Jerome H. Lemelson v. The United States, the Bendix Corporation and Brown & Sharpe Manufacturing Company, Third-Party

BALDWIN, Circuit Judge. This is an appeal from a July 26, 1983 judgment of the United States Claims Court dismissing Lemelson’s complaint under 28 U.S.C. § 1498 and finding non-infringement of claim 1 of U.S. Patent No. 3,481,042 (the ’042 patent), claim 15 of U.S. Patent No. 3,226,833 (the '833 patent), and claim 12 of U.S. Patent No. 3,636,635 (the ’635 patent) by the government’s use of certain coordinate measuring machines (CMMs). We affirm in part, vacate in part, and remand. The judgment issued as a result of a joint motion of defendant United States and third-party defendants Bendix Corporation (Bendix) and Brown & Sharpe Manufacturing Company (Brown & Sharpe) at the close of the 26-day presentation of Lemelson’s case-in-chief. The motion was made under Claims Court Rule 41(b) solely on the ground that Lemelson had failed to establish infringement of the claims in suit. The Claims Court opinion, Lemelson v. United States, 3 Cl.Ct. 161 (1983), deals only with the infringement issue. Description of the patent claims in suit A. The ’833 patent Only method claim 15 of Lemelson’s ’833 patent, entitled “Automatic Inspection Apparatus and Method,” is at issue. The claim, directed to measuring the distance between two surfaces of a workpiece, states (paragraphing and labeling of claim steps supplied): A method of automatically measuring dimensions between surfaces of a workpiece comprising the steps of: (1) relatively prepositioning a workpiece and an automatic measurement device; (2) detecting a first surface of said workpiece by means of surface positional indicating means; (3) thereafter relatively moving said surface positional indicating means and said workpiece and automatically generating signals indicative of the degree of said movement; (4) detecting a second surface of said workpiece with said surface position indicating means when said second surface is aligned therewith; (5) integrating the signals generated during the movement of said workpiece and said surface detecting means between said first and second surfaces; and (6) generating a further signal indicative of the distance between said first and second surfaces. Operation of one embodiment of the claimed invention is illustrated with reference to Fig. 1 from the patent. Apparatus 10 moves along an overhead track 11 and/or a floor mounted track 21 for prepositioning its measuring head probe assembly 31 relative to the workpiece W. Base 9 may be a fixed work table or a conveyor adapted to bring the work into alignment with the measuring apparatus. A programming controller CO controls servo motors (e.g., Mx, Mz, MA2) which preposition the measuring head over the workpiece prior to the beginning of the measuring steps. This description of prepositioning corresponds to step 1 of the claim. For step 2, a surface-sensing means *1542 such as contact probe SWi on jaw 32 detects a first surface of the workpiece. The point at which the contact probe S.Wi engages the surface of the workpiece is the reference point from which measurements are taken. In step 3, the second surface-sensing means (contact probe SW2 on jaw 33) is advanced toward the workpiece causing electrical pulses to be generated which represent incremental units of distance traveled by the probe SW2. The pulses continue to be generated until the probe detects (e.g., contacts) the workpiece as in step 4. Meanwhile the pulses are being summed in a counter which sum will represent the total distance traveled by probe SW2. The final step 6 requires that a further signal be generated representing the dimension D — the distance between the first and second surface. This would be calculated by subtracting the distance traveled by probe SW2 from the original distance between the jaws prior to step 3. B. The ’635 patent Only method claim 12 of Lemelson’s ’635 patent, entitled “Automatic Measuring Apparatus,” is at issue. The claim is directed to measuring dimensions and presenting the measurement obtained in a visually readable form. The ’635 specification and drawings are substantially similar or identical to those of the ’833 patent. Claim 12 states (paragraphing and labeling of claim steps supplied): A method for automatically measuring dimensions comprising: (1) predeterminately locating a member containing dimensions to be measured and a surface-sensing probe; (2) relatively moving said probe and said surface; (3) generating pulse signals with said relative movement and feeding said pulsed signals to a summing means; (4) summing said pulse signals derived from said relative movement; (5) indicating when said probe senses the the [sic] surface by generating a control signal; (6) applying said control signal to control the summing of said pulse signals whereby the pulse signals generated at the time said surface is sensed by said probe are presented in the form of digital information on the output of said summing means; and (7) applying said digital information signals to activate a visual presentation means to visually indicate the distance travelled by said probe in intelligible form. The first five steps are similar to the steps already described for claim 15 of the ’833 patent. In step 6, the summing device outputs the sum in digital format. In step 7, the sum is visually presented to indicate the distance traveled by the probe. C. The ’042 patent The ’042 patent is a continuation-in-part of, inter alia, the application which matured into the ’833 patent. Only apparatus claim 1 of Lemelson’s ’042 patent, entitled “Surface Sensing Apparatus,” is at issue: Surface measuring apparatus comprising: (1) means for sensing a surface to be measured; (2) manipulation means for said sensing means; (3) a variable program control means for controlling the operation of said manipulation means and positioning said sensing means adjacent a surface to be measured, said program control means including: (a) a record member; and (b) means for reproducing positional control signals from said record member; (4) means for using said positional control signals to preposition said sensing means; (5) means for relatively positioning said workpiece and said manipulation means; (6) signal generating means associated with said manipulation means and said sensing means for obtaining signals *1543 representative of the location of a surface of said workpiece relative to said manipulation means; and (7) recording means for recording said signals representative of the surface being measured on predetermined areas of said recording member. Figures 1, 3, 4, and 5 are reproduced below: *1544 The sensing means of element 1 is depicted in Fig. 1 by the probe sensor 23. Manipulation apparatus 11 (including 11A and 11B) of Fig. 4 or base 12 of Fig. 1 correspond to the manipulation means of emment 2, although Lemelson disputes this, *1545 see infra. The manipulation means prepositions the sensing means in alignment with the workpiece W. Element 3, a variable program control means, is depicted in Fig. 5 as a computer or positional controller 49. The means of element 4 which use the positional control signals include the digital positional controller 57 and servo motors 63, 18, and 43 of Fig. 5. The means of element 5 is described in Fig. 1 as work holding feature WP, which is capable of retaining and prepositioning the workpiece W. Once prepositioning is complete, measurement can begin. Element 6’s signal generating means is shown in Fig. 5. Movement of arm 22 causes the pulse generator 55 to generate pulses representing the distance traveled by the probe 25 relative, for example, to the stationary base plate of Fig. 4. The pulses are summed by pulse counter 53. The recording means of element 7 records the signals from the pulse generator onto a channel of recording member 51 (e.g., magnetic tape). The Accused Devices and Methods Lemelson introduced evidence at trial relating to six CMMs: the IOTA 2204 P and IOTA 1205 DNC made by Digital Electronic Automation Inc. (DEA), the Validator 200 and Validator 300 made by Brown & Sharpe, and the Cordax 803 DCC and Cordax 5000 DCC made by Bendix. Some of the CMMs are “bridge”-type machines and others are of the “cantilever”-type as illustrated below: Each of the accused CMMs has a worktable to support the workpiece. Mounted upon the worktable is either a bridge suspended by two struts or a cantilever beam, capable of moving the length (the x-axis) of the table. A carriage is mounted on the bridge or beam so that the carriage can move across the width (the y-axis) of the table. Mounted in the carriage is a verticle shaft designed to move vertically up and down (the z-axis) over the worktable. The probe sensor is attached to the end of the vertical shaft. Movement of the probe along the x, y, and z axes is motorized and computer controlled. According to advertising for the accused devices, the computer may be programmed to automatically select the probe’s path or its next position, or an operator may control movement with a pair of electronic joy sticks. Using joy stick control, the operator can navigate the probe to any point on the surface of the workpiece. When the probe senses the workpiece, it causes a signal to be sent to the computer controller which records the probe’s position. To define a position, the accused CMMs use a three-dimensional co *1546 ordinate system with a reference origin in which the position is expressed by an x, y, and z coordinate. This coordinate position is stored in memory. The probe may then be moved to another point on the workpiece, where a new coordinate position can be recorded. The distance between the two points (e.g., a dimension of the workpiece) along one of the coordinate axes is calculated by subtracting the relevant coordinate of the one point from the relevant coordinate of the other point. To establish infringement of the method claims of the ’833 and ’635 patents, Lemelson put in issue a single specific operating mode for each of the accused devices. On appeal, Lemelson has focused on the operation of the Validator 200 in measuring the wall thickness of a flanged cylinder. The drawing below, made by a witness familiar with the operation of the Validator 200, describes the measurement mode. The rectangular blocks represent the probe at various positions during measurement. The hatched figure represents a partial cross-section of the flanged cylinder (the left surface is the inside of the cylinder, the right surface is the flanged exterior) to be measured. The operator moves the probe to the various positions using, inter alia, a series of “VMOVE” and “VMEAS” computer programmed commands. He begins measurement by calling VMOVE which moves the probe from an initial position at block 1 to block 2. From the position indicated by block 2, the probe is moved to the right in a positive x direction to block 3. With the VMEAS command the probe then contacts the surface of the cylinder and the coordinate position is recorded. Next, the operator backs the probe away from the surface slightly and then with another VMOVE command moves the probe up in a positive z direction clearing the top of the cylinder to block 4. Then the operator moves the probe in a positive x direction to block 5 and then down in a negative z direction to block 6. Finally, the operator moves the probe in a negative x direction to block 7 and then in contact with the exteri- or surface of the cylinder for recording another coordinate position. The wall thickness is then calculated by taking the difference between the x coordinates. Claims Court Decision The lower court held that Lemelson failed to establish, by a preponderance of the evidence, that the inventions claimed in the three patents were used by. or for the United States. The court found that neither claim 15 of the ’833 patent nor claim 12 of the ’635 patent is infringed by any of the accused methods. In construing the method claims, the court found that the first step in each requires prepositioning of the workpiece to be automatic, whereas in the accused methods, an operator manually places the workpiece onto the worktable. Claim 12 of the ’635 patent was found not infringed for the additional reason that the accused methods do not measure and therefore do not visually display the distance traveled by the probe. In the accused methods, the probe can follow any path between the points on a workpiece, whereas a probe executing claim 12 must follow a straight line coaxial to the dimension being measured. As to claim 1 of the ’042 patent, the court determined that the manipulation means of step 2 is directed only to prepositioning of the probe. On the accused devices, how *1547 ever, the workpiece is placed manually onto the worktable, thus, a manipulation means is both unnecessary and absent in the accused coordinate measuring machine. A manipulation means being absent in the accused devices, the court concluded that the accused CMMs did not infringe claim 1. On appeal, Lemelson asserts error in the court’s construction of the claims, refusal to consider the prosecution histories, interpretation of the accused devices, requirement that Lemelson show that the government actually used the accused devices, and exclusion of certain testimony and restrictive admission of certain documentary evidence. Issues 1. Whether the Claims Court erred in dismissing Lemelson’s complaint at the end of his case-in-chief. 2. Whether the Claims Court erred in construing the claims in suit. 3. Whether the Claims Court clearly erred in finding that the accused devices and methods did not infringe the claims in suit. OPINION I. Standard of Review Lemelson has the burden to establish that the Claims Court’s ultimate finding of non-infringement was clearly erroneous, see, e.g., Raytheon Co. v. Roper Corp., 724 F.2d 951, 956, 220 USPQ 592, 596 (Fed.Cir.1983), cert. denied, — U.S. -, 105 S.Ct. 127, 83 L.Ed.2d 69 (1984), and that the court’s claim construction was erroneous, SSIH Equipment S.A. v. USITC, 718 F.2d 365, 376, 218 USPQ 678, 688 (Fed.Cir.1983). These burdens are not affected by the fact that the court below dismissed Lemelson’s complaint in response to a rule 41(b) 1 motion after the close of Lemelson’s casein-chief. Unlike in a rule 50 motion for a directed verdict where the trial judge considers the evidence in a light most favorable to the plaintiff (i.e., to determine if a prima facie case has been established), the trial judge in passing upon a rule 41(b) motion evaluates and resolves conflicts of evidence and credibility. Stearns v. Beckman Instruments, Inc., 737 F.2d 1565, 1567-68, 222 USPQ 457, 459 (Fed.Cir.1984). Findings entered as a result of a rule 41(b) motion are therefore reviewed under the same clearly-erroneous standard as are findings entered at the close of all the evidence. Woods v. North American Rockwell Corp., 480 F.2d 644, 645-46 (10th Cir.1973). II. Burden of Proof in the Claims Court At trial, Lemelson had the burden of proving infringement by a preponderance of the evidence. Hughes Aircraft Co. v. United States, 717 F.2d 1351, 1361, 219 USPQ 473, 480 (Fed.Cir.1983). This burden extends to infringement under the doctrine of equivalents as well as literal infringement. Lemelson brought this action under 28 U.S.C. § 1498 2 seeking just compensa *1548 tion for unauthorized use and manufacture by or for the United States of the accused CMMs. The Claims Court held that the proof required for liability under section 1498 was substantially synonymous with the proof required for a claim of patent infringement under 35 U.S.C. §§ 271 and 281. The court required plaintiff to prove that the patent claims in suit were infringed under the following standard: [I]t is incumbent upon plaintiff to adduce evidence that the Government actually used the accused devices in an infringing manner, or practiced the precise methods claimed. Merely proffering evidence that the measuring devices procured by the Government might be operatively assembled in an infringing mode, or that the machines are capable of executing the claimed methods, does not prove infringement. Lemelson disagrees with the court’s assignment of the burden of proof, contending that the court erred in requiring him to prove actual use of an accused devices in order to prove infringement. In support he cites cases from one of our predecessors, the Court of Claims, which held that “use” under section 1498 could be found even where the government, although not actively using an accused device, had the device available for use. See, e.g., Olsson v. United States, 25 F.Supp. 495, 497, 37 USPQ 767, 769 (Ct.Cl.1938), cert. denied, 307 U.S. 621, 59 S.Ct. 792, 83 L.Ed. 1500 (1939) (howitzers disassembled only for storage purposes, but ready for use should the need arise, constitutes “use”). We reject Lemelson’s argument. Although this court has noted that a section 1498 action and a title 35 action are only parallel and not identical, Motorola, Inc. v. United States, 729 F.2d 765, 768, 221 USPQ 297, 299 (Fed.Cir.1984) (section 1498 does not incorporate 35 U.S.C. § 287), the principles of claim construction and reading claims on accused devices and methods are the same for either type of action. We think the issue properly posed, however, is not whether the infringement analysis for a section 1498 action differs from that of a title 35 action, but rather, whether the burden of proving the structure of an accused device (or its mode of operation) in a section 1498 action differs from the burden in a title 35 action. We conclude that the burden is the same. Olsson and its progeny do not stand for the proposition that infringement for section 1498 purposes is met simply by showing that parts delivered to the government are capable of being configured in an infringing manner. In Olsson, the accused howitzers were completely assembled and later disassembled for storage purposes. There was no question that the assembled howitzers embodied the claimed invention and therefore infringed. Thus, all Olsson held on this point was that an infringing device available for use, like a device actually in use, constitutes infringing use under section 1498. Generally, infringement can occur ony when the claimed combination has been assembled and is used or is available for use. Decca Ltd. v. United States, 640 F.2d 1156, 1168, 209 USPQ 52, 61 (Ct.Cl.1980), cert. denied, 454 U.S. 819, 102 S.Ct. 99, 70 L.Ed.2d 89 (1981); cf. Paper Converting Machine Co. v. Magna-Graphics Corp., 745 F.2d 11, 19-20, 223 USPQ 591, 597 (Fed.Cir.1984) (infringement found in the testing of significant, unpatented assemblies of elements, which enabled the infringer to deliver the patented combination in parts to the buyer, although the entire patented combination together was not tested). In the present case, Lemelson failed to meet his burden, not because he failed to prove that the government “used” the accused devices, but because he failed *1549 to prove that the accused devices embodied the claimed invention. As to apparatus claim 1 of the ’042 patent, the trial court found non-infringement on the basis that certain admittedly present physical components of the accused CMMs did not correspond to claim l’s manipulation means. As to method claim 15 of the ’833 patent and method claim 12 of the ’635 patent, the government did not admit to the modes in which the accused CCMs had been operated or to the peripheral equipment including computer software used in conjunction with the CMMs. In these circumstances, the trial court was correct in requiring Lemelson to prove the manner in which the CMMs were operated. For method claims, a plaintiff must prove the modes in which the accused devices have been operated and the manner in which those modes infringe the patent claims. See, e.g., Roberts Dairy Co. v. United States, 530 F.2d 1342, 1354-55, 198 USPQ 383 (Ct.Cl.1976). III. Infringement Proper claim construction necessarily precedes determining whether the claims in suit read on the accused devices. See, e.g., Fromson v. Advance Offset Plate, Inc., 720 F.2d 1565, 1569, 219 USPQ 1137, 1140 (Fed.Cir.1983). Claim construction generally involves examining the specification, prosecution history, and other claims. Id. at 1569-71, 219 USPQ at 1140-41. A. Claim 1 of the ’042 patent 1. Claim Construction Lemelson argues that the court erroneously construed claim 1, citing numerous errors of construction which admittedly do not bear on the dismissal for non-infringement. Even if we were to conclude that error was made as to these, we would not be compelled to reverse because of the harmless error doctrine. Gardner v. TEC Systems, Inc., 725 F.2d 1338, 1345, 220 USPQ 777, 782 (Fed.Cir.) (en banc), cert. denied, — U.S. -, 105 S.Ct. 116, 83 L.Ed.2d 60 (1984). The thrust of Lemelson’s attack on the Claims Court’s construction of claim 1 is on the meaning of element 2’s “manipulation means for said sensing means.” The court construed the manipulation means as directed only to that part of the apparatus which prepositions the probe prior to measurement (base 12 in Fig. 1 and manipulation arm 11 in Figs. 3 and 4), and not the apparatus parts which are involved in manuevering the probe during measurement. Lemelson argues that other disclosed elements are also necessary to “manipulate” the sensing probe, and cites, for example, elements 31, 33, 34, 40, 41, 43, and 44 (Fig. 3). We do not dispute that these elements “manipulate” the probe. But what the lower court determined, and with which we agree, is that the manipulation means of element 2 relates to prepositioning the sensing probe prior to measurement and not to manipulation of the sensing probe during measurement. This is supported by the language in the claim itself and the specification. Elements 2-4 concern the manipulation means and its function of prepositioning the sensing means of element 1. The driving means of element 4, under element 3’s computer control, operates the manipulation means of element 2. Elements 5-7 are concerned with the measuring function of the apparatus. The specification discloses embodiments of the manipulation means along with its function. For Figure 1, the specification states that “a positional relationship must be established between the carrier or manipulation means, such as the base 12 [of Fig. 1] or a mount on which said base is movable, and the work being measured.” Also, we read that by a combination of longitudinal, lateral, vertical, and rotary motions imparted to the mount 12, the probe assembly 15 may be “automatically positioned at substantially an infinite number of locations in space within a given region to effect predetermined positioning of the probe assembly.” With regard to Figure 3, a manipulation apparatus 11 is described as being able *1550 to position the assembly 30 relative to a workpiece or conveyor. 2. Prosecution Histories Lemelson asserts that the lower court committed reversible error by failing to construe the claims in light of their prosecution histories, citing SSIH Equipment S.A. v. USITC, 718 F.2d at 376, 218 USPQ at 688, where this court said that the prosecution history is always relevant to a proper interpretation of a claim. The court received the prosecution histories into evidence en masse near the beginning of the trial without objection by the defendants. However, in its decision, the court said that “the prosecution histories of the patents in suit and their implications have not been adduced” and that the claims would be construed without their benefit. The appellees suggest that Lemelson is presenting a de novo file history argument on appeal because (1) Lemelson never argued the file wrappers in his pre-trial brief or in his brief in opposition to defendants’ motion to dismiss, (2) Lemelson’s expert witness admitted at trial that he had not reviewed the file wrappers in preparing his infringement opinion, and (3) Lemelson did not refer to, mention, or use so much as one page of any file wrapper in presenting his case in chief. This court will not reach the appellees’ argument that Lemelson “waived” his right to seek review of the prosecution history issue because we are faced with an inadequate record on the point. We assume, therefore, that Lemelson properly raises the issue and we agree with Lemelson that the lower court erred in not considering the prosecution histories of the patents in suit. The prosecution histories being admitted into evidence, the court should have considered them in its construction of the claims. SSIH Equipment S.A. v. USITC, 718 F.2d at 376, 218 USPQ at 688; Fromson v. Advance Offset Plate, Inc., 720 F.2d 1565, 1569-71, 219 USPQ 1137, 1140-41 (Fed.Cir.1983); McGill Inc. v. John Zink Co., 736 F.2d 666, 673, 221 USPQ 944, 949 (Fed.Cir.1984). Nevertheless, after examining the prosecution history as argued by Lemelson on appeal for the ’042 patent in conjunction with claim 1 and the specification, we are unpersuaded that the lower court erred in interpreting the ’042 patent. The court’s failure to consider the prosecution history of the ’042 patent was, therefore, harmless error not meriting reversal. Gardner v. TEC Systems, Inc., 725 F.2d at 1350, 220 USPQ at 786. Lemelson’s arguments that clear factual errors, regarding operation and structure of the accused CMMs, infected the Claims Court’s conclusion of non-infringement of the ’042 patent are similarly insubstantial. Notwithstanding what appears to be the Claims Court’s incorrect labeling of and reference to the probe shaft on the accused devices, we do not see that the court thereby committed clear error in applying claim 1 to the accused CMMs. None of the asserted errors infect the finding that the manipulation means of claim 1, directed to prepositioning of the workpiece, is not present in the accused structures. 3. Doctrine of Equivalents Lemelson also argues that the asserted clear factual errors made it impossible for the lower court to determine infringement based on the doctrine of equivalents. Equivalence is a question of fact for which the burden of proof rested on Lemelson. Hughes Aircraft Co. v. United States, 717 F.2d at 1361, 219 USPQ at 480; Thomas & Betts Corp. v. Litton Systems, Inc., 720 F.2d 1572, 1579, 220 USPQ 1, 6 (Fed.Cir.1983). The doctrine allows a court to find infringement when the accused device and claimed invention perform substantially the same function in substantially the same way to yield substantially the same result. Graver Tank & Mfg. Co. v. Linde Air Products Co., 339 U.S. 605, 608, 70 S.Ct. 854, 856 (1950); Perkin-Elmer Corp. v. Computervision Corp., 732 F.2d 888, 900, 221 USPQ 669, 679 (Fed.Cir.), cert. denied, — U.S.-, 105 S.Ct. 187, 83 L.Ed.2d 120 (1984). *1551 It is also well settled that each element of a claim is material and essential, and that in order for a court to find infringement, the plaintiff must show the presence of every element or its substantial equivalent in the accused device. Interdent Corp. v. United States, 531 F.2d 547, 552 (Ct.Cl.1976); Autogiro Co. of America v. United States, 384 F.2d 391, 403, 155 USPQ 697, 707 (Ct.Cl.1967). Although the trial court incorporated Graver Tank’s tripartite test for equivalence in its general infringement test and did not make separate findings on equivalence, 3 we are nevertheless able to understand the court as finding neither a manipulation means nor its substantial equivalent in the accused devices. See ACS Hospital Systems, Inc. v. Montefiorse Hospital, 732 F.2d 1572, 1582, 221 USPQ 929, 936 (Fed.Cir.1984). Lemelson fails to point to any evidence showing a substantial equivalent for the manipulation means in the accused devices. We give no weight to the series of conclusory statements offered by Lemelson’s expert witness because, among other things, the statements do not even suggest that the accused devices possess a substantial equivalent for claim l’s manipulation means. B. The Method Claims 1. Claim 15 of the ’833 patent Claim 15 of the ’833 patent covers a method of measuring similar to that of claim 12 of the ’635 patent. The trial court concluded that both claims require automatic prepositioning of the workpiece and the measuring device. The court found that each accused method was similar to the others in that (1) the workpiece to be measured is manually prepositioned onto the worktable, and (2) the information displayed by the accused CMMs after completing a measurement is not indicative of the distance traveled by their probes. According to the court, these similarities distinguished the accused methods from the two method claims in suit. Lemelson contends that the court erred in construing the claims as requiring automatic prepositioning. We discuss this contention with reference to claim 15 of the ’833 patent. The court concluded that claim 15 is limited to automatic prepositioning of the workpiece and measurement device for three reasons: (1) the ’833 specification refers to automatic prepositioning in twelve instances whereas no mention is made to fully manual prepositioning; (2) the claim’s preamble refers to the method as one of “automatically measuring dimensions”; and (3) the parts of the specification which disclose manual prepositioning of either the workpiece or measuring device do not teach a completely manual prepositioning step because where manual prepositioning of the workpiece is disclosed, it is accompanied by automatic prepositioning of the measuring device, and where manual prepositioning of the measuring device is disclosed, it is accompanied by automatic prepositioning of the workpiece. We conclude that claim 15 is not limited to automatic prepositioning. We begin by examining the claim itself. When reading the preamble language (a “method of automatically measuring dimensions”) in conjunction with the language of step 1 (“relatively prepositioning a workpiece and an automatic measuring device”) we are not convinced that the “automatic” of the preamble modifies step l’s “relatively prepositioning” language. If automatic prepositioning were intended, we might expect to see the word “automatic” used with “prepositioning” just as “automatic” precedes “measurement device” in the same step. Because the word “automatic” expressly modifies another phrase in the same step, this supports not implying “automatic” for the phrase “relatively prepositioning.” We also note that the phrase “generating signals” found in step 3 is expressly modified by the term “automatically.” The proseeu *1552 tion history, as argued by Lemelson, does not add to our construction of the claim. The trial court placed great emphasis on the numerous instances of automatic prepositioning found in the specification. We emphasize, instead, that nowhere does the specification require automatic prepositioning. In Fromson v. Advance Offset Plate, Inc., this court cautioned against limiting the claimed invention to preferred embodiments or specific examples in the specification. 720 F.2d at 1568, 219 USPQ at 1139 (citing Smith v. Snow, 294 U.S. 1, 11, 55 S.Ct. 279, 283, 79 L.Ed. 721 (1935)). Even if the specification only discloses apparatus directed to executing automatic prepositioning of the workpiece or the measurement device or both, this does not dictate reading such a limitation into the prepositioning step of the claim. The trial court’s erroneous construction of claim 15, which was the sole basis for finding non-infringement, renders clearly erroneous that finding. Amstar Corp. v. Envirotech Corp., 730 F.2d 1476, 1481, 221 USPQ 649, 653 (Fed.Cir.1984). We therefore vacate the Claims Court’s finding of non-infringement as to claim 15 of the ’833 patent and remand for further proceedings consistent with this opinion. 2. Claim 12 of the ’635 patent Claim 12 of the ’635 patent was construed by the trial court as requiring, like claim 15 of the ’833 patent, automatic prepositioning of the workpiece and measuring device. The parties and the trial court agree that the specification and drawings of the two patents are substantially the same. Claim 12’s preamble (“automatically measuring dimensions”) and step 1 (“predeterminately locating a member containing dimensions to be measured and a surface-sensing probe”) are also agreed to be similar to those of claim 15. We see nothing in the prosecution history of the ’635 patent which adds to our construction of the claim. Accordingly, we construe claim 12, like claim 15, as not requiring automatic prepositioning. This error of construction does not mandate vacating the trial court’s finding of non-infringement because, unlike its treatment of claim 15 of the ’833 patent, the trial court based its finding of non-infringement of claim 12 on the additional ground that the accused methods, do not visually indicate the distance traveled by the probe. Crucial to understanding claim 12, according to the court, is that “the distance traveled by jaw 33 is actually being measured, and from that measurement, dimension D of workpiece W is calculated.” We agree with the court’s construction that claim 12 is limited to a method which visually indicates the distance traveled by the probe. Step 7 expressly includes this limitation: “applying said digital information signals to activate a visual presentation means to visually indicate the distance traveled by said probe in intelligible form.” We also agree with the court’s finding that the accused methods do not measure the distance traveled by the probes of the CMMs. On the CMMs, a workpiece dimension is calculated from the coordinate positions at which the probe contacted the workpiece surface. Thus, as shown by the sketch of the measuring operation for the accused Validator 200, supra, the probe can follow nearly any path between the points on a workpiece. By contrast, claim 12 requires the probe to follow a straight line coaxial to the dimension being measured. We are unpersuaded by Lemelson’s argument that step 7 of claim 12 should be construed to include a method which visually indicates the net distance traveled by the probe. The expressed language of the claim, the specification, and the drawings point to the contrary. Nor, by reason of the above, is there infringement under the doctrine of equivalents because the accused methods do not perform substantially the same function in substantially the same way to yield the same result. Therefore, we affirm the trial court’s finding that claim 12 is not infringed by the accused methods. IV. Evidentiary Rulings At trial, Lemelson sought to admit into evidence plaintiff’s proposed exhibit *1553 125 (PX-125) as the manual which describes Bendix’s Cordax 5000. The court restrictively admitted PX-125, allowing it to be relied upon only to prove potential operating modes of the accused CMMs, but not allowing it to be relied upon, absent corroborating evidence, to show that the accused devices were actually operated according to the methods set forth in the manual or to show the actual structure of the accused devices. Lemelson contends that the court erred in restrictively admitting PX-125 because the government and Bendix, in responding to certain pretrial requests for admission, admitted that the document was the service manual for the accused Cordax 5000. We agree with the court’s restrictive admission of PX-125. It is uncontradicted that PX-125 is a general service manual for at least three different Cordax CMMs and that much of the manual’s text, illustrations, and schematics refer to all Cordax models. Thus, the court properly required corroborating evidence linking the material in PX-125 with the specific accused device. The only material that Lemelson can point to as corroborating evidence are those same admissions of the government and Bendix in the request responses. Since those admissions are too general to merit unrestricted admission of PX-125, they certainly cannot constitute the necessary corroborating evidence. Lemelson also urges that the trial court erred in not admitting the testimony of Joseph Watson, a Bendix engineer, which was elicited under an offer of proof. In conclusory language, Lemelson asserts that Watson’s testimony was not hearsay and was relevant. The trial court’s opinion consolidates its discussion of offers of proof from several witnesses without mentioning the witnesses by name, but we assume that Watson’s proffered testimony is embraced by the discussion. The court noted that each of the offers of proof arose from objections granted on relevancy, foundation, and hearsay. The objections were sustained and the testimony was proffered subject to the stipulation that Lemelson submit memoranda controverting the grounds for objection. Although we see nothing in the record before us, Lemelson apparently submitted arguments for admissibility on the ground of relevancy. The court, however, held the proffered testimony inadmissible as being hearsay and lacking proper foundation because Lemelson only addressed the relevancy ground and “did not directly address, nor refate the sustained objections regarding hearsay and foundation defects.” We will not review the trial court’s exclusion of the proffered testimony. Lemelson’s appellate brief fails to comply with rule 28(e) of the Federal Rules of Appellate Procedure which provides in relevant part: If reference is made to evidence the admissibility of which is in controversy, reference shall be made to the pages of the appendix or of the transcript at which the evidence was identified, offered, and received or rejected. In his brief, Lemelson only cites the question and answer text of the proffered testimony. He does not refer us to (nor were we able to find) where in the record counsel made the offer of proof, and where counsel presented any reasons for admissibility. Since Lemelson does not dispute that the exclusion was grounded on relevancy, hearsay, and foundation, and he does not show us that he argued the hearsay and foundation ground to the court below, this court will not entertain his arguments. Conclusion We affirm that part of the judgment based on the Claims Court’s findings that the accused CMMs and methods do not infringe claim 1 of the ’042 patent and claim 12 of the ’685 patent. We vacate that part of the judgment based on the finding of noninfringement as to claim 15 of the ’833 patent and remand this part of the case for further proceedings consistent with this opinion. AFFIRMED IN PART, VACATED IN PART, AND REMANDED. Footnotes [1] Claims Court rule 41(b) provides in relevant part: Involuntary Dismissal: Effect Thereof____ After the plaintiff has completed the presentation of his evidence, defendant, without waiving its right to offer evidence in the event the motion is not granted, may move for a dismissal on the ground that upon the facts and the law the plaintiff has shown no right to relief. The court as trier of the facts may then determine them and render judgment against the plaintiff or may decline to render any judgment until the close of all the evidence. If the court renders judgment on the merits against the plaintiff, the court shall make findings as provided in Rule 52(a). Unless the court in its order for dismissal otherwise specifies, a dismissal under this subdivision ... operates as an adjudication upon the merits. This rule corresponds to rule 41(b) of the Federal Rules of Civil Procedure. [2] The section provides in relevant part: (a) Whenever an invention described in and covered by a patent of the United States is used or manufactured by or for the United States without license of the owner thereof or lawful right to use or manufacture the same, the owner’s remedy shall be by action against the United States in the United States Claims Court for the recovery of his reasonable and *1548 entire compensation for such use and manufacture. For the purposes of this section, the use or manufacture of an invention described in and covered by a patent of the United States by a contractor, a subcontractor, or any person, firm, or corporation for the Government and with the authorization or consent of the Government, shall be construed as use or manufacture for the United States. 28 U.S.C. § 1498(a) (1982). [3] If equivalence is in issue, we think the better practice is to consider equivalence apart from the issue of literal' infringement and to make separate findings of fact on equivalence.