed for failure to timely pay maintenance fees.

Mega and Bartley cross-appeal: (1) the district court's construction of certain claim limitations of the '376 patent; (2) the district court's determination of damages to be awarded to Ferguson for infringement of the '376 patent; (3) the district court's conclusion that Ferguson did not infringe the '048 patent; and (4) the district court's ruling barring the submission of evidence of prior art due to lack of notice under 35 U.S.C. § 282.

Because the district court erroneously construed certain claim limitations of the '991 patent, we reverse the district court's claim construction, vacate the determination that version 3 of Mega's APC 1000 device did not infringe, and remand the issue for further findings consistent with the claim construction set forth in this opinion. We affirm the district court's conclusion that Bartley did not induce infringement based on the correct legal standard applied by the court to the facts in evidence. We reverse the district court's denial of Ferguson's motion to amend the complaint to include an assertion of willful infringement and remand for further proceedings on that issue. We affirm the district court's refusal to consider whether Mega engaged in inequitable conduct when it successfully revived the '048 patent. Further, we affirm the district court's construction of the claims of the '376 patent; however, because we believe the district court abused its discretion in determining the amount of the damage award for infringement of the '376 patent, we vacate that award and remand the issue of damages for determination consistent with this opinion. Finally, we affirm both the district court's determination that Ferguson did not infringe the '048 patent and the district court's evidentiary ruling under 35 U.S.C. § 282.

The patents in this case concern control systems and related methods used in the production of petroleum products from a well. While some wells are capable of producing (or outpouring) liquid petroleum products under naturally induced reservoir pressures, it is more common for wells to require an artificial lift mechanism to be productive. '991 patent, col. 1, ll. 7-11. One such artificial lift system repetitively causes pressure to build by first closing in the well while it is subject to the inflow of liquids. Id. at ll. 14-18. After the proper combination of pressure and liquid develops, the well is opened to a gathering system to collect the liquid produced and expelled by the pressure built up when the well was closed. The cycle is then repeated successively. This cyclical shutting in and opening of the well is known as "intermitting." Id. at ll. 25-31. The gathering system typically includes a gas/liquid separator, one or more sales lines, and a tank or reservoir for collecting the liquids issuing from the well during the open intervals. Id. at ll. 18-24.

Many artificial lift well installations also typically employ a "plunger method" to aid production. Using the plunger method, a piston, or plunger, is installed within the tubing string of the well and travels the entire length of the tubing string during each cycle of the shutting in and opening of the well. The presence of the plunger improves the production and delivery of the liquid from the well to the gathering system. Id. at ll. 57-62.

The timing of intermitting and operation of the plunger mechanism are critical for the success of the well operation. Prior to the inventions involved in this case, control over these timing issues generally had been relegated to crude, clock-operated devices, requiring hand winding and frequent technician intervention. Id. at col. 2, ll. 30-34. All of the patents-in-suit are directed to improvements over these earlier control systems.

Figure 1 of the '991 patent, shown below, is illustrative of the technology at issue and shows "a well installation for plunger lift production according to the method of the invention." Id. at col. 6, ll. 32-34. NOTE: OPINION CONTAINING TABLE OR OTHER DATA THAT IS NOT VIEWABLE

Well installation 10 includes a casing 12, extending from the surface into a gas-oil formation. Wellhead 14 supports a tubing string 16 having an open lower end 18 in the vicinity of the lower region of casing 12. A plunger 22 is disposed in tubing string 16. T-connection 28 provides fluid communication between tubing string 16 and conduit 42, which extends to a separator 46. A motor valve 48 provides control over conduit 42. Valve 48 is controlled to open and close conduit 42 by a microprocessor-driven controller 50. The pressure in the sales line is monitored by a sales line gauge 60. When the gas pressure in the sales gathering system is too high, and a set threshold is reached or exceeded, a "high line contact" is generated as an electrical signal and conveyed to controller 50. A conduit 70 extends from T-connection 32 and is controlled by a second motor valve 72 — a tank valve or tank control valve. Tank valve 72 is also controlled between on and off states by controller 50. Opening tank valve 72 opens tubing string 16 to the low pressure of a tank or reservoir 76. A plunger detector 86 located above plunger catcher 38 provides a magnetic shut-off on arrival signal to controller 50. Id. at col. 7, l. 58 — col. 9, l. 27.

The '376 patent issued on October 5, 1982, and is directed to a microprocessor-based controller that improves upon a controller subject to an earlier patent, U.S. Patent No. 4,150,721 ("the '721 patent"),1 also owned by Delaware Capital and licensed to Ferguson. The '721 patent discloses an electronic well controller system, providing long-term battery-operated control over wells and simplified control adjustment procedures. See '376 patent, col. 2, ll. 36-42. The '376 patent improves upon the system described in the '721 patent by allowing greater flexibility of operation and accommodation for a wide range of parameters, as well as other features, such as the ability to withstand a variety of severe environments and to run on conventional D-cell batteries. Id. at col. 3, l. 60 — col. 4, l. 18.

Ferguson asserts that Mega's APC 1000 controller infringes claims 1, 16, and 35 of the '376 patent. Representative claim 1 recites, in relevant part and with the disputed term underlined:

1. A controller for use in conjunction with the control of well installations of a variety wherein a control valve regulating the flow of fluid hydrocarbon is selectively actuated between an on state and an off state in response to corresponding control inputs thereto, comprising:

* * *

valve means responsive to said actuation signals to derive said control inputs.

The valve assembly, in operation, shifts between an on state, where fluid is produced from a well, and an off state, where the well is shut in. Id. at col. 7, l. 62 — col. 8, l. 45. The on state is defined by a window of time bounded by the time when the valve means is opened and extending until the time when the valve means is closed. Id.

The '991 patent issued on September 15, 1992, and is directed toward a method for use in controllers, such as that described in the '376 patent. '991 patent, col. 8, ll. 25-36. The method includes continuous monitoring and adjustment of well performance by evaluating plunger speed. Id. at col. 3, ll. 13-16. The method allows the duration of the off-cycle time and afterflow cycle time to be changed to optimize performance, by achieving a consistent plunger speed. Id. at ll. 16-23. To achieve plunger speed-based control, a well technician will select a consistent on-cycle interval. Within that interval, windows are then set to reflect fast, good, and slow speeds for the plunger to arrive at the wellhead. The method will make no change to operating times if the plunger continues to surface within the good window. However, if the plunger surfaces within the fast window, the off-cycle time will be decreased and the afterflow cycle time will be increased. Conversely, if the plunger arrives within the slow window, the method will increase the off-time interval and decrease the afterflow time. Id. at ll. 24-46.

Ferguson asserts that the Mega Systems APC 1000 controller infringes claims 1-3 and 5-15 of the '991 patent. Independent claim 1 is representative and recites, with the relevant terms underlined:

1. The method for operating a well installation having a control valve regulating the flow of fluid hydrocarbon from a well tubing string to a sales line which is selectively actuated between an on-state and an off-state, and wherein a plunger is located within the said tubing string of said well for movement between a lower region and a wellhead sensing position, comprising the steps of:

assigning first values corresponding with the rate of movement of said plunger from said lower region to said wellhead which represent normal plunger performance;

assigning second values less than said first values corresponding with the rate of movement of said plunger from said lower region to said wellhead which represent slow plunger performance;

assigning a predetermined value for the time interval of said on-state;

assigning a predetermined value for time interval of said off-state;

actuating said control valve to transition from an off-state to an on-state;

then detecting the arrival of said plunger at said wellhead prior to expiration of said predetermined value for the time interval of said on-state, and determining the time elapsed from said actuation;

determining the presence of any coincidence of said time elapsed from said actuation of said control valve with said assigned second values;

then increasing said predetermined value for the time interval of said off-state by a predetermined first time increment when a said coincidence with said assigned second value is present; and

terminating the said on-state in response to said plunger detection, and actuating said control valve to transition from said on-state to the next off-state in response to said termination of said on-state.

Independent claims 8 and 14 recite similar methods. The relevant portions of each of those claims have been reproduced below (with the disputed terms underlined):

8. The method for operating a well installation ... comprising the steps of:

* * *

assigning select values corresponding with the rate of movement of said plunger from said lower region to said wellhead which represent predetermined plunger performance;

* * *.

14. The method for operating a well installation ... comprising the steps of:

* * * assigning first values corresponding with a said rate of movement of said plunger from said lower region to said wellhead which represent normal plunger performance;

* * *.

The term "values" in these claims was construed by the district court to mean "more than one value" and also to require that the more than one value must be different, "so as to define a window of time, or time interval." Ferguson Beauregard/Logic Controls, Div. of Dover Res., Inc. v. Mega Sys., LLC, No. 99-CV-437, slip op. at 23 (E.D.Tex. Jan. 22, 2001) ("Markman Report"). The district court, however, did not explicitly construe either "normal plunger performance" or "predetermined plunger performance."

Ferguson asserted that Mega's APC 1000 device infringed both the '376 and '991 patents. The APC 1000 controller is used in the production of petroleum products using plunger lift and intermitting technologies. Ferguson Beauregard/Logic Controls, Div. of Dover Res., Inc. v. Mega Sys., LLC, No. 99-CV-437, para. 41 (E.D.Tex. Aug. 31, 2001) ("Special Master's Report"). To use the APC 1000 controller, a user selects a target time, TT, for example, based on well depth. Id., paras. 147-149. The device then automatically adds one minute to the entered target time to derive time TT + 1. Id. The window of the on state is then bounded by the values TT and TT + 1. Id.

The '048 patent, owned by Mega, is also directed to electronic controller systems and related methods for use in oil and gas wells. In particular, the controller of the '048 patent detects the arrival of a cycling plunger at the wellhead and monitors the time required for the plunger to make each round trip. '048 patent, col. 5, ll. 50-54. In one embodiment, the system decreases the off-time interval by a preselected value each time the plunger arrives prior to the expiration of the on-time, and the off-time interval is increased each time the plunger fails to arrive prior to the expiration of the on-time. Id. at col. 6, ll. 26-32. Further, the off-time interval is increased by a second preselected value, greater than the first preselected value, in response to expiration of the on-time prior to the plunger arrival on two successive cycles of intermitting. Id. at ll. 32-36.

Mega originally alleged that Ferguson's AutoCycle controller, which apparently embodies the '991 patent, infringed all claims of the '048 patent, including 20 separate independent claims. However, the district court determined that only claims 6 and 35 were being pursued by Mega and withdrew the remaining claims from consideration. Special Master's Report, para. 194 n.23.

Claims 6 and 35 recite, with the disputed terms underlined:

6. A method for controlling the operation of a plunger completion petroleum production well having a motor valve connected between the tubing of the well and a flow sales line comprising:

decreasing the length of the first time period (off-time-period) for the next cycle of the well in response to the plunger having reached the wellhead before the expiration of the second time period (on-time-period); and

increasing the length of the first time period (off-time-period) in response to the plunger not reaching the wellhead before the expiration of the second time period (on-time-period).

35. A method for optimizing the production from a petroleum producing well having a motor valve connected between the tubing of the well and a flow sales line and a plunger mounted for movement within the tubing of the well from the bottom thereof to the well head to carry liquids from the well to the flow sales line in response to downhole casing pressure when the motor valve is open, the method comprising:

changing the value of either the off-time or the exhaust-time in response to whether a plunger arrival is detected prior to the expiration of the on-time on each successive cycle of the intermitting of the well.

On July 26, 1999, Ferguson brought this action against Mega, and James Bartley individually, asserting, inter alia, that Mega's APC 1000 controller device infringed the '376 and '991 patents. Mega responded, counterclaiming that Ferguson's AutoCycle controller infringed its '048 patent. Delaware Capital was joined and deemed an "involuntary plaintiff" by the district court, in response to claims made by Mega.2

On December 5, 2000, the district court referred the issue of claim construction to Special Master Danny L. Williams, who issued his report on January 22, 2001. See Markman Report. This report was adopted in its entirety by the district court as its Markman findings on March 22, 2001. Ferguson Beauregard/Logic Controls, Div. of Dover Res., Inc. v. Mega Sys., LLC, No. 99-CV-437 (E.D.Tex. Mar. 22, 2001). On April 30, 2001, the district court referred the case to Special Master Gale R. Peterson, who entered his Findings of Fact and Conclusions of Law on September 5, 2001. See Special Master's Report. Among other findings, Special Master Peterson concluded: (1) Mega's APC 1000 device infringed the asserted claims of the '376 patent, over Mega's arguments about various aspects of claim construction and equivalents, id. at paras. 42-83, 268-270; (2) version 2 of Mega's APC 1000 device infringed the asserted claims of the '991 patent, but version 3 of the device did not, because the values selected in the operation of the device "do[] not bear any relationship to `normal plunger performance,'" id. at paras. 157-159; and (3) Ferguson's AutoCycle Controller did not include each element of the asserted claims of the '048 patent, and Mega did not present evidence of infringement under the doctrine of equivalents. Id. at paras. 240, 291. For infringement of the '376 patent, the Special Master found Ferguson to be entitled to lost profits in the amount of $837 for each APC 1000 made, used, or sold prior to October 5, 1999, or, alternatively, to a reasonable royalty in the amount of $100,000. Id. at paras. 294-295.

On December 13, 2001, the district court adopted the Special Master's Report with some modifications. Ferguson Beauregard/Logic Controls, Div. of Dover Res., Inc. v. Mega Sys., LLC, No. 99-CV-437 (E.D.Tex. Dec. 13, 2001) ("Order Adopting the Special Master's Report"). The district court entered its final judgment on February 6, 2002 based on the Special Master's Report and including the following specific changes: (1) an award for lost profits of $408,456 was entered in lieu of a reasonable royalty award of $100,000; and (2) a judgment for $8,370 was added for infringement of Ferguson's '991 patent by Mega's APC 1000 version 2 controller. Judgment at 4-5.

As discussed above, the parties appeal various aspects of the Judgment to this court. We have jurisdiction pursuant to 28 U.S.C. § 1295(a)(1) (2000).

This court reviews a district court's judgment following a bench trial for errors of law and clearly erroneous findings of fact. Manville Sales Corp. v. Paramount Sys., Inc., 917 F.2d 544, 549 (Fed.Cir. 1990); Gould v. Quigg, 822 F.2d 1074, 1077 (Fed.Cir.1987). "A factual finding is clearly erroneous when, although there is evidence to support [the factual finding], the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed." Tegal Corp. v. Tokyo Electron Am., Inc., 257 F.3d 1331, 1338-39 (Fed.Cir.2001) (internal quotation omitted).

In this appeal, we are faced with the question of whether to apply Federal Circuit or Fifth Circuit law to a number of the issues raised by the parties. We answer this question on an issue by issue basis and apply the law of the regional "circuit to which district court appeals normally lie, unless the issue pertains to or is unique to patent law," Molins PLC v. Quigg, 837 F.2d 1064, 1066 (Fed.Cir.1988) (citations omitted), in which case, we apply our own law to both substantive and procedural issues "intimately involved in the substance of enforcement of the patent right." Viam Corp. v. Iowa Export-Import Trading Co., 84 F.3d 424, 428 (Fed. Cir.1996).

Given the number of issues on appeal, the specific standards of review for the individual issues are not set out in this section, but are included individually in the separate sections discussing each issue.

Ferguson first argues that the district court erroneously interpreted the claim terms "normal plunger performance" and "predetermined plunger performance." Ferguson contends that the trial court erred in implicitly giving the term "predetermined plunger performance" the same definition as "normal plunger performance." Ferguson also disagrees with the district court's conclusion that infringement cannot lie because the TT + 1 time calculated by the APC 1000's algorithm "does not bear any relationship to `normal plunger performance.'" Mega, on the other hand, argues that the district court properly construed the "plunger performance" claim terms and appropriately determined that version 3 of the APC 1000 does not infringe. We agree with both of Ferguson's arguments.

The district court did not explicitly construe the terms in question — "normal plunger performance" and "predetermined plunger performance" — but has conflated claim construction and infringement issues in its analysis. In analyzing the claim limitation "assigning first values ... which represent normal plunger performance," the district court noted that Special Master Williams did construe "values" to require at least two different values, defining a window of time or a time interval. See Special Master's Report at para. 156; Markman Report at 23. The district court then turned to an example in the specification and determined that "normal plunger performance" should be equated with a good window of plunger arrival. Special Master's Report at para. 156. From there, the district court examined the accused device, and determined that the APC 1000 operator entered only a single target time, TT, from which the system generated a second TT + 1 time. Id. at para. 158. The district court concluded that the window, bounded by the time entered by the operator, TT, and the generated value, TT + 1, "does not per se represent `normal plunger performance.'" Id. at para. 159 (emphasis in original). Although the district court then found that an operator might select a one-minute good window, it still determined that there was no infringement, because it concluded that "the TT + 1 value does not bear any relationship to `normal plunger performance.'" Id.

The determination of infringement is a two-step process. First, the court construes the claims at issue to correctly determine their scope. Second, it compares the properly construed claims to the accused device. Cybor Corp. v. FAS Techs., Inc., 138 F.3d 1448, 1454-56 (Fed. Cir.1998) (en banc). Claim construction is a question of law that we review de novo. Id. A determination of infringement, whether literal or under the doctrine of equivalents, is a question of fact. Insituform Techs., Inc. v. Cat Contracting, Inc., 161 F.3d 688, 692 (Fed. Cir.1998).

We begin our claim construction analysis with the words of the claim. See Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed.Cir.1996). "In construing claims, the analytical focus must begin and remain centered on the language of the claims themselves, for it is that language that the patentee chose to use to `particularly point[] out and distinctly claim[] the subject matter which the patentee regards as his invention.' 35 U.S.C. § 112, ¶ 2." Interactive Gift Express, Inc. v. Compuserve, Inc., 256 F.3d 1323, 1331 (Fed.Cir.2001). In the absence of an express intent to impart a novel meaning to the claim terms, the words take on the full breadth of the ordinary and customary meanings attributed to them by those of ordinary skill in the art. See, e.g., Teleflex, Inc. v. Ficosa N. Am. Corp., 299 F.3d 1313, 1325 (Fed.Cir.2002). The ordinary and customary meaning of a claim term may be determined by reviewing a variety of sources. Some of these sources include the claims themselves, see Process Control Corp. v. HydReclaim Corp., 190 F.3d 1350, 1357 (Fed.Cir.1999); dictionaries and treatises, Texas Digital Systems, Inc. v. Telegenix, Inc., 308 F.3d 1193, 1202 (Fed.Cir.2002); and the written description, the drawings, and the prosecution history, see, e.g., DeMarini Sports, Inc. v. Worth, Inc., 239 F.3d 1314, 1324 (Fed.Cir.2001).

Words often have different meanings to different people and in different contexts, accounting for the multiple ordinary meanings found in dictionaries. Dictionary definitions, while reflective of the ordinary meanings of words, do not always associate those meanings with context or reflect the customary usage of words by those skilled in a particular art. The words used in the claims must be considered in context and are examined through the viewing glass of a person skilled in the art. Tegal Corp., 257 F.3d at 1342. It is the use of the words in the context of the written description and customarily by those skilled in the relevant art that accurately reflects both the "ordinary" and the "customary" meaning of the terms in the claims of a patent. As set forth in Brookhill-Wilk 1:

In construing claim terms, the general meanings gleaned from reference sources, such as dictionaries, must always be compared against the use of the terms in context, and the intrinsic record must always be consulted to identify which of the different possible dictionary meanings is most consistent with the use of the words by the inventor. "Where there are several common meanings for a claim term, the patent disclosure serves to point away from the improper meanings and toward the proper meanings."

Brookhill-Wilk 1, LLC v. Intuitive Surgical, Inc., 334 F.3d 1294, 1300 (Fed.Cir. 2003) (quoting Renishaw PLC v. Marposs Societa' per Azioni, 158 F.3d 1243, 1250 (Fed.Cir.1998)); see also Tex. Digital, 308 F.3d at 1203. The written description must also be examined in every case, because it is relevant to determine if the presumption of ordinary and customary meaning is rebutted. See Renishaw, 158 F.3d at 1250. Once the court has construed the claim limitations, the second step in the analysis is to compare the properly construed claims to the accused device. Cybor Corp., 138 F.3d at 1454-56.

As noted, the claim limitations at issue are "normal plunger performance" and "predetermined plunger performance." The parties do not dispute the meaning of the terms "plunger performance," so the initial focus of our analysis is on the terms "normal" and "predetermined" within the context of each of the claim limitations in question. Starting with the claim limitation "normal plunger performance," we consider first the ordinary and customary meaning of the word "normal." The first listed dictionary meaning of the word "normal" is "according to, constituting, or not deviating from an established norm, rule, or principle: conformed to a type, standard, or regular pattern: not abnormal." Webster's Third New International Dictionary 1540 (1966) ("Webster's"). Applying this definition to the expression "normal plunger performance" in context informs that the claim limitation refers to performance of the plunger according to a rule or conforming to a standard or regular pattern of operation that is not abnormal. See Brookhill-Wilk, 334 F.3d at 1299 ("While certain terms may be at the center of the claim construction debate, the context of the surrounding words of the claim also must be considered in determining the ordinary and customary meaning of those terms."). Thus, the ordinary meaning of "normal plunger performance" based on the above-noted dictionary definition is performance that is standard or regular and not in need of correction.

The written description is consistent with and supports this construction in the context of this invention. Specifically, the patent describes the operation of the plunger in terms of its arrival time following each round-trip cycle. '991 patent, col. 10, ll. 6-9. If the plunger arrives or surfaces at times that are considered appropriate in terms of effective well production, no change or correction is made. Id. at ll. 14-16. Operation during those times is described as operation customarily considered by those skilled in this art to be within a "good window." Id. at ll. 6-9. If, however, the plunger arrives outside of the good window, adjustments are made to the system to alter the plunger speed and return it to the desired good window of operation. Id. at ll. 17-23. The district court thus concluded from the written description that "normal plunger performance" as used in the claims refers to the described good window of plunger arrival, or performance customarily considered by those skilled in the art not in need of correction. We agree.

The district court recognized that normal plunger performance was performance during a good window of operation, as defined by the upper and lower values assigned to trigger correction of plunger arrival time. While there is nothing in the language of the claims, the written description, or the prosecution history to specify that the upper and lower values be separately settable, the district court implicitly read such a capability as being required for there to be infringement. Because the upper limit of the APC 1000 window is fixed to be one minute greater than the lower limit and cannot independently be set to any other value, the district court incorrectly concluded that the window bounded by TT and TT + 1 "does not per se represent `normal plunger performance.'" Special Master's Report at para. 159 (emphasis in original). The district court's conclusion was in error both because it was influenced by the structure and function of the alleged infringing device, see SRI Int'l v. Matsushita Elec. Corp. of Am., 775 F.2d 1107, 1118 (Fed. Cir.1985) (en banc) (stating that claims may not be construed with reference to the accused device), and because it read into the claim a limitation of independent adjustability of the upper and lower values not specified or required by the claim language, as properly construed. Simply stated, the ordinary and customary meaning reflected in the context of the written description of the claim expression "normal plunger performance" is simply plunger performance not in need of correction. Performance not in need of correction is defined by arrival times within the values set by the upper and lower limits, regardless of whether those limits are independently input by an operator or are separated by a fixed, preset time period automatically calculated once one of the limits is input by an operator.

As to the term "predetermined plunger performance," again the initial focus is on the word "predetermined." The ordinary meaning of "predetermine" is "to determine beforehand." Webster's, supra, at 1786. This is the meaning consistent with that found by Special Master Williams in construing the separate phrase "predetermined ... time increment," Markman Report at 23, and we conclude that it is also the meaning that should apply to the phrase "predetermined plunger performance."

The written description is consistent with this construction, noting that an operator may set a window for operation beforehand. '991 patent, col. 10, ll. 2-6. We disagree with the district court's implicit conclusion that "predetermined plunger performance" necessarily has the same construction as "normal plunger performance." Predetermined in this context simply means determined beforehand, and may include performance that is normal or subject to correction. The coincidence of a particular predetermined window with a good window is irrelevant. There is no requirement, either in the claim itself or in the written description, that the values selected beforehand in the limitation "predetermined plunger performance" must represent the good window, as with "normal plunger performance." There is also nothing in the claim language, written description, or prosecution history to require that the upper and lower values that define the window of "predetermined plunger performance" be set independently of each other. The values must just be determined beforehand.

With these constructions in mind, we now review the district court's determination of non-infringement by the APC 1000 device. The district court held that version 3 of the APC 1000 device could not infringe the '991 patent because "the TT + 1 value [determined by the APC 1000 device] does not bear any relationship to `normal plunger performance.'" Special Master's Report at para. 159.3 As noted above, the correct construction of the term "normal plunger performance" is performance of the plunger that does not require correction, or arrival of the plunger within the good window.

In version 3 of the APC 1000, the operator enters a target time (TT) and the system generates a one-minute window bounded by TT and TT+1. If the plunger arrives within that one-minute window, no correction is performed. The district court made much of the testimony given by Mr. Bartley that a well operator would select a good window of one minute only for a specific depth of well, particularly a 2,400 foot well. Id. For this reason, the district court concluded that "the TT+1 value does not bear any relationship to `normal plunger performance,'" thus foreclosing a finding of infringement. Id.

The district court's conclusion is erroneous for two reasons. First, the fact that the TT+1 limit is tied to the TT limit makes it no less a "value" that, together with limit TT, set the bounds of plunger performance not needing correction, i.e., "normal plunger performance." Second, the fact that testimony was given that a one-minute window would be chosen only for wells of 2,400 feet gives an unwarranted connotation to the term "normal." The plunger performance that a well operator may desire for a well of a particular depth is a matter of subjective choice; the claim language is unconcerned with that choice. As construed by this court, the term "normal" means not requiring correction. Therefore, it does not matter when or why Mr. Bartley would choose to use a one-minute window. If a window of one minute is chosen, for whatever reason, and correction is made whenever the plunger does not arrive in that window, then that window is, by definition, normal plunger performance.

We conclude that the district court correctly construed the claim term "normal plunger performance" to mean plunger performance not requiring correction, i.e., the good window of plunger performance, but incorrectly ascribed the same meaning to "predetermined plunger performance." We also conclude that the district court erroneously required the two values defining the window of normal plunger performance and the two values defining the window of predetermined plunger performance to be independently settable, and further erroneously determined that, since a one-minute window would only be chosen given certain well parameters, the value TT+1 bore no relationship to normal plunger performance. For these reasons, the district court clearly erred in determining that a finding of infringement by the APC 1000 device was barred. Therefore, we vacate the district court's determination of non-infringement as to the asserted claims of the '991 patent and remand the case for further proceedings consistent with the above analysis.

Ferguson next argues that the district court erred, as a matter of law, in determining that Bartley was not personally liable for inducing Mega to infringe the '376 and '991 patents. In particular, Ferguson argues that the district court applied the incorrect legal standard for determining personal liability. Mega argues that the standard that the district court applied was indeed appropriate and that the conclusion that Bartley is not liable for inducing infringement is correct. We agree with Mega.

"[W]hoever actively induces infringement of a patent shall be liable as an infringer." 35 U.S.C. § 271(b) (2000). The district court noted that to be found liable under § 271(b), "a patentee must show that the individual charged with inducement took actions that actually induced infringement and that such individual knew or should have known that such actions would induce direct infringement," citing Micro Chemical Inc. v. Great Plains Chemical Co., 194 F.3d 1250 (Fed.Cir.1999). Special Master's Report at para. 175. The district court then concluded that the evidence proffered by Ferguson did not show that Bartley "knew or should have known his actions would induce infringement," regardless of how actively Bartley was involved with the development and/or marketing of the APC 1000 device. Id. at para. 179.

Ferguson argues that the district court should have applied a less stringent test for inducement liability, requiring that the officer be aware only of his activities, not necessarily aware that his activities amounted to infringement. In support of this less stringent test, Ferguson cites a number of district court cases that are not binding on this court. Not only are the cases cited by Ferguson not binding on this court, they are contradicted by our precedent. See, e.g., Micro Chem., 194 F.3d at 1261 ("Officers of an allegedly infringing corporation can be held personally liable for actively inducing infringement under 35 U.S.C. § 271(b) only if they `knew or should have known [their] actions would induce actual infringements.'" (quoting Manville Sales, 917 F.2d at 553)). Ferguson's position is untenable based on Manville Sales, which makes clear that "it must be established that the defendant possessed specific intent to encourage another's infringement and not merely that the defendant had knowledge of the acts alleged to constitute infringement." 917 F.2d at 553.

Because the district court applied the correct test for determining liability for inducing infringement, and because Ferguson submitted no evidence to show that Bartley "knew or should have known his actions would induce infringement," we affirm the district court's determination that Bartley was not personally liable for inducing Mega to infringe the '376 patent.

Ferguson next argues that the district court erroneously denied Ferguson the right to pursue its claims of willful infringement. Specifically, Ferguson asserts that the district court was wrong in not permitting the plaintiffs to amend the complaint to include an allegation of willful infringement. Our review of the denial of a motion to amend a complaint is determined by the regional circuit's standard which, in the Fifth Circuit, is abuse of discretion. Norman v. Apache Corp., 19 F.3d 1017, 1021 (5th Cir.1994); Avatar Exploration, Inc. v. Chevron, U.S.A., 933 F.2d 314, 320-21 (5th Cir.1991).

In its initial complaint, Ferguson averred that Mega's infringement of the Ferguson patents-in-suit was willful. Mega counterclaimed with allegations of inequitable conduct by Ferguson in the prosecution of the Ferguson patents and with assertions of willful infringement by Ferguson of the Mega patent. The district court reviewed the pleadings and concluded that the willfulness and inequitable conduct allegations did not meet the standards of Federal Rules of Civil Procedure 9(b), requiring that "[i]n all averments of fraud or mistake, the circumstances constituting fraud or mistake shall be stated with particularity." The district court then struck all claims and defenses of willfulness and inequitable conduct alleged by both parties. Fed.R.Civ.P. 9(b); see also Ferguson Beauregard/Logic Controls Div. of Dover Res., Inc. v. Mega Sys., LLC, 6:99cv437, slip op. at 9-10 (E.D.Tex. Aug. 14, 2000) ("Order Striking Affirmative Defenses"). Ferguson subsequently filed a motion to amend its complaint, setting forth the facts surrounding the allegations of willfulness in more detail. The district court, however, denied entry of the amended complaint, stating "[t]he action alleged by [Ferguson] to be evidence of willful infringement by [Mega's] counsel fails to hit the mark." Ferguson Beauregard/Logic Controls Div. of Dover Res., Inc. v. Mega Sys., LLC, 6:99cv437, slip op. at 4 (E.D.Tex. Mar. 22, 2001) ("Order Denying Entry of Second Amended Complaint").

We disagree. Willfulness does not equate to fraud, and thus, the pleading requirement for willful infringement does not rise to the stringent standard required by Rule 9(b). See, e.g., Vulcan Eng'g Co. v. Fata Aluminium, Inc., 278 F.3d 1366, 1378 (Fed.Cir.2002) (stating that willful infringement is a deliberate disregard of another's patent rights); McNeil-PPC, Inc. v. L. Perrigo Co., 337 F.3d 1362, 1371 (Fed.Cir.2003) (separately enumerating willful infringement and fraud); Fed.R.Civ.P. 9(b) ("In all averments of fraud or mistake, the circumstances constituting fraud or mistake shall be stated with particularity."). Because Ferguson's willful infringement allegation should not have been stricken in the first instance under Rule 9(b) as a matter of law, we find the refusal to allow an amended complaint to reintroduce willful infringement to be an abuse of discretion. The district court's dismissal of Ferguson's willful infringement allegation is reversed and the district court is instructed on remand to give Ferguson an opportunity to amend its complaint to plead willfulness by Mega with respect to the '376 patent. Our action on remand should not be construed as an opinion on the merits of Ferguson's willful infringement allegation, which we expressly decline to address.

Ferguson finally argues that the district court committed legal error in declining to consider Ferguson's claim that Mega engaged in inequitable conduct when it successfully revived the '048 patent after it lapsed for failure to pay maintenance fees. Maintenance fees are required to be paid at 3 1/2 years, 7 1/2 years, and 11 1/2 years, each with a six month grace period, after the grant of a patent to keep the patent in force. 35 U.S.C. § 41(b). The PTO has the authority to accept payment of maintenance fees up to twenty-four months following the six month grace period "if the delay is shown to the satisfaction of the Director to have been unintentional." Id. § 41(c).

The '048 patent issued on May 1, 1990 — with the 7 1/2 year maintenance fee thus coming due on November 1, 1997, or within the following six month grace period. See 35 U.S.C. § 41(b); Special Master's Report at para. 241. The maintenance fee was not timely paid, and the '048 patent lapsed accordingly. Special Master's Report at para. 241. On August 9, 1999, a petition to revive the '048 patent was filed, in which it was asserted that the delayed payment of the maintenance fee was unintentional. Id. at para. 244. The PTO granted the petition on or around September 7, 1999, but noted "[i]t is not apparent whether the person signing the statement of unintentional delay was in a position to have firsthand or direct knowledge of the facts and circumstances of the delay at issue." Id. The PTO further stated that "[n]evertheless, such statement is being treated as having been made as the result of a reasonable inquiry into the facts and circumstances of such delay," and requiring that "[i]n the event that such an inquiry has not been made, petitioner must make such an inquiry." Id.

In the district court, Ferguson did not allege inequitable conduct, but merely asserted that the revival of the '048 patent was improper. Id. at para. 243. In particular, Ferguson argued that the petitioner did not have firsthand knowledge of the facts and circumstances surrounding the late payment, and further, that no inquiry was made. Id. at paras. 245-247. Ferguson also asserted that the late payment was not unintentional. Id. at para. 248.

The district court noted that Ferguson had not pled inequitable conduct, id. at para. 251, and concluded that it had no authority to take action based on the allegedly "improper" revival of the '048 patent. Id. at para. 252. The district court's determination that it had no authority to take action in this case is essentially a conclusion that Ferguson has failed to state a claim upon which relief may be granted. See Fed.R.Civ.P. 12(b)(6). This court reviews the dismissal of a claim under Rule 12(b)(6), a matter of procedure, by applying the law of the regional circuit. C&F Packing Co., Inc. v. IBP, Inc., 224 F.3d 1296, 1306 (Fed.Cir.2000). The Fifth Circuit reviews such a dismissal de novo. See Vander Zee v. Reno, 73 F.3d 1365, 1368 (5th Cir.1996); Eason v. Holt, 73 F.3d 600, 601 (5th Cir.1996).

Ferguson argues that the district court should have entertained the "improper revival" claim by inferring inequitable conduct. However, in contrast to the willfulness claim discussed above, inequitable conduct, while a broader concept than fraud, must be pled with particularity. Despite the dicta provided by the district court that "such a conclusion [of inequitable conduct] would be relatively easy based on the present record," Special Master's Report at para. 251, we decline to infer facts to support a claim that must be pled with particularity.

We affirm the district court's decision declining to address the alleged inequitable conduct issue because it was not properly raised.

On cross-appeal, Mega first argues that the district court erred in construing the "valve means" claim limitation in independent claims 1, 16, and 35 of Ferguson's '376 patent. Although the parties did not submit this limitation to Special Master Williams for construction, the district court determined that this limitation was a means-plus-function limitation, governed by 35 U.S.C. § 112, paragraph 6, based on the construction of a similar term in the '721 patent. Special Master's Report at paras. 128-130. The district court then identified the structure for performing the claimed functions of being "responsive to said actuation signals" and "deriv[ing] said control inputs," as the valve shown in Figure 19 of the '376 patent. Id. at paras. 131-132. The valve identified by the district court is "described and claimed in the ['721 patent]." '376 patent, col. 36, ll. 25-27. The '721 patent describes this valve as a pneumatically-actuated valve.

Although couched as an error of claim construction, Mega's argument with respect to the "valve means" of the '376 patent is centered around the factual comparison between the claimed invention and the accused product, and specifically, between the pneumatically-actuated valve described in the patent and the solenoid-actuated valve present in Mega's APC 1000. See Cybor Corp., 138 F.3d at 1467 ("The second step of the infringement analysis requires a factual comparison of the claimed invention to the accused device, which is done by the fact finder."). Mega argues that the solenoid-actuated valve of the APC 1000 operates in a substantially different way to achieve a substantially different result, and thus, is not equivalent to a pneumatically-actuated valve. In support, Mega asserts that the testimony of Ferguson's expert, Mr. Quayle, does not describe "whether the APC 1000 valve performs the same function in substantially the same way to achieve substantially the same result" and moreover "is simply wrong." Mega also asserts that the solenoid-actuated valve of the APC 1000 cannot perform all of the functions described in the '376 patent, in that a solenoid-actuated valve cannot pressurize a container.

The district court found, however, and we agree, that there was sufficient evidence presented that the solenoid-actuated valve of the APC 1000 is equivalent to the pneumatically-actuated valve described in the patent. Mr. Quayle testified that the valve in the APC 1000 was equivalent to that of the "valve means" of the '376 patent in terms of § 112(6). Although Mr. Quayle noted there were differences, he explained why he considered the valves to be equivalent. Special Master's Report at paras. 134-135. Moreover, while Mega argues that the solenoid-actuated valve cannot perform all of the claimed functions, it points only to one function as being missing, namely pressurizing a container. But that is not a claimed function. The claimed functions of being responsive to actuation signals and deriving control inputs were both found by the district court to be performed. Based on this evidence, we find no clear error in the district court's conclusion of infringement. Therefore, we affirm the district court's construction of the term "valve means" and the finding of infringement of the '376 patent by Mega.

Mega also argues that the district court's award of damages for infringement of the '376 patent is in error. In particular, Mega asserts that the award of $100,000 reasonable royalty and the award of $837 per infringing controller for lost profits damages are not supported by substantial evidence.

Determining the amount of damages to award "is not an exact science, and the methodology of assessing and computing damages is committed to the sound discretion of the district court." State Indus., Inc. v. Mor-Flo Indus., Inc., 883 F.2d 1573, 1576-77 (Fed.Cir.1989). This court thus reviews the methodology for computing damages for an abuse of discretion and reviews the damage amount for clear error. Micro Chem. v. Lextron, Inc., 318 F.3d 1119, 1122 (Fed.Cir.2003).

With respect to the reasonable royalty award, Mega asserts that the evidence supporting the award of $100,000 was premised on a finding by the Special Master of infringement of both the '721 and '376 patents, Special Master's Report at para. 182, and is thus clearly erroneous. This argument is moot, however, in light of the district court's Order Adopting the Special Master's Report. In that order, the district court struck the reasonable royalty award and instead awarded Ferguson lost profits in the amount of $408,456 for infringement of the '376 patent. Order Adopting the Special Master's Report at 5.

With respect to the lost profits award, Mega argues that the award was erroneously based on evidence concerning sales of Ferguson's AutoCycle device. While the parties do not dispute that Ferguson's AutoCycle device does embody the '376 patent, found infringed, it also embodies the algorithm that forms the basis for the '991 patent, found not infringed by version 3 of the APC 1000. Thus, Mega argues, the lost profits should not have been based on sales of the AutoCycle device, but instead should have been based on the sales of Ferguson's LiquiLift device, which embodies only the '376 patent. To recover lost profits damages for patent infringement, a patentee "must show that it would have received the additional profits `but for' the infringement." See, e.g., King Instruments Corp. v. Perego, 65 F.3d 941, 952 (Fed.Cir.1995). In this case, the district court based its lost profits award on evidence of sales of a device embodying features in addition to those present in the infringed '376 patent, namely, those features attributable to the '991 patent. The district court therefore failed to distinguish the allocation of profits that would have been made "but for" the infringement of the '376 patent with the profits that could fairly be allocated to customer demand related to the features embodying the '991 patent. Cf. id. at 953 (taking into account differences between the product embodying the infringed patent and the accused product, where the accused product had additional features not related to the patented features); Rite-Hite, 56 F.3d at 1549 (allowing recovery of damages based on the value of a patentee's entire apparatus, containing several features, when the patent-related feature is the basis for the customers' demand).

Because the district court's award of damages for infringement of the '376 patent is based on clearly erroneous factfinding, we vacate the damages award and remand for further proceedings, consistent with this opinion.

Mega next argues that the district court erred in finding Mega's '048 patent not infringed by Ferguson. In particular, the district court held that the '048 patent required an adjustment to be made to the length of the off-time interval based on two claim limitations—"decreasing the length of the [off-time] period for the next cycle of the well" and "changing the value of either the off-time or the exhaust-time... on each successive cycle." Special Master's Report at paras. 231-232 (For example, "Claim 6 is thus viewed as plainly requiring an adjustment in `off-time' ...." (emphasis added)). Mega argues that the district court's infringement analysis of the '048 patent violates this court's mandate in Bell Communications Research, Inc. v. Vitalink Communications Corp., 55 F.3d 615 (Fed.Cir.1995). In that case, this court vacated a grant of summary judgment of non-infringement because the accused device sometimes, but not always, performed a limitation of the claimed invention. On remand, the court stated that "any future infringement analysis ... should be undertaken with due attention to the principle that an accused product that sometimes, but not always, embodies a claimed method nonetheless infringes." Id. at 622-23.

According to Mega, the district court's infringement analysis does not take into account the evidence that Ferguson's device does, on occasion, perform the step of adjusting the off-time. Mega's argument fails, however, due to the plain language of the claim. This is not a case, like Bell Communications, where the accused product sometimes accomplished the entire method of the claim. In order to accomplish the entire method of claims 6 and 35 of the '048 patent, an adjustment must always be made. Thus, if Ferguson's device does not always make a correction, as the district court found it did not, Special Master's Report at para. 233, then it never accomplishes the entire method of the claimed invention. The district court's determination of non-infringement is not clearly erroneous, and we affirm its decision on this issue.

Mega finally argues that the district court erred in prohibiting Mega from introducing evidence showing invalidity of the '991 patent and the existence of an ongoing reexamination of this patent based on evidence of the sale, by another prior to the filing of the application, of a controller using the same algorithm. The district court excluded this evidence based on Mega's failure to give notice under 35 U.S.C. § 282. That statute states in relevant part that for "actions involving the validity ... of a patent, the party asserting invalidity ... shall give notice in the pleadings or otherwise in writing to the adverse party at least thirty days before the trial...." Mega argues that it presented the evidence supporting its claim of invalidity in the filing for reexamination and also in its filing of an "Advisory to the Court re: Patent Re Examination" [sic]. Further, Mega claims Ferguson was on notice because it participated in discovery, including a deposition and interrogatories, related to the prior art that Mega sought to introduce. Thus, Mega argues that Ferguson was served with multiple notices of Mega's intent to assert invalidity of the '991 patent and that the district court therefore erred.

Evidentiary rulings are reviewed under the criteria of the regional circuit, unless the evidentiary issue is unique to patent law or litigation. See ATD Corp. v. Lydall, Inc., 159 F.3d 534, 548 (Fed.Cir.1998). Because 35 U.S.C. § 282 is unique to patent law, our precedent informs that we review this issue for abuse of discretion. Eaton Corp. v. Appliance Valves Corp., 790 F.2d 874, 879 (Fed.Cir.1986). The district court's rulings, rather than being contrary to law, followed the letter of § 282 precisely. The district court simply declined to allow Mega to offer up actions not described in the statute as sufficient notification. On the record before us, we cannot say that the district court abused its discretion. The district court's evidentiary rulings on this issue are affirmed.

AFFIRMED-IN-PART, REVERSED-IN-PART, VACATED-IN-PART, AND REMANDED COSTS.

RADER, Circuit Judge, concurring.

This court often uses the term "ordinary and customary meaning." While the "ordinary" meaning, often represented by the first listing in a reputable dictionary, can occasionally have relevance to construing terms in a patent claim, this court's case law requires primary reliance on the "customary" meaning. The "customary meaning" of a term in a patent claim links the inquiry to the understanding of one of ordinary skill in the art at the time of invention. Multiform Desiccants, Inc. v. Medzam, Ltd., 133 F.3d 1473, 1477 (Fed. Cir.1998) (stating that the "inventor's words that are used to describe the invention — the inventor's lexicography — must be understood and interpreted by the court as they would be understood and interpreted by a person in that field of technology"). Thus, I concur to highlight the chain of reasoning that links this court's use of the dictionary in this case to the proper context of this invention.

The word "normal" that appears in this case has a different "customary" meaning to a psychiatrist in a mental hospital (where "normal" refers to a sane mental state) than to a pharmacist seeking proper dosages (where "normal" doses would vary with body weight and other factors), than to a marriage counselor (where a "normal" marriage probably means an average marriage, i.e., one perpetually in danger of divorce), than to an oil well driller in a pump technology. In other words, dictionaries often supply multiple definitions for a single term with those definitions varying from dictionary to dictionary and varying also over time as the language evolves. Without proper context in selecting a dictionary definition, a court can err by importing a limitation into patent claims from a dictionary as well as from a patent specification. See Rambus Inc. v. Infineon Techs. AG, 318 F.3d 1081, 1088-89 (Fed. Cir.2003). Thus, when a court relies on a dictionary definition, it must include additional reasoning to substantiate its choice amongst many possible definitions in many possible dictionaries at many possible times.

This case illustrates a proper search for customary" meaning. This court does not simply select a definition, but instead supplies the reasons that this court chose a particular definition of "normal" in a particular dictionary at a particular time. This court properly links that definition to the specification's terminology and other indicia of the understanding of artisans of ordinary skill at the time of invention. This court chose a particular "customary" view of the word "normal" based on the usage of this word in this field at the time of invention.

Opinion for the court filed by Circuit Judge LINN. Concurring opinion filed by Circuit Judge RADER.