OPINION
LYDON, Judge:In this patent case, plaintiff, Hazeltine Corporation, seeks “reasonable and entire compensation” pursuant to 28 U.S.C. § 1498 (1982) for the alleged unauthorized use by or for the United States of the invention disclosed and claimed in United States Patent No. 3,836,977, entitled “Antenna System Having A Reflector With A Substantially Open Construction” (hereinafter referred to as the Wheeler or ’977 patent). The ’977 patent, applied for on June 25, 1973, was issued to Harold A. Wheeler on September 17, 1974 and thereafter assigned to plaintiff.
Plaintiff asserts that the ’977 patent covers certain antennas (the accused devices) built and supplied to the government by Texas Instruments (TI) and Radiation Systems, Inc. (RSi) in connection with separate *419contracts between those companies and the government (Contract Nos. DOT-FA77WA-3988 dated June 30,1977 (TI) and DTFA01-81-C-20118 dated September 30, 1981 (RSi)). The accused devices are intended for use in the Air Traffic Control Radar Beacon System (ATCRBS) administered by the Department of Transportation’s (DOT) Federal Aviation Administration (FAA). Plaintiff’s administrative claim, filed on June 14, 1979, that the antennas being manufactured by TI were covered by the ’977 patent was formally denied by DOT on September 10, 1980.
Only the question of liability is now before the court, quantum having been reserved by the parties for further proceedings, if necessary. Defendant denies liability on three principal and separate grounds. First, it maintains that it is permitted, under the terms of an April 25, 1973 contract between Transportation Systems Center (TSC) of DOT and plaintiff (Contract No. DOT-TSC-598, the “ ’598 contract”) to use the invention royalty-free since the invention was developed, i.e., reduced to practice, during the course of the ’598 contract. Next, it maintains that the ’977 patent is invalid on the grounds of anticipation and obviousness. Finally, defendant argues that the accused devices do not infringe the ’977 patent. Plaintiff, of course, vigorously disputes these contentions.
After a 9-day trial, and upon consideration of the pre- and post-trial submissions of the parties, and after oral argument, the court finds that the evidence presented strongly supports the conclusion that the subject of the ’977 patent was first reduced to practice during the ’598 contract as maintained by defendant. Accordingly, since this conclusion serves to dispose of this litigation, it is unnecessary to address the other issues before the court. See McDonnell Douglas Corp. v. United States, 229 Ct.Cl. 323, 325, 670 F.2d 156, 158, 214 USPQ 857 (1982); Mine Safety Appliances Co. v. United States, 176 Ct.Cl. 777, 781, 364 F.2d 385, 387, 150 USPQ 453 (1966). See also Autogiro Co. v. United States, 181 Ct.Cl. 55, 92, 384 F.2d 391, 415, 155 USPQ 697 (1967).
I.
The court makes the following findings of fact, which are presented in narrative form. Additional facts found by the court will be set forth as necessary when discussing the dispositive reduction to practice issue.
An antenna is used either to transmit or receive energy into or from free space. An antenna radiates wave energy in the form of an electromagnetic field in space that is produced by oscillating currents on the antenna. This field propagates outwardly from the antenna in many directions and consists of two components, an electric field and a magnetic field, which are at right angles to each other and which are both perpendicular to the direction of the propagation. The spatial distribution of this radiated energy constitutes the antenna’s “radiation pattern.” The bandwidth of an antenna is the band of frequencies over which the antenna meets the operational requirements of the given system.
The antennas at. issue are used in Air Traffic Control (ATC) systems to assist in determining the location of aircraft in the control area. ATC systems depend on radar to locate the aircraft. Primary radar systems, called Airport Surveillance Radar (ASR) and Air Route Surveillance Radar (ARSR), determine the position of aircraft by transmitting radio signals that are reflected by the aircraft. The direction in which the rotating radar antenna is pointing when the reflected signal is received establishes the aircraft’s azimuth (horizontal) location. The time it takes for the radio signal to reach the aircraft and be reflected back fixes the range (distance) of the aircraft. The antennas at issue are not part of the primary radar systems.
For some time (apparently since the m'id-1950’s), primary radar systems have been supplemented by secondary radar beacon systems. The type of secondary system involved herein is known as the Air Traffic Control Radar Beacon System (ATCRBS). ATCRBS sends (transmits) an “interroga*420tion” signal, which, when received by a transponder in the aircraft, causes it automatically to emit a “reply” (beacon) signal.1 ATCRBS determines the azimuth and range of the aircraft in a manner similar to that of the primary radar systems. In ATCRBS, however, the transponder may also send, with the reply signal, specially coded signals to the ground station to furnish it with additional information, e.g., identity of the aircraft and its altitude. This information appears near the “blip” on the radar screen representing the aircraft. The antennas at issue are used in ATCRBS.
Primary (ASR or ARSR) and secondary (ATCRBS) radar and beacon systems are used at many air traffic control sites. ASR and ARSR use a reflector type antenna that is mounted on a pedestal and rotated at a constant rate. The ATCRBS antenna is then generally mounted on top of, and in alignment with, the primary radar antenna. Thus, the ATCRBS antenna rotates at the same rate as the primary antenna.
Prior to September 1972, most ATCRBS antennas were of the form identified as FAA-7202 or FAA-8043. The FAA-7202 antenna has a vertical dimension (height) of approximately 17.625 inches and a horizontal dimension (length) of approximately 328 inches (27.33 feet); the corresponding dimensions of the FAA-8043 antenna are 15 inches and 336 inches (28 feet). Because of their shape, the antennas are commonly referred to as “hog trough” antennas. Appendix I is a picture of a hog trough antenna mounted atop a primary radar antenna.
The angular width (beamwidth) in a given plane of the signal transmitted by an antenna is inversely related to the antenna’s dimension in that plane. Accordingly, the hog trough antenna’s relatively long length gives it a relatively narrow beam-width in the azimuth direction. Conversely; the hog trough’s relatively small height gives it a large vertical beamwidth.
This large vertical beamwidth causes excessive “illumination” of the angular region below the antenna’s horizon, i.e., of the surrounding terrain and buildings. This illumination results in reflected signals that sometimes reinforce and sometimes cancel the “true” signals. In cases of reinforcement, an additional “false” target appears in the direction of the reflecting object. In cases of cancellation, the aircraft is “missed” and fails to appear on the radar screen. Obviously, these phenomena are undesirable in the ATC systems. On the other hand, however, the hog trough presents a long narrow profile to the wind, and thus offers little wind resistance.
The problems with the hog trough antenna became quite important in the late 1960’s as air traffic increased and the FAA began developing plans to use computers as an aid in ATC. A computer does not have the same ability to adjust or account for the “false” and “missing” targets that a skilled human does.
It was generally believed in the 1970’s that the increased illumination of the hog trough could be reduced to an acceptable level, viz., the desired pattern cutoff slope could be achieved, by utilizing an antenna with an approximately 8-foot high reflector. It was not practical, however, simply to increase the height of the hog trough antenna to obtain a narrow vertical beam-width because the corresponding increase in surface area would result in higher wind loads, which the existing ATCRBS installations were generally incapable of withstanding.
In September 1971, the FAA, as part of its efforts to achieve upgraded ATCRBS performance, awarded a cost plus incentive fee contract to plaintiff for $958,729 to design and build a motionless, electronically scanned, 8-foot high ATCRBS antenna that was fixed in place around an ASR tower and which was known as the “E-Scan” antenna. The E-Scan antenna provided improved ATCRBS performance, but was too expensive to construct and install *421at the numerous airports requiring an improved ATCRBS system.
In late 1971, plaintiffs engineers began studying other approaches to improve ATCRBS that would, hopefully, be less expensive than the E-Scan antenna. The two engineers at Hazeltine principally responsible for the effort were Vincent Mazzola (Mazzola) and Peter W. Hannan (Hannan). In late 1971, Mazzola was an experienced senior antenna engineer at Hazeltine, having both a bachelor’s and master’s degree in electrical engineering and over 10 years of antenna experience. Mazzola received a master’s degree in accounting in 1974 and has been working as an accountant since then. In late 1971, Hannan had bachelor’s and master’s degrees in engineering and over 20 years of experience in antenna and microwave component development. At the time of trial, Hannan had authored approximately 20 published papers, principally related to antennas, and been granted about 20 patents, relating primarily to antenna techniques.
Hannan and Mazzola wanted to improve ATCRBS performance by increasing the vertical height of the antenna. As experienced engineers, however, they knew that the necessary height of a “solid” antenna would “overload” the structure of the primary radar. Accordingly, Hannan postulated an “open array antenna” that would have both a low wind resistance to allow for mounting on the primary radar, viz., as a retrofit for the hog trough, and a high reflectivity, i.e., backlobe suppression, to prevent backward radiation leakage through the structure.2 An array antenna is an arrangement of several individual antennas or radiators so spaced and excited that their individual contributions add energy in some directions while cancelling energy in other directions.
On or about February 14, 1972, personnel from Hazeltine, including Hannan and Mazzola, met with personnel from TSC.3 During this meeting, Hannan and Mazzola discussed various possible solutions to the hog trough problem, including Hannan’s open array approach. The open array approach discussed by Hannan consisted of a planar grid of parallel, equally spaced columns with a planar array of dipole antennas mounted on one side of the columns. A dipole is a radiating element consisting of a thin linear conductor (a straight wire or rod), often approximately one-half wavelength long. An example of a dipole antenna is the “rabbit ear” antenna on a TV set. Hannan left the meeting believing that “the best approach for Hazeltine to pursue for a new beacon antenna would be an open array-type of antenna.” Mazzola seemed to have shared this view.
After the meeting, Hannan and Mazzola investigated various configurations of open arrays in an effort to design one that had sufficient backlobe suppression while remaining open to the wind in an icing situation. These efforts included placing continuous conductors, i.e., vertical metal rods, between the conductive columns supporting the dipoles. Calculations were made and measurements performed on several test *422structures, but none had both the low wind loading and high backlobe suppression required for ATCRBS.
In connection with their investigation, Hannan and Mazzola consulted Dr. Harold A. Wheeler (Wheeler). Wheeler was characterized as an eminent scientist and had honorary doctorates from George Washington University in the District of Columbia and Stevens Institute of Technology in New Jersey. He was active in the field of radio, especially in the field of antennas. Since 1935, he had been a “fellow,” the highest grade of membership (awarded for professional distinction) in the Institute of Electrical and Electronic Engineers. He received other awards for professional achievement, written chapters for antenna handbooks, published numerous articles, and been granted about 180 U.S. patents. As of 1972, Wheeler had been associated with Hazeltine for almost 50 years. In 1971-1972, Wheeler’s responsibilities at Hazeltine were as Chief Scientist under a consulting contract and as Chairman of Hazeltine’s Board of Directors. His time was not charged directly to any of Hazel-tine’s customers, but instead was accounted for as overhead.
Wheeler had a couple of discussions with Hannan and Mazzola regarding the open array problem. Wheeler’s proposed solution to the problem was a reflector structure comprised of conductive columns to support an array of radiating dipoles, and “tuned” reflective elements interspersed between the conductive columns to achieve the desired backlobe suppression. A reflector is an element or structure for reflecting radio waves. A “tuned” element of an antenna system operates at or near its resonance condition. Resonance is a condition associated with a natural frequency of vibration, either mechanical (sound energy) or electrical (radio wave energy). A tuning fork has mechanical resonance at a frequency corresponding to its pitch. A metal rod used as a reflector is most effective when its length is approximately one-half wavelength at the frequency of the radio wave to be reflected. Wheeler described his design to Hannan and Mazzola on or about April 4, 1972; the exact date is uncertain.
The ’977 patent embodies Wheeler’s conception, which was the subject of Hazeltine Patent Department Docket R3909. Neither Mazzola nor Hannan was able to predict confidently that the open array design proposed by Wheeler would satisfactorily solve the ATCRBS problem. Specifically, Hannan’s main concern was whether it would “provide minimum leakage or sufficiently low leakage of RF energy through the ground plane over the beacon system band.”
Consequently, a experimental program was thereafter undertaken at Hazeltine in an attempt to determine whether the tuned reflector ground plane would operate as theorized. Mazzola was responsible for conducting the testing program, and he caused a model to be constructed by a Hazeltine technician. This model only represented a small segment of a proposed open array antenna. The antenna or test structure (vertical ground plane) was constructed of aluminum sheets mounted on a wooden frame with an opening that could be exposed or covered. Appendix II is a picture of the test model structure. The test model constructed was not available for inspection and examination at time of trial, and the parties could not agree on the size specifications of the model. Photographs of the test model were, however, available. There was no satisfactory explanation in the record as to the whereabouts of the model or what happened to it.
In the absence of the model, defendant called on the Federal Bureau of Investigation (FBI). Gerald B. Richards (Richards), an FBI agent and expert in the area of forensic photographic examinations, analyzed available photographs of the test structure and determined that the gross (overall) dimensions of the vertical ground plane were 130 inches wide by 43.7 inches high. He likewise determined the dimensions of the opening to be 31.6 inches wide *423by 21.2 inches high.4 Mazzola testified, looking at one of the photographs, that the vertical ground plane was approximately 6 feet (72") wide by 4 feet (48") high, with an opening of 3 feet (36") wide by 2 feet (24") high. The court concludes that Richards’ scientific analysis and overall testimony was more persuasive than Mazzola’s estimate, which was based on just looking at the photographs. Mazzola had no specific recollection of the size of the test model and, as stated above, based his size determination solely on a photograph of the test model.
To test the tuned array concept, the open space of the test model was fitted with wooden columns (dowels). The tuned rods were formed by cutting gaps in aluminum tape that was wrapped around certain wooden dowels. The conductive columns were apparently modeled by wrapping certain adjacent dowels with aluminum tape to form “wide strips.” Wheeler testified that a wide strip was electrically equivalent to a circular column with a diameter of one-half the width of the strip.
Appendix III presents a close-up picture of the test model opening in the vertical ground plane with the dowels wrapped as described above. The “conductive columns” had an average width, based on Richards’ analysis, of 2.68 inches, and the “tuned reflectors” were approximately 0.54 inches in diameter (including tape).
The vertical ground plane was mounted perpendicular to and diagonally across a somewhat smaller horizontal ground plane, as seen in Appendix II. A short distance in front of the center “conductive column” was a monopole antenna element, which was also mounted perpendicular to the horizontal ground plane. A monopole is another fundamental radiating element consisting of a thin linear conductor that here was approximately one-fourth wavelength long. A monopole is usually used with an “image” plane. A monopole protruding through an image plane produces a radiation pattern approximating that of a dipole if the imaging plane is infinite. This is known as the “imaging principle.” However, if the imaging plane is small, the peak of the vertical radiation pattern does not lie along the ground plane, but rather tends to be tilted upwards toward the zenith. The image plane in the Hazeltine’s test model was the horizontal ground plane seen in Appendix II.
The monopole antenna was intended to simulate the dipole antenna that was ultimately planned for use with the tuned array. Due to the imaging principle, a monopole perpendicular to an infinite ground plane is essentially equivalent to a dipole electrically. However, use of a monopole fails to take into consideration the possible effect, i.e., radiation, of the “feed” lines of the dipoles. This is so because the feed lines of a monopole are located below the imaging (here horizontal) ground plane. Feed lines transmit electrical energy to or from an element. An element in this context is a relatively small antenna, either a dipole or a monopole, which is used as part of a large antenna system.
Mazzola estimated that it took about 8 hours for a Hazeltine technician to construct the test model, which he concluded must have been done on April 5, 1972, one day after the discussion with Wheeler. The test structure was assembled at Hazel-tine’s Smithtown, New York antenna range. Mazzola further estimated that it “might” have cost as much as $200 for materials to build the test structure. It is to be noted that no government employee connected with the ‘598 contract ever saw the actual test structure itself.
According to Mazzola, on April 5 or 6, 1972, a Hazeltine technician used the test structure antenna and another antenna to *424generate data that was recorded and presented in graphical form, i.e., as “antenna patterns.” An antenna pattern is a representation of the antenna’s radiation pattern; it indicates the shape of the antenna beam in the given (tested) plane. The radiation pattern of an antenna refers to the spatial distribution of the energy radiated from the antenna. This pattern is generally one of the most important properties that is specified for an antenna. According to Mazzola, “[w]e took a number of patterns with various gaps in the tuned rods and so on and so forth until we arrived at something that appeared suitable. And, my recollection is that we did this over some frequencies and 180 degrees in azimuth.” Additionally, a “calibration” pattern was taken with the opening of the test structure covered by a solid sheet of metal. Theoretically, it might be added, the solid sheet should completely reflect all energy impinging on it.
All patterns were taken with the other antenna at an elevation angle of 10° above the horizontal ground plane. Mazzola testified that this was done to avoid “contamination” of the test signals from reflection by the horizontal ground plane. Additionally, Mazzola estimated the distance between the two antennas utilized in the test at approximately 15 to 18 feet; in any event, surely less than 50 feet.
Of the patterns taken, only two original ones remained; copies of two other patterns appeared in Hazeltine’s proposal to TSC relative to an ATCRBS open array antenna. From the numbers on the original patterns that do exist, it appears that a minimum of 13 or 14 patterns were generated. One of the original patterns was titled “A-A [wavelength] Tuned Rods.” It is reasonable to conclude that this is the pattern generated by a set up very similar to or identical with that of the test structure in Appendix II. This pattern was taken at a frequency of 1180 MHz (mega Hertz). A mega Hertz is a unit of measurement of frequency equal to 1 million cycles (waves) per second. This pattern had a deep “null” or depression greatly in excess of -35dB, i.e., — 42dB or thereabouts, at 180° azimuth.5 The pattern was somewhat irregular at the “main lobes,” i.e., in the region around zero degrees; it even doubled back at one point. This pattern appeared to be the bottom pattern in Figures 6-6 and 6-7 of the Hazeltine proposal, discussed below, perhaps with some “smoothing out.”
The two non-original patterns were in Figure 6-7 of the Hazeltine proposal. It is reasonable to conclude that these patterns were generated by the same test structure. These patterns were apparently taken at frequencies of 1150 and 1120 MHz. The “null” in these two patterns was only -31 or -32dB in the 180° region. These patterns also contained perturbations in the main lobes. Nonetheless, there was no evidence of “doubling back” in any of the graphs in the proposal.
The other “original” pattern was the “calibration” pattern, viz., the pattern made with a solid sheet comprising the entire test structure. The opening in the test structure was completely closed off for this pattern. This pattern was generated using a frequency of 1060 MHz, and it also had a maximum suppression (deepest null) in excess of -35dB, but it was not as great as that of the other original pattern, i.e., the pattern with the test array in place. Additionally, the main lobes of the pattern were noticeably smoother and more regular than those of the other patterns. Inter alia, this pattern had the following notations on it—“Narda Mixer”; “Absorber Around Mixer”; “Al. Tape on Ant Feed Cable.” Mazzola testified that a Narda *425Mixer is used to convert one frequency to another and that it is sometimes surrounded with an absorber “to prevent spurious signals from entering the mixer” and thereby “to insure that what you were measuring was not leakage but was the signal of interest.” Similarly, he testified that “Al. Tape” was aluminum tape that is wrapped around an antenna cable to prevent spurious signals.
There was no evidence that any further tests on the model were performed. There were no model environmental tests concerning ice, rain or snow, and there was no showing of any patterns having been taken with the grid structure absent and no cover present, viz., with the center portion of the test structure completely open. It is to be noted that no government personnel witnessed the tests of the model structure or ever saw the test model structure.
The results of the tests on the experimental model convinced the involved engineers at Hazeltine that they had a viable solution to the ATCRBS problem.6 Indeed, Hazeltine planned to propose a tuned reflector antenna in response to an RFP that they anticipated would be forthcoming from TSC.
On September 1, 1972, TSC issued RFP No. TSC/R & D TER-0070-RN (the ‘070 RFP) to 21 companies, including Hazeltine, soliciting proposals for “Air Traffic Control Radar Beacon System Modification Kit.” The closing paragraph of the ‘070 RFP transmittal letter provided:
Your attention is invited to the fact that the Contracting Officer is the only individual who can legally commit the Government to the expenditure of public funds in connection with this proposed procurement.
The letter was signed “Daniel J. Kelleher [Kelleher]/Contracting Officer [CO],” but did not explicitly state that Kelleher would be the CO on any contracts issued pursuant to the ’070 RFP. Nonetheless, as an experienced government contractor, Hazeltine should have known that Kelleher was to be the CO, and there was some testimony that some members of Hazeltine did in fact know. In any event, there was no evidence that any government officials misrepresented their position or were aware of any Hazeltine confusion relative thereto.
“Contracting Officer” was defined in clause 7 of the “General Provisions * * * ” portion of the ’070 RFP as:
The person executing this contract on behalf of the Government, and any other officer or civilian employee who is a properly designated Contracting Officer; and the term includes, except as otherwise provided in this contract, the authorized representative of a Contracting Officer acting within the limits of his authority.
Also, clause 37 of the General Provisions set forth the Patent Rights (Title) provisions. This provision read in pertinent part as follows:
(a) Whenever any invention, improvement, or discovery (whether or not patentable) is made or conceived or for the first time actually reduced to practice, by the Contractor or his employees, in the course of, in connection with, or under the terms of this contract, the Contractor shall immediately give the Contracting Officer written notice thereof; and shall promptly thereafter furnish the Contracting Officer with complete information thereon; and the Secretary shall have the sole and exclusive power to determine whether or not and where a patent application shall be filed, and to determine the disposition of all rights in such invention, improvement, or discovery, including title to and rights under any patent application or patent that may issue thereon. The determination of the Secretary on all these matters shall be accepted as final and the provisions of the clause of this contract entitled “Dis*426putes” shall not apply; and the Contractor agrees that he will, and warrants that all of his employees who may be the inventors will, execute all documents and do all things necessary or proper to the effectuation of such determination.
(e) Whenever any invention, improvement, or discovery relating to the work called for or required under this contract is constructively reduced to practice by the Contractor or his employees, during the period of performance of the contract, there shall be a prima facie presumption that such invention, improvement, or discovery was conceived or first actually reduced to practice in the course of, in connection with, or under the terms of this contract, and the Contractor shall immediately notify the Contracting Officer and submit the information required in paragraph (a) above. The Contractor shall have the burden of proving to the Contracting Officer that any such invention was not conceived or first actually reduced to practice in the course of, in connection with, or under the terms of this contract. If the Contractor fails to meet this burden, such inventions will be subject to all of the provisions of this clause.
The Patent Rights Clause was in accord with government policy, which was, at the time of the ’070 RFP and the execution of the contracts resulting therefrom, for the government to acquire principal rights to inventions made in the course of or under a government contract where, inter alia, “[a] principal purpose of the contract is for exploration into fields which directly concern the public health, public safety, or public welfare.” 41 C.P.R. § 12-9.6102 (1973; effective date of regulation was 8 days after March 4, 1972, see 37 Fed.Reg. 4,802 (1972)). See Memorandum & Statement of Government Policy (President Nixon), 36 Fed.Reg. 16,887 (1971). See also 41 C.F.R. § 12-9.107-350 (1984).
The ’070 RFP also contained a Design Constraints Section. Included among the constraints were the following: (1) The pattern cut-off slope of the antenna, i.e., suppression of ground illumination, had to be quite steep, viz., no greater than -21dB at any angle 5 degrees or more below the zero degree reference; and (2) The antenna was required to remain operational “without structural failure” in winds as high as 85 knots with an ice build-up of up to 1.5 inches and be able to survive (continuous operation not required) winds of 130 knots with 1.5 inches of ice.
Hazeltine submitted two proposals in response to the ’070 RFP. The first was fully responsive and involved an antenna 8 feet high. This proposal was not selected for negotiation and is of limited relevance to the issue involved herein. Hazeltine’s other proposal (the “Volume II Proposal”) put forth the tuned reflector antenna as a cost-effective “direct replacement” for the hog trough antenna. Specifically, Hazel-tine proposed a 4 foot high by 28 foot long tuned open planar array that could, inter alia, be mounted directly on top of the primary radar. Appendix IV shows plaintiff’s design conception as set forth in its proposal. The Volume II proposal was not responsive to the RFP because it could not offer the required cut-off slope at the horizon. The Volume II proposal did, however, predict backlobe suppression of at least -30dB for the open array antenna, based on the tests described above.
The Volume II proposal stated that the array would be comprised of 2 inch diameter columns approximately 9 inches apart with a V2 inch diameter tuned rod between the columns. The model structure, on which the tests discussed earlier were conducted, was described as containing a 4-foot X 4-foot open area in which the array was mounted. Although two photographs of the test set-up were provided, there was no information relative to the actual size or spacing of the test elements. Nor did the Volume II proposal report that Hazeltine had achieved -21dB suppression with an untuned reflective structure similar to that of the tuned array test structure. Finally, the proposal noted that “heaters will be used to keep the structure free of ice.” The report did state that *427dielectric tubes would be used in forming the tuned rods, but nowhere indicated that the function of those tubes was other than to provide support.
The proposals received by TSC were evaluated by a Source Evaluation Board (SEB) and submitted to a Source Selection Official (SSO). James Andersen (Andersen) was Chairman of the SEB for the ‘070 RFP. He testified that the SSO was the then Deputy Assistant Secretary for Transportation, William Davis (Davis), who had given him “very specific instructions” that the FAA be able to procure freely for any products developed under the ’070 RFP.
The initial report of the SEB described the Hazeltine Volume II proposal as “[c]on-fusing” and “poorly organized.” Additionally, the report characterized the solution as having a “[f]airly high technical risk” and listed the risk as a “[w]eakness[ ]” of the approach. Nonetheless, the SSO determined that the Hazeltine Volume II proposal, as well as the fully responsive proposals of three other companies, be negotiated. Andersen thereafter appointed four members of the SEB to conduct negotiations; Robert Nelson (Nelson) was appointed Chairman of the “Negotiation Team.” Kelleher, the CO, directed Hazeltine to make all inquiries concerning the negotiations to Nelson.
Walter Tedford (Tedford) was the leader of Hazeltine’s negotiation team. In an in-house memo dated February 2, 1973, Ted-ford summarized the results of “preliminary negotiations” between TSC and Hazel-tine that apparently took place on January 30, 1973. The memo noted that Hazeltine was to submit “a revised letter cost proposal” by February 9, in which “Hazeltine should ask for exclusion of the pending patent application for the open array antenna from the Patents [sic] Rights Clause.”
Hazeltine submitted a “revised cost plus fixed fee proposal” in response to the above on February 9, 1973. One of the conditions contained in the response was that “the ‘Patent Rights’ Clause shall have no application to the subject matter of Ha-zeltine’s Docket R3909, entitled ‘Open Array Antennas,’ which has been developed at Hazeltine’s private expense.”
Nelson referred the patent rights question to Herbert E. Farmer (Farmer), then patent counsel for TSC. At the time, Farmer was TSC’s only patent counsel. Because of the varied responsibilities of his position, Farmer was unable to make an independent determination of whether or not a claimed invention had actually been reduced to practice. Rather, he testified that his advice to a contractor seeking such a determination “always was that this was a business decision that only they could properly make because they had the resources, the knowledge, only they could make it as a business judgment [sic], I could not do it.”
On February 21, 1973, Farmer had a telephone discussion with Peter Calamari (Calamari), a member of Hazeltine’s patent staff. It would seem, from a Hazeltine in-house memo written by Calamari and dated March 2, 1973, that Hazeltine did not present Farmer with any new information, but simply maintained that pages 6-1 through 6-10 of the Volume II proposal “evidenced” reduction to practice of the antenna. The memo also contained the following statement: “While Mr. Farmer refused to formally acknowledge whether or not a reduction to practice had actually taken place, he agreed that if one had taken place, the invention would not be covered by the standard DOT patent rights clause.”
Because of Farmer’s position, the Calamari memo concluded as follows:
It is therefore recommended that a letter be written to the contracting officer, requesting one or more of the following items:
(a) An exemption of the invention from coverage by the standard patent rights clause of any contract awarded: [sic ]
(b) Acknowledgment of the fact that Ha-zeltine has reduced the invention to practice; or
(c) An admission that if the invention was actually reduced to practice under *428Hazeltine’s in-house funds, it is not covered by the standard patent rights clause.
Farmer sent a memo dated February 21, 1973 to Nelson regarding Hazeltine’s February 9 request for exclusion from the Patent Rights clause. The memo provided, in pertinent part:
Both TSC technical monitors, Mr. George G. Haroules and Mr. Bernhard Kulke, state that the open array antenna, disclosed in the 2 paragraphs on page 6-5 starting with “Using this approach ...”, pages 6-7 through the first paragraph on page 6-10 ending with “ ... element pattern).” and page 6-11 of the Hazeltine Technical Proposal, Report 3-5303, Volume II, dated 16 October 1972, has actually been reduced to practice. The Ha-zeltine patent attorney confirms that the antenna disclosed on the cited pages is in fact the subject matter of invention disclosure R3909. Mr. Calamari also states that the invention has actually been reduced to practice. Therefore, the Patent Rights clause is self-deleting as to this invention because it does not apply to any invention which has actually been reduced to practice prior to execution of a contract. Mr. Calamari agrees with my interpretation of the applicability of the clause; accordingly, Hazeltine ' is withdrawing their February 9 request.
Haroules and Kulke both testified at trial. Although both were experienced electrical engineers, neither witness considered himself well versed in the specialty of antenna engineering. Also, Haroules stated that he didn’t know what “reduction to practice” meant. Haroules denied ever having a conversation with Farmer as reported in the memo; Kulke couldn’t remember having such a conversation, but was sure he would have remembered if one had taken place. Haroules pointed out that the form of the memo was inconsistent with administrative policy at TSC, which required that a copy of the memo be sent to Haroules and Dr. Kulke and that Farmer sign the memo using his full name. There was no evidence that copies were sent to the two men, and Farmer signed the memo simply by using his first name, “Herb.”
The only explanation Haroules had for the language of the memo was that Farmer had confused the open array antenna of Hazeltine’s Volume II proposal with another, different open array antenna that he had obtained from the Navy and which Haroules and others had modified for improved performance and possible ATCRBS use. Haroules testified that he had had several discussions with Farmer regarding patenting the modifications to the Navy antenna and had shown the antenna, which was installed on the roof of TSC’s laboratory facilities, to Farmer on at least two occasions.
For his part, Farmer remembered having seen the Navy open array antenna referred to by Haroules, but he did not give any indication that he had based his reduction to practice statements in the memo on discussions regarding that antenna. Farmer denied that the memo was an acquiescence in Hazeltine’s claim that the tuned array antenna had been reduced to practice. Instead, he testified that it was merely a recording of statements made to him, perhaps to serve as a “red flag” to look for inventions under a contract; he testified that he would not give much weight to any statements by Kulke or Haroules on reduction to practice since they were not patent attorneys.
The memo did not, however, qualify the statements made therein in any manner, nor did it advise Nelson not to agree that the tuned array antenna had been reduced to practice. At trial, Nelson could not recollect receiving the memo or its contents, but other evidence, i.e., a negotiation summary dated March 20, 1973, convinces the court that Nelson received the memo and was aware of its contents. On the other hand, it must be pointed out that there was no evidence that Hazeltine knew of the memo or its contents until after institution of the present action.
Nelson and Tedford had a telephone discussion regarding the ’070 RFP on February 23, 1973. The content of this discus*429sion is a matter of dispute between the parties. A March 2, 1973 Hazeltine memo by Tedford summarizing the discussion stated, in pertinent part:
As per the telephone discussions between Messrs. P. Calamari and J. Maune of Hazeltine and Mr. Farmer of TSC, TSC recognizes that Hazeltine has reduced to practice the open array antenna technique for which we intend to file a patent application. Mr. Nelson stated there was, therefore, no need for an exclusion to the patent rights clause in the contract.
Tedford testified he was “certain” Nelson made the above statement. Tedford accounted for the 9 or 10 day gap between the date of the discussion and the memo by explaining that he had probably made a draft shortly after the discussion, but had not had a chance to get it in final form due, for example, to job-related travel. It bears repeating here that Calamari, an employee of plaintiff’s patent department, sent an in-house memorandum, dated March 2, 1973, to a Hazeltine Contracts Department attorney relative to the same telephone conversations referred to in Tedford’s March 2, 1973 memo, wherein Calamari stated “ * * * Mr. Farmer refused to formally acknowledge whether or not a reduction to practice had actually taken place.” See ante pp. 427-28.
Nelson denied ever agreeing with Hazel-tine that the open array antenna had been reduced to practice. He further stated that he would not have had authority to enter into such an agreement or to grant an exclusion or deviation from the Patent Rights clause. He also noted that he had never entered into an oral contract with any contractor while with TSC. Tedford knew that Nelson had limited authority and that some of his acts were subject to ratification. Moreover, Tedford did not rely on Nelson’s statement regarding the need for an exclusion from the clause. Rather, Tedford forwarded the information to Ha-zeltine’s patent attorneys.
Hazeltine prepared a letter, which was dated March 6, 1973, for signature by N. Jacobson, Hazeltine’s department head for contracts, to Nelson. The letter referenced Hazeltine’s Volume II proposal, the February 21, 1973 discussion between Farmer and Calamari, and Hazeltine’s revised February 9, 1973 proposal. The letter requested that any contract with Hazeltine resulting from the ’070 RFP include one or more of the alternatives set forth in Calamari’s March 2, 1973 memo, which was attached to the letter and reproduced in pertinent part ante pp. 427-28.
The draft letter dated March 6, 1973 was never sent, however; each of its pages had a line running from the upper right hand corner to the lower left hand corner thereof, supportive of the view it was not sent. In addition, the draft letter was attached to a cover page that was entitled “Document for Signature” and which contained handwritten notations. Edward A. Onders (On-ders), who was then (1973) General Patent Counsel for Hazeltine, testified that he had marked “Cancel” “73 Mar. 7” on the cover page. Another notation, this one not written by Onders, appeared in the section of the page headed “Remarks.” The notation read: “Per Pete Calamari—letter is can-celled as a result of Tedford’s telcon with DOT wherein the C.O. refused H’s request set forth in the attached letter.” Onders didn’t know the substance of the referenced conversation, but believed that whoever had made the notation believed it to be accurate.
Hazeltine did send Nelson a letter dated March 8,1973 that was intended to confirm negotiations. The letter provided, in pertinent part:
Pursuant to [the February 23, 1973 conversation between Tedford and Nelson], Transportation System Center has acknowledged that the subject matter of Hazeltine’s Docket No. R3909, entitled “Open Array Antennas,” the substance of which is described on pages 6-1 through 6-10 of Hazeltine’s Technical Report 3-5303, has been reduced to practice by Hazeltine with in-house funds and is, therefore, not covered by the “Patent Rights” clause.
*430Tedford testified that this letter “would have given the company’s position” on items under discussion or negotiation.
At trial, Nelson testified that the contractor could “make any statement he wants” in the proposal letter inasmuch as it was just a statement of position that doesn’t require a reply from the CO. Nelson believed that “final resolution” of these matters was found in the contract document. At a deposition, however, Nelson testified that he would have taken some action if he had received a letter indicating something contrary to his understanding of the telephone discussion. Nelson didn’t state precisely what he would have done, but noted that negotiations would not have been closed if he didn’t agree with the letter. Nelson apparently had no discussions of any sort with Hazeltine after receipt of this letter.
Nelson was “primarily responsible” for preparing a memo summarizing negotiations with Hazeltine that was dated March 16, 1973 and addressed to the Chairman of the SEB. The memo contained the following statement:
Hazeltine’s claim that their Docket No. R3909 entitled “Open Array Antennas” has been reduced to practice has been confirmed by DOT/TSC Messrs. Har-oules and Kulke. (See [Farmer’s] memo of February 21, 1973).
This memo became part of a document dated March 20, 1973 that purported to be the “Final Report” of the SEB regarding the ’070 RFP. Unlike an earlier report of the SEB, however, this report was only signed by one of the six members of the SEB.
Another document dated March 20, 1973 was Hazeltine’s “best and final offer” in response to the ’070 RFP. As relevant here, the document noted that all of the “terms and conditions” of Hazeltine’s March 8, 1973 letter applied to the offer.
It is clear that no government personnel witnessed the tests conducted by Hazeltine and upon which plaintiff rests its claim that its invention had been reduced to practice. At best, the only possible indication the government had of plaintiff’s invention was contained in its Volume II proposal in response to the ’070 RFP. A reasonable reading of the record convinces the court that any reduction to practice statements in the various government documents had their origin in plaintiff’s repeated self serving assertions that its open array antenna idea had been reduced to practice. There is no credible evidence in the record as to how government personnel would have known of such a reduction to practice other than by being told it was so by plaintiff’s personnel or by discerning the same from plaintiff’s Volume II proposal. There is no persuasive evidence in the record which would support a finding that government personnel were persuaded by plaintiff's Volume II proposal that its open array antenna had been reduced to practice.7 The record before the court, in any event, will not support a finding that the government had agreed with plaintiff that plaintiff had reduced the Wheeler open array antenna invention to practice prior to entering into contract ’598.
Subsequent to March 20, 1973, Hazeltine received a copy or copies of the '598 contract in unexecuted form. The contract contained the same Patent Rights clause as the ’070 RFP (set forth ante pp. 425-26).
*431The contract also included the following provision:
The rights and obligations of the parties to this contract shall be subject to and governed by the following documents: (a) this award/contract, (b) the solicitation, if any, and (c) such provisions, representations, certifications, and specifications, as are attached or incorporated by reference herein.
It is to be noted that the contract made no mention either of an exclusion from the Patent Rights clause or of any agreement that the open array antenna had been reduced to practice. Indeed, Onders admitted that Hazeltine knew the Patent Rights clause was in the contract “and would have whatever effect it would have.”
Additionally, the contract contained the following clause:
INTERPRETATION OR MODIFICATION
No oral statement of any person, and no written statement of anyone other than the Contracting Officer, or his authorized representative designated in accordance with section 12-1.402-50 of the DOT Procurement Regulations (41 CFR 12-1.402-50), acting within the limits of his authority specified in such designation, shall modify or otherwise affect any provision of the contract.
Hazeltine signed the ’598 contract on April 3, 1973 and thereafter sent it back to TSC.
Andersen, Chairman of the SEB, testified that contracts signed by the four contractors with which the government had negotiations, but not by the government, were presented to the SSO. Andersen did not sign the March 20, 1973 “Final Report” and had no recollection of the language of the March 16 memo that was included with the Final Report and quoted above, ante p. 430, or of an agreement with Hazeltine regarding an exclusion from the Patent Rights clause or regarding a reduction to practice. He further stated that TSC would not have awarded a contract to a company that insisted upon an exclusion to the Patent Rights clause. Andersen presumed, however, that the Final Report would have been part of the file presented to the SSO.
By memo dated April 11, 1973, the SSO informed Andersen that “awards shall be made to” three contractors, including Ha-zeltine. The SSO determined that it was not in the government’s interests to enter into a contract with the fourth contractor. Andersen and Kelleher both stated that the memo merely authorized the CO (Kelleher) to sign the three contracts; it did not mandate the signing thereof.
Kelleher testified that it was his course to review the files before signing a contract on behalf of the government. He noted that the file, if completely assembled, would have contained correspondence from the contractor as well as governmental memos. That is, the file in the instant case should have contained all of the material discussed above with the exception of Ha-zeltine in-house memoranda and the letter of March 6, 1973 that was never sent.
Kelleher did not, however, recall as a matter of fact seeing the portion of the March 20 Final Report dealing with the question of reduction to practice. He stated that he would not have had authority to sign a contract without a Patent Rights clause and doubted that he would have had authority to sign a contract wherein there was a collateral agreement relating to reduction to practice. Kelleher could not remember discussions with anyone at TSC regarding an exclusion to the Patent Rights clause or with anyone regarding a collateral agreement, but thought he would have if they had occurred.
Kelleher signed the ’598 contract on or about April 25, 1973, thereby awarding it to Hazeltine. The parties stipulated that Hazeltine actually began contract performance on March 21,1973. Hazeltine used its own funds prior to award, but apparently suffered no loss as the ’598 contract provided for a reimbursement thereof.
The ’598 contract was a cost-plus-fixed-fee contract. The court believes it is fair to characterize this contract, at least the Phase I portion, as a research and develop*432ment type contract. The ’598 contract called for Hazeltine, inter alia, to “conduct a comprehensive, detailed, engineering study of the * * * ATCRB [sic ] interrogator installation, and concurrently, design [a four foot high open array antenna] in accordance with [Hazeltine’s Volume II proposal and the design constraints of the ’070 RFP, except that the required elevation pattern cut-off slope was reduced].”
The ’598 contract had two phases. As described by Tedford, Phase I was a “study” phase, whereas Phase II, which was optional with the government, was a “hardware” phase. Francis J. LaRussa (LaRussa), one of the government’s expert witnesses, was technical monitor on the ’598 contract. He described Phase I as “study and preliminary design,” with Phase II being to “design and build and test the antenna” if the results of Phase I proved satisfactory. LaRussa had a bachelor’s and master’s degree in Physics, had taken graduate courses in electromagnetic theory, microwave and antennas, and in 1973 had 20 years of antenna experience.
LaRussa and two others from TSC had a “design review meeting” at Hazeltine headquarters on May 3, 1973 to discuss the tasks to be performed by Hazeltine during Phase I. As recorded by LaRussa in a memo of this meeting, “[t]he discussions were chiefly concerned with [Hazeltine’s] plans to prove the feasibility of [the open array] approach during Phase I.” Hazel-tine intended “to demonstrate electrical feasibility” by “continu[ing] the work described in Volume II of their proposal.” TSC had several questions “regarding the electrical performance of the proposed ‘open reflector’ array,” including the following:
What is the bandwidth of the tuned reflector?
What is the effect of mutual coupling on the bandwidth?
What is the magnitude of the cross-polarized component of radiation?
Additionally, TSC was interested in further mechanical and power information regarding the proposed method of deicing, i.e., the heaters mentioned in the Volume II proposal.
Hazeltine conducted a number of tests during Phase I of the ’598 contract to address the questions raised above, as well as other concerns the resolution of which was necessary to provide actual verification of the tuned reflector’s ability to operate effectively in its intended environment. Most of the following information comes from Hazeltine’s interim report on the ’598 contract, detailing the results of work done on Phase I of the contract between April and June of 1973. Hazeltine performed these tests on models of a tuned reflector open array segment.
The principal model utilized in these tests was a 4-foot high by 8-foot wide (long) tuned reflector segment composed of a series of 2-inch diameter columns (4-foot high) spaced 9 inches between centers. In between the columns were V2-inch diameter tuned reflectors constructed by wrapping aluminum tape around plastic rods and cutting periodic gaps in the tape. Cords were used to keep the reflectors straight. A frame was used to align, i.e., hold, the columns. It is to be noted, in contrast to the test model structure described and discussed previously and upon which plaintiff’s reduction to practice claim is based, that no solid ground plane surrounded the array segment of this test structure. A picture of this test structure, with dipoles mounted thereon, appears as Appendix V to this opinion. The test structure was mounted at the Smithtown range approximately 20 feet above the roof and 30 feet from a source antenna at the same elevation. This test structure was used in the manner described below.
First, a single driven dipole was mounted on the middle reflective source column. The “feed” cable for the dipole went down through the middle of the support tube. Patterns were taken with this structure, with solid half-inch rods in place of the tuned reflectors, and with a solid aluminum sheet covering the entire open array.
*433As would be expected, the solid sheet had the best backlobe (radiation leakage) suppression, with a maximum suppression in excess of -40dB. The interim report noted, however, that “because of diffraction around the edges of the ground plane, the backlobe structure is a function of elevation angle.” Nevertheless, at every angle tested, the backlobe was less than -30dB. The backlobe at the horizon was only -14dB with the solid rods; with the tuned reflectors, there was a null of -30dB for approximately 10 degrees at 180 degrees of azimuth.
Thereafter, eight additional dipoles were added to the test structure as shown in Appendix Y. Patterns were taken with the solid sheet ground plane and with the tuned reflector ground plane. Tuned reflector patterns were taken over elevation angles from 0-70 degrees (5 degree increments) at frequencies of 1030 and 1090 MHz. These patterns showed peak back-lobes of -17dB. Hazeltine believed that these high backlobes were “caused by radiation from the stems of the dipoles.” The cause of this unbalanced stem current was “mutual coupling” between the surrounding dipoles, which was apparently not anticipated in the original dipole design, i.e., was “nonexistent.”8
A second “design review meeting” for Phase I was held at Hazeltine headquarters on June 22, 1973. At this meeting, Hazel-tine informed TSC that it planned to solve the mutual coupling problem, which apparently also could cause unacceptable levels of “cross-polarized radiation,” by means of “corrective chokes” on the dipoles. Edward M. Newman (Newman), an antenna engineer at Hazeltine who devoted significant effort to the ’598 contract, didn’t think that Hazeltine was “concerned” about the problem prior to the ’598 contract. LaRus-sa “didn’t expect” that mutual coupling would have caused that particular effect, but “wasn’t sure it was not” the cause in view of the tendency for element interaction in arrays.
Newman had a bachelor’s and master’s in engineering, as well as a master’s degree in business. It is unclear, however, when he received his graduate degrees, viz., whether they were earned before or after the time of the events at issue. Nonetheless, it can be said fairly that Newman had at least a bachelor’s degree and over 5 years of experience in mid to late 1973.
The parties also discussed the possible problem of “detuning” of the tuned reflectors due to the accumulation of water or ice thereon. Hazeltine stated that it planned to cover the tuned rods with a dielectric sleeve that would leave an enclosed air space around each element, i.e., rod. This problem and the solution thereof are discussed infra.
Three days after this meeting, on June 25, 1973, Hazeltine filed the patent application that eventually ripened into the ’977 patent. Hazeltine did not inform Kelleher (the CO) or anyone else at TSC of this filing.
Hazeltine designed a shield for the dipole balun. A balun is a device for connecting an unbalanced transmission (feed) line to a balanced antenna. After the nine baluns had been shielded, another set of patterns was taken with the tuned reflector test structure. Regular and cross-polarization patterns were taken over elevation angles from 0—70 degrees (5 degree increments) *434at 1030 and 1090 MHz. Patterns were also taken at 1000 and 1120 MHz, but apparently only at zero degrees elevation and without measurement of cross-polarization suppression. These patterns showed a suppression of backward (backlobe) and cross-polarized radiation of -25dB or more. Hence, Hazeltine’s initial belief, viz., that mutual coupling was having an adverse effect on backlobe suppression, was correct.
As indicated above, TSC and Hazeltine discussed the problem of ice and rain. It is well known that ice and rain, i.e., precipitation, can have an adverse effect on the operation of a tuned element, viz., the element can be detuned.9 If detuning occurs, the antenna may not have sufficient back-lobe suppression over its operating frequency range.
As stated above, Hazeltine planned to put each tuned rod inside a dielectric shield (sleeve). A dielectric shield does not conduct electricity, i. e., it is an insulator. The theory behind use of a dielectric shield was that it would put sufficient distance between the rod and the precipitation so that any detuning would be negligible.
Hazeltine ultimately decided to use a % -inch O.D. (outside diameter) fiberglass tube to surround the tuned rod. According to Newman, it was Hannan who chose the size of the tube. In its Phase I report, Hazeltine stated that the tube “prevents ice and rain from coming very close to the tuned reflector, and permits the optimum tuning of the tuned reflector to be retained in the presence of ice and rain.”
Hazeltine tested this design for ice effect by surrounding a portion of a tuned rod, including the sleeve, with a nylon block to give an average radial thickness of approximately Vz-inch. Nylon was chosen because it has a dielectric constant similar to that of ice. That is, nylon can be used to simulate ice without appreciable error. The rod portion was tested in a waveguide simulator with and without the nylon. The tests showed that the tuning of the rod was not significantly changed by the presence of the nylon. Thus, the tests allowed one to infer that the backlobe suppression of the antenna would not be adversely effected by the presence of 1/2 inch of radial ice.
To test for rain degradation, Hazeltine constructed tuned rods out of Vi-inch diameter rexolite (plastic) rods wrapped with aluminum tape (with gaps cut therein) and inserted them in a %-inch diameter dielectric sleeve. Unlike the ice test, complete rods were fabricated and made a part of an array structure. The testing procedure consisted of taking a reference pattern and then other patterns while simulating rain. In one test, a sprinkler hose was mounted on top of the test structure. This test simulated a rain rate of 4.5 inches per hour, with the measurements made in a 10-15 mile-per-hour breeze. In the other test, a technician directed a hose at the front of the array. This test was done to simulate heavy rain (12 inches per hour) and high wind velocities. Neither test showed a serious degradation of antenna performance.
Hazeltine prepared the interim report referenced above detailing the work done under Phase I of the ’598 contract. This report contained a detailed discussion, inter alia, of the test procedures described above. The report also contained well over 100 antenna patterns showing the response of different test structure configurations (e.g., one dipole) under various conditions of design (e.g., without balun shields) and operation (e.g., in ice and rain). In the Summary portion of the report, Hazeltine made, in pertinent part, the following characterization of these tests:
*435 Pattern Feasibility Tests
The objective of these tests was to eliminate risks associated with the open-array ground plane design. We have concluded from the pattern feasibility tests conducted that the open-array ground plane is feasible, and that sufficient backlobe suppression can be obtained to achieve the desired system performance.
Newman wrote in his notebook that the objectives of the tests were:
1. Demonstrate feasibility of concept;
2. Reduce risk of development problems;
3. Provide design information to expedite array development.
These objectives were essentially repeated in the Electrical Design Section of the Report:
Objectives. The experimental program described in this section served a threefold purpose. First, and most important, the program has demonstrated the feasibility of the open array comcept [sic] using turned reflectors. * * *
Second, The measurements have served to reduce the risk associated with a new design such as the open array. To minimize the risk, potential trouble areas were defined and these areas of concern were thoroughly investigated during the measurement program. * * *
Finally, the experimental program has provided design information, especially in the area of tuned reflectors, which will expedite the final design of the array in Phase II. * * *
Although the parties did not highlight the following facts, there was a third meeting, apparently on July 12, 1973, between officials of TSC, personnel from Hazeltine and outside consultants retained by TSC. Hazeltine presented the results of its efforts under Phase I at this meeting. By memo dated August 3, 1973, LaRussa recommended that the government exercise its option for Phase II of the ’598 contract.
According to Hazeltine’s monthly progress reports, the government exercised its option on August 22, 1973 with a retroactive effective date of August 3.
Phase II of the ’598 contract, it will be remembered, called for the design, construction and delivery of a 4-foot by 26-foot open array antenna (the prototype) as described and specified in the Volume II proposal and the ’070 RFP, as modified.10 While performing Phase II, Hazeltine requested additional funding to complete construction of the prototype. In a letter in support of its request, Hazeltine asserted that the “major factor” responsible for the increase in cost was the fact that as a result of the work done under Phase I, the prototype “is significantly more complex in detail mechanical design” than the antenna originally proposed in the Volume II proposal. In a later letter reducing the request to $28,360, Hazeltine listed as a contributing factor the “[i]ncrease in the magnitude of the design of the tuned reflectors, resulting from detail design definition during Phase I.” It appears that Hazeltine ultimately received the $28,360 requested.
Hazeltine completed construction of the prototype and tested it at its Smithtown range during January 1974. In March the antenna was taken to another, longer range. LaRussa testified that the longer range was necessary to ensure a plane wave field. LaRussa stated that a plane wave was required to get an accurate description of the radiation pattern. Defendant’s other expert, Phillip N. Richardson (Richardson), whose experience will be related infra, defined a plane wave as a wave having the same phase wherever it is measured on the test screen. Richardson also noted that it is “fairly well understood” in the antenna field that arrays, particularly planar arrays, usually correspond to plane waves. LaRussa stated that the range length necessary to ensure a plane wave *436was given by the formula L = 2D2/[wave-length], where L is the length and D is the maximum dimension of the antenna in one direction.
During March 1974, after the long range, plane wave tests, Hazeltine delivered the prototype to the antenna test facility at Naval Aviation Facilities Experimental Center (NAFEC) in Atlantic City, New Jersey for testing under simulated operating conditions. Basically, the tests at NAFEC consisted of flying aircraft in certain patterns, e.g., radial patterns at varying altitudes, and measuring the performance of the antenna. Dr. Lester Van Atta (Van Atta), whom both parties characterized as highly regarded in- the antenna field, was one of the consultants to TSC, responsible for giving an evaluation regarding the performance of the contractors under the ‘070 RFP. Van Atta examined the Hazeltine prototype at NAFEC and presumably either witnessed the tests or reviewed the results thereof. In a letter to TSC, Van Atta expressed the following opinion with respect to Hazeltine’s open array prototype:
The Hazeltine ATCRBS antenna design is a highly successful solution to the problem of finding a direct replacement of much higher performance for the current “hog trough” antenna. The Hazel-tine design represents an imaginative response, both electrically and mechanically, to the specific requirements of this project. It probably will provide adequate performance for many of the less demanding sites.
The main performance problem with the prototype antenna was its insufficient cutoff slope at the horizon. LaRussa testified that the specification was 2dB per degree, while the prototype was only capable of achieving in the neighborhood of 1.1 dB per degree. The practical effect of this lower performance capability was the increased possibility for “false” or “missing” targets, which effects were in fact observed during testing. The parties did not make clear how the 2dB per degree specification related to or was derived from the ’598 contract requirement of no more than -7.5dB of power (illumination) at -1° elevation, i.e., 1 degree below horizontal, and no more than -21 dB at angles 9 degrees or more below horizontal.
Nevertheless, the government awarded contract DOT-FA75WA-3712 (the '3712 contract) to Hazeltine on September 11, 1975. The ’3712 contract was sole source, i.e., was not competitively bid, and called for Hazeltine to deliver ten (10) 4-foot high by, presumably, 28-foot long antennas. The 4-foot antenna was put through extensive testing at some locations and was found not suitable for “high performance” locations. Hazeltine received more than $1.1 million for its work on the ’3712 contract.
Tedford, who was in charge of negotiations for Hazeltine, testified that a total of $398,000 was spent, and therefore paid by the government, on the ’598 contract. The ’598 contract was of the cost-plus-fixed-fee type and estimated a price of $92,000 for Phase I. Hazeltine’s Monthly Cost Report No. 13, the latest one available for Phase II, notes that over 11,900 hours (were expended at a total cost in excess of $278,000. The report estimated that final costs would total almost $279,500. Also, the '598 contract provided for Hazeltine to receive a fixed fee of $20,087 for performing the work done under Phase II.11
On the other hand, although no direct evidence was presented at trial, the court believes the record supports a finding that plaintiff did not invest more than $20,000, including salaries and overhead, in its pre-contract work relative to the tuned reflector open array antenna. Plaintiff cited a monthly report (April 1972) by Mazzola that it interpreted as reflecting 875 engi*437neering man hours spent. At oral argument, plaintiff claimed that each engineering man-hour represented $20; defendant claimed the actual figure, as of 1972, should be $9-10. Using plaintiffs figure yields an “investment” of $17,500. It is to be noted that plaintiffs figure appears, from the monthly report, to include all effort expended (as of the date thereof) on the ATCRBS problem, excluding, of course, the work done on the E scan antenna. That is, the figure is not limited solely to the time spent during the discussions among Wheeler, Mazzola and Hannan relative to the tuned reflectors or to the time and materials involved in building and testing the model, all of which seem to have taken place over a period of no more than a few days.
In his contractor performance rating of Hazeltine under the ’598 contract, LaRussa made, in pertinent part, the following observations:
9. The contractor did not fully appreciate the difficulties involved and tended to underestimate the time required to solve the problems associated with hardware development.
11. Aside from the novelty of the concept which resulted in the contract award, there were no results that were not originally expected.
14. The contract was for hardware development and was moderately difficult because it involved a novel technique which had never been used for this application.
At trial, LaRussa stated that the second comment, supra (H 11), merely reflected TSC’s prior expectation that the antenna couldn’t meet the cut-off slope at the horizon. He also stated, relative to the comments, supra (H1111 and 14), that he viewed plaintiff’s open array proposal as “novel” because it would mount right on the primary radar, wouldn’t need a separate pedestal, and wouldn’t have a beam “registration” problem, i.e., the primary and ATCRBS beams would point in exactly the same direction. As plaintiff pointed out on cross-examination, however, those same factors apply to the hog trough antenna. The court believes it fair to conclude that the term “novel” was meant, at least in part, as a characterization of Hazeltine’s proposed solution.
By letter dated September 14, 1976, Ha-zeltine advised Gladys M. Frederick (Frederick), Administrative Contracting officer for the ’598 contract, that there were no subject inventions under said contract. Frederick did not testify at trial. The substance of her deposition testimony is that she consulted Farmer, TSC’s patent attorney, with respect to Hazeltine’s negative report of inventions and was, in essence, advised by him that patent issues were not a bar to closing out the ’598 contract. Farmer did not recall reviewing Hazeltine’s final patent report, but agreed there was a policy to review such reports and to consult with the technical monitors to see if the reports were accurate.
Following the tests done during Phase II of the ’598 contract, DOT awarded contract DOT-TSC-1014 (the ’1014 contract) to Ha-zeltine on May 15, 1975. The ’1014 contract, for $220,000, was awarded on a sole source basis and called for the design of a 5-foot open array antenna, incorporating further refinements in the open array design. In a January 10, 1975 memo, Kelle-her justified the sole source procurement on the basis of “Hazeltine’s expertise and operational experience as the only company to successfully apply the ‘tuned reflector’ technique to produce an open array ATCRBS antenna.” The court does not find this particularly surprising or weighty in light of the expensive work done by Hazeltine under the ’598 contract. Indeed, the memo observed that Hazeltine was “presently the most qualified” to do the work required by the ’1014 contract. (Emphasis added.)
During the ’1014 contract, Hannan implemented an improvement to the open array antenna consisting of a vertical bend at the center of the antenna to reduce backlobe levels further. Richardson, one of defendant’s experts, testified that the bend was necessary to meet the ATCRBS specifica*438tions for backlobe suppression. By letter dated May 23) 1977, Hazeltine informed the contracting officer on the ’1014 contract that the bend was a subject invention under said contract.
In late 1976, the government solicited bids for the production of ATCRBS open array antennas. The solicitation was competitively bid, and Hazeltine participated therein. During the bidding process, however, Hazeltine informed the government of its position that the ’977 patent was applicable to the subject antennas. Nonetheless on June 30, 1977, DOT awarded fixed-price contract No. DOT-FA77WA-3988 to TI. Under this contract, TI has supplied, and the government has accepted, 197 open array antennas.
On September 30, 1981, DOT awarded fixed-price contract No. DTFA0-1-81-C-20118 to RSi. This contract called for the manufacture and delivery of 56 open array antennas and had options for 90 more, which have been exercised. Pursuant to the contract, RSi has delivered, and the government has accepted, at least one antenna.
Secondary radars of the type involved, i.e., ATCRBS Interrogators, are used by DOT, the Navy and the Air Force. Since January 1,1977, none of these departments has purchased ATCRBS Interrogators other than open array antennas. As of September 27, 1983 and October 25, 1983, the Navy and DOT, respectively, had no intentions of purchasing anything but open arrays for use as ATCRBS Interrogators. The Air Force has apparently not made a final decision regarding antenna type for other applications. In addition, the three departments plan to use open array antennas at a number of sites. The Navy, for example, plans to equip each of its 47 ATCRBS sites with an open array antenna. The Air Force, on the other hand, has equipped 14 of its 129 sites with open array antennas, and plans to equip only 64 more sites with said antennas at this time. Finally, the government sold three ATCRBS open arrays manufactured by TI to the Canadian government and two, also manufactured by TI, to the British government, at the request of each government.
Hazeltine and Tokyo-Shiburo Electric Co., Ltd. (Toshiba) entered into an agreement described as “Technical Information and Assistance Agreement,” but also characterized as “Know-How and Technical Assistance Agreement.”12 The agreement grants Toshiba the non-transferable, nonexclusive and indivisible right to sell open array antennas in a defined area, and to sell open array antennas “in combination with * * * Toshiba ATC radar” in another defined area, all without fear of an infringement action by Hazeltine in exchange for certain payments. Onders testified that Hazeltine had received about $150,000 as of the time of trial under this agreement. Onders also testified, however, that the agreement was primarily a “technology license” and not a “patent license,” and therefore the payments should not be considered patent royalties.
Additionally, Hazeltine entered into a “Technology Transfer, Technical Assistance and License” with Andrew Antenna Company Limited (AACL), a Canadian company. The parties signed this agreement in 1984 but provided for a 1982 effective date. The litigants had different impressions regarding Hazeltine’s role in the agreement. The dispute is somewhat confused, but fortunately is of no moment to the outcome of this case. For purposes of this opinion, the court deems it sufficient to note that Hazeltine was paid $5,000 and supplied AACL with information that AACL used in preparing a proposal for Canada’s Radar Aids Modernization Program (RAMP). AACL did not receive a RAMP contract, and has not made any further payments to Hazeltine. Whether additional payments are presently due is a matter of dispute between Hazeltine and AACL. Finally, the record dealing with the agreement between plaintiff and Canada relating to open array antennas is rath*439er ambiguous and provides no adequate support for a definitive finding relative thereto.
The ’977 patent was issued on September 17, 1974. Over 11 years has passed since that date. In light of the preceding discussion, the court finds on this record that, as of this date, the only commercial application of the '977 patent appears to be for ATCRBS antennas.
II.
A. Preliminary Statement
As stated above, the critical issue in this case is whether the government has a right, under the Patent Rights clause of the ’598 contract, to procure and use the subject antennas, which plaintiff claims are protected by the ’977 patent, without liability to Hazeltine. Resolution of this issue depends upon when the invention disclosed by the ’977 patent _was first reduced to practice. Plaintiff first maintains that defendant and Hazeltine agreed that the open array antenna had been reduced to practice prior to the ’598 contract. Plaintiff argues that this agreement equitably estops the government from now claiming that there had not been a precontract reduction to practice. Plaintiff next asserts that, in any event, as a matter of fact and law the invention in question was reduced to practice prior to the ’598 contract. Defendant disagrees with all of these contentions.
Before addressing these matters, the court must dispose of two subsidiary matters discussed by plaintiff in its briefs. First, contrary to plaintiff’s assertion, the fact that the Department of Justice (DOJ) rather than DOT first raised the “Government Rights” defense does not in any manner detract from the validity of said defense. Pursuant to 28 U.S.C. §§ 516-20 (1982), DOJ represents the United States in litigation. As “the nation’s litigator,” DOJ is entitled, if not obligated, to raise any meritorious defense of which it is aware in litigation, whether or not prompted by the agency or department involved. Similarly, the fact that the administrative denial of plaintiff’s claim did not rely on the government rights defense is of no import. It has been held that “the government did not have to spell out all its defenses to the administrative claim if it chose to reject the claim on * * * one ground alone.” McDonnell Douglas Corp. v. United States, 222 Ct.Cl. 570, 573, (1980).
Secondly, plaintiff appears to believe there is some significance to the fact that it filed a negative report of inventions, which was accepted by TSC, and which presumably was used by Farmer as a basis for authorizing Frederick to close out the ’598 contract insofar as patent matters were concerned. There is no significance, on this record, in this chain of events. As stated by the Court of Claims in response to a similar contention, “[t]he purpose of the reporting requirement is not to limit the government’s rights to inventions but to apprise the United States of the inventions in which it has an interest.” McDonnell Douglas Corp. v. United States, 229 Ct.Cl. 323, 328, 670 F.2d 156, 160, 214 USPQ 857 (1982). In the instant case, as in McDonnell Douglas, there was nothing in the contract imposing a duty upon defendant to inform plaintiff that the government had rights under the contract—“[Njothing in the contractual provisions called upon the government to disagree with plaintiff’s failure to list the [’977] invention or to set forth defendant’s own claims.” See McDonnell Douglas Corp. v. United States, supra, 222 Ct.Cl. at 572.
B. Agreement/Estoppel
Plaintiff maintains that there was an agreement between the parties that the invention claimed in the ’977 patent, which, for purposes of this discussion, will be referred to as the tuned array or open array antenna, was reduced to practice prior to award of the ’598 contract. This agreement, plaintiff contends, should estop the government from now raising the issue of reduction to practice. After a review of the law, and careful consideration of the facts, the court concludes that plaintiff’s contentions must be rejected.
*440Plaintiff has the burden of proving that an agreement was extant. The genesis of the alleged agreement is the February 23, 1973 telephone discussion between Nelson and Tedford. The court is not convinced, as a matter of fact, that Nelson entered into any agreement with plaintiff.
The evidence directly bearing upon this issue is set forth above, ante pp. 428-30. One telling factor in this evidence is the “Document for Signature” cover sheet cancelling Hazeltine’s March 6 letter, which letter included a request for acknowledgment of reduction to practice, because the “C.O.” refused the requests] of said letter in a “telcon” with Tedford.
Again, it must be noted that defendant had no independent knowledge of the tests relied on by plaintiff as establishing a reduction to practice. No government personnel witnessed these tests, nor were they privy to the complete actual test results. It is in this context that defendant’s conduct must be evaluated. The court believes there is a probability that plaintiff, believing strongly that its position was correct and advocating the same repeatedly, mistakenly perceived statements by defendant as assent to that position, when in fact the defendant did not do so and was in no position to make any informed decision on the question of reduction to practice. Nelson testified he made no such agreement with plaintiff, and the surrounding circumstances support him in this regard.
Here, as in McDonnell Douglas, “[i]t may be that [Hazeltine] always wanted and intended to retain full patent rights to the [Wheeler] invention and the [’977] patent, but there is no adequate showing that the defendant ever agreed with that position.” McDonnell Douglas Corp. v. United States, supra 229 Ct.Cl. at 328, 670 F.2d at 160, 214 USPQ 857.
Assuming arguendo that an agreement was reached between Nelson and Hazel-tine, there remains the issue of what effect such an agreement would have, i.e., are the facts and law otherwise sufficient to invoke an estoppel against the government. This issue involves questions of authority and the conclusiveriess of the language of the contract itself, as well as the traditional elements of estoppel.
It is clear that the United States is not bound by the unauthorized acts of its agents. E.g., Federal Crop Insurance Corp. v. Merrill, 332 U.S. 380, 384, 68 S.Ct. 1, 3, 92 L.Ed. 10 (1947); Urban Data Systems, Inc. v. United States, 699 F.2d 1147, 1153 (Fed.Cir.1983); Porter v. United States, 204 Ct.Cl. 355, 366, 496 F.2d 583, 590 (1974), cert. denied, 420 U.S. 1004, 95 S.Ct. 1446, 43 L.Ed.2d 761 (1975); Prestex, Inc. v. United States, 3 Cl.Ct. 373, 377 (1983). Additionally, the government is not estopped to deny the acts of its agents beyond the scope of their authority. E.g., Urban Data Systems, Inc., supra, 699 F.2d 1154; Putnam Mills Corp. v. United States, 202 Ct.Cl. 1, 9, 479 F.2d 1334, 1338 (1973); Gratkowski v. United States, 6 Cl.Ct. 458, 461 (1984). In equitable estoppel cases, the law requires that plaintiff prove that the official whose act(s) is relied upon had the requisite authority. E.g., Thanet Corp. v. United States, 219 Ct.Cl. 75, 85, 591 F.2d 629, 635 (1979).
Plaintiff did not present any evidence showing that Nelson was authorized to agree, on behalf of the government, that the tuned array antenna had been reduced to practice, nor does it press the point in its briefs. Indeed, one of Hazeltine’s own witnesses admitted knowing that Nelson had limited authority. Under DOT Order 4200.-7, 37 Fed.Reg. 4891 (1972), only the contracting officer, whose own delegation of power had to be written, had the authority to make agreements or commitments on behalf of the government.13
*441Recognizing Nelson’s lack of actual authority, plaintiff argues that the contracting officer, Kelleher, adopted and ratified Nelson’s actions by awarding the ‘598 contract to plaintiff. In order for Kelleher’s award of the contract to be efficacious as a ratification of Nelson’s actions, Kelleher must have known all the facts and circumstances pertinent to the act at issue. See United States v. Beebe, 180 U.S. 343, 354, 21 S.Ct. 371, 375, 45 L.Ed. 563 (1901); Groves v. United States, 202 Ct.Cl. 660, 678 (1973). See also EWG Associates, Ltd. v. United States, 231 Ct.Cl. 1028, 1030 (1982) (ratification must be based on a demonstrated acceptance of the agreement at issue). Plaintiff attempts to distinguish United States v. Beebe, supra, on the grounds, inter alia, that the government was acting in a sovereign, not proprietary, role at the time. It is clear, however, that the holding was not premised on this distinction. See United States v. Beebe, supra, 180 U.S. 354, 21 S.Ct. at 375.
Kelleher did not recall, as stated above, reading the memos regarding reduction to practice or the letter from Hazeltine referencing the “agreement” entered into by Nelson, or having discussions with anyone regarding either subject. The court does not believe it would have been unreasonable for Kelleher not to have made a “searching” review of the other materials available to him in view of the fact that he was also presented with the ’598 contract signed by Hazeltine that included the Patent Rights clause, but which made no mention of an acknowledgment or agreement relating to reduction to practice. Further, even if it was negligence for Kelleher not to have reviewed the file before him completely, that is insufficient to establish ratification. See United States v. Beebe, supra, 180 U.S. at 354, 21 S.Ct. at 375. In sum, plaintiff has not persuasively shown ratification of any agreement by Kelleher. The record, at the very least, preponderates in favor of this finding.14
Also, it is arguable that the court should not look beyond the four corners of the contract document. The ’598 contract contained a provision limiting the rights and obligations of the parties to the terms of the contract, the solicitation, and such other representations as were attached or incorporated therein. The contract also expressly provided, ante p. 431, that “[n]o oral statement of any person * * * shall modify or otherwise affect any provision of the contract.” (Emphasis added.) The alleged agreement that plaintiff relies upon for the estoppel is founded, as stated above, upon the February 23, 1973 telephone conversation between Nelson and Tedford. Again, the ’598 contract evidences neither the fact of this conversation nor the asserted result thereof.
Lengthy and detailed contract documents are generally deemed conclusive as to contractual relationship between the parties and in derogation of negotiations whose resolution is in the contract. Willems Industries v. United States, 155 Ct.Cl. 360, 369, 295 F.2d 822, 827 (1961), cert. denied, 370 U.S. 903, 82 S.Ct. 1249, 8 L.Ed.2d 400 (1962) (citing Holmberg v. United States, 91 Ct.Cl. 501 (1940)). See also SCM Corp. v. United States, 219 Ct.Cl. 459, 464, 595 F.2d 595, 598 (1979) (“Oral understandings which contemplate the finalization of the legal obligations in a written form are not contracts in themselves.”). Here, it bears noting that the parties had specific discussions regarding the operation of the Patent Rights clause.
In some circumstances, on the other hand, “a separate, prior oral agreement may be valid if it serves as consideration or inducement for a subsequent written contract and does not vary the terms of the written contract.” Prestex, Inc. v. United States, 3 Cl.Ct. 373, 378 (1983) (citing *442McCloskey v. United States, 66 Ct.Cl. 105, 127 (1928)). Even there, however, the pri- or oral agreement is invalid and unenforceable if it conflicts with or varies the terms of the written contract. Prestex v. United States, supra, 3 Cl.Ct. at 378.
The court believes that the alleged agreement here should not, in any event, be given effect. The Patent Rights clause of the ’598 contract, ante pp. 425-26, required Hazeltine to notify the contracting officer if there was an actual or constructive reduction to practice of an invention in the course of the contract in order that the contracting officer or the Secretary might make inquiries or reach determinations relative thereto. That the alleged agreement would impact on this clause is plain—the clause would be rendered surplusage with respect to the open array in the context of a precontract agreement of actual reduction to practice. Specifically, the agreement would vary the terms of the ’598 contract by eliminating the possibility that the government could acquire rights under the contract in or to the principal subject thereof, even if it turned out that the open array had not actually been reduced to practice before the contract. In view of the purposes of the Patent Rights clause and the discussions between the parties, the court believes, under the law, and the language of the contract, that the alleged agreement would, at the very least, have to be in writing and be expressly referred to in the ’598 contract. See also Heckler v. Community Health Services, 467 U.S. 51, 65, 104 S.Ct. 2218, 2226, 81 L.Ed.2d 42 (1984) (strong argument exists that an estoppel cannot be erected on the basis of oral advice).15
The final item for analysis here centers on the principles of estoppel. In light of the above discussion and findings, principally that there was no agreement, the discussion here will not be as detailed as might otherwise be required.
As set forth in Emeco Industries v. United States, 202 Ct.Cl. 1006, 485 F.2d 652 (1973) {Emeco), the following four elements must be present in order to establish an estoppel:
(1) The party to be estopped must know the facts; (2) he must intend that his conduct shall be acted on or must so act that the party asserting the estoppel has a right to believe it is so intended; (3) the latter must be ignorant of the true facts; and (4) he must rely on the former’s conduct to his injury.
[Id. at 1015, 485 F.2d at 657 (quoting United States v. Georgia-Pacific Co., 421 F.2d 92, 96 (9th Cir.1970).]
As to the first element, the government did not know the true facts. Hazeltine’s Volume II proposal was the only concrete information the government had regarding the tuned array antenna. The proposal, however, did not reflect the true nature of the test model. For example, the proposal stated that the dimensions of the model tuned reflector (open section) were 4 feet (48 inches) by 4 feet; the actual dimensions were approximately 32 inches wide by 21 inches high. The size of the model, as will be discussed later, would have had an effect on the view of one of ordinary skill in the art relative to whether a reduction to practice had taken place. LaRussa, in his review of the proposal, was not sure the proposal would work and felt that the ‘598 contract was necessary to see if it would work. The inaccurate and incomplete information provided by plaintiff simply cannot form a basis for a finding that the government had sufficient knowledge of the facts involved to make estoppel proper.16 Nor is the court aware of any author*443ity that would place the burden of inquiry upon the government.
The second element goes to the reasonableness of the plaintiff’s reliance upon defendant's action. See Emeco, supra, 202 Ct.Cl. 1017, 485 F.2d 658. Under the facts and circumstances of this case, the court holds that any reliance on the alleged agreement was unreasonable. It seems clear to the court that plaintiff relied more on its own determination that it had reduced the invention to practice, which would place it outside the pale of the Patent Rights clause, rather than on any assertions of Nelson or actions of Kelleher.17 In addition, some of the documents plaintiff relies on to show the basis for an agreement were never disclosed to plaintiff prior to litigation. Cf. Schweiker v. Hansen, 450 U.S. 785, 789-90, 101 S.Ct. 1468, 1471-72, 67 L.Ed.2d 685 (1981) (per cu-riam ) (internal manual does not bind government).
More importantly, the Federal Circuit has stated that reduction to practice is “a legal conclusion,” i.e., a question of law. Barmag Barmer Maschinen-Fabrik AG v. Murata Machinery, Ltd., 731 F.2d 831, 837, 221 USPQ 561 (Fed.Cir.1984) (quoting D.L. Auld v. Chroma Graphics Corp., 714 F.2d 1144, 1151, 219 USPQ 13 (Fed.Cir. 1983)). In a situation such as the instant case, the court believes that the following language of Judge Bennett is particularly pertinent:
An experienced Government contractor like plaintiff cannot seriously come into this court and complain that it was put into a straightjacket because it took its legal advice from the opposing party and it turns out the advice was not very good. [CRF v. United States, 224 Ct.Cl. 312, 344-45, 624 F.2d 1054, 1072 (1980) (Bennett, J., concurring in part and dissenting in part) (citations omitted).]
Indeed, in Airmotive Engineering Corp. v. United States, 210 Ct.Cl. 7, 535 F.2d 8 (1976), the plaintiff argued that defendant’s prior legal interpretation estopped defendant from taking certain action. The Court of Claims replied—“One who relies upon a legal interpretation by a government official [even one with authority] assumes the risk that it is in error.” Id. at 12-13 n. 3, 535 F.2d at 11 n. 3 (citation omitted). Thus, assuming that TSC agreed with plaintiff that its tuned array antenna had been reduced to practice prior to the ‘598 contract, any such agreement, which rests on a legal conclusion, would not be binding on a court subsequently called on to make such a legal conclusion.
With respect to the third element required for estoppel, i.e., lack of knowledge on plaintiff’s part, plaintiff had complete knowledge of the facts relating to the reduction to practice issue. Defendant had only what knowledge plaintiff saw fit to give it, which was very little. As regards the Patent Rights clause itself, it is noted that Onders, then General Patent Counsel for Hazeltine, knew that the Patent Rights clause would be part of the ’598 contract and “would have whatever effect it would have.” Under such circumstances, plaintiff’s plea for estoppel has a hollow ring.
As to the fourth element of estoppel, the court is not persuaded that plaintiff relied on TSC assurances and/or actions in entering into the ’598 contract. Although the plaintiff might have royalty rights vis-a-vis the government relative to the tuned array antenna were it not for the ’598 contract, there is some question as to whether plaintiff would have invested the time and money to develop the tuned array antenna on its own. Indeed, the ’598 contract provided plaintiff with an excellent opportunity to establish in Phase I the viability of its concept and to design and develop the concept into a concrete item in Phase II of the *444contract, all at government expense. Since the only application for the antenna appears to be as part of ATCRBS, there would be virtually no domestic market for such a product once TSC had committed itself to an ATCRBS improvement program pursuant to the ’070 RFP.18
In addition to the four elements of estop-pel discussed above, plaintiff also must demonstrate affirmative misconduct on the part of government officials in order for equitable estoppel to be applied against the government. First National Bank of Louisa, Ky. v. United States, 6 Cl.Ct. 241, 244 (1984). See, e.g., Vickars-Henry Corp. v. Board of Governors, 629 F.2d 629, 635 (9th Cir.1986). Furthermore, mere negligence by government officials is not sufficient to establish affirmative misconduct. See TRW, Inc. v. Federal Trade Commission, 647 F.2d 942, 951 (9th Cir.1981). Plaintiff’s argument in this regard is that “[t]he government has engaged in affirmative misconduct by denying its previous agreement.” This “boot-strapping” cannot suffice as it would clearly have the effect of eliminating the affirmative misconduct requirement. Simply put, the government does not engage in affirmative misconduct by challenging the validity of its employees’ acts or agreements. Nor is the court aware of any other facts in the instant case that could constitute affirmative misconduct on the part of government personnel.
In summary, the court has found that there was no agreement between the parties or acknowledgment by the government regarding reduction to practice. Assuming such an agreement could be found, the court holds that the facts are otherwise insufficient, as a matter of law, to estop the government from contesting the bona fides of such an agreement. Plaintiff is and was an experienced government contractor and had its patent staff involved in this matter throughout the negotiation process. Whether consciously or not, plaintiff did not insist on an inclusion of the alleged agreement in the ’598 contract. The court will not, on the record before it, do now for plaintiff what plaintiff attempted to do before it executed the ’598 contract.19 The court does not believe that equitable considerations support plaintiff’s position, quite apart from the legal deficiencies of its position.
The most that can be said for the pre-contract negotiations is that each party tended to play its cards close to the chest. See Singer-General Precision v. United States, 192 Ct.Cl. 435, 447, 427 F.2d 1187, 1193 (1970). Both parties recognized the import and impact of the Patent Rights clause on the matter at hand. Each, represented by patent counsel, signed the ’598 contract, with said clause contained therein and no exclusion from its coverage and reach provided. In Bendix Corp. v. United States, 220 Ct.Cl. 507, 516, 518-19, 600 F.2d 1364, 1370-72, 204 USPQ 617 (1979), cited by plaintiff, the contractor rejected the boiler plate Patent Rights clause that defendant proposed to insert in the contract the parties were considering. After months of negotiations, the parties entered into a contract wherein they agreed that the government would not receive a license to use the invention in question. As a result, the agreement, reflected in the contract, was precisely the same as that de*445manded by Bendix when it rejected the government’s original proposal to enter into a contract with the possibility of a license to use the invention. Thus, the contract in the Bendix case accurately reflected the intentions of the parties. In the case at bar, the ’598 contract accurately reflected the intentions of the parties that the Patent Rights clause would be effective and controlling. Under such circumstances, it is just to allow the provisions of the contract the parties executed with full knowledge of the circumstances to control, unfettered by any prior perceived oral agreement.
C. Reduction to Practice
Plaintiff claims that the tuned array antenna was reduced to practice prior to award of the '598 contract. Acceptance of such a claim would put the ’977 patent beyond the reach of the terms of the Patent Rights clause of the contract. Plaintiff’s claim in this regard rests on tests performed on an experimental model of the invention during April 5-6, 1972. Defendant maintains that the tuned array antenna was not reduced to practice before the ’598 contract and counters that the invention was actually reduced to practice during the performance of the contract.
After careful consideration of the facts of record and the law applicable thereto, the court concludes that the tuned array antenna had not been reduced to practice prior to award of the ’598 contract, but rather was reduced to practice, both actually and constructively, during performance of the ’598 contract. Accordingly, pursuant to the Patent Rights clause of the contract, defendant is free to procure and use the accused devices, i.e., the antennas manufactured by TI and RSi, without payment to plaintiff.
Before discussing the evidence and the law bearing on this issue, and the conclusions to be drawn therefrom, a few general, preliminary observations regarding the subject at hand are in order. In the case at bar, the question of “reduction to practice” arises in the context of a “Government Rights” defense. However, the issue of whether an invention has been “reduced to practice” can arise in other contexts as well. See, e.g., Lockheed Air Corp. v. United States, 213 Ct.Cl. 395, 405-06, 553 F.2d 69, 193 USPQ 449 (1977) (whether publication was “prior art” in context of 35 U.S.C. § 102(a)); Barmag Barmer Mas-chinenfabrik A.G. v. Murata Machinery, Ltd., 731 F.2d 831, 838-39, 221 USPQ 561 (Fed.Cir.1984) (Barmag) (whether invention had been “on sale in this country” more than a year prior to the patent application so as to make issuance of a patent improper under 35 U.S.C. § 102(b)); Winter v. Lebourg, 55 C.C.P.A. 1212, 394 F.2d 575, 157 USPQ 574 (1968) (passim) (interference action); Kimberly Clark Corp. v. Johnson & Johnson, 745 F.2d 1437, 1444-45 (Fed. Cir.1984) (whether work qualified as “prior art” under 35 U.S.C. § 102(g)). The Federal Circuit has indicated that the term “reduction to practice” has the same meaning in each context, at least with respect to “on sale” and interference cases and insofar as its precedents are concerned. See Bar-mag, supra, 731 F.2d 838 n. 6. The court here can find no principled justification for not turning to “other” reduction to practice cases for guidance, governed by appropriate precedents of the Court of Claims and the Court of Customs and Patent Appeals, absent relevant Federal Circuit holdings. See South Corp. v. United States, 690 F.2d 1368, 1370, 215 USPQ 657 (Fed.Cir.1982).
In dealing with the reduction to practice issue in this case, the burden of proof is on plaintiff. Plaintiff must show that the tuned array antenna was reduced to practice prior to award of the ‘598 contract. This is so as a legal principle because Hazeltine “is the plaintiff and * * * the party asserting the prior reduction to practice.” McDonnell Douglas Corp. v. United States, supra, 229 Ct. Cl. at 329, 670 F.2d at 161 (McDonnell Douglas).
There is some authority, prior to McDonnell Douglas, indicating that defendant bears the burden of proof on the license defense and must establish by a preponderance of the evidence that a reduction to practice occurred in the performance of a *446government contract. Technical Development Corp. v. United States, 220 Ct. Cl. 128, 151, 597 F.2d 733, 746 (1979). See Williams v. Adm’r of NASA, 59 CCPA 1329, 463 F.2d 1391, 1400, 175 USPQ 5 (1972), cert. denied, 412 U.S. 950, 93 S.Ct. 3013, 37 L.Ed.2d 1003 (1973); Hummer v. Adm’r of NASA, 500 F.2d 1383, 1388, 183 USPQ (C.C.P.A.1974). See also Fitch v. Atomic Energy Commission, 491 F.2d 1392, 1394, 181 USPQ 41 (C.C.P.A.1974). In its briefs, plaintiff did not dispute the government’s assertion that it (plaintiff) bore the burden of proof. At oral argument, however, plaintiff stated, without citing any of the above cases, that it believed the burden is on the government to establish a post-award reduction to practice.
The possible inconsistency relative to whom the burden of proof is upon is not critical in this case for two reasons. First, as discussed below, the burden is effectively on plaintiff pursuant to the terms of the ’598 contract. Secondly, even if the burden is cast upon defendant, it carried the burden of proving, by a preponderance of the evidence, that the tuned array was first reduced to practice during the ’598 contract.
As stated above, the burden of proof is also placed on plaintiff pursuant to the terms of the ’598 contract. Paragraph (e) of the Patent Rights clause, see ante, pp. 425-26, declared that there is a “prima facie presumption” that an invention was first actually reduced to practice during the contract if such invention was constructively reduced to practice during the contract. The filing of a patent application is a constructive reduction to practice. See, e.g., In re Katz, 687 F.2d 450, 454, 215 USPQ 14 (C.C.P.A.1982). It will be recalled that Ha-zeltine filed a patent application for the tuned array antenna on June 25, 1973, two months after the ’598 contract was awarded. Accordingly, Hazeltine has the burden, under the contract, of proving that the tuned array antenna was not actually reduced to practice during the contract.20
Relative to the nature of plaintiff’s burden, the Court of Claims stated in Lockheed Aircraft Corp. v. United States, 213 Ct. Cl. 395, 553 F.2d 69, 193 USPQ 449 (1977) that:
* * * it is well established that the burden of proof of an inventor’s alleged conception and reduction to practice is a heavy one requiring full corroboration by other than the inventor’s own self-serving testimony or records.
[Id. at 405, 553 F.2d at 74 (citation omitted)]. See, e.g., Eastman Kodak Co. v. E. I. DuPont de Nemours & Co., 298 F. Supp. 718, 728, 161 USPQ 150 (E.D.Tenn. 1969) (“Proof of an alleged inventor’s conception and reduction to practice requires full corroboration by other than the inventor’s own self-serving testimony.”) (citations omitted).] The Court of Customs and Patent Appeals applied a “rule of reason” with respect to the evidence needed to establish the credibility of the inventor. See, e.g., Reese v. Hurst, 661 F.2d 1222, 211 USPQ 936 (C.C.P.A.1981). Additionally, that court stated that the inventor’s test results may be sufficient corroboration in some cases. See Mattor v. Coolegem, 530 F. 2d 1391, 189 USPQ 201 (C.C.P.A.1976). For purposes of this discussion, the court will assume that Hazeltine’s test results are entitled to “full” weight, i.e., are no more or less credible than any other evidence. The court also has utilized the standard preponderance of the evidence test, since plaintiff’s failure to carry that burden a fortiori would satisfy any stricter standard that might be required by Lockheed Aircraft Corp. v. United States, supra.
*447McDonnell Douglas, supra, held that a reduction to practice occurs “when it is established that the invention will perform its intended function beyond a probability of failure.” McDonnell Douglas, supra, 229 Ct.Cl. at 329, 670 F.2d at 161 (citing General Electric Co. v. United States, 228 Ct.Cl. 192, 204, 654 F.2d 55, 60, 211 USPQ 867 (1981); Eastern Rotorcraft Corp. v. United States, 181 Ct.Cl. 299, 302-03, 384 F.2d 429, 431, 155 USPQ 729 (1967)). Unless a device is so simple that complete construction of it is sufficient to demonstrate its workability, testing of the invention is required. Eastern Rotorcraft Corp. v. United States, supra, 181 Ct.Cl. at 302, 384 F.2d at 431, cited in McDonnell Douglas, supra, 229 Ct.Cl. at 329, 670 F.2d at 161. Plaintiff concedes here that testing of “Wheeler’s Concept” was required to establish a reduction to practice thereof.
Since it is undisputed that the tuned array’s efficacy is not self-evident, “[t]he question is whether the tests done by [Ha-zeltine] prior to [April 25,1973] ‘were sufficiently comprehensive to demonstrate the workability of the device.’ ” McDonnell Douglas, supra, 229 Ct.Cl. at 329-30, 670 F.2d at 161 (quoting Bendix Corp. v. United States, 220 Ct.Cl. 507, 517, 600 F.2d 1364, 1370, 204 USPQ 617 (1979). This view is in accord with that of the Federal Circuit, which has noted that “reduction to practice requires that an invention be ‘sufficiently tested to demonstrate that it will work for its intended purpose.’ ” Barmag, supra, 731 F.2d at 838 (quoting General Electric Co. v. United States, supra, 228 Ct.Cl. at 201 n. 8, 654 F.2d at 60 n. 8). Some of the reduction to practice cases examined by the court can be read to state a somewhat differently phrased test requirement. Close analysis of these cases, the court believes, indicates that the articulation of the test requirement is governed by the particular facts in each case and does not suggest or indicate a conflict among the courts relative to the requirement for or the type of testing needed in support of a reduction to practice claim. In this regard, the court views the following language as especially pertinent—“[T]he nature of the testing which is required to establish a reduction to practice depends upon the particular facts of each case, especially the nature of the invention.” Gel-lert v. Wanberg, 495 F.2d 779, 783, 181 USPQ 648 (C.C.P.A.1974).
At the outset, the court must decide what, if anything, could have been reduced to practice by the precontract experimental model, i.e., the test model built and tested prior to award of the ’598 contract.21 In order for a test structure to constitute an actual reduction to practice of the device claimed by the patent, the structure must include every essential limitation of the claim(s). See, e.g., Fredkin v. Irasek, 55 CCPA 1302, 397 F.2d 342, 347, 158 USPQ 280, cert. denied, 393 U.S. 980, 89 S.Ct. 450, 21 L.Ed.2d 441 (1968), cited in Correge v. Murphy, 705 F.2d 1326, 1329, 216 USPQ 753 (Fed.Cir.1983).22 In this re*448gard, the court finds that a plurality of elements and a mounting of the same on the columns were essential limitations of the claims.
The claims involved in the present suit are claims 1-9 and 11-13 of the ’977 patent; claim 10 is not at issue. Claims 2-12 require, inter alia, “a plurality of [i.e., more than one] * * * antenna elements for radiating wave energy, said antenna elements being mounted on at least some of said columns * * * ”; claim 13 is literally worded differently, but has essentially the same limitations. In claims 6 and 13, the antenna elements are specified, as dipoles; in the other claims, they are simply “linearly polarized.”
Plaintiff proposed no findings supporting a reduction to practice of claims 2-13, and admitted that these claims do not literally read upon the first experimental model. Nevertheless, plaintiff contends that the precontract model was “equivalent in structure to the structure defined by claims 2-13.” The only evidence at trial that supports this contention was Dr. Wheeler’s expression of his belief that the experimental model and claims 2-13 were “equivalent,” which he defined as having “substantially the same equivalent dimensions and substantially the same essential performance.”
Plaintiff did not cite any persuasive authority relating to the reduction to practice of “equivalent” claims in the manner of an exception from the rule of Fredkin v. Ira-sek, supra. The Court of Customs and Patent Appeals (CCPA) long held that reduction of an “equivalent” was insufficient to constitute a reduction to practice of the claimed invention in the context of an interference proceeding. See, e.g., Martin v. Snyder, 41 CCPA 1010, 214 F.2d 177, 180, 102 USPQ 306 (1954); Smith v. Stone, 57 CCPA 884, 420 F.2d 1065, 1068, 146 USPQ 453 (1970).
On the other hand, it also appears that the CCPA also had a rule that, for purposes of antedating a prior reference under 35 U.S.C. § 102(a) or (e) before the Patent and Trademark Office, a reduction to practice as to the basic invention (claim) would also suffice as to claims differing therefrom in details that were obvious to a person of ordinary skill in the art. Application of Spiller, 500 F.2d 1170,1177-78,182 USPQ 614 (C.C.P.A.1974) (Spiller). See Application of Wertheim, 541 F.2d 257, 173,191 USPQ 90 (C.C.P.A.1976) (Miller, J., dissenting in part and concurring in part). The court refused, however, to apply the principle of Spiller in an interference proceeding. Wetmore v. Quick, 536 F.2d 937, 942-43, 90 USPQ 223 (C.C.P.A.1976).
The court believes the case at bar is more akin to an interference proceeding than to a prior reference proceeding because of the burdens of proof involved and therefore concludes that the principle of no equivalence expressed in, e.g., Martin v. Snyder, supra, is applicable. Even if Spiller were applicable, there was no showing that a person of ordinary skill in the art in 1972 would have regarded successful operation with an array of dipoles obvious in light of the precontract testing of the experimental model.23
Indeed, when the post contract model built under Phase I of the ’598 contract (the “second” experimental model) was first tested with an array of dipoles, its backlobe suppression was far less than expected. Although this problem, i.e., “mu*449tual coupling” {see note 8, supra), was not exceptionally difficult to solve, there was credible evidence that this problem was not anticipated by Hazeltine at the time the precontract model was tested. Thus, the precontract model, which did not include the dipole feeds that were the cause of the mutual coupling problem, was inadequate as a prediction of the performance of claims 2-13 and cannot serve as a reduction to practice thereof. See McDonnell Douglas, supra, 229 Ct.Cl. at 330, 670 F.2d at 161. The court concludes that claims 2-13 could not have been reduced to practice before tests proving the compatibility of the components were performed.24 See id. at 332 n. 7, 670 F.2d at 162 n. 7 (“Especially when the invention is so complex, it is difficult to maintain that it was reduced to practice [at the earlier date] when the compatibility of its components remained in question until [the later date].”)
There remains for consideration whether the first experimental model could have served as a reduction to practice of claim 1 of the ’977 patent.25 Claim 1 requires, inter alia, “a reflector with an open construction” and “at least one linearly polarized antenna element for radiating wave energy, arranged with the polarization of said element substantially parallel to [the] columns” and with the location of the tuned arrays “selected to cause suppression of leakage of said radiated wave energy-”
Addressing the limitation of linear polarization and suppression of radiation first, defendant’s criticism of the tests made on the first model centers on its use of a single monopole. Wheeler conceded that the ’977 patent was not directed to a monopole antenna. Defendant maintains that a monopole has a variable polarization—vertical at the horizon (ground plane), but gradually changing to predominately horizontal directly over the end of the monopole. Defendant then contends that the polarization of the element (monopole) in the first model was not parallel to the columns and that the tuned arrays failed to suppress the horizontally polarized radiation, thus failing to meet those two requirements of the claim.
Although a monopole does have a horizontal polarization component at certain locations, the court believes it is fair to say that the monopole’s polarization was substantially parallel to the columns and that the tuned arrays were directed at suppression of this vertically polarized radiation.26
The crux of defendant’s attack on the use of a single monopole is that the single element radiated a spherical wave whereas ATCRBS generates (and receives) plane waves. In a spherical wave, the phase is constant over the surface of the sphere. In a plane wave, the phase is constant over the surface of the plane.27 The response of *450a screen, i.e., the tuned arrays, to a spherical wave i.e., one produced by an monopole, is different from its response to a plane wave, i.e., one produced by an array of dipoles, because with a spherical wave, the “excitation,” of each tuned rod, i.e., the energy impinging upon it, has a different phase.
On the other hand, defendant did not criticize the test parameters of Phase I of the ’598 contract and, in fact, used the results of those tests as the basis for entering Phase II. The Phase I model had a maximum dimension of 8 feet and thus required, using LaRussa’s formula, roughly 128 feet for plane wave response (2 x 82/l foot, the agreed wavelength for ATCRBS). The distance between the two antennas was, however, only 30 feet. The court finds this fact somewhat inconsistent with defendant’s criticism of the precon-tract testing as regards wave type and concludes that the precontract testing was not fatally flawed for failing to provide a plane wave field. *
The court believes defendant’s criticisms regarding the use of a monopole in the testing procedure are more properly directed at the probative value of the tests performed on the precontract model as indicating a reduction to practice of the open array antenna. That is, the court finds that claim 1 of the ’977 patent, insofar as it relates to a linearly polarized antenna element and suppression of radiation leakage therefrom, reads upon the first experimental model. However, the use of the monopole, and the circumstances of its use, in the testing of the precontract experimental model weaken plaintiff’s claim that the invention had been reduced to practice as a result of such tests.
Defendant also contends that the precon-tract test model did not meet the “open” limitation of claim l.28 The overall size of the first model was 130 inches wide by 43.7 inches high; the opening was 31.6 inches wide by 21.2 inches high. Column 2 of the ’977 patent specified that the construction was “open” when “the spacing between the columns and arrays is greater than the cross section of the arrays and columns.” In practical terms, the ’977 patent describes a structure that is at least 50 percent open.
All of the experts at trial agreed that the precontract test model could not properly be characterized as open.29 They further agreed that the test results reflected the performance of the entire structure, not just the opening. One of defendant’s experts, Phillip H. Richardson (Richardson), suggested that it would be possible, however, to conduct tests on the first model in such a manner as to show what the suppression due to the reflective grid, i.e., the tuned arrays, was.
Richardson received a bachelor’s degree in electrical engineering in 1966 and a master’s degree in 1967 with a specialty in electromagnetics, viz., antennas and microwaves. At the time of trial, he had written a few technical articles on these subjects, been granted one patent and had another application pending, and had taught college courses in electrical engineering, generally including antennas and microwaves. Richardson had five years of relevant work experience in 1972 and 18 years at the time of trial. Richardson was employed by TI as an antenna engineer at the time of trial. Plaintiff challenges Richardson’s credibility since TI would have to indemnify the Federal Government relative to any judgment favorable to plaintiff herein growing out of sales of open array antennas manufactured by TI for the government. While these facts do impinge on the weight that one gives Richardson’s testimony, in light of his expertise and experience in the subject matter at hand, together with his demeanor and performance at trial, his employment status and related circumstances do not *451justify total disregard of his testimony and opinions.
The court believes defendant’s criticisms regarding the structure of the precontract test model are more properly directed at the weight to be given the tests performed relative to a reduction to practice determination. Richardson’s testimony suggests that tests could have been performed on the precontract test model in a way that would have allowed one to measure the performance (suppression) of the reflective grid, which, by itself meets the openness requirement. That is, the precontract test model could have been used to determine if the opening containing the tuned arrays, upon which claim 1 reads, achieved its objectives, namely, baeklobe suppression. Whether or not the tests performed on the precontract test model were sufficient to allow one to draw such conclusions is the next topic of analysis.
As stated above, a device is reduced to practice when it is shown that the device will perform its intended function beyond a probability of failure. The tests performed on the precontract test model can only be characterized as laboratory tests, viz., they were not made under actual use or service conditions. Nonetheless, laboratory tests may still constitute a reduction to practice “if the conditions of the tests adequately simulate the conditions of practical use.” Technical Development Corp. v. United States, 202 Ct.Cl. 237, 308-09, 179 USPQ 180 (1973) (citation omitted), cert. denied, 416 U.S. 983, 94 S.Ct. 2384, 40 L.Ed.2d 759 (1974). The efficacy of the tests as constituting an actual reduction to practice is determined by examining what one of ordinary skill in the art would conclude from the test results. Winter v. Lebourg, supra, 394 F.2d at 581. Thus, the question is whether the test results garnered from the precontract experimental model would have convinced a person of ordinary skill in the art that the device would work beyond a probability of failure.
Dr. Jean-Claude Sureau (Sureau) was plaintiff’s expert witness at trial. Sureau received a bachelor’s degree in electrical engineering in 1961, a master’s degree in electromagnetics in 1964, and a Ph.D. in electromagnetics in 1971. He had relevant work experience from 1960 and had specialized in microwave antennas throughout his career.
Sureau was asked, on cross-examination, whether it would have been obvious to someone of ordinary skill in the art in 1972 that the precontract test model would operate in the manner intended in the ’977 patent based on the test results. He responded as follows:
Someone I guess in the context of ordinary skills in 1972 * * * might have been less optimistic then [sic] I would have been at that time or now in reviewing those results. There might have been some—given the results I think that were shown, I think that someone—that person would have been at least encouraged. Whether that person would have considered it to be definite proof or not [; it] is certainly not definite proof that the demensions [sic] are the current dimensions but that the principle [is] at least of some—some merit and it’s probably valid, you know I can’t put a probable figure on it—it could have been. But again, it’s just a—, I would say it’s a gray case.
Shortly thereafter, Sureau testified further, “I’m not sure that it would have been obvious to someone of ordinary skill”; he also admitted having “some doubts” as to whether one of ordinary skill would have thought it would work.
Such testimony, coming as it does from plaintiff’s own expert, makes it exceedingly difficult to conclude that the ’977 patent was reduced to practice prior to the ’598 contract, even without considering defendant’s criticisms of the test methodology. Results “of an encouraging nature” that “may have justified a prediction that the invention would probably be successful” are not sufficient for establishing a reduction to practice—“[w]hat is required is not a mere basis for prediction but an actual demonstration.” See Elmore v. Schmitt, 47 CCPA 958, 278 F.2d 510, 513, 125 USPQ *452653 (1960). Stated another way, “when the invention has not quite passed beyond experiment and has not quite attained certainty and has fallen short of demonstrating the capacity of the invention to produce the desired result, the invention itself is still inchoate.” 1 Deller’s Walker on Patents § 46 at 202 (2d ed. 1964), quoted in Kimberly-Clark Corp. v. Johnson & Johnson, supra, 745 F.2d at 1445.
Although Hazeltine personnel, namely Wheeler, Hannan and Mazzola, testified that the tests convinced them that the tuned array antenna would work, their testimony cannot overcome the court’s conclusion. Presumably, it was only natural for them to concentrate on the positive indications; Wheeler, after all, was the inventor of the tuned array, and the other two were closely associated with him. Accordingly, there is reason to view their testimony with some hesitancy. See Lockheed Aircraft Corp. v. United States, supra, 213 Ct. Cl. 405, 553 F.2d 44. See also, e.g., Eastman Kodak Co. v. E.I. Dupont de Nemours Co., supra, 298 F.Supp. 728. See generally ante p. 446. In addition, they were men of greater than ordinary skill in the art; they did not contradict Sureau’s opinion relative to the view of someone of ordinary skill. Finally, except for Mazzola, there was no strong evidence that they viewed the model or were aware of the possible shortcomings, e.g., the relatively small size of the opening, in the test set-up.
Sureau’s hesitancy regarding what a person of ordinary skill would have concluded from the tests can perhaps be explained by considering the shortcomings of plaintiff’s test as pointed out by defendant’s experts. Consideration of their testimony leads the court to conclude that the ’977 patent was not reduced to practice prior to the ’598 contract.30
The '977 patent is directed at an antenna with “improved suppression of radiation leakage through the reflector.” Sureau testified that the “critical area” in which the suppression had to be verified was 180° 4- 2V20 (azimuth), i.e., over a 5° horizontal angle centered directly in back of the antenna. Sureau further stated that the vertical area of interest (elevation angle) was 0-40°.
In order for tests to be acceptable as proof that the tuned reflector concept would work, two requirements are crucial. First, the test model must have been of sufficient size to exhibit the resonance that is central to the ’977 patent. Secondly, the tests must have been performed on the model in such a manner as to allow one to conclude that the tuned rod structure actually does suppress radiation. Plaintiff’s precontract testing fails to satisfy either requirement.
The portion of the precontract test model representing the tuned array was 31.6 inches wide by 21.2 inches high; it contained four tuned rods and three conductive columns. LaRussa testified that the width of the opening was “too small to characterize the operation of the grid.” LaRussa didn’t give an opinion relative to the minimum size required, but did state that additional columns would have been necessary “to say with reasonable certainty that the suppression was due to the grid.”
Richardson testified that the screen would need to be on the order of eight wavelengths wide to represent its response to a plane wave. The precontract model was tested at a frequency of 1180 MHz, which translates into a wavelength of 10 *453inches.31 Thus, the width of the opening would have to have been approximately 80 inches to simulate plane wave response. In actuality, the width of the test model grid was only about 32 inches, which is less than 3V2 wavelengths wide. Richardson did note, however, that if the data were accurately recorded, “the experiment shows what it shows.”32 Whether the data were accurately recorded and just what the experiment shows are matters of dispute between the parties.
In a deposition, Sureau testified that the “absolute minimum” dimensions of the screen to exhibit the requisite resonance with one element, i.e., the monopole, were 5 wavelengths wide by 2V2 wavelengths high; he would have been “more comfortable” with an opening 10 wavelengths wide by 10 wavelengths high. At the test frequency, these standards would correlate to openings 50 inches wide by 25 inches high and 100 inches wide by 100 inches high, respectively.33 During the deposition, defendant’s attorney emphasized the statement regarding minimum size, saying “I don’t want there to be any confusion about that,” to which Sureau responded “Correct.”
At trial, Sureau attempted to qualify his earlier testimony. He characterized the size of the opening in the precontract test model merely as “an unfavorable condition” that “would tend to diffuse the effect” of the structure. That is, at trial, Sureau felt that the size of the structure only made it less likely that the suppression would be very pronounced; in his view, one would anticipate that any positive results would be even better on a larger scale. He admitted, however, that a person of ordinary skill in the art might have been concerned about the size of the test reflector in the precontract model.
Finally, it cannot be overlooked that when Hazeltine built a model of the open array under Phase I of the ’598 contract to prove the feasibility of the concept, it chose a size of 8 feet wide by 4 feet high. At the ATCRBS frequency, the width satisfies Richardson’s 8 wavelength criterion; the' height was equivalent to final design. Moreover, in contrast to Sureau’s prediction, the large scale model did not perform better than the precontract test model, which plaintiff’s counsel frankly admitted was “crude.” These later tests can be used to judge the efficacy of earlier ones as establishing a reduction to practice. See Brown-Bridge Mills v. Eastern Fine Paper, 700 F.2d 759, 766, 217 USPQ 651 (1st Cir.1983).
The court finds, on the basis of the foregoing, that the size of the opening in the precontract test model, i.e., the reflector structure, was not large enough to demonstrate that any observed suppression was due to resonance of the tuned rods. The testimony of all three experts and the approach taken under Phase I of the ’598 contract strongly supports a conclusion that a person of ordinary skill would have required a larger reflector structure for a test designed to prove the efficacy of the reflector.
The second element necessary to prove that the tuned reflector operated as desired is a test methodology that generated results which could reasonably be construed as proof that the grid suppressed radiation. Because of the manner in which plaintiff performed the tests on the precontract model, and the conditions under which such tests were made, the court concludes that *454there is no way to determine, with any degree of certainty, whether the reflector structure actually provided the requisite radiation.
As noted above, all three experts agreed that the test results represented the performance of the entire precontract experimental model and not just the reflector portion thereof. That is to say, part of the observed suppression was attributable to the solid portion of the model and part to the grid. Additionally, there is no intrinsic way, using only one pattern, to “separate” the grid suppression from the solid suppression.
Plaintiff asserts, however, that the test results can be properly interpreted as proof that the tuned reflectors provided suppression. Plaintiff maintains that the two available original antenna patterns, taken together, prove a reduction to practice of the tuned array concept. As will be recalled, see ante pp. 424-25, one of these patterns was taken with the reflector structure in place (the first pattern) and the other with the opening covered, i.e., with the test structure being comprised, essentially, of a single, solid metal sheet. The latter pattern was denominated the “calibration” pattern, and it theoretically represented complete reflection of all energy impinging upon the metal sheet. Thus, the calibration pattern evidenced the theoretical maximum suppression obtainable from an antenna with gross (overall) dimensions equivalent to those of the precon-tract model.
Since the first pattern showed suppression performance similar to that of the calibration pattern, plaintiff reasons that the tuned array was equivalent to a reflector. In other words, plaintiff believes the tuned reflectors suppressed radiation because the pattern with the grid in place (the first pattern) corresponded closely to the calibration pattern (ideal performance). Under the facts of this case, the court cannot accept plaintiffs reasoning.
In testing the precontract model, the purpose of the large sheet surrounding the opening, and of the horizontal ground plane, was to reduce substantially the diffraction of energy around the edges of the structure, as well as the effects on the radiation pattern introduced by the edges of the structure. With the solid ground plane in place, the opening was closed off; there was, therefore, no leakage through the closed opening. The energy in the 180° area of the calibration pattern (solid vertical ground plane) was the ambient level of energy resulting from range reflections and diffraction around and off of the structure (including the horizontal ground plane). The measurement at 180° was approximately -34dB. The pattern varied rapidly in this area, however, ranging from approximately —31.5dB to —40.5dB for azimuth locations within 2.5° of 180°.
The first pattern, and the two patterns in Hazeltine’s Volume II proposal, were taken with the opening exposed and the tuned reflector structure in place. The first pattern showed a deep null of approximately —45dB at 180°. The null was somewhat symetrical about 180°, but the radiation level did increase approximately 12dB within 5° of 180°. Defendant prepared an overlay using the two original patterns that clearly illustrated the differences between the patterns in the 180° region.
The overlay showed the tuned reflector purportedly providing greater suppression than the solid sheet over much of the critical area of 180° + 2.5°. Such a result could not possibly have been due to the reflective nature of the tuned array inasmuch as the calibration pattern represented zero passage of energy through the open section. That is, even if the tuned array operated ideally, i.e., with 100 percent reflection, it could not have suppressed more of the energy impinging directly upon it than did the solid sheet. Consequently, the drop in the backlobes must be attributed to effects other than the reflective nature of the grid.
The drop is perhaps explained by edge effects introduced by the borders (edges) of the opening or by actual leakage through the reflector, either or both of which could have resulted in destructive interference *455canceling the ambient energy. Such interference suppression would have tended to mask (hide) the actual reflective character of the grid. This factor is significant as it possibly resulted in as much as some — 13dB of additional suppression.34
Plaintiff had no evidence definitively quantifying the amount of interference suppression; its expert estimated that the radiation leaking through was on the order of about — 28dB to — 32dB. In essence, plaintiff’s view was that the calibration pattern established the threshold of credibility for suppression measurements, i.e., suppression in excess of the calibration could be recorded but wouldn’t represent grid reflection. Plaintiff did not appear to be concerned with the effects of interference on suppression levels above the calibration levels.
Both of defendant’s experts, on the other hand, viewed the test results with some skepticism in light of the seemingly better performance of the tuned reflector over the totally closed segment opening. LaRussa’s opinion was that a “significant uncertainty” in the measurements was presented by the differences in the patterns. He did not believe that the test results could be used as showing that the grid suppressed radiation. LaRussa explained his answer as follows:
You haven’t shown me that there is radiation—You haven’t told me what the radiation impinging on the grid is. You have just showed me that for some condition you get a null back there. I don’t know what has caused that null.
LaRussa had some 19 years of experience testing antennas and testified that he “would surely make an effort” to see what was causing a situation such as that found here.
In Richardson’s opinion, a reference measurement should have been taken without any vertical ground plane structure at all in order to determine the signal strength reaching the receiver. Using Hazeltine’s precontract test model, the next step would have been to generate an antenna pattern with the opening free of all columns but not covered over. Richardson testified that this pattern “would show the reference level against which you would compare the suppression that is provided by the screen.” Finally, the grid would have been inserted and the entire structure, i.e., the precontract test model, tested to determine its suppression.35 Then, the suppression due to the screen under the given test parameters, i.e., frequency and elevation angle, could presumably be determined by comparing the patterns.
The court finds that the skepticism of defendant’s experts was warranted and that the available patterns did not demonstrate that the tuned reflector itself provided suppression of radiation. It seems almost axiomatic that one cannot tell if, or how much, the tuned array grid suppressed radiation without knowing how much radiation the grid was exposed to in the first instance. It is clear that the grid itself could not have provided all of the suppression evidenced by the patterns. Hazeltine, however, presented no evidence of an empirical effort to separate out the other factors involved, e.g., destructive interference. Accordingly, on the facts presented, the court is unable to find that the efficacy of the tuned reflector concept was shown beyond a probability of failure on the basis of tests which merely showed that a large metal sheet containing a relatively small section of a tuned array, which section constituted a small portion (less than 12 percent) of the gross (overall) area of the structure, performed better than a solid sheet of the same gross dimensions.
There are two additional reasons to question the weight to be given the test patterns. The first concerns the use of a *456monopole in the precontract test structure, and the second focuses upon the equipment used in generating the test patterns.
As noted above, Hazeltine used a monopole antenna in the precontract test model to simulate the dipole that was ultimately planned for use with the tuned array. Dr. Wheeler testified that one of the reasons for using a monopole was to avoid, in the initial test, any effects of the dipole feed lines. See ante _p. 424. The court has previously discussed other implications of using a single monopole. See ante pp. 449-50.
In an infinite ground plane, a monopole produces a radiation pattern approximating that of a dipole due to a principle known as “imaging.” In a finite ground plane, however, the peak of the beam is tilted upwards toward the zenith; it does not lie along the ground plane. Also, the signal level (strength) varies unpredictably in the area of tilt. The degree of tilt is a function of the size of the ground plane; LaRussa recalled studies that indicated a 15° tilt for a circular ground plane with a radius of 3 wavelengths. Moreover, the uncertainties described herein are compounded when other than a circular ground plane is used. Finally, if the ground plane is not symetrical about the monopole, different levels of radiation may be found in the rearward and forward directions.
Hazeltine used the horizontal ground plane shown in Appendix II as the imaging plane and to suppress radiation leakage below the test structure. The exact size and shape of the horizontal ground plane, and the relative location of the monopole, are unknown. From available pictures, the ground plane appeared to have been square, and it definitely had a diagonal width less than the width of the vertical ground plane (130"); the precise location of the monopole could not be determined.
Hazeltine measured its antenna patterns with a second antenna at an elevation angle of 10° relative to the horizontal ground plane of the precontract test model. Maz-zola testified that this was done “to avoid any problems with edge effects from that ground plane.” Mazzola knew that the size of the ground plane affected the radiation pattern of the monopole, but he didn’t agree, to the best of his knowledge at the time of trial, that the pattern fell off in the manner illustrated by defendant. He nonetheless agreed that the validity of the measurements taken presupposed that there was no fall as illustrated by defendant, viz., the measurements would have to be taken in another region (angle) to be valid if defendant was correct.
On cross-examination of LaRussa, plaintiff seemingly attempted to show that the size of the ground plane did not have an appreciable effect on the results, but the data it presented were for circular ground planes and therefore not applicable to the case at bar. The court is obviously unable, on this record, to quantify precisely the nature and magnitude of the uncertainties introduced by the non-circular, finite and possibly non-symetrical horizontal ground plane. Defendant has, however, persuaded the court that there was some likelihood that the patterns were inaccurate due to a tilting of the beam. In the face of this evidence, the court believes it was incumbent upon plaintiff to show affirmatively that the patterns were reliable since it was in a better position to know (and control), at the time of testing, the facts pertaining to this matter. See Williams v. Administration of NASA, supra, 463 F.2d at 1401.
The other factor lessening the weight to be ascribed to plaintiff’s test patterns is the conditions under which the patterns were taken. The antenna pattern for the tuned array (the first pattern) was noticeably more erratic in the main lobes than the calibration pattern; indeed, the first pattern doubled back on itself at one point. Plaintiff suggested that the irregularities were due to a shaky pen, but the court is of the view they were the result of stray radiation.
As noted above, ante p. 424, the calibration pattern was taken with the mixer wrapped with an absorber and the cable with aluminum tape. These steps were taken to reduce the possibility of interfer*457ence from spurious signals and are done if one suspects the presence of extraneous signals. In addition, the common practice is to note the changes, i.e., the addition of the absorber and the wrapping of the cable, on the first pattern taken with them.
From other notations that appear on the two patterns, i.e., the numbers, the court finds that the first pattern was taken before the calibration pattern, and, consequently, that it was not taken with the absorber on the mixer or the cable wrapped. Since these are the only apparent changes in the test regime, the court further finds that stray or leaking radiation most probably was a cause of the irregularities in the first pattern.36 The effect this radiation had on the measurements in the critical region around 180° is unknown here also, but its presence does raise a question as to the credibility of the measurements.
It is the court’s opinion that if the results of tests are to be used for comparison in support of a proposition, it must be shown that the tests do not differ in any material way. Plaintiff has not done so here, and, moreover, there is credible evidence supporting the opposite conclusion.
In addition to the shortcomings inherent in the tests that were performed, a conclusion of a precontract reduction to practice is precluded by the fact that certain other tests were not performed. These tests fall into two categories—rain and ice tests, and additional patterns. The court will address the rain and ice issue first.
In order to be reduced to practice, a device must be shown to be operable “in the environment of its practical contemplated use.” See, e.g., Technical Development Corp. v. United States, supra, 202 Ct.Cl. at 308 (citing Elmore v. Schmitt, supra); Eastern Rotorcraft Corp. v. United States, supra, 181 Ct.Cl. at 302, 384 F.2d at 431 (citing Elmore v. Schmitt). In analyzing an asserted reduction to practice under this standard, the first step is to determine the environment in which the device is intended to function.
A fair reading of the ’977 patent makes it clear that the disclosed device is intended for use in an atmosphere including ice and rain, i.e., exposed to the elements, especially precipitation. It is difficult to conceive of how a device the key feature of which is “a reflector which is open to the passage of wind but closed to passage of electromagnetic wave energy” can be said to have an environment of practical application that does not include precipitation. Stated another way, the court does not believe the device disclosed in the ’977 patent would be useful if it couldn’t operate in rain and ice. See 35 U.S.C. § 101; Brenner v. Manson, 383 U.S. 519, 528-29, 86 S.Ct. 1033, 1038-39, 16 L.Ed.2d 69 (1966).
Having defined the intended environment of the invention, the next inquiry is whether the tests performed on the precontract test model support a conclusion that the device was operable in such environment. As the Court of Claims stated in McDonnell Douglas, supra, 229 Ct.Cl. at 331, 670 F.2d at 162, “[t]ests which fail to simulate the varying and multiple conditions of the invention’s intended environment will not serve to prove the operability, stability and reliability of the invention for practical use.” (Citations omitted.)
There is no dispute that the tests performed on the precontract model did not include tests for the effect of rain or ice, either directly or by simulation. Indeed, Dr. Wheeler frankly admitted that the pre-*458contract test model was not designed to take precipitation into account. Additionally, it is clear that precipitation can “de-tune” the array, thereby causing a degradation in antenna performance. See ante pp. 433-34 & n. 9. Plaintiff contends, however, that the nature of the problem was such that no tests were needed.
In support of its position, plaintiff cites the testimony of Hannan, who was listed as “Consultant, Antenna Design,” on the report submitted by Hazeltine pursuant to Phase I of the ’598 contract. Hannan did not think that the structure of the antenna was particularly susceptible to rain or ice and accordingly couldn’t “think of any reason why such a test would be really needed.” Although rain or ice would have some adverse effect on performance, he felt that it was just “an ordinary problem” that occurred frequently in antenna design and which could be solved “very simply.”
Newman, one of plaintiff’s employees, who was extensively involved in the design, fabrication and testing of the open array antenna under the ’598 contract, testified that he didn’t have “serious doubts” about the antenna’s ability to work in rain and ice, but admitted it was “prudent” to test the antenna “to verify that it worked as advertised.” Additionally, in his engineering notebook, Newman listed “rain and ice effects” as a “risk” that had to be addressed.
Dr. Sureau, plaintiff’s expert witness, reviewed the proposals that Hazeltine and others made in response to the ’070 RFP in connection with a position he held at the time with Lincoln Laboratory of the Massachusetts Institute of Technology. Upon reading Hazeltine’s proposal, Dr. Sureau “was concerned that the icing would de-tune the elements and deteriorate the performance of the system beyond acceptable limits.” Before advising the government to purchase an open array antenna for ATCRBS use, Sureau would have required some sort of icing test to be sure that the antenna “would have met the actual environmental requirement as an operational piece of equipment.”
Finally, LaRussa based his opinion that the open array antenna was not reduced to practice prior to the ’598 contract on the fact that the precontract model “was neither built nor tested to demonstrate that it could operate in [its] intended environment,” which he defined as exposed to “[w]ind, rain, ice, snow, sleet, precipitation and a combination of those things.” Clearly, LaRussa viewed testing under precipitation conditions as necessary.
Based on the foregoing, it is the court’s finding that a person of ordinary skill in the art would have required testing for precipitation effects before concluding that the open array antenna would function in its intended, practical environment beyond a probability of failure. Although Sureau and LaRussa stated their opinions in the context of the actual specifications for ATCRBS, the court does not believe that fact lessens the weight to be accorded their testimony. The conclusion is inescapable both that they had doubts that the open array would work in the presence of precipitation and that they would have required tests in support of a conclusion that it would so work.37
Further, the court does not believe that the tuned array can be regarded as reduced to practice, notwithstanding the precipitation problem, under the principle of Farrand, Optical Co. v. United States, 325 F.2d 328, 135 USPQ 249 (2d Cir.1963) (Farrand), upon which plaintiff relies. Briefly stated, Farrand held that changes or improvements necessary for a device to be practically applied did not foreclose an earlier reduction to practice if the earlier device “demonstrated that [the] concept provided the answer” and if it was “ ‘probable’ that the problems remaining [for practical use] could be solved by application of the *459existing art.” Farrand, supra, 325 F.2d at 333. The record here does not support either element.
First, as was set out in great detail above, the court does not believe that the tests of the precontract model demonstrated that the tuned reflector concept would work in any kind of weather, i.e., there was no demonstration that Wheeler’s concept provided the answer.38
In addition, the record does not support the position that the icing problem could have been solved using the existing art.
Although the Hazeltine Volume II proposal stated that a dielectic sleeve would be used in forming the tuned rods, it did not indicate that its purpose was to prevent detuning. Indeed, the proposal stated that heaters would be used to keep the structure free of ice. It was not until the. second design review meeting under the ’598 contract that Hazeltine informed TSC that it intended to use an enclosed air space in the sleeve to protect against precipitation.
It is noteworthy that Hannan, the most experienced engineer working on the ’598 contract, was responsible for designing the dielectric sleeve. Newman, who put significant effort into the ’598 contract and who appears to have fit the description of a person of ordinary skill in the art as of 1972, thought that “[pjerhaps” he could have performed the calculations necessary to design the shield. Richardson testified that the idea of surrounding antenna elements (dipoles) with a dielectric shield was known in the art during the relevant period, but he didn’t think a person of ordinary skill would have known how to “size” the gap so that the tuning would be preserved in the presence of ice. In view of this evidence, the court concludes that a person of ordinary skill in the art probably could not have designed the sleeve necessary for the tuned array to perform as intended.
Plaintiff also relies upon Bendix Corp. v. United States, supra, 220 Ct.Cl. 507, 600 F.2d 1364, and Eastern Rotorcraft Corp. v. United States, supra, 181 Ct.Cl. 299, 384 F.2d 429, in support of its general proposition that the open array was reduced to practice prior to award of the ’598 contract. These cases are, however, easily distinguished from the situation at bar.
In Bendix, the device at issue was intended for use in jet aircraft engines. Plaintiff performed extensive “ground” tests on the device. Notwithstanding that no flight tests were performed, the government purchased the device, in completed form, and promptly flew with it in an aircraft without further ground testing. Not surprisingly, the Court of Claims held that the ground tests demonstrated that the device “could perform its intended purpose beyond a probability of failure,” especially in light of the fact that the government based its decision to purchase the device upon the results of those tests. See Bendix Corp. v. United States, supra, 220 Ct.Cl. 516-18, 600 F.2d 1370-71. Likewise, in Eastern Rotorcraft, plaintiff provided a “sample” to the government, which thereafter awarded a production contract to plaintiff; it does not appear to have been seriously disputed that the device was reduced to practice before the production contract. See Eastern Rotorcraft Corp. v. United States, supra, 181 Ct.Cl. at 305-06, 384 F.2d at 431-32.
Here, on the other hand, there was no completed device for the government even to view, much less test. Furthermore, it was only after the development work was done under Phase I of the ’598 contract that the government agreed, in Phase II, to purchase an actual open array antenna. The simple fact is that the device here was not reduced to practice before the government contract, whereas the devices in Bendix and Eastern Rotorcraft were in corn-*460píete or virtually completed form—they could be used successfully for their intended functions as they stood at the time of contract and thus were deemed to have been reduced to practice.
The court is aware that perfection or lack of need for improvements are not prerequisites to a reduction to practice conclusion. See, e.g., Barmag, supra, 731 F.2d 838-39. The court believes, however, that the modifications required for precipitation protection were not mere improvements or “minor adjustments,” see Bendix Corp. v. United States, supra, 220 Ct.Cl. 518, 600 F.2d at 1371, but rather were “of critical importance to achieving the objectives of the invention,” see Leesona Corp. v. United States, 208 Ct.Cl. 871, 895, 530 F.2d 896, 910, 192 USPQ 672 (1976), and thus preclude reduction to practice before proof of their efficacy. See McDonnell Douglas, supra, 229 Ct.Cl. at 332, 670 F.2d at 163.
The second category of tests needed to be performed for a reduction to practice is additional antenna patterns. As will be recalled, only four test model antenna patterns were available—the two “original” patterns and the two additional patterns found in the Volume II proposal. Again, all four patterns were taken at an elevation angle of 10°. Any estimation of the antenna’s performance at other angles must therefore be based on or derived from these patterns.
The court finds that knowledge of the suppression ability of the precontract test model at other angles was essential to conclude that the tuned array would perform its intended function. Plaintiff never contended that any practical application of the invention included operability at a single, fixed elevation angle of 10°, and it is exceedingly difficult to imagine one. Further, Richardson estimated that greater than 90 percent of all “traffic,” i.e., aircraft, was under 5° elevation in ATCRBS; the specifications delineated the elevation sector at 0°—42°. Since the tuned array was conceived in response to ATCRBS needs and since that is the only application of the invention of which the court is aware, the court is of the view that it is fair to define a range of elevation angles, especially low ones, as part of “the environment of its practical contemplated use.” See, e.g., Technical Development Corp. v. United States, supra, 202 Ct.Cl. at 308.
Sureau thought that one tuned array pattern was the minimum necessary to show backlobe suppression capability over the required elevation sector, viz., 0°—42°. He based his opinion on the fact that the structure of the tuned rods was “periodic,” i.e., recurring at regular intervals, with a “periodicity of roughly a quarter wavelength” and therefore had “performance that tended to be independent of elevation angle.” Sureau admitted, however, it would have been desirable to test as much as possible between 0° and 40°.
Defendant’s experts, LaRussa and Richardson, did not think the patterns could be used to predict performance in the critical 5° region. Both experts essentially agreed that it would not be very reliable to extrapolate down from 10° because of the rapid excursions of the data in the 180° (azimuth) area of the existing patterns.
The court concludes that additional patterns were necessary to show that the tuned array would work as intended.39 Buttressing the court’s conclusion in this regard are Richardson’s testimony that a person of ordinary skill would have tested, inter alia, over a variety of elevation angles and the fact that such a large number of patterns were generated during the '598 contract, upon which tests Hazeltine relied in its interim report as proof that “the open-array ground plane is feasible.”
In sum, the evidence of precontract testing that Hazeltine presented to the court cannot be construed, in any reasonable fashion, as supporting a conclusion that the tuned array was thereby reduced to practice. The test model was incapable of simulating the performance of an “actual” ar*461ray, the test set-up was of questionable reliability, the test methodology did not support the proffered conclusion, an insufficient number of patterns were generated, and absolutely no provisions or tests were made with respect to performance in the presence of precipitation. The two facts inveighing most heavily against a conclusion of a precontract reduction to practice are plaintiff’s own expert's indecision regarding what a person of ordinary skill in the art would have concluded from the tests and the fact that the tuned array evidenced better suppression than the solid sheet, a result that simply cannot be explained by the reflective properties of the invention.
In contrast to the precontract activity, the court concludes that the open array was reduced to practice during performance of the ’598 contract. It was during the contract that a model of sufficient size to simulate tuned array performance was fabricated and put through extensive testing. A test structure with a plurality of dipoles was first constructed, tested and corrected for unanticipated problems during Phase I. In addition, it was not until the ’598 contract was undertaken that a solution to the problem of precipitation was designed and tested to prove its efficacy.
Plaintiff argues that the work done under the ’598 contract merely constituted application of the open array concept to ATCRBS and that defendant evidenced its belief that the concept could be so applied by awarding the ’598 contract to Hazeltine. Although it is true that there is a distinction between a reduction to practice of an invention and subsequent performance testing of the invention to ensure that it meets a customer’s unique specifications, see General Electric Co. v. United States, supra, 228 Ct.Cl. 205-06, 654 F.2d 63-64, the facts of this case do not support such a distinction. The court does not hold that the precontract test model had to have met the range of ATCRBS specifications in order to constitute a reduction to practice of the tuned array concept. Rather, the court holds that the tests performed did not show that the open array would meet its objective, i.e., improved suppression, in pertinent situations. Nor can it fairly be said that the government awarded the ’598 contract in a belief that the concept was reduced to practice and only needed minor refinements to be applied to ATCRBS. The government characterized the approach as a “high risk” one. There was no evidence of an informed reduction to practice conclusion at TSC, and LaRussa, with some 20 years of antenna experience, could not tell whether or not the antenna would work merely from the information contained in Hazel-tine’s Volume II proposal. In addition, the nature of the tests done during Phase I of the ’598 contract works against plaintiff’s “specific application” argument. These tests were of general utility and import, viz., to prove the suitability of the tuned array generally, not just for ATCRBS, and they were performed, using Hazeltine’s own words, to “demonstrate[ ] feasibility of the open array comcept [sic] using tuned reflectors.” Had plaintiff, prior to award of the ’598 contract, performed tests in the nature of those done under Phase I, a different result might well be warranted.
III. Conclusion
The court has found as a fact that there was no agreement between Hazeltine and Nelson regarding a precontract reduction to practice of the open array antenna. Further, the court has found the facts legally insufficient to justify invoking an estoppel against defendant’s raising the Government Rights defense.
The court has also concluded that the precontract test model failed to reduce to practice any of the claims of the ’977 patent in issue. The model was lacking essential elements (limitations) of all the claims in issue but one, and it was not constructed in a manner that would have allowed it to simulate the performance of that claim satisfactorily to a person of ordinary skill in the art. The tests that were performed on the model were inconclusive and incomplete. They did not demonstrate what portion of the observed suppression was due to the reflective nature of the array, which *462the court regards as crucial, when it is obvious that other factor(s) were involved. Also, plaintiff failed completely to perform precontract testing for precipitation effects.
Under the ’598 contract, on the other hand, considerable time, effort and money was devoted to proving the efficacy of the tuned array concept. The money funding this investigation was provided by the government which, rather than Hazeltine, thus bore the risk that its investment could ultimately be for naught. The court believes the following language from McDonnell Douglas, as appropriately modified, covers the situation herein and supports a holding in favor of defendant—
One major aim of the [Patent Rights clause] is to protect the Government from having to pay for using an invention which it significantly helped to develop. See Mine Safety Appliances Co. v. United States, 176 Ct.Cl. 777, 789, 364 F.2d 385, 392 (1966); Ordinance Eng’r Corp. v. United States, 68 Ct.Cl. 301, 353 (1929), cert. denied, 302 U.S. 708 [58 S.Ct. 28, 82 L.Ed. 547] (1937); Techni-trol, Inc. v. United States, 194 Ct.Cl. 596, 613, 440 F.2d 1362, 1372 (1971); Technical Development Corp. v. United States, 202 Ct.Cl. 237, 255-56 (1973), cert. denied, 416 U.S. 983 [94 S.Ct. 2384, 40 L.Ed.2d 130] (1974). That type of significant aid was plainly given here by the [’598 contract]. While [Hazeltine] seems to have spent less than [$20,000] on its development of the device before that contract, the Government expended [almost $400,000] (under the contract) in building, testing, and modifying the [antenna]. Most important, that post-contract development led to essential changes in the device and first proved it workable. As the [Patent Rights] clause contemplated, the United States has and should have the right to procure and use the [open array] invention without payment. [McDonnell Douglas, supra, 229 Ct.Cl. at 333, 670 F.2d at 163.]
On the basis of the foregoing, the court concludes that plaintiff is not entitled to recover and that its complaint is therefore to be dismissed. No costs are to be assessed.
*463Appendix I — “Hog Trough” Antenna Atop Primary Radar (Rear View)
[[Image here]]
*464Appendix II — Test Model Structure
[[Image here]]
*465Appendix III — Close-up (rear view) of Opening in Test Structure
[[Image here]]
*466Appendix IV — Open Array Antenna As Proposed by Hazeltine
[[Image here]]
*467Appendix V — Test Structure Used During ’598 Contract
[[Image here]]
. Because the Air Traffic Control Radar Beacon System (ATCRBS) does not rely on a reflected radio signal, it appears that, strictly speaking, it is not a true radar system.
. Backward or “backlobe” radiation refers to radio signals emitted in a direction opposite to the pointing or main lobe hemisphere direction of an antenna, i.e., the hemisphere behind the antenna. The backlobe contains the back radiation or "leakage” radiation. Backward radiation is undesirable in a radar antenna as it can result, inter alia, in a false target from an aircraft behind the antenna. The "front-to-back ratio,” in this context, is a measure of the level of the radiation leakage in the rearward or backward direction relative to the level of the intended main lobe, forward if you will, radiation. The terms "backlobe” or "radiation leakage" “main lobe” or “front-to-back ratio” are often expressed in decibels (dB). See infra note 5 (explanation of decibel). In this case, the parties were greatly concerned about "suppression of backlobe radiated wave energy.”
. There was testimony from two Hazeltine attendees and substantially contemporaneous documentation by Mazzola to the effect that Frank LaRussa (LaRussa), an engineer with TSC, was present at the meeting. LaRussa denied having been present. The court does not resolve this issue as it finds, in the absence of a similar denial by George Haroules, another TSC engineer who also attended the meeting, that a meeting between TSC and Hazeltine personnel undoubtedly took place in this time period, with or without LaRussa.
. These and other measurements by Richards were premised on the assumption that the actual width measurement of a 2 inch X 4 inch piece of dimensional lumber in the photographs of the model was l‘A inches. If the width of said 2x4 piece of lumber is taken as l9As inches or l5/s inches, the figures should be mul-tipled by 1.04 and 1.08, respectively. Even if these latter figures are used, they do not significantly change the size of either the structure or the opening.
. A decibel (dB) is a logarithmic power ratio. With each increase of 3dB, power is doubled; conversely, power is halved with each decrease of 3dB. As relevant hereto, dBs are used to describe the ratio of the power level at the peak of a radiation pattern to the power level at some other point. For example, if a point on the pattern is at minus 21dB, then the power level at that point is '/us of the maximum power level of the beam. For present purposes, the important point to remember is this fact—the lower the dB number, the greater the ability of the antenna to suppress radiation. In this regard, —35dB is lower than — 21dB.
. The crucial issue in this case is whether the tests described above constituted a reduction to practice of Wheeler’s conception or invention. In order that a somewhat chronological approach may be followed in presenting the background of this case, further discussion of this issue, in the context of the other facts adduced at trial, appears below rather than here.
. Indeed, as noted above, supra, p. 427, the initial report of the SEB characterized the Volume II proposed solution as having a "[f]airly high technical risk" and listed the risk as a "(w]eakness[ ]” of the approach. Similarly, Francis J. LaRussa, Technical Monitor for the '598 contract, described the proposal in an early memo under the '598 contract as "a moderately high risk technical approach.”
Although LaRussa also described the concept as "novel," the other evidence before the court makes it clear that TSC personnel were not convinced, by the Hazeltine Volume II proposal, that the tuned array antenna model would work as described. LaRussa, with his 20 years of experience in antennas, testified that when he read the Volume II proposal he did not know whether or not the open array described therein would work. Indeed, the Phase I and Phase II requirements of the '598 contract were in essence an effort to see if “the novel concept” would work as intended.
. When the radiating currents on a dipole or monopole, both elements, are the result of connecting an antenna to a source of radio frequency energy, e.g. a transmitter, the element is said to be "driven.” Radiating currents in a dipole or monopole may also be created simply by locating the element near a driven element. When this occurs, the radiation from the driven element induces currents in the nearby element, which is said to be "parasitic." Parasitic elements intercept and radiate energy impinging on them. This process is termed "mutual coupling.” In an array of driven elements and parasites, there is a mutual coupling effect that alters the magnitude and phase of reradiated energy from the value it has when only a single driven element and a single parasitic element are present. Indeed, when more than just a few driven elements and parasites are arrayed, the mutual coupling is so complex that analytical design techniques are not viewed as practical.
. Stated simply, the key feature, according to plaintiff, of the antenna disclosed in the ’977 patent is the "selective tuning" of the reflector rods to make them electrically “larger” than their physical size would indicate. This is possible because of the electrical principle of resonance. The tuned rods are sized so that they resonate at the desired frequency. If the electrical configuration of the rods is changed, e.g., by encasing them in ice, the resonant frequency will also change. Thus, the structure becomes: "detuned” and does not work as well at the design frequency.
. The size of the prototype as actually built is not clear. Hazeltine’s monthly progress reports contain the dimensions set forth above; a TSC Program Progress Report for March 1974 describes that it as 4 feet by 27 feet, and LaRussa testified it was 4 feet by 28 feet. The size of the prototype was not in dispute, and it does not appear material to the issue under consideration.
. Close examination of the figures shows a discrepancy of approximately $6,400 ($398,000 - $391,587 (92,000 + 279,500 + 20,087)) between Tedford’s figure ($398,000) and the total otherwise indicated, viz., from the '598 contract terms and Hazeltine’s cost reports ($391,587). The parties have not suggested that the discrepancy is important, and the court does not view it as particularly relevant in any event.
. For some unexplained reason, the parties did not reach agreement regarding the proper title of the agreement, an issue that has no bearing . on the outcome of the case.
. An agreement of the type asserted herein can reasonably be viewed as a deviation from the Patent Rights clause since it amounts to a de facto exemption of the tuned array antenna from the clause’s reach and coverage. Under such circumstances, not even the contracting officer would have had authority to enter into such an agreement without approval of higher authority. See 41 C.F.R. § 12-1.009-2 (1972).
. Other courts have applied a more stringent test for proof of ratification. See, e.g., Rivercity v. American Can Co., 600 F.Supp. 908, 916 (D.La.1984) (facts must indicate clear and absolute intention to ratify), aff’d. per curiam on other grounds, 753 F.2d 1300 (5th Cir.1985); Thomson McKinnon Securities v. Moore's Farm Supply, 557 F.Supp. 1004, 1012 (W.D. Tenn. 1983) (substantial evidence required to show ratification).
. The court does not view Hazeltine’s March 8, 1973 letter to Nelson as sufficient written evidence of an agreement, as both parties agreed the letter was merely a statement of Hazeltine’s position and there was no direct evidence of assent thereto on Nelson’s part, much less on Kelleher’s part. Again, there was, of course, no reference to the letter or the alleged agreement in the '598 contract.
. The court concludes that the "internal” statements of government officials, e.g., Farmer’s February 21, 1973 memo, are entitled to little weight under the circumstances of this case. The factual predicate of the statements is very *443weak. Defendant’s knowledge of the tests and the results thereof was based solely on what plaintiff told it.
. It was clearly unreasonable for Hazeltine to rely on a statement by Nelson regarding reduction to practice in the face of Farmer’s ”refus[al] to formally acknowledge whether or not a reduction to practice had actually taken place," especially when the memo relied upon for Nelson’s acknowledgment referenced the phone call that gave rise to Farmer’s statement.
. It should not be overlooked that Hazeltine received over $1.7 million from the government for its total contract efforts in developing the tuned array antennas (4 foot and 5 foot models). Cf. Heckler v. Community Health Services, 467 U.S. 51, 63, 104 S.Ct. 2218, 2225, 81 L.Ed.2d 42 (1984) (plaintiff'must show it is "significantly worse off’ than it would have been if the challenged action had never occurred in order to support an estoppel claim).
. The record strongly suggests that TSC would not have awarded the '598 contract on the terms, i.e., no Patent Rights clause, sought by plaintiff. Rather, TSC might have purchased, on a fixed-price basis, a completely developed open array antenna for evaluation rather than enter into a cost-plus-fixed-fee contract where developmental costs were borne by the government. Plaintiffs reformation argument must therefore fail as it is central to such a remedy that it be shown that "the Government would have agreed to the contract if worded in accordance with the plaintiffs intention.” Fraass Surgical Mfg. Co. v. United States, 215 Ct.Cl. 820, 825, 571 F.2d 34, 37 (1978) (citations omitted).
. Hazeltine argues that it met the requirements of paragraph (e), viz, it proved a prior reduction to practice to the contracting officer, before it entered into the '598 contract. In essence, this is simply a restatement of Hazeltine’s prior argument that there was an agreement regarding a pre-award reduction to practice. That argument was carefully considered and rejected by the court earlier in this opinion. Even assuming arguendo that Kelleher, the contracting officer, had authority to make such a determination before the contract was awarded, the evidence shows conclusively that he was not aware of the critical factors pertinent thereto sufficient to support plaintiffs position.
. Plaintiff sometimes refers to the "first” experimental model as a "prototype open array reflector." This is somewhat misleading as only a small portion of the model was open. The model cannot reasonably be deemed "open” as that term is defined in the patent (see column 2, Is. 37-42 of '977 patent) because the open area of the model is not greater than the closed area of the model. But see infra pp. 450-51. In any event, the "experimental model”, was also referred to as the "experimental setup,” the "test setup," the "open array antenna model,” the “precontract test model” and the “breadboard model" (plaintiffs term, which it defined, at oral argument, as an engineering term describing a model that is capable of being changed easily and which is typically used in the early stages of an experiment). The use of. these different terms to describe the same test fixture produces some confusion in the record and in the briefs. This is especially so when the first open array antenna (four foot) designed and fabricated by plaintiff under Phase II of contract '598 was also referred to as a “prototype tuned open array" antenna. For purposes of clarity, the court deems use of the term "precon-tract" proper to identify the model upon which the tests were performed prior to the '598 contract.
. Plaintiffs reliance on Hummer v. Adm’r of NASA, 500 F.2d 1383, 183 USPQ 45 (C.C.P.A, 1974), is misplaced. In Hummer, the court found, for purposes of analysis, that one claim was sufficiently representative of all the claims. Since this claim had not been reduced to prac*448tice, neither could the other claims have been reduced to practice. Moreover, the court in Hummer cited Fredkin v. Irasek, 55 CCPA 1302, 397 F.2d 342, 158 USPQ 280, cert. denied, 393 U.S. 980, 89 S.Ct. 450, 21 L.Ed.2d 441 (1968), for the proposition, without qualification, that every element of a claim must be found in the tested device.
. Plaintiffs expert described a person of ordinary skill in the art in 1972 as having a bachelor’s degree in electrical engineering, a few graduate courses, perhaps a master’s degree in the same specialty, and from 5-10 years of experience as a design engineer in the particular area of specialty. Defendant's experts generally agreed that such a person would have had a master’s degree in electrical engineering and about 5 years of experience. The court cannot discern any meaningful distinction between the two standards.
. It is to be noted TSC had questioned the possible effect of mutual coupling on the electrical performance of the tuned array relative to Phase I work under the '598 contract. See supra p. 432.
. Notwithstanding that claims 2-13 were not reduced to practice before the '598 contract, the government could still be liable for an infringement of claim 1, if that claim were reduced to practice prior to the '598 contract, but plaintiffs recovery would be adjusted for the developments made under the contract at government expense. See Farrand Optical Co. v. United States, 325 F.2d 328, 334, 135 USPQ 249 (2d Cir.1963).
. It is to be noted that one of defendant’s own experts stated that the elevation limitation for ATCRBS was 42°, viz., the antenna’s performance apparently wasn’t specified for elevation angles greater than 42°. Accordingly, defendant’s concern with the precontract model’s characteristics at steep angles is somewhat unjustified.
. As the radius of the sphere becomes larger relative to the size of the antenna, the behavior of the wave approaches that of a plane wave. LaRussa testified, see supra p. 435, that the minimum distance, L, to assure a plane wave is given by the formula L = 2D2/[wavelength], where D is the maximum dimension of the antenna in one direction. LaRussa estimated, using this formula, that 200 feet would be needed to get plane wave illumination of the precon-tract test model. Mazzola testified, however, that the distance between the test structure and the other antenna used in the tests was approximately 15 — 18 feet, and definitely less than 50 feet. See supra p. 424. Thus, it is clear that the tests performed on the precontract test model were not taken under plane wave conditions.
. This discussion is equally applicable to claims 2-13, which also contain an openness requirement.
. The "opening” of the first model represents only 12 percent of the total model area ((31.6 X 21.2)/(130 X 43.7)), without taking into account the columns.
. Plaintiff attempts to characterize defendant’s criticisms as an analysis of whether or not the precontract model would satisfy the ATCRBS specifications. Although defendant, perhaps, in its zeal to prevail, did occasionally "paint” its criticisms with a broad brush, the court does not believe they can be disposed of so easily. In this opinion, the court considers only specific assertions and only insofar as they relate to the test for reduction to practice used by the court. Moreover, to the extent the ATCRBS specifications are construed as defining "the environment of [the invention’s] practical contemplated use,” see Technical Development Corp. v. United States, 202 Ct.Cl. 237, 308, 179 USPQ 180 (1973), cert. denied, 416 U.S. 983, 94 S.Ct. 2384, 40 L.Ed.2d 759 (1974) (citation omitted), it is arguable that some broad based criticisms are valid.
. The wavelength, WL, at a given frequency, F, is given by the formula WL = C/F, where C is the speed of light. The speed of light is 186,300 miles per second (11,803,968,000 inches per second). There was no suggestion during trial that either party ever used an unacceptable wavelength value.
. Regardless of how his testimony is construed, it is clear that Richardson viewed the opening as too small.
. Sureau originally defined the minimum height at 5 wavelengths, but reduced it to 2Vz wavelengths because a monopole was used. Presumably, this is acceptable because of the “imaging" principle discussed earlier. If the use of a monopole generally allows the height to be halved, then the latter height would only need to be 50 inches.
. The word "possibly” is used in the text as the tests were not made under the same conditions, and therefore the results may not be directly comparable. This item is discussed below.
. Substantively, of course, the order in which the tests are performed has no effect on the analysis inasmuch as the tests are independent of one another; it is results, not processes, that are to be examined.
. In Richardson’s opinion, the configuration of Hazeltine's antenna range was such that its limit of credibility was —25dB, viz,, observed suppression levels below — 25dB could be due to interference suppression from range reflection rather than any feature of the tested antenna. A finding to this effect would essentially preclude use of the test results as proof of a reduction to practice. The basis for Richardson’s opinion was the "ripples” of the first pattern and 10 CI.Ct.—11 the irregular nature of the backlobes in both patterns. Unfortunately, Richardson did not explain why the "ripples,” if due to range reflections, were not present on the calibration pattern. Additionally, TSC apparently found the range acceptable, for the most part, during the ’598 contract. Accordingly, the court prefers not to give controlling weight to his testimony in this regard.
. Likewise, the fact that part of Sureau's opinion was arguably phrased in terms of commercial marketability, which is not a requirement for a reduction to practice, Barmag Banner Maschinenfabrik A.G. v. Murata Machinery, Ltd., 731 F.2d 831, 838-39, 221 USPQ 561 (Fed. Cir.1984), does not appreciably lessen the weight given thereto.
. This situation should be compared to that of Farrand Optical Co. v. United States, 325 F.2d 328, 331, 35 USPQ 249 (2d Cir.1963). In Far-rand, there was a working prototype with "nothing conjectural” about its being able to perform its intended function. The contract was not awarded to "demonstrate feasibility,” as was the case here, but rather merely to adapt the developed and proven technology to defendant’s specific setting; in Farrand, the government never doubted the efficacy of the device as a solution to the particular scientific problem at hand.
. This discussion assumes that the other shortcomings in the test procedures, e.g., too small opening, inadequate calibration, etc., were eliminated.