*352OPINION
BRUGGINK, Judge.This is an action brought under the Contract Disputes Act, 41 U.S.C. §§ 601-613 (1982) (“CDA”). Transtechnology Corporation, Space Ordnance Systems Division (“plaintiff” or “SOS”) alleges that the technical data package for M206 infrared countermeasure flares which it used in its contract with the United States Army Armament Material Readiness Command was defective. It also alleges constructive changes to the contract. In its complaint SOS seeks a total of $7,710,519 in damages, plus interest in accordance with the CDA. Its damages include alleged delay and inefficiency costs. After trial, the court concludes that plaintiff is entitled to recover part of its claimed costs.
GENERAL FACTUAL BACKGROUND
The following facts are generally applicable to all of SOS’s claims. Because the facts and discussion as to each of the seven counts are best treated separately, additional facts and discussion appear in the context of the particular claims.
On March 1, 1982, plaintiff SOS, as low bidder, was awarded Contract No. DAAA09-82-C-7040 in the amount of $6,282,500 by the United States Army Armament Material Readiness Command (“AMCCOM”). The contract called for the production of 875,000 M206 infrared countermeasure aircraft flares at a unit price of $7.18. Thereafter, pursuant to Modification P00005, defendant was given an option to purchase an additional 347,250 flares at a unit price of $7.82, increasing the total quantity to 1,222,250 flares and the total contract price to $8,997,995. As of April 6, 1989, SOS has produced and shipped 1,162,-520 acceptable M206 flares.
SOS was not the first M206 contractor to use this technical data package (“TDP”).1 The first was Morton Thiokol Inc., Longhorn Division (“Thiokol”). Thiokol is the operator of the Government-owned Longhorn Army Ammunition Plant (“Longhorn”) in Marshall, Texas. Its contract, unlike that of SOS, was a cost reimbursement contract. Thiokol’s contract, further, was with the U.S. Army Research and Development Command (“ARRADCOM”) rather than AMCCOM. ARRADCOM, which is part of the Picatinny Arsenal, is the research and development group that originally developed the M206 TDP. According to plaintiff, AMCCOM generally takes over contracting when the TDP has been proven acceptable and is ready for mass production.
The first wholly private M206 contractor was Hycor, Inc. It contracted with ARRADCOM in 1979. SOS was the next private M206 contractor. SOS was the first to contract with AMCCOM, rather than ARRADCOM.
Hycor bid on the contract won by SOS. While SOS bid a price of approximately $7.00 per flare, which was consistent with other bidders, Hycor, the only private contractor with M206 experience, bid over $15.00 per flare. Three other M206 contracts have subsequently been let, two with Tracor, Inc., and one with Bermite Division of Whittaker Corporation.
The M206 consists of three basic parts, the pellet, the canister from which the pellet is ejected, and the M796 impulse cartridge which ejects the pellet from the canister. The canister is rectangular, about seven inches long, one inch wide, and made of aluminum. At one end of the canister is inserted the impulse cartridge or “squib.” On top of the squib is a plastic piston, upon which the flare pellet rests. The flare canister is capped with an end cap that is held in place with two shear pins.
The flare pellet is composed of magnesium, teflon and hycar. ' Magnesium is the fuel which causes the pellet to burn. Teflon is the oxidizer, and hycar is a rubber *353material which acts as a binder. The percentage of each ingredient is classified.
SOS mixes the ingredients for the flare composition in a Cowles Dissolver. Joe Cabaret, President of SOS during the M206 contract, describes the Cowles Dissolver as “a big pot” three feet in diameter and four or five feet high. Before ingredients are mixed, the hycar must be dissolved using a solvent, generally acetone or methylethylchloride (“MEK”). SOS uses MEK. The hycar, in solution, is put into the dissolver first, followed by very finely ground teflon. Before the magnesium is introduced, additional solvent is added to the mixture. While various ingredients are added, the Cowles dissolver, using what Cabaret described as a motor-driven blending wheel, continually mixes the various ingredients. According to Cabaret, the mixing process is very hazardous because there is always a danger that volatiles will be ignited. The Cowles Dissolver is designed to minimize such risks, and is operated remotely.
Once the flare constituents are mixed together in solution, the mixture must be turned into a dry powder. The solution is solidifed again through removal of the hycar. According to Cabaret, hycar can be removed through a process called “shocking,” which causes the solid compounds to solidify and sink to the bottom. Cabaret stated that shocking can be done with water or hexane. SOS uses water.
After the solids in solution settle, the remaining liquid is drawn off. The solid composition is then put into a centrifuge to force out remaining liquid. What remains is a thick mass of flare composition which resembles dough. The composition is put into pans, placed in ovens, and dried to rid the composition of any excess liquid. After drying, the composition is a fine powder.
Flare composition is then pressed into pellets, using a die exerting between 10 and 11 thousand pounds of pressure per square inch.2 Using a milling machine, two longitudinal grooves are then cut into the pellet. These grooves are four inches long and one-eighth inch in depth and width. Into these grooves are placed first fire and intermediate mix, highly combustible materials which aid in igniting the flare. The intermediate mix and first fire are put into the grooves using a SEMCO gun, which, in Cabaret’s words, “extrude[s] [the material] just like [ ] cake icing.” After application of the first fire and intermediate mix, the pellet is dried. Cabaret testified that, according to the specification, SOS had the option of drying the pellets in ovens for a relatively short period of time or at ambient temperature for eight hours. He did not state which option SOS chose.
After drying, the pellets are wrapped in aluminum tape. According to Cabaret, the purpose of the aluminum tape is to confine ignition of the first fire and intermediate mix in the pellet groove. Finally, under the original TDP, three felts, called obturator felts, are placed around the pellet one at either end and one in the middle. According to Cabaret, these felts, which are .031 inches thick, are supposed to act as a seal. The completed pellet is then inserted into the canister.
The M206 infrared flare, is deployed from several different attack-fighter aircraft models as a decoy to enemy heat-seeking missiles launched to track and destroy the aircraft. Typically, the M206 flares are deployed from M130 Payload Modules which hold 30 flares each. The number of flares and the payload module, however, may differ depending upon the nature of the mission and the type of aircraft.
As a quality control measure, sample flares from all flare lots produced by SOS were tested to assure conformance with various performance parameters. In addition, before SOS could go into production, the contract required SOS to produce a first article. Generally, the quality control standards for the first article sample flare were more stringent than for flares tested for quality control during production. Relevant testing requirements will be dis*354cussed in connection with plaintiffs specific allegations.
On June 13, 1985, SOS submitted six claims to the contracting officer (“CO”) seeking equitable adjustments to the contract. The claims were certified on November 14, 1985 in accordance with 41 U.S.C. § 605(c)(1) and the contract’s disputes clause. The CO denied the claims in their entirety on September 19, 1986. SOS filed its complaint here on November 3, 1986.
In Count I of the complaint, SOS alleges that defendant improperly disapproved the use of paper in lieu of film for the required quality control measure of radiographically x-raying the M206 flares before they are shipped, and that this disapproval constituted a constructive change in contract requirements. In Count II, SOS alleges that the Government changed its interpretation of radiographic examination specifications to include first fire flakes as unacceptable “foreign material” in the flare and that this also constituted a constructive change. Counts III — V allege defective specifications. Count III alleges that the TDP was defective in that flares manufactured in strict compliance with the package did not pass certain environmental testing requirements specified in MIL-F-63107A(AR). SOS contends that the process of oven-drying flares which it employed in an attempt to prevent environmental test failures was therefore beyond contract requirements. Count IV alleges that flares made in conformance with the TDP could not pass static tests. In Count V, SOS claims that defects in the TDP caused additional costs in meeting dynamic testing requirements. Count VI is a claim for delay and inefficiency damages which SOS alleges were the result of the defective TDP. Finally, Count VII seeks claim preparation costs and consulting fees that resulted from alleged constructive changes and defective specifications.
A. Count I — Failure of the Army to Approve SOS’s Non-film Radiographic Technique.
1. Factual background and contract provisions.
Paragraphs E.3 and E.4 of the contract incorporate by reference Defense Acquisition Regulation (“DAR”) §§ 7-103.5(a) and (d), and 7-103.24 (1984). Generally, these regulations assign to SOS the responsibility for inspecting the M206 flares. The contract also incorporates military specification MIL-F-63107A(AR) by reference. Two provisions of MIL-F-63107A(AR) relating to radiographic (x-ray) examination of the M206 flares are pertinent:
3.4 Radiographic examination. Radiographic equipment procedures and evaluation shall be in accordance with MIL-STD-453. Radiographic personnel shall be qualified in accordance with MIL-STD-410. Equipment and procedures will be submitted for approval to the Government (see 6.3). Flare samples shall not exhibit any indications of loose, deformed or missing components, foreign material, or defects within the flare pellet i.e., cracks, voids or discontinuities. Minimum optical density of each radio-graph in the area being inspected shall not be less than 1.25 and shall be within a density range that will permit proper identification of defects.
4.5.2 Radiographic examination. Following each environmental test (first article only), and prior to functional testing, the flare samples shall be examined in accordance with MIL-STD-453 procedure. Two exposures shall be made of each flare, each forming a 90 degree angle with the other and with the longitudinal axis of the flare. Examination of the radiograph shall be made to determine compliance with the applicable requirement (see 3.4). Any flare deemed unacceptable shall be cause for lot rejection.
MIL-STD-453, referenced in the above specifications, provides in relevant part:
4.3.1 Film. Production radiographs shall be made on safety film____ The requirements of this standard are based on radiographic film. Production radio-graphs on photographic paper or nonfilm methods shall be correlated with. film produced in accordance with this standard.
*3555.3 Nonfilm radiographic techniques. The use of nonfilm radiographic techniques is permitted, provided the method used is sufficiently sensitive to resolve the required quality level. However, pri- or approval must be obtained from the contracting agency on the detailed inspection and evaluation procedure.
Following contract award, by letter dated April 30, 1982, SOS requested permission to use paper in lieu of film for the radio-graphic examination of the M206 flares and requested that use of paper be one of the items for discussion at the post-award conference scheduled for May 10, 1982. Plaintiff wrote:
Specification MIL-F-63107A requires x-ray of acceptance samples in accordance with MIL-STD-453. SOS performs this operation using paper in lieu of film. On other contracts specific approval was required by DCAS/QA3 to use paper. Permission is hereby requested.
Defendant concedes that the paper radio-graphic procedures proposed by SOS are encompassed by the language of 1T 5.3 of MIL-STD-453, “nonfilm radiographic techniques.”
At the post-award conference, DCAS quality assurance specialists expressed concern that the paper would deteriorate in storage and that images were less clear on paper than on film. The parties agreed that SOS would submit its test methods and procedures for use of paper radiograph techniques to the Quality Assurance Office at Picatinny Arsenal by May 17, 1982, and that it would submit its formal request to use paper instead of film by June 4,1982 to the Quality Assurance Office at Rock Island Arsenal, Illinois. SOS also was to submit a written procedure for the use of film, in the event that the paper method was not approved.
On June 9, 1982, SOS submitted a radio-graphic inspection procedure to the Rock Island Arsenal. Defendant contends that this submission was not a formal request to use paper and was not in accordance with the agreement made at the post-award conference because it did not contain a sample paper radiograph, was not explained by a cover letter, and did not contain a formal request to use paper radio-graphs. Plaintiff argues that neither MIL-STD-453 nor the agreement reached by the parties at the post-award conference required submission of a sample radiograph. In any event, as plaintiff points out, by letter dated September 16, 1982, plaintiff submitted to the Picatinny Arsenal a written procedure for the use of radiographic film and included with that submission x-ray samples of both film and paper.
Marvin Elmowitz, an electronics engineer at the U.S. Army Armament, Munitions, and Chemical Command, Product Assurance Directorate (“PAD”), in Dover, New Jersey, was responsible for technically evaluating SOS’s design for radiographic test equipment. He obtained the assistance of Emett Barnes, a physicist in PAD, to evaluate SOS’s submission. Barnes is a principal reviewer of radiographic specifications and contractor proposals for radio-graphic equipment and procedures for PAD. Although Elmowitz was responsible for deciding whether SOS would be allowed to use paper in place of film radiographs, it was Barnes who reviewed and evaluated SOS’s submissions for ultimate recommendation to Elmowitz.
Testifying for defendant, Barnes stated that plaintiff’s September 16, 1982 submission included samples of both paper and film radiographs, but only included a written procedure for the use of film radio-graphs. Barnes testified that because he was given a sample film radiograph and a procedure for film, he assumed that SOS was requesting the approval of its film procedure only. He stated that he could have telephoned Elmowitz to inquire about the missing procedure for paper, but never did. For his review of SOS’s submission, therefore, Barnes only had the sample paper radiograph with no accompanying procedure.
*356In evaluating the submission overall, Barnes admitted that he concentrated primarily on the film radiograph to determine whether it conformed to MIL-STD-453. In evaluating the particular paper radio-graphs, he “didn’t do much of anything, I just didn’t like the idea that they wanted to use paper, in fact, I didn’t know that it was paper, but I wanted to preclude them from using paper if that was the case.” Regarding film, Barnes determined that there was a problem with the penetrameter4 selection for the film, and that the film did not meet the necessary optical density requirements. After his review, Barnes gave Elmowitz a handwritten recommendation suggesting that SOS improve certain aspects of its film procedure, and that SOS not be permitted to use paper. He stated at trial that he made the recommendation regarding paper because his “experience with paper was not good,” and that he did not base the recommendation on his examination of the paper radiograph.
Barnes’ poor experience with paper did not involve an actual contract or contractor. Rather, Barnes had participated in a study which had compared film and paper. The study involved Kodak paper. According to Barnes, Kodak had asked ARRADCOM to test a new x-ray paper. Barnes stated that “we did a fairly extensive study of it and we tested it alongside standard x-rays, ... and found that it was good for certain types of applications, but it was a general purpose type of product.” Barnes stated that the sensitivity and resolution was not as good as conventional x-ray film and that the paper could not achieve the same sensitivity as film. In its procedure, plaintiff stated that the paper it would use is “Film-type Industrial 600,” which Barnes stated is a type of Kodak paper. Barnes stated that he did most of the work related to this study himself. When asked if there was any paper that he would have approved for flares at that time, Barnes stated, “No, not to my knowledge.”
In 1989 Barnes reviewed SOS’s paper procedure. He stated at trial that based on the 1989 review, if he had reviewed the prócedure in 1982, he still would have rejected SOS’s request to use paper. He stated that he conducted a penetrameter analysis and found that the penetrameter requirement of the specification was met. Barnes did not elaborate on why he found SOS’s paper procedure unacceptable in 1989,5 although he did note that the procedure failed to address paper storage and handling or image fading. Barnes instead testified generally as to why paper was inferior to film. He stated that film gives a greater optical density range, which provides greater clarity in reading radio-graphs. He also stated that paper radio-graphs have more glare than film and are therefore difficult to read, that paper radio-graphs stick together easily, making storage difficult, and that paper radiographs fade quickly.
Harriet Beard, a radiograph inspector who has been employed by SOS for 22 years, 18 as a radiologist, has used paper to radiograph flares. She stated that she used to use paper “all the time” at SOS. She stated that she never had difficulty reading the paper radiographs. Nor did she ever remember the Defense Contract Administrative Services (“DCAS”) quality control inspectors ever having difficulty reading them. She stated further that she never had difficulty with glare or with the radiographs sticking together. She acknowledged that handling the paper is very *357critical, but stated that SOS has specific procedures to be followed, and she and her co-workers were trained to follow them. Beard does not recall DCAS inspectors or anyone else ever requesting to review radiographs which had been placed into storage.
The Army disapproved SOS’s radiographic procedure for reasons expressed in DRDAR-QA Form 337 (“Evaluation of Acceptance Inspection Equipment Designs”), dated October 18, 1982: (1) “Radiograph paper or a print made from a radiograph is not an acceptable medium for radiographic inspection of [the M206 flare]”; and (2) “Radiograph film furnished does not meet the 1.25 minimum optical density requirement____ Measured densities are 1.10 and lower.” Elmowitz prepared this form based on Barnes’ recommendations. After SOS’s subsequent submissions, the Army approved its radiographic procedures for film.
2. Discussion.
As noted above, MIL-STD-453 115.3 states that if a contractor elects to use nonfilm radiographic techniques to evaluate flares, the following two conditions must be fulfilled: (1) the method used must be sufficiently sensitive to achieve the required quality level, and (2) prior approval must be obtained from the contracting agency on the detailed inspection and evaluation procedure. The Government had a right to insist on compliance with MIL-STD-453. See Maxwell Dynamometer Co. v. United States, 181 Ct.Cl. 607, 628, 386 F.2d 855, 868 (1967); H.L.C. & Assocs. Constr. Co. v. United States, 176 Gt.Cl. 285, 306, 367 F.2d 586, 598 (1966). It must exercise its discretion reasonably, however. See W.G. Cornell Co. of Wash., D.C. v. United States, 179 Ct.Cl. 651, 672, 376 F.2d 299, 313 (1967) (citing Fox Valley Eng’g, Inc. v. United States, 151 Ct.Cl. 228, 236 (1260)). The Government’s defense to Count I rests on its contention that paper is an unacceptable medium for radiographing flares on this contract and that it therefore properly denied plaintiff’s request.
It is plain that Barnes did not give SOS’s request to use paper any real consideration. Barnes, further, did not attempt to conceal his predisposition against paper radiographs in general, stating that there was no paper procedure that he would have approved. For all intents and purposes, the portion of MIL-STD-453 which provided for nonfilm radiograph techniques did not apply to paper on this contract. The Government therefore can hardly be found to have exercised its discretion fairly and reasonably. There was a de facto bar on the use of paper radiographs, and this bar constructively changed the contract.
Before plaintiff can recover, however, it must establish that, had the Government considered SOS’s paper procedure, it would have complied with MIL-STD-453 115.3. In this regard, Harriet Beard’s testimony that SOS has successfully used paper radiographs for flare contracts in the past and that she has never had any difficulty, either reading them or handling them, persuades the court that plaintiff has made an initial showing that SOS’s paper procedure complied with MIL-STD-453.
While Barnes testified that paper is less desirable than film, there was no evidence that paper in general, or plaintiff’s paper in particular, could not be effectively used. It may not have been as good as film, but the court accepts Beard’s testimony that it was good enough. Barnes’ general aversion to paper led him to conclude that there was no procedure he would have accepted.6 That attitude is inconsistent with defendant’s TDP, which plainly implies that paper can be successfully used for radiography. The court cannot accept Barnes’ generalized dislike as a rebuttal to plaintiff’s prima facie showing that its paper procedure complied with MIL-STD-453. It is disingenuous for the Government to allow paper, and then, in essence, *358refuse to consider any paper radiographs submitted for approval. Plaintiff has therefore made a sufficient showing that its procedure complied with MIL-STD-453, and that it is entitled to the difference between the cost of using paper and the cost of using film.
3. Damages.
Plaintiff prepared its bid on the M206 contract based upon a paper radiograph procedure. According to plaintiff, shooting, processing and developing film radio-graphs require more labor than processing and developing paper radiographs. It alleges that using film radiography required approximately five more hours of labor per M206 inspection lot than its paper procedure. In addition, plaintiff alleges it suffered increased costs using film because flares, which are manufactured at its Mint Canyon facility, had to be transported to its other facility in Placerita to be radio-graphed. The Mint Canyon facility, according to plaintiff, was only equipped for Industrial 600 paper radiographs. Plaintiff alleges that the total labor cost differential for using film as opposed to paper was $23,221 and that the total cost for transporting the M206 flares to Placerita for radiographing was $104,212.
Beard testified regarding the time it takes to radiograph flares using paper as well as with film. She stated that for a variety of reasons it takes longer to work with film than with paper. As to exposure time, Beard testified that film requires one minute, using a Paxitron, whereas paper requires eight to 10 seconds. As to developing time, she testified that with film it takes approximately 12 minutes from the time the unexposed negative enters the machine until it can be read, whereas paper takes 10 seconds.
Beard stated that to radiograph 90 flares, which was typical on the M206 contract,7 would require 18 exposures. This is because 10 flares can be radiographed at one time, and that each flare is radio-graphed twice, once in the “zero degree” position, and once in the “90 degree” position. With paper, using two people, these 18 exposures would take 15 to 20 minutes, from beginning to end — shooting, processing, reading, and paperwork. With film, again using two people, Beard testified three hours “is the least you could get by with.” Beard testified that the people working on radiography must be at least Level II technicians.
Emett Barnes testified for defendant regarding the time it takes to expose and develop flares with paper as opposed to film. He agreed that it would require one minute to expose film. As to developing, he stated that with most modern processors it requires approximately 4.5 to 12 minutes to process, or develop, the exposed film. He noted, however, that once the first film is put into the machine, the next film can be put in the machine within the next 30 seconds to 60 seconds, depending on the particular machine used. Assuming a delay of one minute for feeding exposed film into the machine and a development time of 12 minutes per film, he opined that nine exposures could be developed in 20 minutes. Barnes further testified that paper would require additional time because of the special care required to handle it.
Beard also testified regarding the transportation of the flares from Mint Canyon to SOS’s Placerita facility. Based on her own driving experience, she stated that the trip required 35 minutes driving time each way. She stated that the flares were transported by SOS drivers, who came to the Mint Canyon facility “sevén out of ten” days, and that loading of the flares required five minutes. Safety regulations dictate that two people had to be in the trucks transporting the flares. Dean Johnson is an employee of the Defense Contract Management Command (“DCMC”) who had been assigned to SOS while SOS was working on the M206 contract. DCMC is responsible for monitoring compliance with quality standards and with specific contractual obligations. Johnson testified that it only *359required two to three minutes to load the flares.
Masaru Aratani, a Certified Public Accountant with the firm of Price Water-house, prepared a calculation of SOS’s damages with respect to its various claims. He testified for plaintiff, and his work papers are part of the record. One of his primary tasks was to determine the hourly labor rates that should be applied to the work tasks for which SOS is claiming damages. He determined the following labor rates for tasks related to this count: Labor rate for x-ray technicians, $9.68/hr.; labor rate for truck drivers, $6.45. Defendant does not dispute these rates. In addition, Aratani determined an hourly supervision and support (“S & S”) rate, general and administrative expenses (“G & A”)8, and overhead associated with these rates. Defendant does not dispute these as they are applied to Count I damages. Rather, defendant attacks the base number of excess hours alleged by plaintiff in its damages claim under Count I.
In its Post-Trial Brief, defendant sets forth, based on its review of the testimony, the number of excess hours plaintiff has established. Based on the testimony of Beard and "Barnes, it allows 1.3 hours per lot of additional x-ray time incurred by SOS because it used film instead of paper. Damages, as calculated by Aratani and as reflected in plaintiff’s Proposed Findings of Fact, are calculated based on five additional hours per lot. Plaintiff has not established five additional hours, however. Beard testified that the entire radiographic inspection procedure for film for a lot of 90 flares (18 exposures) would require three hours, and with paper would only require 15 to 20 minutes. This translates to a difference of two hours and 40 minutes. The court has difficulties even with this figure, however.
Beard's testimony that three hours is required with film is based on her experience. However, it is not clear that Beard’s estimate of three hours is consistent with her other testimony. Beard also testified that film requires 12 minutes per exposure, and that 18 exposures are required per lot. This translates to 216 minutes, or three hours and 36 minutes, per lot, for development time only. With the additional one minute per film for exposure time, a total of three hours and 54 minutes, exclusive of reading and paper work, would be required with film. Barnes testified, however, that if exposures were fed into the machine at one minute intervals, nine exposures could be developed in 20 minutes, and therefore 18 could be done in 40 minutes, instead of 216 minutes, a saving of 176 minutes. This would reduce the three hour and 54 minute figure by 176 minutes, and the total would therefore be one hour and 58 minutes. It is not clear whether SOS used a machine that could be fed film continuously, nor is it clear whether, if this were possible, Beard based her estimate of three hours in part on this procedure. In sum, plaintiff’s testimony does not coherently establish a stable figure on which to base damages.
Because defendant concedes that 1.3 additional hours were required to use film, and because the parties are in agreement as to this amount, the court will award damages based on that figure. At the conceded hourly rate of $9.68, with an estimated 136 lots radiographed,9 plaintiff is *360entitled to $1,711.42 in damages for additional labor involved in using film. With respect to transportation costs, defendant allows 35 minutes per trip (70 minutes per roundtrip) and the five minutes loading and unloading time to which Beard testified. For 136 lots, given two drivers at an hourly rate of $6.45, as developed by Aratani, plaintiff is entitled to $2,193 for transportation costs resulting from use of film instead of paper.
S & S for x-ray labor amounts to $14.08 and for transportation $322.26. G & A and overhead expenses amount to $4,093.15 for x-ray dollars and $5,964.44 for transportation. Plaintiffs request for 10 percent profit is reasonable. Plaintiff’s total damages for Count I are $15,727.70.
B. Count II — Alleged Constructive Change Related to Radiographic Inspection Requirements.
1. Factual background and contract provisions.
As a quality control measure, SOS was required under Paragraph 3.4 of MIL-F-63107A(AR) to radiograph, in accordance with MIL-STD-453, sample flares selected by DCAS personnel. SOS followed the test procedure of paragraph 4.5.2 of MIL-F-63107A(AR):
4.5.2 Radiographic Examination____ Two exposures shall be made of each flare, each forming a 90 degree angle with the other and with the longitudinal axis of the flare. Examination of the radiograph shall be made to determine compliance with the applicable requirement (see 3.4). Any flare deemed unacceptable shall be cause for lot rejection.
Paragraph 3.4 of MIL-F-63107A(AR) states that “[f]lare samples shall not exhibit any indications of loose, deformed or missing components, foreign material, or defects within the flare pellet i.e., cracks, voids or discontinuities.” MIL-F-63107A(AR) does not define “foreign material.” Paragraphs 4.4.2.1 through 4.4.2.9 of MIL-F-63107A(AR) contain charts la-belled “Classification of Defects and Tests.” These charts list 50 separate defects under the category, “Major,” and one defect in the category, “Critical.” First fire flakes or specks are not included in either defect category.10
In Count II, plaintiff contends that the Army changed its radiographic inspection requirements beginning in December 1983 to classify first fire flakes as “foreign material.” According to SOS, first fire flakes are not a defect under the contract because they are not enumerated as such in the list óf 51 defects (major and minor) found in Paragraphs 4.4.2.1 through 4.4.2.9 of MIL-F-63107A(AR). Furthermore, SOS maintains that under traditional principles of contract interpretation, the Army’s acceptance of flares that contained first fire flakes before December 1983 reflects concurrence in SOS’s interpretation that first fire flakes are not defects. Defendant counters that its position has always been that it is not possible to ascertain with any degree of certainty that a particular indication of a defect represents first fire flakes and that, even if it were possible to detect first fire flakes with any accuracy, first fire flakes, if found on the body of the flare pellet, are foreign material.
2. Discussion.
The central issue in Count II is whether first fire flakes are “foreign material.”11 This is a primarily a matter of contract interpretation. First fire flakes on the pellet body are not listed explicitly as a defect among the more than 50 defects *361listed in Paragraphs 4.4.2.1 through 4.4.2.9 of MIL-F-63107A(AR). Paragraph 4.4.2.6 does, however, note as a major defect “[f]irst fire comp, above flush with respect to grooves,” but it is plan from the testimony that there are two different phenomenon. First fire flakes are a random slight splatter on the body of the pellet, while the listed defect appears to be concerned with overfilling a groove. In this regard, it is noteworthy that Richard Ames, presently Project Leader of the decoy group at ARRADCOM, testified that the Army has never done a study to determine the effect of first fire flakes on the performance of the M206 flares.
However, if the Army before December 1983 concurred in SOS’s interpretation that first fire flakes did not constitute a defect, the court would grant substantial weight to the parties’ contemporaneous construction. See Jet Forwarding, Inc. v. United States, 194 Ct.Cl. 343, 345-46, 437 F.2d 987, 989 (1971); Embassy Moving & Storage Co. v. United States, 191 Ct.Cl. 537, 543-44, 424 F.2d 602, 606 (1970); Maxwell Dynamometer Co., 181 Ct.Cl. at 630, 386 F.2d at 870.
The evidence suggests that prior to December 1983 the Army ignored first fire flakes on the pellet body. It is undisputed that between May and December 1983, SOS did not reject any flare lots because of the presence of first fire flakes even though, as Beard testified, she found a total of 79 first fire flakes on flares sampled from those lots. Further, DCAS inspectors were present during her examination and approved her review determination. It is difficult to believe that DCAS inspectors missed every one of the 79 indications which Beard testified were first fire flakes. Further, even if DCAS inspectors were not able to determine conclusively that these indications were first fire flakes, this does not explain why DCAS inspectors, if in fact they then believed that first fire flakes were a defect, did not note the indications as potential defects or make some further attempt to identify them.12 The court is left to conclude that DCAS inspectors must have seen and ignored potential first fire flakes before December 1983.
Defendant’s suggestion that inspectors in the early stages of the contract were not performing their inspection tasks properly is not plausible. The court concludes that DCAS inspectors prior to December 1983 concurred with SOS that first fire flakes were not “foreign material” and thereafter changed their interpretation.
Given that the contract does not explicitly list first fire flakes as a defect and that DCAS inspectors before December 1983 apparently concurred with SOS’s interpretation of the contract to that effect, the court concludes that the Government constructively changed the contract in December 1983 when it began to treat first fire flakes as “foreign material.”
3. Damages.
In its Proposed Findings of Fact, plaintiff alleges that in order to assure that there were no first fire flakes on its flares, it employed four assembly personnel to brush the flare pellets after first fire had been applied. It also asserts that it employed two inspectors on its assembly line to inspect the flares after brushing to make sure that all first fire flakes were removed. Aratani, in his calculation of damages, assumes four inspectors, rather than two, and accounts for an additional category of labor — Production Technicians. He calculates damages using four full time Production Technicians. Therefore, a total of 12 full time employees were needed, according to Aratani, in the operation of brushing first fire flakes. Aratani’s calculations are based on these procedures being in place for six months.
Per Aratani, plaintiff’s cost for these procedures was $271,186. This includes S & S, G & A and overhead costs at rates which defendant does not dispute. SOS also contends that Project Manager time for investigation into the first fire problem imposed a cost of $29,722, and that Project Engineer time investigating the problem imposed costs totalling $45,026.
*362In its Post-Trial brief, defendant does not dispute Aratani’s determination that four persons were employed in each of the three categories — Production Technicians, Assembly Personnel, and Inspectors. Defendant argues, however, that there was no testimony regarding the hours spent by any of these employees in remedying problems with first fire flakes, and that plaintiff therefore has not proven any damages.
There was no documentary evidence as to the number of employees involved, nor as to the amount of time spent by them. The testimony is also sparse. With respect to brushing the flares, Cabaret testified that, during a six month period (until the first fire problem was solved by reversing the intermediate and first fire) every flare produced by SOS was inspected and brushed. Cabaret, however, did not testify as to how many employees were involved or what portion of their time was involved. Aratani could not consult payroll or other records for this information because the individual time cards did not identify what function had been performed. As noted, Aratani calculated damages assuming 12 employees working full time for six months. His figures were based in part upon discussions with Mike Murphy, Director of Operations at Mint Canyon. According to Aratani, Murphy told him that there were four Production Technicians, four Inspectors, and four Assembly Workers involved in brushing first fire flakes and that the procedure was employed for six months. The court notes that Aratani did not testify that Murphy stated each was working full time during those six months. Reynolds and Winn testified that there were “usually” three work stations and two shifts a day. Winn testified that when the operation was at full force, there were three workers and one material handler, and inspectors were “around.” When asked if they had other functions, Winn stated that they “stayed pretty close to that.” The beginning and ending periods of the work are also very imprecise. The number of shifts fluctuated between one and three.
As to the research and investigation costs (Project Manager and Project Engineer), the only direct testimony was from Dennis Talle, Manager of Contracts and Programs at SOS during the M206 contract. Talle testified that Barney Downing, a Quality Engineer, William Cornell, the Program Manager, and Earl Miner, a Project Engineer, and a Manufacturing Engineer, whose name he could not recall, spent 100 percent of their time investigating the first fire problem during the period when SOS was investigating it. Talle did not state how long that period was, but referred to it as “months.” There was evidence that for essentially two months four individuals worked full time on the issue. Aratani calculated damages based on one Project Engineer and one Project Manager for approximately five months and three months respectively. As to how he determined the number of hours these employees worked, Aratani explained: “Basically, the hours were calculated by determining the incurred hours for the respective category, deducting the amount of the labor category bid, and allocating the difference over various aspects of the claim.” He did not elaborate.
Aratani's explanation of how he arrived at hours is incomprehensible. While it is possible that Aratani’s work papers, which are part of the record, may substantiate the hours he used, his work papers are handwritten and difficult to read, and they are virtually impossible to follow without a better explanation than he provided. Aratani’s handwritten work papers cannot by themselves form the evidentiary support for this part of plaintiff’s claim.
Damages do not have to be proven with absolute certainty. It is sufficient if plaintiff furnishes a reasonable basis for computation, even if it is approximate. F.H. McGraw & Co. v. United States, 131 Ct.Cl. 501, 510, 130 F.Supp. 394, 399 (1955). Nevertheless, the evidence has to be sufficient to enable the court to make a reasonable approximation. Specialty Assembly & Packing Co. v. United States, 174 Ct.Cl. 153, 184, 355 F.2d 554, 572 (1966). Plaintiff retains the “essential burden of establishing the fundamental facts of liability, causation, and resultant injury.” Wunderlich *363Contracting Co. v. United States, 173 Ct.Cl. 180, 199, 351 F.2d 956, 968 (1965). Plaintiff urges throughout its damages presentation that if the court is unpersuaded by the overall calculation of damages, it take a jury verdict approach. The court has used this approach in limited circumstances if there is clear proof of injury, if there is no better way to calculate damages, and if there is some method for making a reasonable approximation. See Delco Elecs. Corp. v. United States, 17 Cl.Ct. 302, 323 (1989). The court declines to adopt such a methodology here. Although there was clearly injury, there was also a better way to both calculate and present damages. In view of what was presented, the best approach in the court’s view is to utilize only those elements of damage which are certain. Although this penalizes plaintiff, the penalty is attributable to a lack of bookkeeping and weak testimony.
The court will allow recovery for one shift of three Technicians, one Material Handler, and no Inspectors, for a six month period. The court also allows one Project Engineer and one Project Manager for two months. The recovery for Technicians is $79,272; for the engineer and manager it is $36,952. With profit the total is $127,847.
C. Count III — Breach of Implied Warranty of Specification Related to Oven Drying.
1. Factual background and contract provisions.
Count III relates to problems SOS experienced during the environmental testing phase of first article tests. The flares were subjected to five types of conditions— humidity, transport vibration, aircraft vibration, temperature-altitude, and shock. These tests are described in Paragraphs 3.3.1-3.3.5 of MIL-F-63107A(AR). Paragraph 3.3 states that following these environmental tests, the “flares shall comply with the visual inspection criteria of paragraph 4.4.2.6 and shall be x-rayed for defects.”
SOS submitted its first article sample flares for environmental testing in July 1982. Testing was performed by Reliant Testing Laboratory, Inc. Humidity testing occurred on July 23, 1982. Although the test results were considered satisfactory, after approximately 24 hours in storage, squibs were ejected from the pellet casing. Plaintiff was notified and testing was stopped. During aircraft vibration testing on July 27, 1982 an end cap dislodged from the flare housing. Again SOS was notified, and all further testing was stopped. Temperature-altitude testing was conducted on August 3, 1982. After fifteen hours it was noted that the squibs had ejected from the flare housing. In addition, one end cap had ejected. Once again, SOS was notified and the testing was stopped.
By letter dated August 20, 1982, the CO wrote SOS, stating that the aircraft vibration test resulted in one defect and that the temperature-altitude test resulted in two defects. The letter did not specify the defects,13 but wrote that the “tests indicated that the drying operation used on the M206 Flare Assemblies may be inadequate to prevent outgassing.” The letter also stated:
When your firm establishes a modified drying process that is considered, through preliminary tests, to be a superi- or production operation in removing excess volatiles, and in order to verify existing data, an additional 30 units are required to undergo Aircraft Vibration and Temperature-Altitude Tests____ Failure to furnish additional units may result in rejection of the First Article Environmental Test Sample.
SOS responded to the Government’s letter by letter dated September 20,1982. By this time SOS had already produced 30 additional flares using an additional drying process and those 30 flares had already passed the temperature-altitude test. They later passed the other environmental tests. The September 20 letter stated that SOS disagreed with “virtually the entire content of [defendant’s original] letter.” Specifically, SOS wrote that it believed the defects occurred during the humidity and temperature altitude portions of the environmental tests (rather than aircraft vibra*364tion). It also stated that there were no excessive volatiles, that a check of the volatiles on the basic composition indicated less than half the amount considered “generally acceptable,” and that the “[specification does not control the volatile content of the compositions.” With respect to the flares which had failed the environmental tests, SOS wrote that every requirement of the TDP had been met and that government inspectors could so attest. SOS stated that it believed the outgassing problem to be inherent in the TDP; that the oven drying step it had added was an additional step not required by the TDP and was not within the scope of the contract.
The oven drying process which SOS adopted involved drying the flares without endcaps, but fully assembled and in packaging for shipment, at 190 degrees fahrenheit for eight hours. According to the September 20 letter, SOS adopted this drying process because it believed that the out-gassing problem it was experiencing was the same outgassing problem as it had experienced on the previous MJU-7 contract.
Cabaret testified regarding the outgassing problem with the earlier MJU-7 contract. He stated that those flares were packaged in cardboard containers without their endcaps, and then heat sealed in a plastic bag that was put around the outside of the containers. Outgassing occurred during storage and was evidenced by the plastic containers “swelling up like balloons.” According to Cabaret, this took six to 18 months to develop. The Air Force ran tests on the gas causing the swelling, and it was found to be hydrogen. Water and magnesium will react to produce hydrogen, and the Air Force determined that the hydrogen gas was being produced by moisture which was trapped in the cardboard packaging material and reacting with the magnesium in the flares. The problem was solved, according to Cabaret, by drying the flares in their packaging material before they were heat sealed. This drove out the moisture in the cardboard. It is undisputed that after SOS implemented a similar eight hour, 190 degrees fahrenheit oven drying of the fully assembled and packaged M206 flares, no more flares were rejected because of out-gassing.
The Army’s reaction to SOS’s September 20 letter is expressed in an internal document dated October 5, 1982. In it, the Army concurred with SOS that the problems experienced during the environmental tests were caused by outgassing. Otherwise, the Army objected to the contents of SOS’s letter. It did not agree that the problem was inherent in the TDP. It agreed that the flares submitted to the environmental testing were acceptable to the resident government inspectors as meeting all dimensional requirements of the TDP but stated that because the environmental tests are described fully in the contract documents, a contractor “should be able to conduct, in their facility, tests of equal or greater severity regarding high temperature exposure (the temperature phase where outgassing occurs most rapidly)____ [S]OS did not adequately test their preliminary flare samples at extreme temperature to prove the adequacy of their drying process.” The Army also noted that SOS uses a “hydro-shocking” as its method of removing MEK, one of the solvents used in the flare composition mix and stated: “[b]y introducing the large quantity of water in the mix, they risk leaving a significant amount of water in the mix that can cause outgassing.....” In conclusion, the Army recommended that SOS’s request for reimbursement of costs for the additional oven-drying process be denied:
The cost that SOS wants to receive for including the additional drying cycle in the production process should not be allowed. It is a cost that any producer should absorb if their drying cycle proves to be inferior. The outgassing problem SOS contends is inherent to the TDP is a totally inadequate explanation. As advisory note 3 in drawing 9311625 forewarns, “Residual amounts of solvents prior to taping may affect performance requirements of MIL-F-63107(AR).”
(Emphasis in original.) The court notes that drawing 9311625 is entitled “Pellet *365Assembly” and describes the application into the pellet groove of intermediate mix and first fire, as well as the taping of the flare with aluminum tape. The drawing does not address the flare pellet composition, either as to ingredients or procedure.
It appears that the Army was aware, even before SOS’s experience during the first article phase of the M206 contract, of a problem known generally as outgassing, though there is some confusion as to the various types. The Army was also aware that oven drying of some form might be necessary to prevent outgassing. As discussed above, SOS had experienced out-gassing on its MJU-7 contract, and out-gassing was not unique to SOS. For example, a January 24, 1979 Department of the Army letter addressed to the Commander at ARRADCOM discusses outgassing which occurred during environmental tests of M206 flares produced by Hycor, the first M206 contractor. According to that document, during temperature-altitude tests the flares produced gas “due to inadequate curing of the pellet.” During 10 day humidity tests, “[p]ost-test visual inspection revealed a continuation of pellet out-gassing,” and after a 28 day humidity test “some out-gassing was still evident.” In a letter from Hycor to the Army dated September 22, 1980, Hycor summarizes a technical meeting called by ARRADCOM on the M206 flare production program. It describes a problem caused by MEK being driven out of the pellet at high temperature, and states that, according to ARRADCOM, the solution was to bake the flare pellets for 16 hours at 95 degrees fahrenheit to drive off the MEK.
Tracor, another M206 contractor, also experienced outgassing during environmental tests. A document dated February 16, 1983 entitled “First Article Lot Test Report” describes Tracor’s outgassing problem. During humidity tests, squibs indicated some extrusion from the pellets, and in 12 instances they were completely ejected. During the temperature-altitude portion of the test, 25 flares had completely ejected end caps, two of which also ejected the pellet itself, and three pellets had partially ejected end caps. The document also stated:
It appears that other flare manufacturers had experienced this same problem and had eliminated it with a high temperature heat cure of the flare pellet wrapped in the aluminum foil and/or after sealing in case____ This temperature cure was 180 [degrees F] to 200 [degrees F] for a minimum of six (6) hours.
William Cornell, SOS’s project engineering manager on the M206 contract, testified that the “other contractors” referred to here had to include SOS.
There was no direct evidence that Thiokol, the contractor which operated the Government-owned ammunitions plant at Longhorn and first produced M206 flares, experienced outgassing problems. However, it is noteworthy that Thiokol employed two 16 hour oven-drying processes, one of the mix composition at 160 degrees, and one of the pellet after application of first fire and intermediate mix at 200 degrees. Wayne Sommerford, the Program Manager for infrared flares at Thiokol’s Longhorn division, testified that the purpose of the 160 degree drying was to remove MEK, and the purpose of the 200 degree drying after application of first fire and intermediate mix was to remove acetone. The court notes that Drawing No. 9311625 directs drying the pellet for five minutes at ambient temperature after application of first fire and intermediate mix.14
Not only does it appear from the record that outgassing was a problem which the Army had confronted before, it also appears that the Army anticipated that out-gassing might pose a problem with SOS’s M206 contract. An Army disposition form dated February 24, 1982, regarding a pre*366award survey visit to SOS, discusses the Army’s concern that using hydroshocking to dispel MEK leaves a large amount of residual water in the mix which can react with magnesium to produce hydrogen gas. The Army appeared to be concerned with the type of outgassing which SOS had experienced on the MJU-7 contract: “This reaction, outgassing, is a concern ... since instances have been reported where the shipping containers were bulging due to the internal gas pressure.” An internal ARRADCOM memorandum from G.R. Baskerville, Chief, Artillery & Mortar Branch, dated two days earlier and addressed to the Commander at ARRADCOM, also discusses outgassing:
Based upon a Pre Award Survey conducted at Space Ordnance Systems, it is requested that your office investigate the need to require a maximum allowable moisture level in the pellet mix drawing. As discussed at the aforementioned survey, reports from depots concerning “out-gassing” of containers were experienced in the past____
If your office believes a max level is justified from engineering tests, ... DRDAR-QAT-M should be contacted in order to incorporate the findings in Sections 3, 4, and 5 of MIL-F-63107A.
The Army was still concerned with out-gassing at the time of SOS’s Post Award Quality Assurance meeting in May 1982. The minutes from that meeting state under “Action Item: # 1” that SOS agreed to a no-charge moisture analysis on its initial production of flares and state that “SOS may be asked to extend drying time thru Contract Modification.” An internal memorandum from J. Digiovannantonio, Acting Chief, Artillery & Mortar Branch, to the Commander at ARRADCOM, also dated May 1982 discusses outgassing more broadly than the MJU-7 type outgassing:
A total volatile requirement of the pellet composition of 0.10% maximum is necessary to prevent outgassing (formation of Hydrogen gas), which results in the bulging of metal shipping containers, popping end caps off from the flares, a fire and/or explosion hazard, and a possible reduction in the flare’s ability to achieve required intensities to decoy missiles.
SOS’s TDP did not have a maximum moisture content for the flare composition mix.
It is therefore clear that the Army was concerned with outgassing on the M206 contract even before SOS’s first article environmental tests. The court notes, however, that there is some inconsistency and apparent confusion on the part of both the Army and SOS regarding outgassing. It appears that there are two types. The type experienced by SOS on the MJU-7 contract is caused by moisture in packaging material reacting with magnesium in the flares. It is manifested by swelling of the packaging and takes six to 18 months to develop. It is cured by drying the packing material. There is no indication in the record that SOS experienced this type of outgassing on the M206 contract. The type of outgassing SOS experienced on the M206 contract occurs during environmental tests over a matter of hours and is manifested by ejection of end caps and squibs. As to its cause, there is no consensus. Although there were two different types of outgassing with apparently different causes, neither SOS nor the Army appeared to make such a distinction when addressing the outgassing problem SOS experienced during first article. Only a few days after it had been informed of the problems encountered during environmental changes, SOS concluded that the outgassing it was experiencing was the same as that it had experienced on the MJU-7 contract despite seemingly obvious differences in the character of the two types of outgassing. When SOS employed the same drying process it had employed on the MJU-7 contract, the Army did not object even though that drying process was meant to control moisture in the flare packaging, and the Army believed that outgassing in this case was caused by solvents in the flare pellet, in first fire, or in the intermediate mix.
2. Discussion.
Count III is one of three counts (along with Counts IV and V) in which SOS alleges a breach of an implied warranty of specifications. SOS contends that the TDP is a *367design specification and that to recover for breach of an implied warranty it need only establish that 1) it was bound by the contract to perform in accordance with the TDP, 2) it manufactured its flares in strict conformance with that TDP, and 3) the flares so manufactured did not perform successfully. See United States v. Spearin, 248 U.S. 132, 39 S.Ct. 59, 63 L.Ed. 166 (1918). Citing R.E.D.M. Corp. v. United States, 192 Ct.Cl. 891, 898, 428 F.2d 1304, 1308 (1970), plaintiff urges that it is not necessary to establish the specific defect which caused the flares to fail performance requirements. Defendant concedes that plaintiffs claim can be maintained if the TDP is a design specification, but argues that it is not. The question of whether the TDP is a design or performance specification resurfaces with respect to Counts IV and V. The common aspects of this issue will be discussed here.
It is well-settled that when the Government contracts for materials to be made in accordance with its own specifications, there is an implied warranty that following the specifications will lead to a satisfactory product. Spearin, 248 U.S. at 132, 39 S.Ct. at 59; Hol-Gar Mfg. Corp. v. United States, 175 Ct.Cl. 518, 525, 360 F.2d 634, 638 (1966). If the specifications are defective the plaintiff is entitled to recover the amount expended trying to comply with them. Bethlehem Corp. v. United States, 199 Ct.Cl. 247, 253, 462 F.2d 1400, 1403 (1972); Hol-Gar, 175 Ct.Cl. at 525, 360 F.2d at 638. A distinction is drawn between specifications, however, in terms of whether they are design in nature or whether they merely require certain performance. J.L. Simmons v. United States, 188 Ct.Cl. 684, 412 F.2d 1360, 1362 (1969), discusses the distinction between design and performance specifications:
The specifications [at issue] ... are a classic example of “design specifications,” and not “performance specifications” ____ [T]hese specifications ... set forth in precise detail the materials to be employed and the manner in which the work was to be performed, and [the contractor] was not privileged to. deviate therefrom, but was required to follow them like a roadmap. In contrast, typical “performance” type specifications set forth an objective or standard to be achieved, and the successful bidder is expected to exercise his ingenuity in achieving that objective or standard of performance, selecting the means and assuming a corresponding responsibility for that selection.
Accord Stuyvesant Dredging Co. v. United States, 834 F.2d 1576, 1582 (Fed.Cir.1987). The relevant question is whether it was expected that the contractor would use its own expertise to determine how the performance level required by the TDP would be achieved. See Stuyvesant Dredging, 834 F.2d at 1582 (performance specifications “leave it to the contractor to determine how to achieve [the results to be obtained]”).
This court and its predecessor have recognized that some specifications can have the characteristics of both performance standards on the one hand, and composition or manufacturing requirements on the other. See, e.g., Bethlehem, 199 Ct.Cl. at 254, 462 F.2d at 1404; Norwood Mfg., Inc. v. United States, 21 Cl.Ct. 300 (1990); Haehn Management Co. v. United States, 15 Cl.Ct. 50, 56 (1988). For the reasons set out below, it is apparent that the TDP has elements of both a performance standard and of a design blueprint.
The TDP is a detailed blueprint for the pellet's physical construction. The composition of the flare is also strictly circumscribed in terms of constituent elements and mix-range tolerance. Making the pellet is only generally described, however, and plaintiff clearly had flexibility in mixing, shaping, drying and choice of solvents. The TDP also has a performance element, however, in that there was “an objective or standard of performance” which the M206 flares were required to meet. The TDP therefore requires certain performance and leaves manufacturing largely up to the contractor, but gives great detail about materials and assembly.
Plaintiff introduced into evidence several of the drawings and parts lists for the *368M206 flare. An example will be helpful in exhibiting the level of detail specified in the TDP. Drawing No. 9342955 is entitled “Pellet Assembly.” It describes the application of first fire and intermediate charge into the flare grooves, as well as the wrapping of the pellet with aluminum tape. Some of the materials specified in this drawing are: first fire, intermediate charge, butyl acetate, acetone, and aluminum tape. Each of these materials has a part or specification number assigned to it. Butyl acetate, for example, has a specification number, TT-B-838a. Specification TT-B-838 is five pages in length. First fire is part number 9311655. It has a separate sheet describing its own constituent parts. Each of those constituents in turn has an associated specification or part number. The court is satisfied that the specification sets forth in “precise detail” the materials to be used.
A design specification, however, requires that the TDP specify in “precise detail” the procedures to be used. The court acknowledges that the TDP does address some procedures. It does so, however, only on some drawings, and then only in the most general sense. Some examples will illustrate. Drawing No. 9344023, entitled “Cap, End Assembly,” states in the notes accompanying it, “Coat O-ring 9344021 generously with grease ... prior to assembly,” and, “There shall be no grease in shear pin holes.” The note accompanying drawing 9311630, entitled “Pin Shear,” states, “May be cut to length at next assembly.” Drawing No. 9342952, entitled “Piston, Cushion Assembly,” states “Piston Cushion shall be centered so as not to block piston hole with adhesive side against piston.” Occasionally more detail is specified. Drawing No. 9344024, entitled “Flare, Aircraft: Countermeasure, M206 (Alternate)” gives an “Advisory Assembly Procedure” in note 3 accompanying the drawing:
NOTES:—
******
3 — ADVISORY ASSEMBLY PROCEDURE
A — INSTALL CUSHION PISTON ASSEMBLY-9342952
B — INSTALL PELLET ASSEMBLY-9342955 OR 9347036
C — INSTALL FELT SPACER-9311631 (NOTE 4)
D — INSTALL ALUMINUM END TAPE — 9327200 WITH ADHESIVE SIDE AGAINST SPACER.
E — INSTALL END CAP ASSEMBLY-9344023 INTO CASE.
F-INSERT SHEAR PINS-9311630 COATED WITH
SEALANT, SILICONE, RTV, TYPE I, SPEC MIL-A-46106 OR SEALANT, SILICONE, RTV, TYPE I, SPEC MIL-A-46146. (NOTE 6)
Even this drawing, however, provides no direction on how to perform any of the steps listed here. The TDP relies on the contractor’s experience and knowhow to fill in the production details. It cannot be said that there is a “road map” to follow, as the TDP does not address all the procedures that would be necessary to produce the M206 flare.15
The TDP is therefore a design specification as to materials and most aspects of assembly, and it is a performance specification with respect to procedures. Plaintiff’s reliance on Spearin, 248 U.S. at 132, 39 S.Ct. at 59, for the assertion that SOS need not identify the precise defect in the TDP is misplaced because this is not fully a design specification. This court has said that where, as here, a defective specification is alleged and the specification is a mixed design and. performance type, the defect must be found in the design portion of the specification. Johns Manville Corp. v. United States, 13 Cl.Ct. 72 (1987), vacated on other grounds, 855 F.2d 1571 (Fed.Cir.1988).
The court concludes that because the TDP partakes of elements of both design and performance, plaintiff must either isolate the defective element of the TDP, or must affirmatively demonstrate that it did not cause the failures. In other words, to *369eliminate the potential causation factors introduced by manufacturing steps left to its own ingenuity, plaintiff has to isolate those processes or compositions mandated by the Government as the only possible causes of failure. Each count, therefore, has to be separately viewed to consider whether plaintiff has created a plausible link between the failure and the design portions of the TDP.
With respect to Count III, plaintiff refers to Drawing No. 9342955 to support its defective specification argument. The relevant portion is set forth below:
NOTES:-
* * * * * *
3— ASSEMBLY INSTRUCTIONS:
A — PILL LONGITUDINAL [sic] GROOVES (4) AND END GROOVES WITH INTERMEDIATE CHARGE-9311655 BY COMBINING WITH SUFFICIENT BUTYL ACETATE SPEC TT-B-838 OR ACETONE, SPEC 0-A-51 TO MAKE A LIGHT PASTE. GROOVES TO BE APPROXIMATELY HALF FULL. DRY AT AMBIENT TEMPERATURE FOR 5 MINUTES MINIMUM.
B — APPLY A BEAD OF FIRST FIRE —9311656 APPROXIMATELY .125 WIDE OVER THE SURFACE OF INTERMEDIATE CHARGE — 9311655. DRY ASSEMBLY ONE HOUR MINIMUM AT 250 [+ OR -] 10 [DEGREES F] WHEN BUTYL ACETATE IS USED IN FIRST FIRE AND AT AMBIENT WHEN ACETONE IS USED IN FIRST FIRE.
# * * * * *
4— ADVISORY:
RESIDUAL AMOUNTS OF SOLVENTS PRIOR TO TAPING MAY AFFECT PERFORMANCE REQUIREMENTS OF MIL-F-63107(AR).
Specifically, plaintiff points to this drawing in response to the Government’s argument that the TDP did not address drying requirements. Defendant, noting that the drawing specifies only minimum drying times, argues that the drawing leaves the contractor free to dry for a longer time and at a greater temperature as necessary. According to the Government, it should have been clear that plaintiff was to use its expertise and ingenuity in determining a drying process sufficient to assure removal of all solvents. Defendant also suggests that the advisory note should have been a warning to SOS that a procedure to remove residual solvents was necessary.
Drawing No. 9342955 does not help plaintiff's position. This count relates to an oven drying process employed after flares are fully assembled and in their packaging. Drawing No. 9342955, however, involves application of first fire and intermediate mix into the groove of the naked pellet and before the flare is assembled. It is important to remember, however, that the drying process at issue in this count is one employed to remove solvents (and hence prevent outgassing). The drying procedures specified in Drawing No. 9342955, however, do not appear to relate to solvent removal in the body of the pellet.
The drying times specified in the drawing are of almost negligible intensity and duration compared to the one plaintiff (and indeed others) employed to cure outgassing. The evidence suggests that the Government was not so ignorant of drying necessary for the pellet as to specify ambient drying for only a matter of minutes. The court concludes that the dryings specified in Drawing No. 9342955 were merely intended to assure that first fire and intermediate fire had dried, and not as a means to remove residual solvents in the pellet composition.
Outgassing was not a wholly understood phenomenon. It is also true that there was no minimum moisture content requirement in the contract. Still, it is clear that the flare-making community was aware of out-gassing and was aware that a high temperature cure for several hours was necessary to eliminate it. Cabaret in fact testified that SOS employed an eight hour, 180 degree oven drying of the flare pellet composition for which it is not claiming compensation. This suggests that SOS anticipated some drying steps not specified or alluded to in the TDP. It is also worth noting that *370SOS instituted the very same process on the MJU-7 contract for which it is now claiming compensation. SOS clearly was not ignorant of the problem, and, given the sparse procedural direction in the TDP, it is not unreasonable to conclude that SOS should have anticipated that an oven drying process was necessary.
In sum, plaintiff has failed to identify any defect in the design portion of the TDP. The TDP was not defective with respect to the drying process. Plaintiff was given discretion in implementing drying procedures by the failure of the TDP to address that known need. Plaintiff is not entitled to recover on this count.
D. Count IV — Breach of Implied Warranty of Specification Related to Static Testing.
1. Factual background and contract provisions.
In this count, SOS alleges that the TDP is defective because flares manufactured in accordance with it continually failed to meet static output requirements.16 As with Count III, it urges that the TDP is a design specification and that it therefore need not identify the defect or defects in the TDP. There is sufficient latitude in the way SOS could manufacture pellets that, as discussed above, plaintiff must point to the particular part of the TDP to which it attributes its difficulties.17 Plaintiff devoted substantial time at trial presenting evidence in two broad areas as to why the TDP was defective. First, it alleges that the Government should not have qualified ground magnesium as an alternative fuel to atomized magnesium because it results in unpredictable performance. Second, plaintiff alleges that the TDP was defecfive because it failed to specify certain static testing procedures without which the static test parameters were unattainable.
Originally, the contract did not permit use of ground magnesium. On September 28, 1981, Armament Material Readiness Command Configuration Control Board Directive No. AID 2001 (the “Directive”) was issued with respect to the M206 TDP. It addressed Drawing No. 9311654, the M206 pellet composition drawing. The purpose of the Directive was to add several new alternative M206 flare compositions to the TDP, including a formulation utilizing ground magnesium manufactured in accordance with MIL-M-382C(AR). Prior to the Directive, the Army had considered only atomized magnesium as suitable for M206 flares. The Directive stated that all the new M206 flare compositions, including the ground magnesium composition, had met the M206 TDP requirements during testing by defendant and Thiokol.
It is apparent from the record that the Army was anxious to qualify ground magnesium because of a perceived shortage of atomized magnesium. A December 1979 correspondence from Major General Light to Major General Eicher, referring to the “critical shortage” of atomized magnesium, states that it is “imperative that the test programs ... be accelerated and monitored closely in order to meet urgent U.S. requirements for decoy flares____” Another internal Army correspondence from December 1979 reflects similar concern. It again refers to the “critical shortage” of atomized magnesium and continues: “[o]ne solution to the problem would be the prompt qualification of the fine ground magnesium as an alternate. Work in this *371direction is ready to begin on a crash basis both at this command and LHAAP [Longhorn]....”
The ground magnesium testing performed at Longhorn, referred to in the Directive, involved two phases, both of which were part of the Alternate Fuel Study (“AFS”) program. The AFS was being conducted under ARRADCOM direction in Dover, New Jersey. The first phase of the AFS was conducted at ARRADCOM, beginning in 1979. The purpose of the first phase was to test ground magnesium samples, which ARRADCOM received from various suppliers, in order to find a blend of ground magnesium suitable for full-size production and capable of production in large quantities.
Patricia Famell, an ARRADCOM chemist, tested ground magnesium samples using 40 millimeter infrared flare test vehicles. (The 40 mm test vehicles are less than one-third the size and weight of M206 flares.) Farnell ultimately recommended BS-2 ground magnesium, manufactured by Reade Manufacturing Company. Static output results of 40 mm test vehicles are only an indication of actual flare performance, but based on the 40 mm tests, ARRADCOM recommended that six formulations of BS-2 be tested in full-size flares.
Full size M206 pellets were tested in February 1981 at Longhorn. Tests were performed on variations of the six different formulations of BS-2 ground magnesium identified by ARRADCOM as possible alternatives. Each variation was designated an AF (Alternate Fuel) number.
These tests involved three stages. Initially, small scale mixes were produced of each of the six formulations, and 10 pellets from each of the six formulations were pressed. Half were finished with a four inch groove configuration and half with a “full groove” configuration. These pellets were tested for bum times, action times, and output, and the results were compared to the MIL-F-63107 requirements.
Based on these preliminary tests, the formulations of the various compositions were adjusted to maximize static output performance. Tests were also performed to verify that no problems existed with breakup during ballistic testing. The most favorable formulations were selected for full-size production, final qualification testing.
Finally, tests were performed on production-sized M206 flares made with certain formulation mixes and blended on ThiokoPs production equipment. These tests are the “Final Qualification Tests.” During this testing phase, all test pellets were made using Thiokol’s then-current production equipment. Thiokol used a mixer known as a Simpson Mull-Mixer. Regulations defining pyrotechnic mixing procedures prohibit manual scrapedown. Consequently, the wheel of the Simpson Mull-Mixer was removed to prevent material buildup. This left the mixing ability of the Simpson Mull-Mixer dependent upon the ability of the mixer’s plows to deflect and fold the material.
The first ground magnesium mix tested during the Final Qualification Tests was designated as AF-18. It is important to note that Thiokol had mixing problems with AF-18 although defendant contends that these problems can be explained by removal of the mixer’s wheel. AF-18 failed several of the static performance parameters. Specifically, AF-18 pellets failed the requirement of 1.25 seconds minimum time above 8800 watts per steradian. They also failed the requirement of 2.3 seconds minimum time above 900 watts per steradian. Finally, the AF-18 pellets exhibited an average total bum time of 2.47 seconds. AF-18 eventually was disqualified.
Thiokol prepared a second ground magnesium mix, AF-21. This mix was made with an increased amount of solvent to prevent mixing problems. This batch, however, still did not mix properly, and large amounts of teflon were left undissolved. Pellets made with AF-21 exhibited a very fast burning rate and fell well below acceptance static testing limits. Specifically, flares made with AF-21 failed to stay above 900 watts per steradian for a minimum of 2.3 seconds. AF-21 pellets also failed to remain above 1800 watts per steradian for a minimum of 2.3 seconds. AF-21 pellets exhibited a total bum time of *372only 2.255 seconds. While there is no requirement for total burn time, the total burn time for AF-21 pellets was less than the “time-above” requirements for 900 and 1800 watts per steradian. The AF-21 pellets therefore could not meet those TDP requirements.
Following the AF-21 test, Thiokol concluded that the failure of that mix to meet M206 burn duration requirements was possibly due to one of two factors. The first explanation was small particle size of the magnesium powder due to variation within the drum. A pellet composed of smaller particles of magnesium will burn more quickly because of greater total pellet surface area. The AF-21 magnesium had been taken from the bottom of the magnesium drum, where the average magnesium particle was approximately 17 microns. The average particle size of magnesium taken from the top of the magnesium drum was 28 microns. A second possible reason for the failure was that, because of poor mixing action, the magnesium was not uniformly coated with the binder, resulting in fast and erratic burning.
In order to determine the effect of particle size upon static output, a third mix, AF-22, was processed using magnesium from the bottom of the drum. This mix was compared to a preliminary lab mix, AF-8, which was composed of magnesium taken from the top of the same drum. Results indicated that there was a slight decrease in burn time in the AF-22 mix, but not enough to explain the inadequate bum time in the AF-21 mix. Thiokol interpreted these results as indicating that particle size did not have an appreciable effect upon bum time.
Thiokol prepared a fourth ground magnesium blend AF-23. For that blend, Thiokol concentrated not on particle size but on mixing. A special request was made to put the wheel back on the Simpson Mull-Mixer for this mix. After discussions regarding the waiver of safety restriction, permission was granted to put the wheel back on the mull-mixer. The mix was manufactured using a new shipment of 190 pounds of BS-2 magnesium from Reade.
Thiokol divided the ground magnesium barrel used for AF-23 into nine equal parts and placed each part into separate blend buckets. AF-23 was manufactured with equal amounts of magnesium from each blend bucket. During mixing, the wheel of the Simpson Mull-Mixer functioned satisfactorily for approximately five minutes. After five minutes, the mix tended to form a lump that would not dissipate, and the mix began to slide in the bowl, making further mixing ineffective. Also, as with the other Thiokol blends, mix AF-23 would not pour from the blender using remote control equipment because it was too thick. It had to be manually removed from the blender using wooden paddles.
The AF-23 composition was pressed into flares and passed all static test requirements. It is plaintiff’s position that AF-23 was an anomalous mix. Dr. Alan Snelson, a scientific advisor who has been employed with ITT Research Institute in Chicago for 29 years, testified for plaintiff as an expert in the physical properties of the M206 flare, including magnesium, ground and atomized. Snelson has a PhD in physical chemistry. According to Snelson, AF-23 met the specifications “very handily.” He also stated that, with respect to the “integral,” which measures the total energy contained in the flare, he had not seen those kind of BS-2 results in any of the documents he had seen. When asked if this mix was an aberration, he replied: “It is difficult to predict what flares are going to do in the way of output____ The number has turned out to be almost amazingly good. That is about as much as I can say.” Snelson also commented on the magnesium that went into AF-23. He stated that there is no indication that Thiokol did any mixing of the particles, and that the particles were of a very consistent size throughout the nine parts into which the drum was separated. In the previous barrels, there had been segregation of particle size.
The results of the AFS are collected in a document entitled the M206 Flare Alternate Fuel Study, dated October 1981. One of the conclusions reached was that:
*373[t]he formulation utilizing the Reade BS-2 Magnesium did not process satisfactorily in the wheelless mix mullers utilized at Longhorn. However, once a satisfactory mix was finally obtained, the flare composition resulted in acceptable end item performance. Although the Reade BS-2 Magnesium is not immediately usable at Longhorn with existing mixers and mixing procedures, it may be suitable for use by other manufacturers with different mixing equipment and procedures not currently available at Longhorn. (For example, manual scrapedown of mixers is not permitted at Longhorn but may be used by the Private Sector.)
The AFS Final Report also concluded that ground magnesium would be satisfactory for M206 flare production and that it would be listed as an alternate on applicable drawings. The parties do not dispute that defendant based its decision to qualify ground magnesium for the M206 TDP based on the results of the AFS.
Plaintiff contends that using ground magnesium in flares will inevitably lead to erratic and unpredictable static outputs of M206 flares. It alleges that the Army qualified ground magnesium without adequate study and did so because it was anxious to develop an alternative to atomized magnesium.
In addition to problems allegedly inherent in the use of ground magnesium, plaintiff contends that it should have been given information on three aspects of testing procedures which affect static performance measurement: test blocks, test hooks, and reflection within the test tunnel. A test block holds secure the electric match which ignites the flare during static output testing. It is undisputed that the Government was aware that test blocks had been used at Longhorn during development of the M206 static output requirements. Thiokol had run tests on the effect of test blocks on static performance measurements and had determined that test blocks can improve “rise time” performance by approximately 65 percent. The Government asserts that there is no evidence that “rise time” performance is increased consistently by 65 percent when test blocks are used, though it concedes that it was aware that rise time” is improved when test blocks are used.
The Government nevertheless contends that use of test blocks was left up to the ingenuity and expertise of the contractor. However, on June 1, 1982, Frank Klosowski, an ARRADCOM employee of the defendant, recommended that the M206 TDP be amended to add the test block procedure. Test block procedures were also specified in the original M206 contracts awarded to Bermite and Tracor, both awarded after the contract with plaintiff, and the test block procedure was added through contract modification to the Thiokol contract. The test block procedure drafted by Klosowski is now included as a permanent part of the M206 TDP. Revisions B and C of MIL-F-63107, dated May 11, 1983, and January 6, 1986, respectively, now include that test block procedure at If 45.4.2.
Plaintiff also contends that the method specified in the M206 TDP for hooking flares to the test stand led to cracking of the flares, causing them to prematurely fall from the test stands. When flares fell from the stand, they were obscured from the radiometer’s measurement, which ultimately led to readings indicating static output failures. MILr-F-63107A(AR) ¶ 4.5.42 describes the method for attaching the flares to the test stands:
The non-ignition end of the pellet will be prepared for mounting by drilling a Vs" diameter hole, %" deep. Hardware with No. 10 coarse wooden thread, %" thread length on one end and opposite end suitable for mounting to the test stand shall be carefully threaded into the Vs" drilled hole. Prior to assembly, screw threads shall be coated with RTV-106 [glue] or an approved substitute to reduce friction and possible ignition.
Cabaret testified that when SOS would attach hooks to the flares using this procedure, the flares would crack and fall prematurely from the stand. According to Cabaret, SOS eventually solved the problem by requesting permission to use a *374smaller hook, a larger hole, and a different glue.
A Value Engineering Change Proposal (“VECP”) submitted by Thiokol in June 1983 describes a similar problem. It states, under II15, entitled “Need for Change,” that “[t]he minor diameter of a standard No. 10 screw metal thread is about 0.134 inches. Using the current Vs (0.125 inches) drill diameter, the pellet must expand to some degree to accommodate the screw. This may result in longitudinal cracking with subsequent early drop off on the test stand.” After describing tests it had conducted to find a solution to this problem, Thiokol proposed that.paragraph 4.5.4.2 be changed to require a %4" diameter hole instead of a Vs" hole and suggested RTV-106, Devcon five minute epoxy instead of the epoxy suggested in 4.5.4.2. The VECP was granted in July 1983. According to Cabaret, the change described by this VECP “is exactly what we did.” Although the VECP related to extruded, not pressed flares, there was no suggestion that the shaping had any effect on the need for a larger diameter hole.
Defendant concedes that during tests of plaintiff’s flares at Longhorn, Thiokol attempted to apply test screws to SOS flares and that six of these flares subsequently revealed cracks. To avoid cracking plaintiff’s flares, Thiokol drilled a larger hole before inserting the screw.
The third aspect of static testing which plaintiff alleges should have been addressed by the TDP is the effect of reflection characteristics in a test tunnel on static output measurements. Specifically, plaintiff contends that the M206 static output requirements were developed at Longhorn’s tunnel, where unique reflection characteristics inflated static output. The result was that the static output requirements were unnecessarily difficult to meet elsewhere. SOS’s tunnel did not have Longhorn’s output-enhancing reflectivity. The issue was resolved during the “round robin” tests conducted in March 1983, which are discussed below.
Cabaret testified that SOS had static output problems from the very beginning, and that after SOS had failed two first article tests for static output failure, it was decided to have “round robin” tests to determine whether SOS’s static output failures were due to testing anomalies resulting from differences between SOS’s testing facility and the Longhorn test tunnel. According to Cabaret, the decision to have these round robin tests was a mutual one between the Army and SOS. Gene Venable testified regarding the round robin tests. Although Venable had “done everything [he] could” and had “great confidence in [SOS’s] infrared measurements,” flares were still failing at SOS. Whether SOS was making a good flare could not be determined until the Army was satisfied that unfavorable results were not due to disparities in testing facilities.
SOS was allowed to select flares from its best lot for the round robin tests. According to Cabaret, SOS flares from the same lot tested at Longhorn were then tested at SOS, and comparisons were made between the two measurements. Longhorn flares were also tested at SOS.
SOS’s third first article test was conducted at Longhorn on March 8-10 using Longhorn test procedures. SOS failed this first article. Based on results obtained, the Army determined that SOS flares exhibited an energy level 15 percent below those of Longhorn. Another first article test was conducted on March 22, 1983, this time at SOS, again as part of the round robin tests. Longhorn flares were also tested at SOS. Based on both the March 8-10 and the March 22 tests, the Army concluded that the peak static output values for both the Longhorn and SOS flares were 20.9 percent greater when tested at Longhorn.
Plaintiff was given an upward adjustment of 20.9 percent on the March 22 first article test. With that correction, plaintiff passed first article. Plaintiff was subsequently allowed a 20 percent correction factor for its flares during production at its own facility.
It is undisputed that in early 1984, Thiokol, under the Government’s direction, conducted tests to develop standard static test procedures and facilities, and, as a result, *375the test facilities at Longhorn were changed to arrive at a standard setup and procedure. These changes reduced the reflective energy in Longhorn’s tunnels by 21.4 percent. Because the static output requirements for the M206 TDP were derived at Longhorn, the TDP was overstated by 21.4 percent. On May 21, 1984, by Notice of Revision A4Q 2041, the M206 TDP was amended to reduce static performance requirements by 21.4 percent and to remove certain static output requirements associated with the “narrow band” of infrared energy (900 watts/steradian). These changes were never incorporated into plaintiff’s contract because, according to defendant, they were not necessary in light of SOS’s 20 percent correction factor.
Plaintiff continued to have static output difficulties even after it passed first article and went into production with the 20 percent correction factor. It is stipulated that the following production lots failed static output testing: Lot 10 of Interfix 001 in July 1983; Lot 027 of Interfix 001 in September 1983; Lot 032 of Interfix 001 in September 1983; Lot 039 of Interfix 001 in October 1983; Lot 045 of Interfix 001 in October 1983; Lot 003 of Interfix 002 in November 1983; Lot 021 of Interfix 002 in January 1984; Lot 014 of Interfix 002 in February 1984. Plaintiff and defendant disagree on the proper characterization of the problems which caused static output failures. Generally, plaintiff characterizes the failures as resulting from short bum time. Defendant characterizes the failures as infrared intensity problems. Almost all of the lots discussed above failed because flares within the lots could not bum long enough to meet certain “time above” requirements.18
Following the failure of Lot 014 of Inter-fix 002 in February 1984, plaintiff changed in several respects the way it manufactured flares. As of May 1984, SOS began to apply first fire into the grooves before applying intermediate mix and also began to fill the flare grooves completely with intermediate mix, rather than only filling them half way.19 Plaintiff also began to prepare its flares using percentages of magnesium at the low end of the tolerance allowed by the TDP. Defendant agrees that by decreasing the percentage of magnesium in proportion to the flare pellet’s other constituents, bum time would be increased, but that decreasing by too great a percentage results in problems reaching other static output requirements such as rise time and peak output. Despite these changes, it is undisputed that time above continued to be a problem, as did rise time to 7500 w/s, and that plaintiff experienced 19 other lot failures, each of 10,000 flares, after Lot 014 of interfix 002.
These changes were not the only attempts plaintiff made to eliminate static output problems. It pretested blends of pellet composition, pressing approximately six to 10 M206 flares from each blend and static testing them to determine blend burn parameters. Only if the blend burned adequately was it processed into finished flares. Blends which were inadequate for fast or slow burn times were set aside and stored later for “cross-blending” — that is, combining blends of opposite characteristics to produce an acceptable blend.
Plaintiff also instituted a process known as screening. When a lot failed static output testing, SOS would test the individual blends that were used in making the flares for that lot, removing blends which exhibited poor static output performance. The lot would then be retested for static output requirements. Twice as many flares were then sampled for retests.
Plaintiff also undertook engineering investigations to determine the cause of the static output problems. These investigations included altering the formulas for basic composition, first fire, and intermediate mix, as well as varying the length of the flare grooves and the amount of first fire and intermediate mix in the grooves.
*376Defendant does not dispute that SOS undertook the measures described above to improve static output performance. It urges, however, that such measures were required by the contract. With respect to pretesting, defendant argues that such procedures are standard practice among manufacturers of infrared flares. It also points to MIL-I-45208A (made applicable to SOS’s contract by Paragraph E.9), which provides:
Contractor Responsibilities. The contractor shall provide and maintain an inspection system which will assure that all supplies and services submitted to the Government for acceptance conform to contract requirements____ The contractor shall perform or have performed the inspections and tests required to substantiate product conformance to drawing, specifications and contract requirements____
Defendant also points to this MIL specification to support its contention that cross-blending was required by the contract. Defendant states that cross-blending is also a standard practice among manufacturers of infrared flares.
With respect to screening procedures, defendant argues these were required by MIL-I-45208A, If 3.7, which provides: “The contractor shall establish and maintain an effective and positive system for controlling nonconforming material, including procedures for the identification, segregation, presentation and disposition of reworked or repaired supplies.” According to defendant, this section also required SOS to perform its engineering studies. Defendant’s response begs the question, however, of whether these steps were occasioned, as plaintiff alleges, because of the use of ground magnesium and the confused testing procedures and facilities.
It is not clear whether the remedial measures, individually or together, ever solved the static output problems completely. Plaintiff continued to use ground magnesium, and it does not contend that there is some point after which static output no longer was a problem.
Defendant argues that SOS’s problems were the result of several factors unrelated to the TDP. Generally, defendant argues that SOS had poor quality control, which was aggravated by personnel turnover. Defendant also avers that the use of hydroshocking to remove MEK contributed to SOS’s problems because the introduction of water with magnesium causes magnesium to oxidize, which reduces energy output. Additionally, defendant points out that as of April 6, 1989 SOS had produced and shipped 1,162,520 flares using the TDP it alleges is defective.
SOS makes one allegation in Count IV unrelated to the TDP. It argues that it was delayed in completion of first article by 120 days because Marvin Elmowitz, an engineer employed by ARRADCOM, withheld approval of SOS’s inspection equipment and procedures until plaintiff made numerous changes, none of which were required by the contract. SOS contends that it would have been ready for first article testing by June 14, 1982. SOS was allowed to and did undergo the environmental portions of the first article test in July and August of 1982.
Gene Venable, a Research Chemist for ARRADCOM, advised Elmowitz on SOS’s test equipment and procedures, and testified for defendant. He stated that SOS did initially have difficulty regarding certain aspects of testing, including operating the testing equipment. For example, he stated that in taking static output measurements SOS was using an inappropriate calibration distance between the “black body”20 and the radiometer and that using an improper distance would result in inaccurate readings. In addition, Venable stated that SOS was not keeping proper logs of data obtained during their daily calibrations and that this made it impossible for SOS to keep track of its calibration equipment. He stated that malfunctioning of the calibration equipment could therefore go unnoticed. According to Venable, SOS was also *377using formulas which were not appropriate to the type of output measurements it was taking. He stated that when SOS testing personnel attempted to explain the calculations they were making, he noticed that they were using formulas for extended sources when point source calculations were called for. Venable stated that Jay Reynolds admitted, when he (Reynolds) was performing the calculations, he “got up about halfway through and said that he really didn’t know what he was doing.” According to Venable, the calculations SOS was using would lead to erroneous data.
Venable testified that all corrections were made “within a couple of months.” There was no testimony from plaintiff that its calibration methods were in fact correct and proper before these adjustments. Venable was a straightforward and credible witness. There was no indication that his criticisms of SOS were contrived or unduly harsh. The court rejects plaintiff’s allegation that defendant unreasonably delayed completion of its first article by unreasonably insisting on correct testing procedures.
Plaintiff claims that it is entitled to a total of $3,272,079 in damages as a result of added costs incurred in coping with the defective TDP and as a result of the delay suffered because allegedly defendant unreasonably withheld approval of test procedures.
2. Discussion.
The court held in connection with Count III that the TDP for the M206 flare is both a performance and a design specification. To recover on its claim, plaintiff must demonstrate that the alleged defect relates to the design portion of the specification. Plaintiff has introduced evidence that the TDP was defective because it improperly qualified ground magnesium as an acceptable fuel and because the testing procedures described in the TDP would lead to artificially low static output readings.
The AFS was the culmination of three years of research and testing. In that time, the Army was able to produce only one acceptable blend. No attempts were made to determine what precisely accounted for AF-23’s success or to assure that the performance of that blend could be repeated consistently with different barrels of ground magnesium.21
In this regard, plaintiff's expert on the M206 flare components, stressed repeatedly that the type of reaction involved in the M206 flare is, in most respects unknown, that predicting how a particular mix of flare composition will perform is very difficult, and that it is very difficult to repeat test results. He further stated:
I think the situation with respect to the M206 flare is that one is trying to obtain the maximum amount of energy out of the flare in a very carefully defined way. It appears that trying to meet those specifications is not easy. It appears that if you tailored the flare to very easily meet a particular criteria, in let us say the amount of energy above 8800 [w/s], then you may penalize yourself in the ease with which you can attain some other parameters maybe the time at 1800 or 900. It appears that you are walking a tightrope in trying to get the performance out of this particular material.
I believe looking at all of the data that I have seen that maybe it is asking too much, that is the performance specification, for this particular material____
Snelson was a believable witness. He was obviously not attuned to the legal issues and simply answered candidly. The court accepts his testimony which supports the conclusion that the Government’s qualification of ground magnesium was improper.
The court agrees with plaintiff that the Government should not have qualified ground magnesium. Not only was the AFS of dubious reliability and thorough*378ness, but the results were inconclusive at best. At worst, it showed that virtually no blend was capable of producing the required results. The Army never really understood what it did to produce one anomalous, small, non-production set of flares that met those requirements. The court is persuaded that even Longhorn could not have duplicated those results with any consistency, and that it was only the perceived urgency of the need for an alternative fuel that caused these apparently serendipitous results to be blithely adopted. The Government did not fully address the mixing problem. The AFS merely stated that AF-23 “may be suitable for use by other manufacturers with different mixing equipment and procedures.” This suggests that the study of ground magnesium was cursory, incomplete, and targeted for a result.
SOS’s only discretion within the TDP was to vary the percentages22 of the flare’s composition, but then only within certain tolerances. Cabaret testified, and the court accepts, that SOS tried to make flares to the limit of every tolerance but still could not get the flares to burn acceptably in a consistent manner.
Defendant argues that SOS’s static output problems were caused in part by manufacturing procedures which the TDP leaves up to the contractor. Specifically, defendant points to SOS’s use of hydroshocking to dispel MEK during mixing.23 The court disagrees. First, SOS’s static output problems continued long after SOS had instituted its additional drying steps to alleviate outgassing. Second, Dr. Snelson performed a study at SOS’s facility in which he compared flares made using the hydroshocking process with flares made using a hexane shocking procedure. Based on that study, he stated that “there does not seem to be any evidence that the water-shock procedure is in any way degrading the magnesium in the flare to any significant degree.” Finally, if defendant were concerned with hydroshocking, it should have denied plaintiff’s request to use this process.
The qualification of ground magnesium as an alternative pellet fuel was a defect in the design portion of the TDP. Plaintiff reasonably relied on the TDP in preparing its bid. Using ground magnesium resulted in unnecessary expense to the plaintiff, and SOS is entitled to recover for damages which flow from this defect.
The TDP was also defective with regard to testing procedures and to the requirements for static output, drawn from Longhorn’s eclectic test setup. Plaintiff used the procedure specified in the TDP. The procedure was eventually approved by the Government. Venable testified that, after SOS’s early difficulties were eliminated, he had complete confidence in SOS’s test procedures. However, these procedures resulted in artificially low readings for three reasons — the flares fell off the test hooks; the absence of a test block led to premature triggering of the sensors; and SOS’s tunnel did not have the same reflective characteristics as the tunnel in which the output requirements were derived. It is undisputed that the TDP was revised to require test blocks, to correct for the problem of flares falling off the test hooks, and to correct output requirements to account for Longhorn’s tunnel reflectivity. Defendant contends that these are procedures with which plaintiff should have been experienced. No one could have anticipated the anomalies created by the Longhorn tunnel, however, and if the test procedures were either correct from the outset or self-evident, the Government would not have agreed to incorporate the changes to the TDP.
With respect to both the decision to qualify ground magnesium and the failure to inform bidders that the static output re*379quirements were based on possibly unique procedures and facilities, these circumstances suggest similarities to Helene Curtis, Indus. v. United States, 160 Ct.Cl. 437, 312 F.2d 774 (1963). The allegation there was that a specification for manufacturing a chemical disinfectant was misleading without additional information available to the Government but not shared with contractors. An excerpt from the decision shows the points of similarity:
The disinfectant was novel and had never been mass-produced; the Government had sponsored the research and knew much more about the product than the bidders did or could; it knew, in particular, that the main ingredient, chlormelamine, was a recent invention, uncertain in reaction, and requiring extreme care in handling; it also knew that the more costly process of grinding would be necessary to meet the requirements of the specification, but that in their understandable ignorance the bidders would consider simple mixing adequate; and the urgency for the disinfectant was such that potential contractors could not expend much time learning about it before bidding. In this situation the Government, possessing vital information which it was aware the bidders needed but would not have, could not properly let them flounder on their own. Although it is not a fiduciary toward its contractors, the Government — where the balance of knowledge is so clearly on its side — can no more betray a contractor into a ruinous course of action by silence than by the written or spoken word.
Id. at 444, 312 F.2d at 778. The court went on to hold that “Specifications so susceptible of a misleading reading (or implication) subject the defendant to answer to a contractor who has actually been misled to his injury.” Id. at 445, 312 F.2d at 778. No recovery was allowed for a follow-on contract, however, because the plaintiff should have been knowledgeable about the problems.
The lesson to be drawn from Helene Curtis is that the Government cannot launch a technical data package into the procurement process if it has reason to know that the results called for are problematic, unless it discloses the knowledge available to it concerning problems and possible solutions, or unless that information is already generally known by contractors in the field. A similar analysis was used by the Court of Claims in Ordnance Research, Inc. v. United States, 221 Ct.Cl. 641, 609 F.2d 462 (1979). That case is factually similar to the one at bar. There, the contractor had been awarded a contract to mass produce fire bomb igniters, which used atomized magnesium. Plaintiff alleged, and the court held, that the specifications were design in nature and implicitly warranted that the igniters could be mass produced in a safe manner. Id. at 670, 609 F.2d at 479. The plaintiff showed that by changing the specifications, safety could be increased. And although not necessary to the result, it was able to demonstrate the source of the explosion risk. Plaintiff was allowed to recover for costs associated with explosions that occurred at a time when it was unaware of a solution to the problem. The court held that
[Specifications furnished plaintiff by the government were design specifications for a mass production of igniters. They described in detail the materials and the manufacturing processes to be used, and required compliance with [safety manuals].
When the government issues design specifications of a detailed nature, as in this case, it warrants the sufficiency and efficacy of those specifications to produce the desired product in a satisfactory manner.
Id. at 670, 609 F.2d at 479.
Also instructive is R.E.D.M. Corp., 192 Ct.Cl. at 891, 428 F.2d at 1304. The manufacturer in that case was unable to arm artillery fuses consistently. The failure rate was high. Although it was unable precisely to explain why it helped, R.E. D.M. was able to improve performance by reducing the thickness of one of the fuse components. The court allowed recovery for the additional costs.
*380The present circumstances partake elements of all three of these cases. Here, the Government, in its eagerness, represented a procurement as ready for large-scale production when the technology had not been adequately developed. The Government wanted a fixed-price production contract when what it was really offering was, in part, a research and development project. It was misleading to hold out ground magnesium as fully suitable when the Government had reason to know success would be highly elusive. It was, at a minimum, unimaginative, and in fact unintentionally deceptive, not to inform bidders that the performance requirements were difficult to meet and were based on the peculiarities of the Longhorn test tunnel and procedures. While the present circumstances are not entirely like those of Ordnance in that manufacturing processes here were virtually uncontrolled, a critical flaw in the design portions of the TDP was the implied warranty that, whatever reasonable manufacturing processes were used, ground magnesium would be suitable.
The court therefore finds that the TDP was defective in not including procedures upon which performance requirements were based and in qualifying ground magnesium. Plaintiff is entitled to recover for any damages which it can show resulted from this defect.
3. Damages.
Plaintiffs damages claim for Count IV is segregated into discrete areas. They will be addressed individually.
a. First article testing
Plaintiff alleges that, had it not been unreasonably refused approval for its testing equipment and had it not experienced difficulties with testing, it would have finished first article testing on time and within the bid estimate. It asserts that unreasonable delay in the approval of test equipment and delay caused by the testing problems resulted in expenditures of $318,658 more than included in the bid.
In arriving at this figure, Aratani consulted an accounting record, referred to as a “Sale Order/Work Order,” to determine the number of labor hours incurred by SOS from the beginning of the contract through May 1983. Aratani testified that, according to SOS personnel, all labor effort at SOS for the M206 contract was devoted to the first article requirement. There was also testimony to that effect by Cabaret and Reynolds. With minor adjustments, Aratani determined that the labor cost through May 1983 was $108,637. He added overhead and material cost and then subtracted the amount bid for first article labor and materials of $43,750. He then added G & A to arrive at the $318,658 figure claimed.
Some important questions remain unanswered regarding Aratani’s calculations. Initially, the damages calculation does not segregate the impact of the various elements which constitute damages. Specifically, the cost of the allegedly unreasonable delay in approving test equipment is not isolated. Plaintiff is not entitled to recover for delay caused by the initial failure to approve test procedures. However, the court is left without any basis for redacting this portion of the claim. Additionally, plaintiff has not recovered on Count III, related to outgassing. These are other possible causes of delays independent of Government fault. While this is not the same type of delay claim as Count VI, plaintiff nevertheless has the burden of establishing a reasonable basis for calculating damages, and is bound by the rule that in a delay claim the claimant must segregate delay costs attributable to the Government from those independently caused. See Wunderlich Contracting Co., 173 Ct.Cl. at 199-200, 351 F.2d at 969; Commerce International Co. v. United States, 167 Ct.Cl. 529, 543, 338 F.2d 81 (1964). Plaintiff has not met this burden. In addition, plaintiff’s contention that SOS would have been within bid costs is an essential one which has not been supported. This defect is one reason total cost claims are disfavored. See WRB Corp. v. United States, 183 Ct.Cl. 409, 426 (1968); J.D. Hedin Constr. Co. v. United States, 171 Ct.Cl. 70, 86-87, 347 F.2d 235, 246-47 (1965). The court concludes that plaintiff *381has not provided a reasonable basis to calculate damages as to test procedures and that it is not entitled to recover.
b. Pretest procedures
Plaintiff alleges that its pretest procedures resulted in increased costs, totalling $500,491.24 Defendant contends that plaintiff has only established $258,-158.33. Defendant does not dispute the labor rates, G & A rates, overhead rates, or S & S rates used by plaintiff in calculating damages for this claim. The difference between plaintiff’s and defendant’s computations rests in the number of labor hours involved in the pretest procedures. The dispute over labor hours required depends almost entirely on the number of flares assumed to have been subject to the pretest procedures.
Plaintiff alleges that 20,550 pellets were pretested. Plaintiff’s figure is based on 3,425 blends having been produced. This figure is taken by Aratani from SOS personnel who utilized actual blend records. This translates to a total of 1,712,500 pressed flares.
Defendant uses 16,314 as the number of pellets tested. Cabaret stated that 500 flares were produced per blend, and defendant alleges that 2,719 blends were subject to pretesting. Reynolds testified that an average of six pellets per blend were pretested. Therefore a total of 16,314 pellets were subject to the pretests, according to defendant. Defendant does not explain the *382connection between Cabaret’s testimony that 500 flares were produced per blend and its assertion that 2,719 blends were produced.
Defendant arrives at its figure for the number of blends apparently from its assumption regarding the total number of acceptable flares produced. However, more flares were produced than were accepted. Aratani received his information, though indirectly, from actual records. The court accepts plaintiff’s figure for the number of blends produced. Using plaintiff’s figure, damages related to pretesting are as follows.
Using plaintiff’s figure for the number of blends, material costs total $81,584. Plaintiff also seeks to recover labor costs for time incurred setting up the pretest procedures. Defendant allows $9,173.91, based on 16,314 pellets tested. Using plaintiffs number of blends yields $11,-555.95. With S & S hours, the total recoverable cost is $13,814.64.
Plaintiff also claims for test technician and material handler labor time. Defendant does not dispute that one of each was required to perform a blend check. Here, however, the difference between defendant’s and plaintiff’s calculations lies not only in the number of pellets assumed tested, but also in the number of hours required to do a blend check. Plaintiff claims that it required two hours to do a blend check, relying on support by Scott Wiitila, a Test Technician for SOS. Defendant allows two hours for set up and test of the first blend of the day, but only one-half hour for “the remaining four blends.” In calculating damages, it assumes one hour per blend. Defendant points to Wiitila’s testimony. Wiitila testified that it took 1.5 hours setup time to do a blend check and that in total the blend check took two hours. There is the clear suggestion, therefore, that once a blend check has been set up, subsequent blends can be tested in only one-half hour, and the court accepts defendant’s use of one hour per blend as the more accurate estimate.
Using plaintiff’s figure for the number of blends, and assuming one hour average testing time per blend, plaintiff recovers $26,509.50 for Test Technician labor, plus $173.20 for S & S, and $16,508.50 for Material Handlers, plus $2,577.16 for S & S.
The next labor category for which plaintiff seeks recovery is driver hours transporting blends from Mint Canyon to Placerita for testing. Plaintiff bases damages on 2,228 hours. Defendant allows 935.25 hours. Aratani arrives at this figure as follows. He assumes 45 minutes each way, plus 30 minutes loading and unloading time, for a total of 120 minutes per trip. With two drivers, as required by safety regulations, the total is 240 minutes per trip. Aratani assumes the trip was made 796 times, for a total of 3,184 hours. He discounts this figure by 30 percent for scrapped blends, arriving at 2,228 hours. Defendant allows 75 minutes and two people per trip. It assumes 374 trips, rather than 796.
Defendant allows 75 minutes per trip based on Beard’s testimony that the trip from Mint Canyon to Placerita took 35 minutes. Dean Johnson testified that if all paper work were ready, picking up and loading flares would take two to three minutes. Defendant allows five. Plaintiff, in support of its figure for transportation hours, relies primarily on the testimony of Pamela Winn, an employee of SOS who has held various jobs at SOS, including senior production technician and foreman. She worked at both the Mint Canyon and Placerita facilities. Winn testified that packaging, loading, and transporting the flares from Mint Canyon to Placerita required 1.5 to 2 hours. Relying on Beard’s testimony that a one way trip from Mint Canyon to Placerita required 35 minutes, plaintiff concludes that the total time was “at least 2lh hours.” As to the total number of trips, plaintiff notes Reynolds’ testimony that at least three trips were made per week, and that for safety reasons flare pellets could not travel in the same vehicle as completed flares. Plaintiff does not explain how this justifies the assumption that a total of 796 trips were made. The court notes that, according to Beard, trips were made between Mint Canyon and Placerita seven out *383of 10 days. The number of weeks of the contract is 159, or 1113 days; this number multiplied by 70 percent would yield 779.1 trips.
The court favors defendant’s figure for the number of trips. Reynolds testified that three trips were made per week. Over 159 weeks, this would be a total of 477 trips. However, according to Reynolds, not every one of those trips could transport flares. Defendant’s figure of 374 trips suggests roughly 2.4 trips per week. While the court is unclear how defendant arrives at its figure, it better conforms to the testimony. The court accepts defendant’s figure for the number of trips made.
Regarding the number of hours required per trip, Winn did not elaborate on why it required 1.5 to 2 hours to package and load the flares and travel one way from Mint Canyon to Placerita. The driving time is only 35 minutes, leaving between one and 1.5 hours for packaging and loading. The court cannot accept, without any explanation, that one to 1.5 hours were required for the seemingly uncomplicated task of packaging and loading the flares. The court therefore rejects Aratani’s assumption that each trip involved two hours of labor. Damages will be calculated based on defendant’s assumption that the round trip, including loading and unloading, required 75 minutes. With two drivers, and a total of 374 trips over the life of the contract, plaintiff is entitled to compensation for 935.25 hours. At $6.45 per hour labor rate, this amounts to $6,032.36; adding $886.46 for S & S, the total is $6,918.82.
Plaintiff is entitled to a total award of $66,501.82 for labor costs incurred in connection with pretesting. Adding overhead and G & A yields $221,448.93. With material costs of $81,584 and profit, plaintiff’s total recovery is $333,336.22.
c. Cross-blending
Plaintiff originally claimed $28,578 for cross-blending. Defendant concedes $28,-222.33. Both these calculations are based on the time required for actual blending, on the assumption that a cross-blend requires approximately 25 percent as much time as does a regular blend. According to plaintiff, Aratani misunderstood SOS personnel; in actuality, cross-blending requires 25 percent less time than a regular blend. According to Aratani’s work papers, he received the 25 percent estimate from Bill Campbell, an SOS employee who held a variety of positions during the M206 contract, including manufacturing manager and production manager. At trial, however, Campbell testified that the ratio of time between a cross-blend production and a regular blend production is 75 to 80 percent. Winn testified that approximately the same number of people and hours are required in cross-blending as in regular blending.
The testimony supports a finding that a more accurate estimate for cross-blending is 75 percent of the time required for regular blending. The difficulty the court has with this aspect of damages, however, is that plaintiff makes no attempt to account for the number of hours of cross-blending which, had the TDP not been defective, it would have performed. Several of defendant’s witnesses testified that cross-blending is a common practice in the pyrotechnic industry. Wayne Sommerford, the program manager for infrared flares at Thiokol, stated that Thiokol always presses sample pellets from new mixes, tests them, and cross-blends with mixes of opposite characteristics if necessary. He stated that, because of the money invested in preparing a mix, he believed most manufacturers would have to engage in a similar procedure. Farnell also described cross-blending and stated that, because pyrotechnic results are not necessarily reproducible, one would have to expect such a process. Lastly, Richard Ames testified that in his experience cross-blending is a common practice among flare manufacturers. While the 25 percent figure used initially may not be wholly accurate, the failure to account for some cross-blending is no less an error. Plaintiff and defendant are in agreement to the minimum amount of $28,222.33, which includes overhead and G & A. With profit, plaintiff will be awarded damages in the amount of $31,044.56.
*384d. Screening
Plaintiff asserts that the cost of implementing and performing the screening process was $115,386, which includes overhead and G & A. Assuming liability, defendant concedes $71,405.70 is established based on $24,711 in direct labor, as revealed in SOS’s records, plus overhead and G & A. The difference between plaintiff’s and defendant’s calculations arises because defendant does not include any hours for Test Technicians, Test Engineers, or consultant costs.
Plaintiff claims 422 hours were spent in the engineering effort expended on screening. No records support this figure. Plaintiff points to the testimony of Reynolds. He stated that 10 to 11 weeks was a reasonable estimate for the amount of time he spent on screening between November 1984 and April 1985. Assuming he worked full time (eight hours per day), 10 weeks would be 400 hours. Because Reynolds personally confirmed this claim, and because it seems conservative in view of the six-month period during which he worked, 400 hours are allowed. With over head and G & A, the recovery for Reynolds is $27,-061.05.
The remaining damages claimed for screening costs arise from hours spent by Test Technicians and Material Handlers in 1989. There is no testimony on the hours used in calculating damages other than from Aratani. He stated that they were estimates, based on conversations with Stan Barlog, but gave no details of how he arrived at these hours. The court, for lack of proof, cannot accept them as a solid basis for assessing damages.
The court rejects the $2,556 allegedly paid to consultants. The only testimony on the consultant costs was from Aratani. He testified that $556 was paid to Tracor as a consultant fee and that $2,000 was paid to another consultant to verify the test equipment that was used in the screening process. There was no documentary support. Plaintiff has not properly established those fees.
In sum, the court concludes that plaintiff is entitled to damages in the amount which defendant concedes has been established, plus Reynolds’ time. Plaintiff is entitled to $98,466.15, plus profit, for a total amount of $108,313.42.
e. Rejected flares
Plaintiff urges that it is entitled to be paid for all flares which were rejected for static failures and which were not accepted after retest because these flares conformed to all TDP design requirements. Plaintiff’s position, though not explicit, is that all static output failures were due to the defect in the design portion of the TDP — namely, the use of ground magnesium.
Plaintiff’s calculation of damages is simple. It states that of the 248,000 flares originally rejected 207,648 flares remained rejected after retest. Using the option price of $7.82 per flare, plaintiff seeks $1,623,807 in damages.
Defendant’s primary dispute is with the number of flares for which plaintiff should be credited. In addition, it applies the option price to 28.41 percent of the flares, the proportion which was purchased at the option price.
Plaintiff’s figure for rejected flares comes from Aratani. He relied on information in a Government technical audit report performed by Zack Brown, who Aratani stated was with the Van Nuys DCAS office. Brown’s work papers recorded that 178,000 flares were rejected before the environmental shutdown in 1985. Aratani also consulted Barlog, who identified seven additional rejected lots of 10,000 flares, bringing the total rejected up to 248,000. Aratani reduced this number by 12,500 for flares which Barlog stated were still in hand, and not actually rejected. This correction was not explained in any further detail. Aratani also reduced his number by 27,852, the number of flares which Brown indicated were accepted on retest.
Defendant begins with the total number of flares produced on the contract to determine the number of rejected flares. According to the work papers of Kyle Owens, plaintiff’s accounting expert on the delay and inefficiency claim (Count VI), a total of 1,359,728 flares were produced through Oc*385tober 29, 1989. Defendant corrects this downward for a three percent scrap factor, which is also noted in Owens’ work papers, leaving 1,318,936 available flares. According to Owens’ work papers, the number of flares shipped is 1,162,400. Defendant therefore concludes that the total number of flares not shipped is 156,536.
Defendant refines this figure as follows. It first determines the number of flares used in testing. Assuming 136 lots were produced,25 and 180 flares were tested per lot (90 static and 90 dynamic), defendant arrives at 24,480 for the number of flares which were expended for dynamic and static testing. This leaves 132,056 unshipped flares. Defendant subtracts from this figure 5,800, the number of flares which, according to Aratani, were rejected for dynamic testing,26 and 2,600, the number of flares which according to Aratani were used for engineering studies. This leaves, according to defendant, 123,656 rejected flares. Applying the original price of $7.18 to 71.59 percent of these and $7.82 to 28.41 percent, defendant’s figure for total allowable damages is $910,335.45.
There is no question that flares were improperly rejected. The court has misgivings about the rejected flare figures offered by both parties. Aratani’s figure was based on a Government audit not before the court and on a conversation with Barlog. Aratani did not state whether Bar-log revealed the source of his data. In addition, Aratani’s number for rejected flares exceeds the number of flares which, according to Owens’ records, were produced but not shipped. There was no explanation for this inconsistency. Defendant’s figure is based on estimates both as to the number of lots and the number of flares tested per lot. It is also unclear whether the reduction of 5,800 flares reflects the same flares that were tested or, rather, were from lots which were rejected based on flares which were tested. Accordingly, it will use the average between the two numbers, 165,652, as a way of minimizing potential error. The court adopts defendant’s methodology of assessing damages based on weighted percentages of flares originally purchased at the lower price and those falling under the option price. Plaintiff is entitled to $1,219,500.80. No adjustments are necessary for indirect costs or profit.
f. Engineering effort
Plaintiff seeks to recover $158,262 allegedly expended in the engineering effort involved in solving blend problems. Again, plaintiff provided no documentary evidence of labor hours. Aratani relied on estimates provided by Cabaret and Reynolds. Reynolds testified that he spent a total of four to six months on the testing effort. A total engineer’s time of 314 hours is claimed. This is less than half of the six months to which Reynolds referred. Cabaret testified that test technician time “would probably be two to three times as much.” The testimonial evidence thus presents a range from 1360 to 3096 hours for test technician time. For no apparent reason 156 hours are claimed.
Because Reynolds presented first hand evidence as to engineering time, and because that amount is significantly more than the hours claimed, the claim is allowed. The evidence as to test technician effort is wholly unsatisfactory however. Evidence as to the Project Engineer, other engineers, chemists and material handlers was equally vague. Cabaret was asked for his “feel” for how many hours were involved and gave a "fair estimate.” The amount claimed did not “sound unreasonable” to him. If this testimony were being used to support conclusions from raw data, it would have some weight. But it is not. Plaintiff is offering these guesses as the only evidence. They are simply inadequate.
*386Damages are awarded based on 314 hours of Test Engineer time, at a labor rate of $14.90. After inclusion of overhead and G & A, and uncontested amounts for materials, the recovery on this item is $25,-195.26, plus profit, for a total of $27,714.78.
g. Lot acceptance tests
Plaintiff produced more lots than called for by the contract due to difficulties generated by the qualification of ground magnesium. These lots had to be dynamically and statically tested. In calculating additional costs attributable to tests on those lots, Aratani used 19 as the figure of additional lots. It is not clear where he derived that figure. Plaintiffs witness Owens calculated total flare production at approximately 1,360,000, or about 136 lots. The amended contract called, in effect, for 122 lots. The difference between these two figures, 14, is the basis for defendant’s calculations, and the court adopts it for both static and dynamic testing. Defendant’s labor calculations are utilized, with the exception that hours for quality control/safety are increased from four to eight per test day. Total labor costs are $4,139.84. Overhead and G & A bring the labor total to $21,598.90. Rental of the Boron testing facility for seven days adds $11,900. Total, including profit, is $36,-848.79.
h. Intermediate mix, labor and material
The parties stipulate that plaintiff increased the quantity of intermediate mix in the pellet grooves in order to boost static output. The court finds that this was occasioned by the uncertainties and poor performance resulting from qualification of ground magnesium. In terms of material costs, defendant’s figure for the number of flares produced with increased intermediate mix is more credible. It is drawn from Owens’ trial exhibit. Aratani’s calculations are less direct, and he was unable to explain the source of some of his figures. Based on a total of 820,985 flares, plaintiff is entitled to reimbursement for materials costs, plus profit, of $105,480.15.
Plaintiff also seeks to recover $119,622 in labor costs. This represents an asserted 3,900 hours in additional labor. The provenance of that figure is shrouded in mystery, however. Aratani testified that he used a figure of .0046 hours as the bid time for applying intermediate mix. That is apparently based on his understanding that it took two passes with a mix-filled syringe to fill the groove. That figure is questionable because it is based on the bid and because it presumes that Aratani was able to isolate from the bid the precise effort involved in filling and making one pass with a syringe, and that it took exactly twice as much time to fill the groove. Setting aside the issue of how that figure was derived, it is also not clear how Aratani uses it. During trial he indicated that he multiplied .0046 “against the units affected.” He testified, presumably erroneously, that the product was the “added blended hours,” against which he applied labor rates. It would appear to represent instead the amount to be deducted from actual time spent. Finally, it simply appears incredible that it would take 486 extra man-days to fill as opposed to half-fill the grooves. For lack of any clear proof on additional labor, that element of the claim is denied.
i. Second first article tests
Plaintiff seeks to recover for additional first article costs associated with tests conducted after the long plant shutdown due to environmental problems. Plaintiff contends that, but for static testing problems, the shutdown would have occurred after completion of contract performance, and there would have been no necessity for a new first article. There are two problems with that analysis. First, the environmental shutdown is a supervening event for which regulatory agencies blamed SOS, and the court will not undertake a collateral review of the validity of their actions. Second, as discussed in connection with Count VI, there were too many variables in the production process to say with any confidence that plaintiff would have completed before the environmental shutdown. No costs are awarded.
E. Count V — Breach of Implied Warranty of Specification Related to Velocity Testing.
This is the third count alleging a breach of warranty of specification. Plaintiff con*387tends that given the tolerance range of the various M206 components and the allowable tolerance for output of the M796 impulse cartridge which ejects the flare, flare velocity will randomly fall outside the MIL-F-63107A(AR) performance requirements. Plaintiff argues that velocity problems are therefore inherent in the TDP. At trial, plaintiff did not rely exclusively on the occurrence of velocity problems and compliance with the TDP, but also offered evidence in an effort to isolate the source of its difficulties.
1. General factual background.
Velocity testing standards for the M206 flare are set forth in paragraphs 3.5.2 and 4.4.2 of MIL-F-63107A(AR). Paragraph 3.5.2 requires that M206 flares be ejected from their case with an initial velocity between 75 and 165 feet per second (“fps”). For lot sizes between 1,201 and 3,200 units, Paragraph 4.4.2 allows a maximum of five velocity test samples to fall outside the required velocity range. For lot sizes of between 3,201 and 10,000 units, a maximum of six units may fall outside the range specified in Paragraph 3.5.2.
Plaintiff submitted flares for first article velocity testing in November 1982. As with all SOS flares, the squibs used in those flares were furnished by the Government. The squib is the primary source of ejection velocity. Specifically, the impulse cartridges were from CFM (Customer Furnished Material) Lot 001 M796. The flare sample failed the velocity requirements. Later, in January 1983, plaintiff submitted a second first article test sample using CFM Lot 002 M796 impulse cartridges. These flares passed first article velocity requirements.
The first eight lots of production flares, Lots 001-008 of Interfix 1, were submitted by SOS beginning in late May and ending on June 10, 1983. These flares were made with CFM Lot 002 impulse cartridges. Although Lots 001 and 002 initially failed velocity testing on June 3, 1983, these two lots were retested on June 23-24,1983, and passed. Lots 003-008 all passed velocity testing.
With Lot 009 of Interfix 001, however, plaintiff began to experience velocity problems. Plaintiff began to use Impulse Cartridges from CFM Lot 001 in producing its flares. Lots 009-029 of Interfix 001 were submitted for velocity testing during the period June 10-September 30 of 1983. Of these, Lots 009, 010, 012, 013, 015, 016, 026, 027, and 028 failed velocity testing. Because the inspection forms are not in evidence for Lots 009, 010, 012, and 013, the court is unable to determine whether these lots failed for high or low velocity or both. However, the inspection forms beginning with Lot 014 are in evidence and indicate the nature of the velocity problems experienced. Lot 016 failed because seven flares exhibited velocity over 165 fps. In Lot 026 one flare exhibited velocity over 165 fps, four exhibited velocity under 75 fps, and two exhibited velocity under 60 fps. Lot 027 had six flares with velocity below 75 fps. In Lot 028, six flares exhibited velocity under 75 fps. The inspection forms, therefore, indicate a mixture of high and low velocity problems.
Plaintiff switched back to the CFM Lot 002 impulse cartridges for Lots 030 and 031 to determine whether the impulse cartridges from CFM Lot 001 were the cause of the failures which began with Lot 009. Lots 030 and 031 were tested in September 1983, and both passed velocity testing. However, Lot 030 had one high velocity defect, and Lot 031 had one high and one low velocity defect.
With Lot 032, plaintiff again used impulse cartridges from CFM Interfix 001. Lot 032 failed velocity testing. Plaintiff prepared Lot 033 of Interfix 001 using impulse cartridges both from CFM Lots 001 and 002.27 Although Lot 033 passed velocity testing, three of the flares failed one of the velocity testing requirements — one flare failed for high velocity, two for low velocity. The three flares that failed used CFM Lot 001 impulse cartridges.
According to the inspection forms, the only other lots to fail velocity testing were Lots 35 and 46 of Interfix 001, and Lots *388001 and 002 of Interfix 002. Lot 035 was tested in late September 1983. It had one flare which exhibited velocity below 75 fps and one which exhibited velocity below 46 fps. There is no indication in the record what impulse cartridges were used with Lot 035. Lot 046 was tested on October 21,1983. It failed because six flares exhibited velocity greater than 165 fps. According to Reynolds, these flares were tested with impulse cartridges from CPM Lot 001. Lots 001 and 002 of Interfix 002 were tested in late October and early November, respectively. They also used impulse cartridges from CPM Lot 001. Flare Lot 001 had eight flares with velocity over 165 fps and one with velocity under 75 fps. Eighteen flares in Lot 002 exhibited velocities over 165 fps, and one flare exhibited velocity below 75 fps.
During the time SOS was experiencing velocity problems, it conducted a variety of investigations to determine the cause of the problems. In July 1983, plaintiff manufactured flares with first fire in the cross grooves at the ignition end of the flare pellet. This was done in response to defendant’s suggestion that low amounts of first fire and intermediate mix might be accounting for low velocity failures. Plaintiff, however, determined that reduced first fire had no effect on velocity.
Later in July plaintiff undertook tests to determine whether the tolerance ranges in the size of the flare were responsible for velocity failures. It manufactured two groups of flares from the same mixes, one group having a width and depth at the lower end of the TDP tolerance, the other having a width and depth at the upper end. According to plaintiff, the tests indicated that flares which were larger and heavier tended to achieve somewhat greater velocities than smaller flares. However, both groups had an average velocity which was well within the requirements of the TDP.
Sometime in July plaintiff conducted tests to study the effect of the number of obturator felts on M206 flare performance. Specifically, plaintiff investigated whether using one rather than the required three obturator felts would affect velocity. Plaintiff tested 20 flares, 10 with three obturator felts, and 10 with one felt. The felts used were .031 inches thick, as required by the TDP. The report on these tests indicated that there was no effect in the flare’s velocity performance based on the number of obturator felts used.
In August plaintiff conducted tests to investigate whether using shear pins at the non-ignition end of the M206 flares, as required by the TDP, had an effect on velocity. Plaintiff manufactured and tested flares both with and without the shear pins and concluded that flares with shear pins achieved a higher rate of speed and performed more consistently in velocity tests than flares without shear pins. In its report to defendant, however, plaintiff stated that “no effect in velocity on performance of the flare based upon shear pins mounted in the end cap.” There is no indication in the record that plaintiff took any action to change its manufacturing of flares as a result of these investigations.
In September plaintiff conducted tests to determine whether the velocity test equipment and procedures it was using (procedures which had been approved by defendant) were adversely affecting performance readings by prematurely sensing the flare. Plaintiff mounted a 10 inch by four inch plate one-half inch below the infrared detector to prevent premature sensing. Plaintiff tested 10 flares using the shielding plate. Eight flares exhibited acceptable velocity performance. One flare exhibited unacceptably high velocity (166.6 fps) and one unacceptably low velocity (74.0 fps). Although plaintiff determined that the tests were inconclusive, the shielding plate became a permanent part of SOS’s velocity test procedures.
Sometime before October 1983, defendant furnished plaintiff with 300 impulse cartridges from CFM Lot 006, suggesting that SOS compare flare velocity performance of flares using CFM Lot 006 impulse cartridges with flares using CFM Lot 001 impulse cartridges. This investigation, conducted during October, was inconclusive because the flares using the CFM Lot 006 impulse cartridges did not perform as well *389as those which used CFM Lot 001 impulse cartridges.
SOS also began testing all flare pellets for conformance with dimensional requirements. An internal government memo of November 4, 1983 suggests that one possible explanation of low velocity failures is “snug-fitting” flares. That same memo also suggests that every lot of squibs will have an “inherent small quantity of ‘mavericks.’ ”
In October plaintiff focused again on whether the obturator felts used in the flare were affecting performance. This time, in addition to testing the effect of the number of obturator felts on performance, plaintiff tested the effect of felt thickness. Plaintiff tested three sets of flares. In the first set, plaintiff used three obturator felts which were .038 inches thick. In the second set, it used one obturator felt .024 inches thick. In the third set, it used three obturator felts .024 inches thick. The felts specified in the TDP are .031 inches thick. Plaintiff investigated the amount of pressure which would be required to eject the flare from its casing using each of the three configurations. Plaintiff determined that the .038 inch thick obturators should not be used in the flares. It also decided to perform further tests on the single obturator variations.
Plaintiff performed tests on the single obturator version beginning on October 20, 1983. The purpose of the tests was to compare thin and thick obturators in the single obturator design. Thirty flares were prepared with a single thin felt, and 26 with the thicker felts. Although plaintiff determined that the results were inconclusive regarding the effect of felt thickness on performance when the single obturator design was used, it concluded that low velocity problems were virtually eliminated when one, instead of three, obturator felts were used.
In October 1983, plaintiff submitted Engineering Change Proposal (“ECP”) 7040-E-001, Control No. A3Y3026, suggesting reduction of the number of obturator felts used in the M206 flare from three to one, and reduction of obturator thickness from .031 to .023 inches. Although defendant disapproved the ECP, it did permit plaintiff to incorporate the changes from the ECP into its production process by Deviation No. A3Y7031, dated October 11, 1983.
Cabaret testified that after SOS incorporated this deviation, it experienced no further velocity test failures. The record does not entirely support this statement. Lot 001 and 002 of Interfix 002 were made with only one obturator felt, and both lots failed velocity testing, on October 28 and November 2, respectively. It is true, however, that according to the inspection forms, these were the last velocity failures experienced by plaintiff.
Defendant’s position is that plaintiff’s initial failures were due to faulty velocity testing equipment. Defendant points out, and plaintiff does not dispute, that in November 1983 plaintiff analyzed its testing equipment to determine whether faulty readings from equipment were the cause of inconsistent velocity performance. Plaintiff compared readings from its Hewlett Packard HP-5233L Counter with the readings obtained using a Nicolet digital oscilloscope. These tests indicated that plaintiff’s HP-5233L gave inconsistent and inaccurate results, and SOS’s test procedures were modified to use the Nicolet oscilloscope. It is undisputed that all of the lots which failed velocity testing were tested before SOS changed its testing procedures.
There is evidence that SOS’s equipment was not in fact faulty. SOS conducted a test on its HP-5233L in which a 2,000 frame per second high speed camera was used to determine whether SOS’s equipment was triggering properly. That test suggested that SOS’s equipment was functioning correctly. SOS also sent the HP-5233L to a calibration lab for testing, and it found no problems. SOS later tested the HP-5233L against the Nicolet for a second time, and the two correlated. Reynolds testified that the reason why the first comparison test with the Nicolet suggested that the HP-5233L was functioning improperly is “still a mystery.” According to Reynolds, DCAS later allowed SOS to use *390the HP-5233L because it took less time to record information gathered by the Nicolet.
Nevertheless, by letter dated January 5, 1984, SOS requested a retest of Lots 009, 010, 012, 013, 015, 016, 026, 027, 028, 032, 035, 046 of Interfix 001 and Lots 001 and 002 of Interfix 002, stating that the request for a retest was made “on the basis that Space Ordnance System’s response equipment set-up for obtaining the dynamic velocity was faulty and did not provide a true result of the time event for use in the calculation of velocity for the dispersed flare.” In addition, the letter stated: “[T]he analysis which resulted in the submittal of Waivers 7040-W-00-1 thru -0011 which attributed the dynamic velocity variations to the variable output of the M796 Impulse Cartridge (“CFE”) was incorrect and Space Ordnance Systems requests that these waivers be rescinded.” Reynolds testified that SOS blamed velocity failures on test equipment so that they could obtain permission for a retest. He stated that had SOS blamed velocity problems on the impulse cartridges or the TDP, SOS would have had to submit the failed lots on waivers. If the test equipment were blamed, however, SOS could “normally get right back into retest after going back through a short investigation period.”
SOS was granted permission to retest the lots which had failed velocity testing. The retesting was conducted in March 1984, and all the lots which were retested passed.
At trial, plaintiff presented the testimony of Arne H. Wiederman of ITT Research Institute of Chicago to establish that velocity test failures were inherent in the TDP, specifically, that the defect was in the impulse cartridges. Wiederman has been an employee of ITT Research Institute for a total of 32 years, though not continuously. He holds a Master of Science in mechanical engineering, and his experience includes solid mechanics, transient gas dynamics, and thermodynamics.
Wiederman attempted to isolate the variables that could affect velocity for flares which were made in conformance with the TDP. Using the military specification for the M796 impulse cartridge and various drawings from the TDP, Wiederman determined that there were eight factors within the TDP which could affect velocity: the performance of the impulse cartridge; the initial volume of air at the base of the flare assembly; the weight of the flare assembly; the resistance force of the two plastic pins (shear pins) that hold the end cap in place; the flare canister length; the flare canister area; the leakage area of the gas that can leave the system during the launching process; and the friction caused by obturator felts (both the three and one felt variations).
Flares made in strict conformance with the TDP would exhibit some flare to flare variation in the above factors. With the exception of the impulse cartridge, the variation is due to tolerance ranges. For example, flare canister length must be 7.825 inches long, with a tolerance of —.020 inches. The effect of canister length on velocity would therefore vary from flare to flare depending on where the length fell within the tolerance range. With the impulse cartridge, Wiederman believed that there would be flare to flare variation in performance because, in his opinion, when impulse cartridges are fired, sometimes all of the material burns, and sometimes only some of it burns, depending on a variety of chemical and physical influences. He believed this would occur even when the impulse cartridge was made in conformance with specifications. To study the effect of variation in impulse cartridge performance, Wiederman constructed numerous possible “burn paths” — relationships between the amount of propellant burned and time. How a particular impulse cartridge performed would depend upon which burn path its propellant happened to follow.
Wiederman attempted to isolate which of the eight factors, if any, was dominant in determining the flare’s ultimate velocity. To examine the effect of a particular factor on velocity, Wiederman would hold seven of the variables constant within the tolerance range and vary the eighth factor. Wiederman studied these effects through modelling. Modelling, as Wiederman de*391scribed it, is “establishing] the physical effects that occur; dealpng] with the laws that govern those effects, and dealpng] with the interaction of various effects or parameters that are associated with particular phenomena that are involved.” These effects are evaluated mathematically or analytically, without actually firing any flares.
Wiederman’s calculations suggested that flare velocity could vary by as much as 75.4 percent due to variability in acceptable burn paths. The next largest percentage variation in flare velocity was 3.8 percent. This, according to Wiederman, would be caused by variation in the effect of friction when three obturators are used. Wiederman’s conclusion was that the dominant parameter causing variation in flare launch velocity was the response of the impulse cartridge. According to Wiederman, “It wasn’t even close.”
Plaintiff also pointed to the inspection forms to demonstrate that impulse cartridges caused velocity problems. It notes that the first eight lots of flares, which used impulse cartridges from CFM Lot 002, all passed velocity testing, but that when SOS switched from CFM Lot 002 to CFM Lot 001 impulse cartridges beginning with flare Lot 009, it experienced a string of failures. Plaintiff also introduced evidence to suggest that other contractors had difficulty with the M796 impulse cartridge. Plaintiff did not attempt to explain, however, why the flares all passed when they were retested using CFM Lot 001 impulse cartridges.
2. Discussion.
Plaintiff offers two possible design defects — the obturator felts and the Government-furnished M796 impulse cartridges. Regarding the obturator felts, SOS experienced no more velocity problems after it switched from three to one obturators. Lots 001 and 002 of Interfix 002, however, both of which utilized only one obturator felt, failed velocity testing. Also, the obturator was meant to eliminate only one type of velocity problem — low velocity — by reducing friction. SOS had failed velocity testing because flares exhibited velocities which were both too high and too low to meet specification.28
As to the squibs, there is evidence that they were an appropriate suspect, and government personnel apparently shared those suspicions. However, there is also no question that evidence implicating the squibs is equivocal. The court notes that velocity problems ended abruptly and permanently on November 2, 1983. If the M796 impulse cartridge behaves as Wiederman suggests, it is curious that SOS only experienced problems initially. Plaintiff also attempted to demonstrate that the impulse cartridges were defective by pointing to its switch from CFM Lot 002 to CFM Lot 001 impulse cartridges, beginning with Lot 009. It notes that Lot 009 marked the beginning of velocity problems. All the lots which failed were retested, however, using impulse cartridges from CFM Lot 001, the same lot which had been used when the flares were tested initially. All these lots passed.
The squib is the logical first place to look if velocity is erratic, however. It exerts by far the greatest influence on velocity. Other factors, such as ignition gas, inertia, and friction, are of secondary importance. Reynolds testified, and the court believes accurately, that the contractor’s only controls over velocity are the pellet dimension and the x-ray check to make certain the flare is properly assembled and has no defects. Although the defendant intimates that the plaintiff may have affected velocity by poor assembling or improperly sized or weighted pellets, there is no evidence to support such a conclusion. Unlike the defects alleged in Counts III and IV, there is no serious question that plaintiff followed the TDP, that it had no relevant flexibility, and that there were failures.
Here, there is only one suggestion by defendant that anything within plaintiff’s discretion could have caused the velocity *392failures. That has to do with the possibility that test equipment gave false readings. This is certainly a suggestion that SOS was willing to adopt when it was convenient to do so. The court does not treat that as an admission, however. Cabaret and Reynolds explained the reason why SOS took that position, and it was never rebutted. More importantly, there is no real evidence that the HP-5233L was defective. Subsequent tests suggested the contrary.
In sum, the court is confronted with a phenomenon that has no clear cause. What is apparent, however, is that plaintiff followed the specifications and that there are no reasonable alternative causes for failure attributable to plaintiff. Under these circumstances the general proposition applies, that the Government warrants a satisfactory result if design specifications are followed. It is no answer to suggest that the relatively small number of failures should be written off to chance. Those failures precipitated a large-scale, non-frivolous investigation by SOS for which it should be compensated.
One of the elements of the plaintiff’s claim for damages, however, is denied. SOS asserts in its proposed findings of fact that the elimination of velocity failures is attributable to pellet rejection at the upper and lower limits of the dimensional tolerance. The plain suggestion is that the standard for acceptability was narrowed as an aid to solving the velocity problem. The evidence cited by SOS does not support such a finding. What plaintiff established is that flares were run through “go, no-go” gauges to assure compliance with contract dimensional requirements. There was no proof that tolerances were narrowed, or, if that were the case, that it helped test results. To the extent that SOS ran all flares through go, no-go gauges to assure consistency with the contract, no recovery is warranted.
Damages are awarded as follows. The only direct evidence of additional labor was the testimony of Reynolds concerning the time he invested as Test Engineer. He stated that it took three to four months of his time. The lack of payroll evidence as to other workers is fatal considering the lack of anything other than Reynolds’ concurrence in the claim estimates. Based on 322 hours of test engineering time, supervision and support, materials and overhead, damages total $17,877.68. Plaintiff’s recovery for flares rejected is included in damages under Count IV, supra. Plaintiff is also entitled to the expense of hiring Spin Physics, $13,000. With G & A, the total for Spin Physics is $15,379. Total recovery on this count, including profit, is $36,582.35.
P. Count VI — Delay and Inefficiencies Associated With the Defective Specifications.
1. Factual background.
This is a claim for damages caused by delay and inefficiency due to SOS’s inability to consistently produce acceptable flare blends. Therefore, this claim flows directly from (and is dependent upon), success in Count IV. According to SOS, its bid price was premised upon its being able to maintain a full production force throughout the term of its contract. SOS estimated that with a full production force it could produce approximately 27,500 flares per week. SOS alleges that due to the unanticipated need for cross-blending, blend testing, and outright rejection of blends made in accordance with the TDP, its production line was forced to remain idle. It is the inability to maintain a full production line which allegedly resulted in delays and inefficiencies. Plaintiff claims a total of $3,124,511 in this count.
Kyle Owens is a principal and director of the government contract consulting practice in the accounting firm of Arthur Andersen. He testified for plaintiff regarding calculation of the damages alleged in this count. Because of the intricate nature of the calculations, Owens’ direct testimony was supplied to the court in written form.
There are two elements of damages — delay and inefficiency. The inefficiency claim, reduced to its essence, is that if the work schedule of SOS’s production line had not been so erratic (because of the erratic supply of acceptable blends), plaintiff would have been able to produce flares *393more quickly, and thus labor costs would not have been as high as they were.
In the delay claim, plaintiff asserts that if the production line could have produced flares more quickly, it would have finished the contract in less time, and the labor hours expended on the contract would not therefore have been affected by inflation. Owens' task was to compare labor rates for the period during which the contract theoretically could have been performed with the labor rates for the period during which the contract was actually performed.
For the inefficiency claim, Owens performed what he described as a “worked versus earned analysis.”29 Owens testified that this is a standard type analysis for manufacturing work, particularly in the defense industry. A work versus earned analysis compares actual efficiency, expressed in terms of hours needed to produce a single unit, with a potential efficiency. According to Owens, a contractor will perform initially at below 100 percent (of the hypothetical) efficiency, but will eventually perform at over 100 percent. The average therefore should be roughly 100 percent efficiency.
Owens turned to SOS production charts and daily production records to determine the number of hours SOS employees spent on the various manufacturing steps for the M206 flare. The production records show the number of flares processed through each manufacturing operation on a daily basis. Owens, however, only had production records for March 1983 through December 1984. All the concrete data in this count flows from records for this period, which account for approximately 75 percent of production.
The potential efficiency which Owens used for comparison was reflected in SOS’s bid. The bid, however, was not divided into the same operations as were production records.30 According to Owens, he reviewed with Stan Barlog each manufacturing step described in the bid and, based on Barlog’s knowledge of SOS manufacturing operations, determined how many hours SOS had budgeted for each manufacturing operation described in the production records.
Owens added the number of hours budgeted in the bid for all the steps associated with the production of flares and divided this number by the number of flares expected to be produced on the contract (less a three percent “scrap factor”) to calculate what he termed “budget earned” for each flare. For example, if SOS had budgeted one hour per flare and if SOS produced 10 flares in a day, it would have a budget earned of 10 hours for that day. Note that budget earned is simply the amount of time it should have taken according to the standard measure of efficiency (the bid) to produce a given flare. To determine SOS’s efficiency on that day, the actual number of hours required to produce these flares is compared with the budget earned. If, for example, production records indicated that it took 20 hours to produce the 10 flares produced on the day in question, SOS’s efficiency would be 50 percent for that day.
Owens analysis of efficiency was on a weekly basis over the period from March 1983 to December 1984.31 He testified that *394efficiency went up and down dramatically from week to week and averaged 54.1 percent over that period. According to Owens, such experience is unusual because “one would normally expect the efficiencies on a high volume contract to be relatively stable.” Owens testified that he spoke with “a number of people” at SOS regarding the reason for the seemingly random efficiency experienced and learned from them that the availability of acceptable blended material paced the production line, and that if material was available, people could work efficiently.
Owens could not study the effect of availability of blend material on efficiency directly because he did not have empirical data on availability. Owens testified, however, that according to Stan Barlog and Bill Campbell, blended material was put into production as soon as it was available. Owens therefore treated information on flares pressed, which he did have, as indicative of the availability of an acceptable blend on that day.
The result of Owens’ study is reproduced in the chart below, copied from Tab B, Page 1 of Px 263:
NUMBER OF FLARES PRESSED AVERAGE EARNED BUDGET EFFICIENCY EARNED BUDGET EFFICIENCY RANGE NUMBER OF WEEKS REACHED PRODUCTION LEVEL
0 - 2,499 22.4% 15.3 32.4% 4
2.500 - 4,999 38.4% 31.7 47.3% 6
5,000 - 7,499 42.5% 30.6 46.7% 9
7.500 - 9,999 53.1% 36.2 62.4% 9
10.000 - 12,499 52.7% 34.9 63.2% 15
12.500 - 14,999 60.3% 49.8 68.3% 6
15.000 - 17,499 60.0% 47.9 65.3% 11
17.500 - 19,999 67.8% 51.6 88.8% 9
20.000 - 22,499 71.4% 56.8 92.7% 8
22.500 - 24,999 80.6% 80.6 80.6% 1
25.000 - 26,634 76.1% 69.8 86.9% 6
84
Owens interpreted the results of this study as demonstrating that as the number of flares pressed increases, the average efficiency increases, indicating a positive relationship between flares pressed (and therefore availability of blend) and efficiency.
Owens’ study suggested that when SOS pressed 25,000-26,634 flares per week, it could operate at an efficiency level of 76.1 percent. Owens therefore studied the cost impact of producing flares at a rate of 54.1 percent, rather than 76.1 percent. To perform this calculation, Owens took the total hours expended by SOS’s Department 55 (which handled production) and subtracted from these the hours claimed in Counts I-V. He arrived at a figure of 338,897 hours total expended. From SOS’s bid, Owens had previously determined that SOS budgeted .135 hours per flare. Multiplying this by the total number of flares produced gives the total budget earned of 183,563. Note that 183,563 divided by 338,897 is 54.1, SOS’s actual efficiency. If SOS had been operating at 76.1 percent, which Owens is assuming would have been possible had SOS been able to produce enough acceptable blends to press 25,000-26,634 flares per week, the total hours SOS would have expended would have been 241,214 hours (183,563/241,214 = 76.1 percent efficiency).
Subtracting 241,214 hours from 338,897 hours gives a total of 97,683 extra hours spent as a result of operating at 54.1 percent rather than 76.1 percent. A similar calculation was performed for support labor (as opposed to direct labor). According *395to these calculations, 10,469 excess support labor hours were expended.
To calculate the total excess labor costs, Owens broke down the total excess hours by year. He did this so that the actual labor rates experienced in each year could be applied to the excess labor in that year. Based on these calculations, the labor impact was $668,598. The total claim for reduced efficiency, however, is $2,685,-228.32 This figure includes the increased labor burden (or overhead) and increased general and administrative expenses,33 and 10 percent for profit.
Owens next turned to the related delay claim. As discussed earlier, this claim attempts to identify the increased costs due to inflation associated with performing the contract over a time period longer than anticipated. Specifically, the delay claim attempts to calculate increased costs in three areas: labor, overhead, and G & A expenses. To calculate increased labor costs, Owens began with the total number of labor hours incurred on the contract.34 (These hours were incurred from October 1982 to October 1989.) He subtracted the hours claimed in Counts I-V and those claimed for the inefficiency portion of Count VI. The figure he arrived at, 269,-503 hours, thus represents the total hours for which no claim is made. SOS acknowledges it was responsible under the contract for these hours. What the delay claim attempts to do is allocate the cost of incurring these hours over a longer period than anticipated in SOS’s bid.
Owens’ next step was to determine when these hours actually would have been incurred. To do this, Owens looked at Department 55 production records for March 1983 to December 1984. They revealed that the average hours per month expended on Department 55 were 13,938. Taking from his calculations on the inefficiency claim the number of hours which should have been expended on the contract for production, 241,214, and dividing this by 13,938, Owens arrived at 17.3 months as the amount of time it should have taken to perform the contract.35 Based on 17.3 months as the actual time the contract should have taken to perform, and assuming the contract would have begun in August 1982, as anticipated in SOS’s bid, Owens determined that production labor would have been 111,507 hours in fiscal year (FY) 1983 (only eight months of FY 1983 were spent in production) and 129,707 hours in fiscal 1984 (only 9.3 months of which would have been spent in production), the total being 241,214. Support labor would have been 12,879 hours in fiscal year 1983 and 14,981 in fiscal year 1984. In total, therefore, Owens determined that of the 269,074 unclaimed hours, 124,386 would have been spent in fiscal year 1983 and 144,688 in fiscal year 1984.
To determine what the total labor costs would have been, Owens had to first determine what average labor costs were for the fiscal years 1983 and 1984. He did this for 1984 by referring to SOS’s records. For 1983, Owens utilized the actual average labor rate for a manufacturing job which took place at Mint Canyon during fiscal *396year 1983 because “the actual FY 1983 average labor rate for [the SOS job] reflects a higher than normal mix of support labor to Department 55 labor which results in an actual average labor rate which is much higher than what would have been incurred during normal production.” The labor rate used for FY 1983 was $5.52/hr. and for 1984 it was $5.84/hr. The average for these two years is $5.69.
The weighted36 average labor rate over the entire term of the contract was $6.34. The difference between the average for the two years which SOS allegedly should have been in production and the average labor rate over the years SOS actually was in production is $0.65. This figure, multiplied by the total of 269,503 labor hours, gives a figure of $175,177 for the impact of delay on labor costs.
To calculate the increased labor burden (labor overhead) rates experienced due to delay, Owens went to a chart labelled “Indirect Rate Recap” prepared by Price Waterhouse during its earlier study. That chart gives the labor overhead rates during each year of the contract, as calculated by Price Waterhouse. The average labor overhead according to that chart was 185.1 percent. Owens estimated what labor overhead rates would have been had SOS not experienced delays by calculating labor overhead rates using as the cost of labor $5.69/hr., rather than $6.34, as calculated above.37 If the cost of the total labor incurred on the contract had been $5.69, the average overhead rate, according to Owens, would have been 135.52 percent.38 The difference between the 185.1 percent experienced and what overhead rates supposedly would have been is 49.6 percent.
To translate this figure into a dollar figure, Owens multiplied 49.6 percent by the total labor cost impacted by the increased labor burden rate. The total labor cost impacted was obtained by taking the total labor cost for the contract, $2,983,930,39 and subtracting from it the excess labor costs identified in Counts I-V ($663,395) and the labor component of the inefficiency claim ($668,598) to arrive at $1,651,937. The cost of increased labor burden due to delay is therefore 49.6 percent of $1,651,-937, or $819,361.
The final expense allegedly affected by delay was the G & A expense. The first step in calculating the effect of delay was to determine the costs subject to the increased G & A expense. This was determined by adding the total materials, labor, and labor burden rates, as given in the Price Waterhouse “Indirect Rate Recap,” and adjusting these costs for claims against them made in Counts I-V and in Count VI. The total cost subject to G & A expenses was determined to be $10,573,795. Price Waterhouse determined a weighted average G & A for the contract of 18.3 percent. By reflecting in the computation of G & A the labor, overhead, and materials which SOS should have used if 76.1 percent efficiency had been achieved, Owens determined that the average G & A for March 1983 to December 1984 was 14.6 percent. Applying the difference between the two G & A rates, 3.7 percent, to the $10,573,795 in expenses subject to G & A, Owens arrived at a figure of $391,230 for cost of increased G & A expenses due to delay.
The total claimed due to delay is therefore $1,385,768. The total due to inefficiency is $2,685,228, bringing the total claimed in Count VI to $4,070,996. The Amended Complaint seeks $3,124,551. No explanation is provided for the difference.
2. Discussion.
Plaintiff must establish a causal connection between costs allegedly incurred and the actions of the Government and must provide a reasonable and ascertainable basis for the estimates of such costs. Cos*397mic Construction Co., ASBCA Nos. 24014, 24036, 88-2 BCA (CCH) ¶ 20,623 at 104,240; R.W. Contracting, Inc., ASBCA No. 24627, 84-2 BCA (CCH) ¶ 17,302 at 86,216; Fermont Division, Dynamics Corporation of America, ASBCA No. 15806, 75-1 BCA (CCH) ¶ 11,139 at 53,004. For the reasons which follow, the court concludes that plaintiff has failed to meet that burden,
a. Efficiency
The premise of Owens’ analysis is that the effect on efficiency of blend availability can be studied using the number of flares pressed as an index. In other words, the number of flares pressed is used as a direct measure of inefficiency introduced by the Government’s breach of warranty with respect to its specification of ground magnesium. It is critical to observe that by using the number of flares pressed as a measure, plaintiff has not only simplified its calculation of damages, but has assumed its conclusion of compensable inefficiency. This assumption is based on Owens’ discussions with Barlog and Campbell regarding SOS’s policy of pressing flares as soon as acceptable blends were produced. Without any real data on the impact of ground magnesium on efficiency, plaintiff merely assumes, without any empirical evidence, that there are one to one correlations between use of ground magnesium, flares pressed, and inefficiency. Owens does not support the corollary assumption that when flares were not pressed, this was only because blend material was not available. In using the number of flares pressed as the sole index for efficiency, Owens ignores the possibility that there may have been times when flares were not pressed even though acceptable blend was available. To establish his premise, Owens would have to demonstrate that availability of blend, and only availability of blend, paced the production line. Using the number of flares pressed as an index for availability of blend does not supply this proof; rather, it assumes what is to be proved. As discussed in connection with the related delay claim, there were numerous other potential sources of delay which cannot be attributed to the Government. See discussion supra at 399.
Even assuming that the number of flares pressed is an accurate index of blend availability, and thus efficiency, Owens does not support another critical assumption in his analysis — that SOS could consistently produce 25,000 to 26,634 flares per week if blend were available. Once Owens determined that 76.1 percent efficiency was achieved when 25,000 to 26,634 flares were pressed, he moved on to evaluate what would have happened had SOS performed at this level throughout the contract. Owens’ study in and of itself does not support a conclusion that sustained performance at that level was feasible.40 The study deters mined only that when in fact 25,000 to 26.634 flares were pressed, efficiency was 76.1 percent. While SOS’s bid projected 27,500 flares per week, the bid itself could have been inaccurate. Owens’ only testimony that the production level of 25,000 to 26.634 flares per week is a reasonable assumption was that "[w]e know that they actually did achieve that level of production for six weeks.”
Another weakness in Owens’ efficiency analysis is that he was forced to rely on incomplete data. Owens only had production records from May 1983 through December 1984. Nor did he have the actual hours SOS employees spent producing M206 flares. To estimate the number of labor hours, he compared SOS’s distribution of labor report with SOS’s job cost ledger for the M206 contract. Owens did not, however, explain the comparison he *398made, and the court has no way of evaluating the accuracy of the very critical determination of how many hours SOS employees spent in production.
To calculate budget earned, Owens again needed to compensate for incomplete data. Specifically, SOS’s bid was not broken down into the same production steps as were the production records. To compare production labor hours with labor hours bid for the various production steps, Owens reviewed with Barlog each manufacturing step on the original bid and reclassified the production step given in the bid into one of the steps listed on the production record. Owens did not give a specific example of how this correlation was determined. While the reasonableness of the hours estimated in the bid is not relevant in this context,41 it is important that actual hours are consistently compared to the same hypothetical hours for each task. The court has no basis to evaluate these comparisons.
An expert can draw upon a wide variety of reasonable sources in developing an analysis. But here, virtually none of the data is from empirical “first sources.” The bid was used to generate the hours earned per flare and the projected division of labor. The Price Waterhouse Report, which Owens drew from heavily, is only before the court in the form of a brief excerpt. The critical connection between non-availability of blends and production slowdowns is based on the subjective evaluation of others. There is no testimony of a single specific interruption or slowdown, and therefore nothing concrete in terms of length of interruption, what happened on the production line, who made decisions about whether to keep workers in place, whether other tasks could be performed, or how often interruptions occurred.
Owens’ analysis is impressive in terms of its logic, complexity, and comprehensiveness. But after following each step to his ultimate conclusion, the impression is left that what has been constructed is an elaborate accounting trompe I’oeil. While plaintiff's analysis of the effect of erratic blend supply on efficiency flows logically from its assumptions, the assumptions themselves are poorly supported and are not based on reliable empirical data. While the lack of acceptable blend may have affected production, the court has no confidence in Owens’ figures. He has not isolated with reasonable certainty the effect of erratic blend supplies on efficiency. See G.M. Shupe, Inc. v. United States, 5 Cl.Ct. 662, 719 (1984); Wunderlich Contracting, 173 Ct.Cl. at 199, 351 F.2d at 968.
b. Delay
The delay portion of Count VI is built upon the same information as that used in the inefficiency claim. The level of attenuation from original data is therefore great. There was no isolated period of delay. Owens had to estimate delay occurring primarily during periods when production was occurring. Owens first assumed that production would begin, as estimated in the bid, in August 1982. He stated that this assumption was reasonable because “[w]e understand from Joe Cabaret that the first article was ready to be tested only one week behind schedule.” To estimate when production should have ended, Owens determined that the average number of hours worked by the employees of Department 55 (the department which pressed flares) during the period April 1983 to December 1984 was 13,938 hours. Taking from the efficiency calculation the figure 241,214 for the total hours which would have been expended to complete the contract if SOS had operated at 76.1 percent efficiency and dividing it by 13,938 hours/month, Owens *399estimated that 17.3 months would have been required to complete the contract. Thus production, according to Owens, should have ended in December 1983.
Due to the nature of Owens’ delay analysis — i.e., there are not discrete, separate points of delay — it is virtually impossible to quantify delays resulting from erratic blend supply. As the Court of Claims taught in Commerce International, “[N]o matter how unreasonable the Government’s delay, there can be no recovery without proof that the delay caused material damage.” 167 Ct.Cl. at 542, 338 F.2d at 89. Here there is an even more fundamental lack of precision. Even though the difficulty in producing acceptable blends is attributable to the breach of warranty, it does not automatically follow that plaintiff should be compensated for the difference in production time between what it anticipated in its bid and what it actually experienced. SOS has to put on some concrete proof that it is entitled to anything other than the additional direct costs of producing those blends.
Owens did not explicitly account for any “learning curve,” which he testified all contractors experience. Instead of accounting for reduced efficiency at the start of SOS’s contract, Owens assumed that from the onset SOS would be at 76.1 percent efficiency and would continue at that level. He felt that 17.3 months was reasonable, however, because the bid estimated 14 to 15 months for production. Owens provided no independent support that the bid was reasonable, however.
Plaintiff has the burden of making a reasonable effort to differentiate between delays attributable to defendant and delays caused by extraneous circumstances. William F. Klingsmith, Inc. v. United States, 731 F.2d 805, 809 (Fed.Cir.1984); Wunderlich Contracting, 173 Ct.Cl. at 199-200, 351 F.2d at 969; Commerce International, 167 Ct.Cl. at 543, 338 F.2d at 89-90; Fermont, 75-1 BCA at 53,004. Aside from the learning curve, there are other causes of delay of which Owens took no account. The court rejected Counts III and V, relating to outgassing and velocity testing problems, both of which could have affected completion of first article or production. Defendant also points to testimony which suggests that there are other possible causes of delay independent from any action by the Government: days of low humidity; in-process production problems, including lack of material; first fire above flush; rescreening; equipment calibration; fires in press bay and mill room; lack of cleanliness and/or careless assembly; marginal test preparation; and use of the wrong impulse cartridges.
Another delay independent of the Government is the shutdown of SOS for environmental violations beginning in April 1985. According to Owens, if efficiency had been 76.1 percent SOS would have ended production in December 1984, before the shutdown. Counsel for plaintiff explained that this justified inclusion of the shutdown as part of the delay for which the Government is held responsible. As discussed above, however, Owens’ estimated production period does not take account of other causes of delay, such as testing problems not attributable to the Government, which Venable testified took a “couple of months” to clear up. Owens has not proven that, at 76.1 percent efficiency, production would necessarily have been complete by April 1985.
In sum, while Owens’ analysis does have some logical and factual validity, it fails in the end because it attributes all delay to the Government, and assumes that all of SOS’s delay and inefficiency on the M206 contract can be traced back to the unavailability of blend. The court is ultimately left with the impression that the figures Owens arrives at to account for delay damages are a guess. The attenuation of proof is simply too great, and the court concludes that plaintiff is not entitled to recover for delay and inefficiencies.
G. Count VII — Claim Preparation Costs.
In Count VII, SOS seeks to recover claim preparation costs and consulting fees. These costs allegedly arise from SOS’s “good business decision to hire outside professional legal and accounting consultants to assist in ongoing administration *400of the Contract and in the hope of avoiding contract disputes.” Defendant correctly argues that these type costs are normally nonrecoverable. DAR § 15-205.31(d) provides in part: “Costs of legal, accounting, and consulting services and directly associated costs, incurred in connection with ... the prosecution of claims against the Government are unallowable.” One narrow exception to this rule was recognized by the Court of Claims in Singer Co., Librascope Div. v. United States, 215 Ct.Cl. 281, 568 F.2d 695 (1977). Claim preparation costs and attorney fees are recoverable if (1) the contractor presented its claim prior to completion of work, (2) Government liability was clear, and “they bore ... [a] beneficial nexus either to contract production or to contract administration.” Id. at 328, 568 F.2d at 721. The costs sought must be related to contract performance. If the costs were incurred only in connection with the preparation and presentation of SOS’s current claims against the Government, they are not recoverable. Plaintiff made no effort to fit within the exception, and the claim for these expenses is denied.
CONCLUSION
Plaintiff is entitled to recover on Counts I, II, IV and V in the total amount of $2,042,395.77, plus interest pursuant to 41 U.S.C. § 611 from November 14,1985. Recovery is denied as to the other counts. The Clerk is directed to enter judgment accordingly.
It is so ORDERED.
ORDER ON RECONSIDERATION
Plaintiff has filed a motion for amendment and reconsideration of portions of the judgment of October 15, 1990. For the following reasons, the motion is granted in part, and denied in part.
1. Plaintiff requests that the total amount of the judgment be amended to reflect an error in addition in the original opinion. Plaintiff is correct. The court inadvertently omitted in the total recovery the amount allowed as damages in connection with increased use of intermediate mix and related labor costs, $105,480.15. The total judgment, therefore, should be for $2,042,395.77.
2. Plaintiff is also correct that the conclusion on page 80 [page 400, supra ] incorrectly reflects nonrecovery on Count II, relating to radiographic inspection requirements. Plaintiff did recover on Count II.
3. Plaintiff also points out that the statement on page 48 [page 380, supra ] that plaintiff was denied recovery on Count V is incorrect. Plaintiff did recover on that count. *
4. The final element of the motion for reconsideration is considerably more problematic. The court found for plaintiff on Count IV in which it claimed that the Government breached an implied warranty by specifying that ground magnesium could be used to make the flares at issue. Pages 55 and 56 of the opinion address that element of damages under Count IV associated with payment for improperly rejected flares. The court reviewed the parties’ competing figures for damages and concluded that neither side’s calculations were compelling and compromised on a figure between the two amounts proposed. Plaintiff now, in its motion for reconsideration, asserts that the court’s calculations were erroneous because certain evidence was overlooked. It is important to bear in mind that the evidence the court considered, however, was in part that to which it was specifically directed by plaintiff in its post-trial brief. Plaintiff now asks the court to ignore the only testimonial evidence it adduced at trial on this issue and consider certain other parts of the record. This shift in strategy normally would not be a proper basis for a motion for reconsideration.
Before evaluating what plaintiff offers in its pending motion, the court will set out the proposed findings of fact and argument on this issue advanced by plaintiff after trial, along with the documentary and testi*401monial evidence relied upon. Plaintiffs post-trial brief on this major issue consists of two sentences:
Plaintiff documented that 207,648 flares remained rejected for static failures after screening. Aratani, pp. 1362-66; P-261, p. 278. These flares conformed to the technical data package; SOS is entitled to the contract price of $7.82 per flare, for a total of $1,623,807, which matches the finding of fact.
The reference to P-261, p. 278 is to a brief excerpt of a handwritten worksheet of plaintiffs expert accounting witness, Mr. Masaru Aratani. Mr. Aratani's testimony is consistent with the exhibit. He relied on two sources for his total figure of 248,000 initially-rejected flares. For flares rejected prior to the 1985 environmental shutdown, he used a report of a technical audit done by the DCAS which reflected 178,000 flares. For rejections after re-start, he relied on statements by Mr. Stan Barlog, an employee of plaintiff, that seven lots of approximately 70,000 flares total were rejected. From the total of 248,000 Mr. Aratani made deductions for units on hand and units subsequently accepted on retest. He testified to a total of 207,648 flares improperly rejected. At a unit price of $7.82, plaintiff sought, in its post-trial brief, $1,623,807.
The findings of fact to which the brief adverts are those found at Count IV, numbers 96 and 97:
96. Of the 250,000 flares that originally were rejected due to the static failures, 207,648 remained rejected after the screening process. These flares conform to all design requirements of the technical data package.
97. Plaintiff is entitled to be paid the contract price of $7.82 per unit for each of these flares for a total of $1,623,807.
The findings rely for support exclusively on the same Aratani testimony and exhibit referred to above. The term “stipulated,” used for numerous other Plaintiff’s Proposed Findings of Fact (“PPFF”), did not appear.
The following observations initially can be made. First, the number of improperly rejected flares reflected in the post-trial brief and in the findings referenced above is not the same as the figures advanced by plaintiff in its motion for reconsideration. The number of flares which plaintiff now advances for flares originally rejected is 277,400. That figure did not appear in any of plaintiff’s pre- or post-trial filings. The amount now claimed, $1,815,491.87 (discounted for use of the weighted average flare pricing method), compares with $1,623,807 originally sought. Second, finding 96 is itself somewhat inconsistent with the evidence relied upon in the findings. PPFF No. 96 refers to the number of flares originally rejected as 250,000. Aratani in fact testified to 248,000. Third, Aratani deducted from this figure 27,852 flares that were later accepted. Plaintiff now abandons that number and uses defendant’s estimate of 30,791.
Plaintiff now contends that the court should have reached the higher figures by extrapolating from other PPFF, not referred to in the post-trial briefing in connection with this damage issue. The following findings and responses are relied upon:
—PPFF 52, 53, 56, 58, 60, 62, 64, 66, 67, 68, 70, and 71 as agreed to by defendant, establish that 102,400 flares were rejected out of interfix 1 and interfix 2 through lot 16.
—PPFF 77 states that “Following rejection of Lot 14 of Interfix 002, plaintiff experienced fourteen additional lot failures based upon static performance. Each of those lots consisted of 10,000 flares. [Defendant’s Exception: Number of lots rejected was at least 19, rather than 14.]”
From these data, plaintiff concludes that, in addition to the 102,400 flares covered by the first series of PPFF, defendant’s response to PPFF 77 stipulates “in general terms” to an additional 190,000 static test failures. After deducting 15,000 for duplicate testing of previously rejected lots,1 *402plaintiff now asks for calculation of damages based on a figure of 277,400 flares (292,400 less 15,000 tested twice). A number of observations can be made, however.
First, plaintiff did not stipulate to defendant’s response to PPFF 77. Rather, its response to DPFF refers to defendant’s proposed number of lots (19), as an “Exception.” Second, PPFF 77 states that 14 lots failed for static performance after Lot 014 of Interfix 002. Two lots after Lot 014 were Lots 015 and 016. Those were already included in the 102,400 rejected flares. Thus, looking at PPFF 77 from plaintiff’s standpoint, it stipulated only to a total of 222,400 flares initially rejected (102,400 plus 12 lots of 10,000 flares each). Third, as plaintiff candidly admits, there is a large difference between the quality of the evidence as to the first 102,400 rejected flares, and as to the last 190,000 flares. Plaintiff refers, correctly, to the latter figure as being more “general.” It is that generality and lack of comprehensiveness which persuades the court that it would be improper to find that the defendant has consciously stipulated that 292,400 flares were originally rejected. The difficulty that existed at the time the court examined the evidence initially, and which still remains, is that there is no single document or set of records or stipulation which consecutively and coherently captures all the lot failures for static testing and the number of flares that remained rejected. Instead, the court is left to draw inferences from partial data and incomplete proposed findings. The inconsistencies between the post-trial brief and the motion for reconsideration, between PPFF 77 and the motion for reconsideration, and between PPFF 96 and the testimony, further highlight the lack of precision and coherence in the evidence. Fourth, plaintiff’s motion makes no adjustment for units currently on hand, which Aratani calculated to be 12,500.
It is no answer, as plaintiff now writes, that this confusing state of affairs is the result of stipulations entered into between the parties. For the reasons discussed above, the PPFF are not a clear answer to the question of how many flares remained improperly rejected. More importantly, if, as plaintiff now contends, the answer is contained in the agreed-upon PPFF, it had the obligation to rely on those and point them out to the court in its post-trial brief, rather than point the court to figures which it now has abandoned.2
It may well be that the figure adopted by the court for flares that were improperly rejected, 165,652, shortchanges plaintiff. In the final analysis, however, the plaintiff had the burden of proving a higher figure by a preponderance of the evidence. It must accept the responsibility for a failure of proof due to insufficient records, or its method of capturing and presenting damages. Plaintiff has not presented sufficient grounds for reconsideration of the findings relative to this aspect of damages.
CONCLUSION
Grounds 1, 2, and 3 of plaintiff’s motion for reconsideration are well-taken. Ground 4 is denied. The attached pages 48 and 80 are substituted for those same-numbered pages in the opinion of October 12, 1990. The judgment of October 15, 1990 is vacated, and the Clerk is directed to enter a new judgment in accordance with the terms of the opinion as corrected.
. The TDP for the M206 flare consists of Military Specification MIL-F-63107A(AR), other Military Specifications and Standards incorporated by reference into MIL-F-63107A(AR), and the drawings and specifications contained in TDP Listing 9311623, Revision D, dated May 30, 1981. While the basic TDP.has not changed, there have been many modifications and revisions.
. Extrusion is another method of pressing flares. This process was later available for the M206, but was not available at the time SOS bid.
. Contract administration on SOS’s contract, including inspection and quality control, was delegated to the Defense Contract Administrative Services ("DCAS”). QA stands for "quality assurance.”
. A penetrameter measures optical density. Paragraph 4.5.2.1 of MIL-F-63107A(AR) states: ‘The penetrameter used in the application specified in MIL-STD-453 shall be designed and constructed to insure a 2.0 percent sensitivity or better.”
. As to SOS’s paper procedure, Barnes concluded:
Well, as far as I could tell, this paper procedure was consistent with the paper radio-graph that I had seen in 1982. I saw no reason to believe that it was anything other than the procedure that that radiograph had been made with. Not only was the paper identified properly, but the time of exposure which was listed in the fourth column of the technique card is five seconds. That would be consistent with paper radiograph using fluorescent screens. Subsequent to that, I believe I’ve heard that they may expose up to ten seconds. But the difference would not be very substantial.
. Plaintiff’s failure to timely detail a procedure for paper did not prejudice Barnes' review. It was clear from the April 30 letter and the post-award conference that SOS had requested permission to use paper. Barnes was familiar with the sample submitted and could have asked for an explanation of how it would be used if he felt that were uncertain.
. MIL-F-63107A(AR) requires that when inspecting a lot of 10,000 flares, which SOS typically produced, a sample of 90 flares must be radiographed and inspected.
. In other counts, defendant contests the correctness of plaintiff’s calculation of G & A for 1985. It is not entirely clear to the court whether the disputed additions to the G & A pool were legal costs or expenses of physical efforts at cleaning up hazardous wastes, or both. In any event, defendant points to 48 C.F.R. § 31.205-15 (1985) which disallows “fines and penalties” resulting from violations of law. Neither legal expenses nor cleanup costs are fines and penalties. The court adopts plaintiffs composite G & A rate of 18.3 percent for all years.
. Aratani testified that he based his calculations on 835 roundtrips. He derived this figure from conversations with SOS employees and concluded that there were an average of 5.25 trips between the facilities per week, and that production lasted 159 weeks. Apparently realizing the lack of connection between general trips between the facilities and the need for paper radiography, plaintiffs post-trial brief attempts to build from a base of 141 flare lots, plus additional radiography occasioned by smaller lots or repeat x-rays. It is patent that the figure Aratani uses is a complete guess. It is based on second-hand data that was not at all focused on the particulars of this claim. As to the number of flare lots which had to be sampled, which is a *360logical beginning point, the court accepts the figure 136. Plaintiffs figure of 141 is an estimate. The Government’s figure is taken from a document prepared by Kyle Owens, an accountant hired by plaintiff. See Plaintiffs Exhibit 263. There is no meaningful way to quantify the number of additional trips occasioned, so the court will use 136.
. First fire "above flush” is listed as a defect.
. While Paragraph 3.4 of MIL-F-63107A(AR) states that flares shall not exhibit indications of "loose, deformed or missing components, foreign material, or defects within the flare pellet i.e., cracks, voids or discontinuities,” and while first fire flakes arguably could fit under more than one of these categories when DCAS began noting first fire flakes on the inspection forms, they were listed under the category of "foreign material.”
. Indeed, after December 1983, SOS was asked to cut open flares to confirm that the defects noted as first fire flakes were actually first fire flakes.
. Ames testified that a popped squib was not a defect.
. Thiokol submitted a request for “[c]ost increases associated with the end item redesign to resolve TDP problems of pellet breakup and velocity.” One such cost item was the 200 degree drying process: "Increased (recurring) manufacturing costs have resulted from the more stringent process/QA controls and tests ... e.g., 200 [degree F] drying of flare composition____” Thiokol’s was a cost plus contract.
. As a conspicuous example, and one which defendant points out, there is no discussion of the procedures for mixing the flare composition.
. Section 4.5A2 of MIL-F-63107A(AR) outlines the static output testing to be performed on sample flares. Section 3.5.3 specifies the static output requirements that must be achieved for both individual flare samples and lot averages during Section 4.5A2 testing. Individual flare samples are required to achieve decoy intensity within .14 seconds and must remain at that intensity for 2.3 seconds. In addition, each sample flare must reach a burn intensity of 7500 watts per steradian within 0.5 seconds and must remain above that intensity for 1.25 seconds. Finally, each sample flare must reach a maximum intensity of 8800 watts per steradian.
. The Government contends that plaintiffs specificity with regard to the cause of the failures — the approval of ground magnesium — constitutes a separate claim which has not been certified and submitted to the Contracting Officer. The court disagrees. The assertions concerning ground magnesium are explanations of the existing claim and do not constitute a separate item of relief.
. One of Farnell’s comments regarding ground magnesium was that it burned very quickly and at a steadily increasing rate. She found this unusual.
. SOS was granted a Request for Deviation, no. 7040-D-0002, in May 1984.
. Venable described the black body as a heated cavity which emits radiation. It is not clear from the record how precisely it is employed in static output testing.
. In 1983, Thlokol attempted to produce M206 flares using ground magnesium. Plaintiff contends, and defendant agrees, that of the 64,523 M206 flares manufactured with ground magnesium, only 14,971 were successful. Defendant maintains, however, that Thiokol’s flares failed because of short burn times, whereas SOS’s flares failed to maintain high intensity. In any event, it is undisputed that Thiokol, in order to salvage the ground magnesium, cross-blended with atomized magnesium.
. These percentages are classified and were not discussed in documents or testimony. Defendant did not allege that plaintiffs problems were due to improper exercise of its discretion in this regard.
. The court dismisses any implication that poor quality control and management turnover caused the static output problems. While these may have contributed to delay, the record does not establish a nexus between turnover and the prolonged problems SOS experienced with static testing.
. Defendant alleges that this is a new claim, not previously submitted to the CO. It makes the same contention with respect to cross-blending and screening. The issue is whether screening, cross-blending, and pre-testing are new claims or whether they can be viewed as a logical extension of the certified claim as amended. Although the test is not precise, the decisions of this court and the Federal Circuit create a coherent framework. The factual basis of the contested claim must be the same as the original claim. J.F. Shea Co. v. United States, 4 Cl.Ct. 46, 55 (1983). There is no bar to merely increasing the damages flowing from the original claim. Glenn v. United States, 858 F.2d 1577, 1580 (Fed.Cir.1988); Tecom, Inc. v. United States, 732 F.2d 935, 937-38 (Fed.Cir.1984). Other factors or ways of expressing the same considerations are captured in Cereberonics, Inc. v. United States, 13 Cl.Ct. 415, 417 (1987) (footnotes and citations omitted):-
In determining whether a claim is properly before this court for the purpose of de novo review under the Contract Disputes Act (CDA), the Claims Court has generally looked to its "operative facts.” If the complaint brought here is based on the same set of operative facts underlying the claim presented to the contracting officer, then this court has jurisdiction under the CDA. Agency boards of contract appeals, to whom negative decisions of contracting officers can be appealed under 41 U.S.C. § 607 and 608, have looked variously to the "ultimate character” or "essential nature” of the original claim, or whether the operative or legal facts have been changed on appeal. The critical test appears to be whether the scheme of adjudication prescribed by the CDA is undermined by the contractor’s claim on appeal — that is, by circumventing the statutory role of the contracting officer to receive and pass judgment on the contractor’s entire claim.
The issue thus becomes whether the three challenged claims are distinct from the original claim, or whether they can fairly be characterized as mere factual, legal or monetary augmentation by degree and not a change in kind. There is a factual link between testing, screening, and cross-blending on the one hand, and the narrative statements supporting the amended claim on the other. The claim, originally filed February 13, 1985, and amended on June 13, 1985, put the defendant on notice that the plaintiff was alleging serious problems in achieving static performance because of asserted defects in the TDP. Plaintiff recited the steps it had taken to achieve required standards. The pre-testing claim, for example, is based in part on labor involved in installing test hooks, test blocks, tape and the electronic match. Those same steps are identified in the claim. SOS also recites that the cost overrun to date was due to the additional "labor, management, and material costs for the constant tests, research and development that SOS was required to undertake.” It also claimed that SOS had to alter the TDP in order to get satisfactory production and that this resulted in "extra steps in the manufacturing process requiring additional production and assembly personnel." Although screening and cross-blending are not itemized in the claims, they can fairly be viewed as additional costs identified to the same set of facts originally complained of. That is a particularly appropriate result here, where the amended claim was filed when SOS had not developed a link through discovery between performance problems and the qualification of ground magnesium. Plaintiffs understanding of the causes of the problem had not crystallized. It was not entirely clear which corrective measures were responsible for making production possible. There is no lack of jurisdiction over the claims.
. The figure is an estimate based on a total production of roughly 1.36 million and 10,000 flares per lot.
. These flares, presumably, were not the same flares which were dynamically tested. Rather, they were from lots which were rejected based on data obtained from those flares which were tested.
. Defendant avers that there were 21 impulse cartridges from Lot 001 and 39 from Lot 002.
. Lots 001 and 002 experienced both high and low velocity problems. However, the majority of the defects were for high velocity — eight flares in Lot 001 and 18 flares in Lot 002 exhibited velocity over 165 fps, whereas only one flare in each lot had velocity below 75 fps.
. Owens testified that he first attempted to evaluate the learning curve experienced by SOS under the contract. The accounting firm of Price Waterhouse had previously performed a learning curve analysis based on the number of flares end-capped per week. That study considered the average number of flares produced per hour over time. It found that SOS experienced no learning curve. Owens decided that a "more detailed" analysis was necessary.
. SOS’s bid was broken down by manufacturing steps, but the breakdown did not overlap with production records.
. Owens used information in SOS’s daily production records regarding the number of flares pressed. Owens obtained information on the total number of hours required in production on the SOS contract by referring to SOS’s distribution of labor report and job cost ledger, both of which, according to Owens, were kept by SOS in the regular course of business. The figures he used, however, were corrected for the hours identified in claims I-V as resulting from additional work steps added to the production process. According to Owens, this assured that the comparison was between budgeted hours and hours actually incurred.
. Owens did an alternative analysis to determine the efficiency SOS would have experienced had they produced 27,500 flares per week, as reflected in their bid. That analysis involved first plotting the relationship between the number of flares pressed and the efficiency achieved at that level. Both the number of flares and the efficiency values were statistically "smoothed” to account for variation at that level. Using regression analysis, Owens determined that at 27,500 flares per week, the efficiency would have been 82.9 percent. The R-squared factor was .869258. If SOS could have produced 27,500 flares per week, and if the corresponding efficiency would have in fact been 82.9 percent, the total savings due to decrease in required labor hours would have been 53,179,005.
. There was no testimony on calculation of the overhead or general and administrative expenses as they were affected by inefficiencies.
. The delay claim covers all labor hours incurred, not merely hours for production.
. Owens did not divide the 13,938 figure into 269,503 presumably because the 13,938 hours were related to production hours, not total hours. This assumes that the amount of time it would have taken for the production aspects of performance is an accurate indication of what it would have taken overall.
. It is not clear from the record how the average was weighted.
. The exact calculations are described in Owens’ testimony and in a document prepared for SOS by Arthur Andersen.
. This is an average of the estimates for fiscal years 1983 and 1984, the two years which, according to Owens, SOS would have had to dedicate to the M206 contract.
. This figure was from Price Waterhouse.
. Indeed, defendant questions whether Owens demonstrated a correlation between efficiency and number of flares pressed. Pointing to the data which generated Owens' weekly summary of flares pressed and corresponding efficiencies, defendant notes that for the week ending February 19, 1984, in which 19,984 flares were pressed, there was only 51.4 percent efficiency, whereas for the week ending January 15, 1984, when 19,318 flares were pressed, the efficiency was 88.78 percent. Owens acknowledges that there is some fluctuation, but points to his table, which demonstrates that when specific production levels are considered over all the weeks the levels were reached, a correlation between the production levels and the average efficiencies is revealed. Owens does not, however, attempt to account for the variation.
. One of defendant’s objections to Owens’ analysis is that because it is based on SOS’s original bid, it stands or falls with the accuracy of SOS’s bid. The Government attempted to undermine the bid. The court, however, does not fault Owens' analysis as to comparative efficiencies because it is based on the bid. The bid is merely a reference point against which efficiency is measured. The efficiency analysis merely expresses numerically the observation that as more flares are produced fewer labor hours are needed. Any hypothetical efficiency could be used as a comparison point. The court notes, however, that a separate issue exists with respect to the validity of drawing upon the bid for hours earned and for calculating delay. See discussion supra at 399.
. Plaintiff’s data suggests a deletion of 16,000 from the total of 292,400. It is not clear why only 15,000 were deducted.
. Plaintiff now discounts its own witness’ testimony because it was an estimate based on another estimate "that was not part of the record, added to an estimate prepared by a witness who did not testify.” Those are the very factors that persuaded the court to discount Aratani’s testimony. It is at a minimum puzzling, however, for the plaintiff itself now to question the court’s use of that testimony in arriving at a figure. If Mr. Aratani’s estimate is unreliable because it was “prepared long before the stipulation was reached,” it was nevertheless the most current assertion of plaintiffs position by virtue of its presentation at trial.