OPINION
BRUGGINK, Judge.This is an action brought under the Contract Disputes Act, 41 U.S.C. § 601-13 (1982). The case arises out of a contract awarded by the United States Army Corps of Engineers, Ft. Worth District (“Corps” or “Government”) to Servidone Construction Corporation (“Servidone”) for the completion of an earth fill dam approximately 4 miles long. Servidone has filed a complaint alleging primarily a differing site condition. It seeks recovery of $41,877,029.00, plus interest and costs. The case was tried in Washington, D.C. and in Dallas, Texas. The court conducted a site visit on the parties’ agreement. After post-trial briefing and oral argument, the court concludes that Servidone is entitled to recover on part of its claim.
BACKGROUND FINDINGS
The project in question involved the impoundment of Mountain Creek, a tributary of the West Fork of the Trinity River. The completed dam is located approximately ten miles southwest of Dallas, Texas. The dam is intended to provide flood control, water supply, and recreation and to promote fish and wildlife. Originally the project was described in the specifications as Lakeview Lake. It is known today as Joe Pool Reservoir.
The lake and dam were constructed in 2 phases. The initial phase was awarded to Lane Construction Company (“Lane”) and required construction of an outlet works and placement of approximately 2,000,000 cubic yards of embankment. The second phase required completion of the embankment, construction of a spillway, completion of the outlet works, and building of roads. Completion of the embankment required placement of approximately 10,000,-000 cubic yards of fill. At the time the solicitation for the second phase was issued, work had not yet been completed on the Lane contract.
When bids were opened for the second phase contract, the low bidder was John Carlo, Inc., with a total bid price of $23,-844,078.64. Servidone was the next low bidder at $25,781,338.18. At $27,168,-724.80, Resource Construction Corp. had the third lowest bid. Lane bid $34,103,-831.00. The J.D. Abrams Construction Company (“Abrams”) bid $33,961,358.77. Other contractors bid higher amounts. The highest bid was approximately $46 million. The government estimate, which does not include profit, was $28,671,127.05. No bid verification was requested. The Contracting Officer (“CO”) declined to award the contract to John Carlo, Inc., on the ground that its bid was nonresponsive, and award was made to Servidone on September 30, 1981.
The major elements of construction under the contract were the embankment items. Plaintiff’s bid price for that portion of the work was $14,767,424.00. The government estimate for the embankment items, not including profit, was $19,326,-845.00. Abrams’ estimate for this portion of the work was $24,029,883.00.
Because of a pending bid protest, notice to proceed was not issued to Servidone until May 18, 1982. Servidone promptly began work. The project completion date *350was initially December 31, 1985, but various time extensions were granted by the Government. Servidone finished its work within the adjusted period.
The project was a “production job,” that is, the contractor was required to create the equivalent of an assembly line for the excavating, processing, and compacting of fill material to create the embankment. The contractor needed to achieve and maintain a sufficient rate of production to complete the embankment within the contract period.
Dirt had to be obtained for the embankment from borrow areas or “pits” designated by the Government. Lane had utilized borrow areas Bl, Cl, E3, and E4. By the time the second phase was advertised, a portion of borrow area Cl had been permanently inundated and stocked with fish. All the remaining areas were available to Servidone. Borrow Area A is closest to the east embankment, the principal area of embankment work. At Servidone’s request, the perimeters of borrow area A were eventually expanded, and Servidone ultimately excavated approximately two thirds of its material from that borrow area.
Servidone estimated that it had to place 10,000 to 15,000 cubic yards of fill per work day in order to complete on time, and prepared its project schedule on that basis. Plaintiff’s estimate of equipment costs was based on the use of 2 different types of excavation equipment in the borrow areas: a 631-D scraper and a Holland loader. A 631-D scraper is a self-loading piece of equipment. The bottom, forward edge of its container or pan opens and is equipped with a blade so that, as the unit moves, it cuts a horizontal ribbon of earth that is forced into the pan. When the pan is full, the operator raises the pan, closing the opening. When the unit is loading, it is normally pushed by bulldozers, commonly referred to as pushcats. The scrapers should then be able to haul the material under their own power from the borrow to the fill, where they are unloaded by opening the pan and engaging a hydraulic plunger that pushes the material toward the opening.
Unlike the scraper, which makes a horizontal cut, the Holland loader excavates material from a vertical face. The dirt is cut by a vertical blade and dropped onto the unit’s conveyor belt and then into a separate hauling unit. The operator of the hauling unit positions his equipment under the Holland loader’s conveyor until he has a full load, and then carries the dirt to the embankment. The hauling units used by Servidone in connection with the Holland loader are referred to as “belly dumps.” A Holland Loader can load a belly dump in approximately 45 seconds. The belly dump unloads material on the embankment by opening doors in the bottom, or belly, of the container. The dirt is expected to drop out by gravity.
In preparing its bid, Servidone estimated that 21 hauling units (11 scrapers and 10 belly dumps) would be required to haul the material an average of 3 to 4 minutes of haul distance from the borrow area to the embankment. The scrapers have a heaped capacity of 31 cubic yards. The belly dumps have a heaped capacity of 54 cubic yards. Servidone estimated a shrinkage factor for both its scrapers and belly dumps. Thus, its bid assumed that a 31-cubic-yard scraper would average 21 cubic yards per load and a 54-cubic-yard belly dump would average 42 cubic yards per load.
Dirt was placed on the embankment in “lifts”. After it was dumped, the dirt had to be processed before it was incorporated into the dam. The material initially would be spread by bulldozers equipped with blades. The dirt was then plowed by bulldozers pulling disc harrows. In its bid, Servidone anticipated using 270-horsepow-er D8-H bulldozers pulling 3 Rome disc plows. The 270 horsepower D8-H bulldozer is more powerful than the D8 bulldozers reflected in the Government estimate. The discing served not only to break up clods and chunks in the soil, but also to adjust the moisture content. As the disc harrow aerated the fill material, the consequent evaporation reduced the moisture content of the soil. If a higher moisture content were required, the contractor would wet *351the soil using trucks equipped with water tanks and spray nozzles. Subsequent passes of the disc harrows would mix the water more evenly throughout the soil.
Discing was to be accomplished with a 36-inch diameter plow which had to be able to penetrate the full depth of the lift being processed until a uniform distribution of moisture was obtained. If the disc did not achieve the cutting depth, the contract required that it be ballasted or replaced with a heavier disc.
After other processing was complete, the contract required a minimum of 6 passes of a tamping or sheeps-foot roller for most areas of the fill. In the semi-compacted zones, 2 passes of a rubber tired roller satisfied the compaction specification.
Prior to placement of a subsequent lift, the compacted surface would be scarified, which required using a disc harrow to loosen the surface of the compacted lift. Scarification promoted effective bonding between lifts. The contract also required that whenever fill placement or compaction operations stopped, the fill surface was to be smoothed using bladed equipment, then sealed by running rubber-tired equipment over it. This helped prevent moisture change due to evaporation. For most zones, the contract specified that lifts be no thicker than 8 inches prior to compaction with a tamping roller. The CO could direct lifts to be thinner.
The basic design of the dam called for 4 main “zones” of material, each with different moisture content requirements. Viewed in cross-section, the central core of the dam was “impervious fill.” On both the upstream and downstream slopes of the impervious fill lay a zone of “random fill,” and similarly outside the random fill lay a zone of “semi-compacted fill.” Finally, “select impervious fill” covered the upper slopes and crown of the dam.
The contract did not specify that Servidone achieve a certain degree of compaction with respect to most of the embankment. Instead, the Corps tried to achieve desired compaction by specifying that soil be processed as described above, and that the contractor achieve a certain moisture content. Any given sample of soil will have a particular ability to retain water. For each sample there is a percentage of moisture, as a proportion of the weight of the solid material, at which it is at its maximum compactibility, as determined by a standard compaction test. That point of maximum compactibility is expressed as a figure known as the “optimum moisture content.” This figure is critical to compaction and thus to the overall structural strength of the dam. It will vary from soil to soil.
In order to determine optimum moisture, a compaction test known as the Standard Proctor Compaction Test is run. It involves taking a sample of soil, dividing it into uniform batches, adjusting the moisture content of the batches to create a spread of moisture contents, and then subjecting the batches to a fixed number of blows of a hammer of a certain weight within a cylinder. The optimum moisture content is then extrapolated from the data points created by the different batches of the same soil sample.
The required moisture content for the embankment soils varied with each of the 4 zones. The select impervious zone, which comprised approximately 5.6 percent of the embankment, was required to be at or within 3 percent above optimum moisture content (0 to +3). For the impervious zone, which comprised approximately 21 percent of the embankment, the moisture requirement was also 0 to 3 percent above optimum (0 to +3). The random zone, which comprised approximately 40 percent of the embankment, had to be between 2 percent below optimum and 3 percent above optimum ( — 2 to +3). The semi-compacted zone, which comprised approximately 30.6 percent of the embankment, had to be within 5 percent below optimum to 4 percent above optimum ( — 5 to +4). It thus had the greatest range of acceptability-
Although the Standard Proctor test provides the most direct evidence of the optimum moisture content of a given sample of *352soil, determination of the optimum moisture against which Servidone and Lane were judged was not based directly on Proctor test results for a given sample. Instead, the Fort Worth Division of the Corps used a short-hand method of determining the target or optimum moisture of a given sample. That method is known as the Liquid Limit Correlation Curve (“LLCC”). To describe the curve, however, it is first necessary to set out some of the methods used to characterize clay.
The liquid limit (“LL”) of a clay material is the moisture content at which it ceases to perform as a plastic and begins to perform as a liquid. The higher the LL, the greater its ability to hold water, but still perform as a plastic rather than a liquid. At the lower end of the moisture spectrum, the plastic limit (“PL”) of a clay is the moisture content at which clay ceases to act as a plastic and begins to act as a solid that cannot be deformed without cracks forming in the surface. These tests are normally performed manually by rolling out threads of clay.
Clays are classified into 2 categories— CL and CH. CH clays are highly plastic with a LL of 50 or more. CL clays are of low plasticity and have a LL of less than 50.
By subtracting the PL from the LL, another measure of the character of clay is obtained — the plasticity index (“PI”). The PI is the spread of moisture content which the clay can sustain while still remaining in a deformable, or plastic, state. These characteristics of clay soils are collectively referred to as the “Atterberg limits.”
In order to generate a LLCC, a significant number of Proctor tests is first run on the soil. This establishes the optimum moisture content at particular liquid limits. Figure 1 shows, for example, the Proctor points furnished to bidders on the first (Lane) phase of construction. Superimposed on Figure 1 is the LLCC included in that bid package. The vertical axis of the graph reflects optimum moisture. The horizontal axis shows the LL. Proctor data was not part of the bid package on phase 2.
[[Image here]]
*353The LLCC is a line drawn as a visually estimated average of the plots of optimum moisture content readings based on the Proctor tests. Theoretically, there should be as many Proctor data points above a line as shown below. The net result is that the LLCC is an approximation. The optimum moisture reading shown on the vertical bar of the graph for a soil of a particular LL is not necessarily the true optimum. It is an assigned optimum. Figure 2 shows the LLCC included in the bid materials on phase 2 at Sequence 100 of the contract drawings. It initially was used to test acceptance. While the bid documents state that this curve was established during the Lane contract, none of the contract materials reflect the Proctor data that the Corps relied on in establishing this curve.1
[[Image here]]
The variation in Optimum moisture content for a given LL is a result of 2 phenomena. First, different soils having the same LL may have different optimum moistures because of differences in the mineralogy of the soil itself. Second, there is a human element in the generation of Proctor results. The actual Proctor optimum moisture is frequently a visual extrapolation between individual points.
Despite the inevitable lack of precision built into the LLCC, the Corps in Texas has used it because it is relatively accurate, and because results can be obtained more quickly. LL results can be generated faster than optimum moistures developed by the Proctor test (normally 2 days versus 4). The purpose of the curve is to take advantage of that fact by using a known LL to estimate an unknown optimum moisture. If the LL is known, one can find the corresponding point on the LLCC, and then read across to find the anticipated optimum moisture.
Acceptance of the completed embankment was based on government testing of the compacted soil for water content and LL. As discussed above, each zone of ma*354terial had its own range of acceptable deviation from optimum moisture. The LL of the sample being tested would be determined. By using the LLCC, the tester could determine the assigned optimum moisture content. This percentage would be compared with the actual moisture content of the sample. If the actual moisture content was within the range of deviation for that zone, the sample passed.2 If it did not, the sample failed. The contract provides that acceptance testing will use the optimum moisture content reflected on the LLCC included in the bid documents, although it indicates that the curve might be adjusted. It also required that test samples were to be taken from the lifts below compacted lifts.
Since the LLCC is by nature an estimate, and since it is based on a finite number of Proctor points, subsequent compaction tests on the same general area of soil may demonstrate that the original curve needs to be moved. The contract drawings supplied in the bid package bear the notation that the CO “will periodically make adjustments to these curves.” Two adjustments were made to the LLCC during contract performance. The first of these adjustments was made on August 9, 1982. This adjustment moved the curve upward slightly. The second adjustment, made 10 days later, on August 19,1982, moved the curve downward slightly. These adjustments were based on the results of periodic Proctor compaction tests run on the fill. Approximately 100 such tests were run during the course of the contract. These tests did not directly affect acceptance, but were used to evaluate the curve. In the judgment of the Corps, the additional compaction test results indicated that the optimum moisture content for the materials being utilized by Servidone was slightly different from that reflected by the initial curve.3
Figure 3 reflects a superimposition of the LLCC used after August 19, 1982 on all Proctor data generated during contract performance.
*355[[Image here]]
Bidders were furnished with data about soil characteristics based on borings. Borings were taken in various years, and the data for each boring was shown on the contract plans under each boring log. Among other factors, the drawings indicated the classification, moisture content, LL, PL, bar linear shrinkage (from which plasticity may be approximated), approximate gradations, and visual description of the soil. Classification, water content, bar linear shrinkage, and visual description were provided for every sample.
By using the LL and water content information provided in the boring logs and by consulting the LLCC, a bidder could calculate the amount of moisture change, if any, that was required, assuming the soil was at the same moisture level during actual compaction. Servidone undertook that exercise and concluded that the soil would require very little manipulation of moisture content to be at optimum.
Additionally, the borings indicated the water table and date of determination for groundwater encountered in the boring. The drawings indicated that the absence of an indication of groundwater did not necessarily mean that groundwater would not be encountered at the location or within the vertical reaches of the borings.
Servidone filed its action in this court on June 4, 1984, prior to completion of the contract. It filed a first amended complaint on April 2, 1986, and a second amended complaint on July 20, 1987. The complaint, as amended, alleges that Servidone encountered numerous problems during construction of the dam that more than doubled its anticipated cost and necessitated borrowing most of the cost of completion. As discussed more fully below, Servidone attributes its increased cost primarily to the nature of the soil. It contends that the soil was so difficult to excavate, unload, and process that many additional and more powerful pieces of equipment were required to maintain a sufficient rate of production. Servidone asserts government *356liability for these costs primarily on the basis of the differing site condition clause of the contract. However, Servidone also argues in the alternative that the Government breached implied duties to provide Servidone with information in the Government’s possession which was material to performance of the contract, and to provide the contractor with specifications that were suitable for their intended purpose.
Other claims made by Servidone in its complaint, as amended, relate to the Government’s administration and execution of quality assurance testing, flooding at the project site, difficulties Servidone encountered in placing and compacting lime stabilized subgrade for the roadway on top of the dam, and excavation of material from a portion of Mountain Creek in the closure area of the dam.
DISCUSSION
I. EMBANKMENT CLAIM
A. Differing Site Condition
Servidone claims that the nature and composition of the soil excavated from the Government-specified borrows created a differing site condition as defined by the differing site condition clause of the contract, General Provision (“GP”) 4(a). That clause reads in part as follows:
The Contractor shall promptly, and before such conditions are disturbed, notify the Contracting Officer in writing of: (1) subsurface or latent physical conditions at the site differing materially from those indicated in this contract, or (2) unknown physical conditions at the site, of an unusual nature, differing materially from those ordinarily encountered and generally recognized as inhering in work of the character provided for in this contract. The Contracting Officer shall promptly investigate the conditions, and if he finds that such conditions do materially so differ and cause an increase or decrease in the Contractor’s cost of, or the time required for, performance of any part of the work under this contract, whether or not changed as a result of such conditions, an equitable adjustment shall be made____
The purpose of such a clause is to provide the Government with more accurate bids by discouraging inflation for risks that may not eventuate. The clause eases the contractor’s burden of risk assessment and site investigation. Foster Constr. C.A. v. United States, 193 Ct.Cl. 587, 613-14, 435 F.2d 873, 887 (1970). Differing site conditions thus can occur in 2 circumstances: conditions differing from the contract indications and those differing from conditions normally encountered. These are commonly referred to as Type I and Type II differing site conditions, respectively. Servidone contends that both types were present here.
1. Type I Differing Site Condition Claim
To prevail on its Type I differing site condition claim, Servidone must prove that it encountered subsurface or latent physical conditions at the site that differed materially from those expressly or impliedly indicated in the contract. See id. at 435 F.2d at 881. Servidone asserts theories of liability based on both express and implied indications in the contract. However, the court finds that neither theory has merit.
a. Express Contract Indications
i. moisture levels
Prior to post-trial briefing, Servidone advanced 2 ways in which conditions differed from those indicated in the contract drawings. The first allegation is that the soil was wetter than indicated. This claim was not addressed in Servidone’s post-trial briefing, and the evidence on that issue is equivocal at best. Servidone’s expert, Professor George Sowers, testified that, based on samples taken during construction from locations near the site of the borings described in the contract, “there was a consistent pattern ... moisture contents were one, two, or three percent more” during construction. While defendant did not show Prof. Sowers’ figures to be inaccurate, the court is persuaded for several reasons that the difference is not legally significant.
*357First, the contract data was not inaccurate. As conceded by Prof. Sowers, “I don’t see anything that’s inaccurate about the borings. But, the conditions change, they do change.” As Dr. Steven Poulos, defendant’s expert, explained, however, seasonal variations occur in moisture levels. This should have been evident from the differences in the ranges of moisture percentages in borings made in 1968 compared with those made in 1977.
Second, Servidone’s analysis is flawed because it is predicated on the use of averaged water data. The alleged harm to Servidone from higher moisture percentages in the field is that it was falsely lulled into a belief by the lower contract readings that very little manipulation of water content would be necessary to achieve the content required for any LL given by the LLCC. Larry Fantozzi was Servidone’s Project Engineer. He also did most of the bid workup for plaintiff. He testified that the assumption of little moisture manipulation was based on averaging of moisture readings in the borrow areas. The flaw in that method is apparent, and in any event confirmed by Dr. Poulos. For example, at a LL of 54, the original curve dictated an optimum moisture content of 20 percent. If the moisture content shown in the borings for point A was 15 percent, and for point B was 25 percent, the average would be 20 percent. Apparently Servidone would have posited no extra moisture manipulation. In fact the moisture at both points is 5 percent off.
Third, Dr. Poulos’ comparison of boring results compared with actual experience in the pits reflects a close approximation in the range of moisture contents to that shown in the bid documents. He opined there was no significant difference between the contract documents and Servidone’s test data. Even the maximum variation of approximately 3 percentage points in moisture contents cited by Servidone is less than one fifth of the 17-point range of moisture contents suggested by the boring logs. Also, as noted by Prof. Sowers and Fantozzi, the average values for moisture content in the borrows, as shown on the drawings, were near optimum. Thus an increase of 3 percentage points would still place the average near the range of moisture required by the most demanding requirements in the contract, namely 0 to 3 percent above optimum for the impervious fill. To adopt Fantozzi’s characterization, the moisture contents actually encountered were still “within that range or ... close to that range.”
The moisture content indications were not sufficiently different from actual experience to induce a reasonable reliance by Servidone “that subsurface conditions would be more favorable than those encountered.” Pacific Alaska Contractors, Inc. v. United States, 193 Ct.Cl. 850, 436 F.2d 461 (1971); Framlau Corp., ASBCA Nos. 14,205, 14,474, 71-2 B.C.A. (CCH) 118,989.
ii. reliability of the liquid limit correlation curve
Servidone also argues that the bidding documents inaccurately represented conditions on site because Servidone personnel did not understand that the LLCC was only an estimate — that is, they did not know that the optimum moisture content obtained by a Proctor test for a given sample might deviate from the assigned optimum taken from the LLCC. This phenomenon would have been apparent, Servidone argues, if bidders on the Joe Pool project had been furnished the Proctor points from the Lane contract. By overlaying the curve on those points, the phenomenon of averaging becomes apparent. See Figures 1 and 3 above. It points out that such data is regularly furnished by the Galveston District of the Corps on its dam projects. Servidone contends that it relied on the LLCC as an indication that soil at its LLCC optimum moisture was also at its optimum point of workability, and that no further moisture change would be required. For Servidone’s argument to succeed, it must show that the Corps in effect warranted that use of the curve would ensure maximum ease of workability of the soil. The record does not support plaintiff’s argument.
*358Servidone concedes, as it must, that the target was the optimum moisture assigned by the curve, regardless of what a Proctor-dictated optimum might be. As stated by plaintiffs own expert, Prof. Sowers: “In terms of complying with the written specifications the liquid limit curve of course ... would be the more accurate in terms of complying with the [specifications].” All that the contract indicates is that the curve will be used as the test for acceptance.
Servidone argues that the bidding documents represented the results of the LLCC to be true optimum moisture, because the vertical axis of the graph is labeled “optimum moisture.” A necessary corollary of this argument is that the curve was in fact inaccurate — that is, that a bidder would not have realized that it would have taken additional unnecessary effort to meet the curve when material was not at its true optimum moisture for compaction. The court finds this element of plaintiffs argument unfounded.
Liquid limit correlation curves have been used for approximately 12 dams built in the Ft. Worth and Galveston districts. Use of the curve is apparently unique to the Corps’ Texas districts. Prof. Sowers, with extensive dam construction experience, characterized the curve as a crude measure and is not normally used in dam construction. Even Dr. Poulos, defendant’s expert witness, testified that in the 50 to 100 compaction projects with which he had been involved, he had never seen the LLCC used. He was aware, however, that the curve was used in parts of the Southwest where highly plastic clays are encountered.
By plotting actual Proctor results on the LLCC, a visual assessment can be made of the degree to which the Proctor results deviate from the projected optimum, represented by the curve. The actual deviation, or “scatter,” of the Proctor test results on the Joe Pool project are shown on Defendant’s Exhibit (“DX”) 36. At certain liquid limits, the difference between the lowest point below the curve and the highest point above the curve is as much as 10 points of moisture content. For example, at a LL of 57, there is a Proctor point at approximately 17.5 percent optimum moisture. At the same LL, there is also a Proctor point of approximately 27 percent. The optimum dictated by the LLCC at that point is approximately 22 percent. This means, as Servidone points out, that it was required to aim for 22 percent moisture, when the actual optimum could be 4 or 5 percent wetter or drier.
There are a number of responses to this phenomenon. The first is apparent from a visual examination of the exhibit. It is obvious that the LLCC line is drawn through the middle of the spread of Proctor points. It is also obvious that the bulk of Proctor points is bunched near the curve. The second response was provided by Dr. Poulos. In support of his primary thesis that the LLCC is a reliable method of construction control, he explained that the scatter around the LLCC is in part due to a certain degree of error in the Proctor tests themselves. Plotting the compaction results and drawing a curve through them to estimate the point of maximum compaction involves a degree of speculation. In other words, two people might not agree on how to read the same results. In addition there can be variations in the preparation or compaction of the sample. Dr. Poulos introduced one series of tests of the degree of error in Proctor results which reflected an error rate of 13.1 percent.
The implication of this for plaintiff’s theory is obvious. Plaintiff assumes that the Proctor results are inherently more reliable because they are original data. The curve, on the other hand, is an extrapolation from that data. The court finds, however, based on Dr. Poulos’ testimony and on the work done by Schmidt in reevaluating Proctor results, that use of the LLCC was proper because it has the advantage of ameliorating the risks inherent in use of raw Proctor data. By creating an average, it compensates for results which are peculiar due to error.4
*359Another reason the curve is desirable is that when the Standard Proctor test is run in the normal fashion, results take 96 hours. A LL test, on the other hand, can be completed in 48 hours. Testimony by Fantozzi and others working on the embankment persuade the court that a 2-day time savings is a significant benefit to the contractor.
Dr. Poulos also performed a statistical analysis of the scatter of Proctor data generated during the Servidone project and concluded that the standard deviation (the amount of deviation in both directions from the LLCC which includes 68 percent of all Proctor points) was 1.86 percent. In other words 68 percent of the Proctor results were no more than 1.86 percent water content above or below the curve, a figure Dr. Poulos viewed as low.
Finally, and most persuasive, are the results of Dr. Poulos’ comparisons between the Proctor test and the LLCC with respect to their ability to predict whether the soil tested during construction would pass or fail. For every zone of the dam, plaintiff would have had more failures if the soil samples during construction were tested by the Proctor method. Dr. Poulos explained that this was due in part to the fact that the LLCC by its nature eliminates the error inherent in exclusive reliance on Proctor testing.
The court was unpersuaded that this phenomenon should be discounted, as Servidone suggests, by the fact that plaintiff benefited from a learning curve while working on the dam. Servidone suggests that it attempted to accommodate itself to the LLCC, and therefore the fact that there were fewer failures only means that it became used to the effect of using the curve. There was ample evidence that Servidone’s personnel tried to develop a subjective feel for when soil reached optimum compactibility. If the curve was truly quixotic, and if the possibility of straying from true optimum included going both above and below optimum, that would not have been possible. The learning curve Servidone developed must have been independent of the LLCC, because even LL tests had a 48-hour turn-around. Fill foremen had to make judgments about moisture content and disc passes on a continuous basis without the benefit of immediate test results.
Servidone’s real problem is that its personnel did not understand the LLCC. Servidone had no experience with a LLCC pri- or to the Joe Pool project. Given Servidone’s admitted lack of understanding about how the curve was generated, it was less than reasonable for its personnel to make assumptions about how the curve worked in preparing the bid. The court concludes that the Corp’s furnishing of the LLCC without backup data did not constitute a differing site condition.
b. Implied Contract Indications
Servidone also contends that, by failing to mention anything specifically to the contrary, the contract implicitly represented that the material could be worked by “normal” means. The conditions principally complained of by Servidone were the slickness of the soil, its cohesiveness, and the difficulty in adding or removing moisture during processing. These characteristics were referred to collectively in both parties’ testimony, albeit loosely, as “toughness.” Servidone argues that the contract documents suggested that the soil would be of normal or typical toughness, but that Servidone actually encountered extreme toughness.
The gravamen of a Type I inquiry is whether the condition differed materially from the affirmative representations in the contract. See Pacific Alaska, 193 Ct.Cl. at 864-65, 436 F.2d at 469. Contract indications may be implicit, Foster Construction, 193 Ct.Cl. at 604, 435 F.2d at 881, but there must be sufficient indications of the condition to induce a reasonable reliance in the *360bidder “that subsurface conditions would be more favorable than those encountered.” Pacific Alaska, 193 Ct.Cl. at 864, 436 F.2d at 469.
Testimony from both parties’ geotechnical experts established that there is no precise quantitative measure of toughness. Rather, the term is used to reflect a qualitative assessment of the difficulty of working with a particular soil. Defendant’s expert Dr. Poulos, however, noted that both the LL and PL of the soil are some predictors of toughness, particularly when both values are high. Those characteristics could be obtained from the borings.
Servidone does not point to affirmative contract indications; rather, it points to the lack thereof. For example, the contract does not indicate how many disc passes will be required to process the soil. Servidone contends this suggested that a “typical” value of 1 to 3 passes would be sufficient. Irrespective of what a typical value would be or what was actually performed, it is clear the contract is essentially silent on the issue, and the court holds there was no indication in the contract that conditions would be “normal.” See Ragonese v. United States, 128 Ct.Cl. 156, 159, 120 F.Supp. 768, 769 (1954); cf. United Contractors v. United States, 177 Ct.Cl. 151, 161-64, 368 F.2d 585, 595-97 (1966). There are certainly no affirmative indicators that the soil would not be tough, and the court declines to hold that the absence of an indication of toughness is an implied assurance that the soil was not tough. Servidone is essentially asking the court to compare the actual conditions to what they argue were the normal or average conditions typically encountered in similar projects. Such a comparison, however, is more appropriately advanced in a Type II, rather than a Type I analysis. Defendant’s response to the claim, however, has primarily been that the contract contains adequate indications of the unusual character of the soil, a defense normally more relevant in a Type I claim. Because the court finds below that a Type II condition exists, and that the= contract drawings and specifications were not adequate warning, the difference is probably academic. In any event the claim will be addressed as one for a Type II differing site condition.
2. Type II Differing Site Condition
To prove a Type II differing site condition, the contract requires Servidone to show that it encountered an “unknown physical [condition] at the site, of an unusual nature, differing materially from those ordinarily encountered and generally recognized as inhering in the work of the character provided for in this contract.” GP 4(a). See generally Vann v. United States, 190 Ct.Cl. 546, 572, 420 F.2d 968, 982 (1970); Kaiser Ind. v. United States, 169 Ct.Cl. 310, 315, 340 F.2d 322, 325 (1965); Loftis v. United States, 110 Ct.Cl. 551, 76 F.Supp. 816 (1948). The burden Servidone must carry is heavy because in a Type II case the standard is somewhat amorphous. Unlike in a Type I case, where the contract serves as the basis of comparison, in a Type II case, there is no clear written point of reference. See Charles T. Parker Constr. Co. v. United States, 193 CtCL 320, 333, 433 F.2d 771, 778 (1970).
Servidone alleges that the fill materials at the Joe Pool Lake site were of a highly unusual character; that they were extremely tough in a way that Servidone could not have anticipated; and that this toughness added dramatically to the cost of completion. It also asserts that part of the cause of the toughness of the soils is that they contain poorly crystallized or non-crystalline montmorillonite and calcium carbonate, which, when wet, form a highly unusual inorganic jelly-like substance. Defendant concedes to some extent that the materials were difficult to work with. It contends, however, that the soils were typical of those encountered in the Dallas/Ft. Worth area. Defendant further contends that even if the soils could be characterized as unusual or materially different from those normally encountered, the contract documents provided sufficient information for a reasonable contractor to have predicted the material’s behavior prior to submitting a bid.
a. Was the soil unique in general?
Prof. Sowers described the fill material as behaving like paraffin or grease. *361This is borne out by other testimony, by a video tape prepared by Servidone during construction, and by the numerous photographs taken during construction. Prof. Sowers also stated in his report: “CH soils from the borrow pits proved extraordinarily difficult to place, process and compact. At or near optimum the materials are tough and sticky. They resist all mechanical operations and require extraordinarily large equipment for manipulation.” He testified that despite working with high PI soils before, he had never seen anything like the soil conditions present at Joe Pool. Further, he noted that the earthwork industry, generally, “[tries] to avoid the CH materials in construction where [it] can avoid them.”
Lem Hughes, a fill foreman for Servidone, and earlier for Lane on the initial embankment, had a similar impression. Hughes was shown to have extensive experience in earthwork and construction in Texas, and further experience in Oklahoma, Mississippi, Alabama, Florida, and Georgia. He stated that he had never encountered such material. He noted specifically that while his work on Granger dam in the Ft. Worth district also involved “fat” or high plasticity clays, at Granger they processed out well, requiring roughly 6 passes of a disc plow, versus up to 25 passes for some of the material at Joe Pool.
George McKee and Arland Decker are civil engineers working for Contract Surety Consultants. On the request of St. Paul Fire and Marine Insurance Co. (“St. Paul”), Servidone’s surety, they were sent by their firm to assess the Servidone job to determine whether to recommend that St. Paul provide Servidone with funds to complete the project. Both individuals had significant experience with roadway, dam, heavy construction and other projects involving earthmoving. Each testified that the material at Joe Pool was different than anything he had encountered. Decker, who was allowed to testify by deposition, described the material as both slick and sticky. Both men noted some similarities between the difficulties Servidone was experiencing at Joe Pool and those occurring on the Aquilla dam project, also in the Ft. Worth district. However, both viewed the two materials as distinct, and McKee attributed some of the problems at Aquilla to pre-wetting the material in the borrow areas, which was a procedure that was not used by Servidone on the Joe Pool project. Decker had trouble distinguishing the material at Joe Pool from “gumbo” clay, which he had encountered. He noted such clays can be extremely difficult or impossible to work, stating, “a lot of gumbo clay I’ve seen, you could never push scrapers through the bottom of it. You just couldn’t get enough dozers through it.” He testified that in his experience gumbo clays and “unworkable high plasticity clays” are not used in construction, rather they would be “wasted or something rather than building something stable with them.”
Richard Pierce of Law Engineering was hired by Servidone to consult on problems Servidone and its subcontractor, Downing Brothers, were having in achieving the necessary compaction for the lime stabilized subgrade for the roadway on top of the dam. His analysis and subsequent reports focused on the nature of the soil being used in the subgrade and provided, in the court’s view, the most in-depth assessment of at least a portion of the material in the borrows.5 Pierce stated in his report and testimony that the clay soils at the site were unusual with respect to their toughness and their ability to hold water. He testified he had never seen soils this tough in the Dallas/Ft. Worth area, even though he had performed other jobs in the Eagle Ford Shale Formation, which is the particular geologic formation in which Joe Pool is located. He testified there was only one job he had worked on in which the soils were close to the toughness of those at Joe Pool. That was in Huntsville, Texas, in the Vicksburg Clay Formation.
Defendant’s response to the above evidence consisted primarily of the testimony *362of Dr. Poulos, who was shown to have extensive experience in geotechnology and dam design. Dr. Poulos dug a test trench at the Joe Pool site and performed laboratory tests on the material. He testified the soil was very stiff and that by picking it up one could tell it was very slippery. He further stated that the soil was “very tough.” At trial he viewed a videotape representing site conditions during Servidone’s performance. When asked if he had ever seen similar conditions, he could name only 1 project, the San Justo Dam in California, where the material was of a similar toughness. He estimated that in all his experience he had seen “such clays,” meaning high plasticity clays used in dams, on only 1 or 2 occasions. None of the 5 dams he had designed had used CH clays.
The toughness of the soil is also indicated by the effect on plaintiff’s performance. As noted earlier, the embankment operation was a production job. Servidone’s estimate was based on excavating, transporting, and processing a particular amount of fill material in a given time period. The contractor introduced convincing evidence that the nature of the material, particularly, the toughness of the material, slowed its performance, forcing Servidone to add more and larger equipment to compensate. Fantozzi testified that it is common for a 631-D scraper to be pushed by a bulldozer when loading. At Joe Pool, however, Servidone occasionally used 2 D-9 bulldozers to push the Scrapers through the material. William Gibbons, a government inspector, and acting project engineer for the Corps after October 1984, confirmed that this occurred. Prof. Sowers testified he has rarely seen 2 tractors pushing a scraper, particularly when the tractors were the larger D-9 model. He noted that even with 2 pushcats the scraper was moving at a slower speed than normal. Defendant introduced a Corps engineering manual, Construction and Control for Earth and Rock-Fill Dams, EM 1110-2-1911 (Jan. 17, 1977), DX 12, which pictures a scraper similar to the type used by Servidone on the project. In discussing the merits of the equipment, the manual notes that one of the disadvantages of the equipment is that a “pusher tractor is sometimes required when loading.” The manual clearly anticipates 1 pushcat might sometimes be needed. The court finds use of 2 pushcats is unusual and differs materially from the norm for such work.
Servidone also introduced testimony to the effect that when the scrapers excavated the fill material, the material came into the scraper pan in large curls or slabs due to its cohesiveness. Similar problems, though less severe, were alleged by Servidone to have occurred in the Holland loader operation. Fantozzi testified such problems were typical on the project. Servidone contends this was a result of the unusual toughness of the material and that the effect was to reduce load size and rate of production. Fantozzi indicated that during periods of the contract he counted haul unit loads and compared them to the quantity of fill for which Servidone was subsequently paid. The average load for scrapers and belly dumps was 14.8 and 27 cubic yards, respectively. He asserted Servidone reasonably expected payloads of 21 and 42 cubic yards. Fantozzi, Joe Servidone, and Prof. Sowers testified they had never seen material load in this manner. The court is persuaded that plaintiff’s contentions as to actual payloads and the cause are correct and supported by a preponderance of the evidence.
Gibbons and Coleman, testifying for the Government, both observed the curling phenomenon. Gibbons suggested the curling was due to the natural plasticity and moisture of the material. He suggested that load size could be increased if a thinner cut was taken; however, he conceded this would slow loading significantly by requiring a proportionally longer cut.
Richard Barth, a vice president of Abrams, a competing contractor for the Joe Pool job, who had extensive experience contracting in the Texas area, testified that his company had done an analysis of load capacity relative to LL of clay materials using a 651 scraper. The test was performed on the Granger dam project in the Ft. Worth district. The results showed that a 651, which has a struck capacity of 32 cubic *363yards, produced payloads of 34 cubic yards in low LL clays and 27 cubic yards in high LL clays. Such figures, combined with testimony of both experts that clays with higher liquid limits are generally tougher, supports the argument that Servidone achieved significantly lower load sizes as a result of the toughness of the clay. The court further finds the load sizes materially differ from what would ordinarily be expected.
Servidone used the struck capacity of the 631 scraper (21 cubic yards) in its estimate. John McCreery of Johnson Brothers Corporation, another competing contractor, also used the struck capacity of 21 cubic yards as his estimate on the Joe Pool job. Sears, the Government’s estimator, also used approximately 21 cubic yards in his estimate. He further testified that he had never experienced payloads of 30 percent less than struck capacity.
Defendant offered some evidence that these loads were not unusual. Gibbons had seen similar problems at Carswell Air Force Base in Ft. Worth. However, a scraper was only used on that project a “time or two.” Further, he noted that on other projects on which he had worked, namely those at Reese Air Force Base in Lubbock, and at Segoville, southeast of Dallas, the material had “just drift[ed] into the [belly] pan.” Id. Coleman indicated that he had seen similar curling problems at the Granger, North Fork, and Aquilla dam projects, all in the Ft. Worth district. However, no estimate of load sizes at these projects was offered by defendant. The evidence establishes that scraper load sizes at Joe Pool significantly deviated from the norm for such work and materially affected plaintiff’s rate of production.
Servidone also introduced evidence that the slick nature of the material when moist or wet caused scrapers and belly dumps to have difficulty moving in the borrow areas and the embankment. At times the equipment became stuck and required assistance from pushcats or spreading dozers. McCreery testified that in his experience equipment occasionally gets stuck due to slippery conditions. The evidence showed, however, that throughout the contract period the contractor had difficulty at Joe Pool moving its equipment through the material in several of the borrow pits. Gibbons testified that scrapers didn’t have significant problems moving on the embankment; however, he testified belly dumps had trouble getting on the embankment 25 percent of the time and had to be pushed on the embankment approximately 80 percent of the time. Coleman testified such problems on the embankment occurred quite often, yet when asked if he had seen such problems previously, noted only the Lane contract at Joe Pool. Arland Decker confirmed the above observations regarding belly dumps and noted it was very unusual for an empty belly dump to have problems getting off the fill. He indicated, however, he did not see scrapers having trouble getting off the fill.6
Defendant attempted to show that slipping problems around the borrows were at least partially attributable to the steepness of the ramps leading from the borrow to the haul roads. Decker noted that in the immediate area at the end of the strip the scrapers he observed were running “steep.” However, he noted the haul roads were excellent. In rebuttal, Fantozzi noted that the contract prescribed a maximum grade for such ramps of 8 percent, and no showing was made that Servidone’s ramps were steeper than prescribed.
The court finds that the evidence supports plaintiff’s assertion that mobility over this soil differed materially from what would ordinarily be expected in that its slickness reduced the speed of the equip*364ment in the borrow pits and on the embankment and often caused additional pieces of equipment to be used to push hauling units. The court also finds that there was insufficient evidence that Servidone’s construction of its borrow pits contributed to the difficulty of movement.
Servidone also showed that this material was significantly more difficult to unload from hauling units than would ordinarily be expected. Fantozzi testified that scrapers had trouble unloading on the fill and often needed to be bumped by a bulldozer. This occurred in spite of the fact that the scrapers used were equipped with a hydraulic ejector plate. The hydraulic ejector would be unable to discharge the soil, and the operator would have to back up and try again, sometimes while a bulldozer was bumping the unit in order to dislodge the dirt. This is confirmed by the video tape. Fantozzi testified that he had never seen this occur before. Prof. Sowers also noted such incidents on his visit to the site. Fantozzi further testified that the belly dumps had similar problems. Photographs and the videotape show the belly dumps carrying a full load of fill material with the bottom doors wide open. Fantozzi testified bulldozers would have to strike the belly dumps to get the material to discharge and at times a backhoe dislodge was used to dig the material out. Such action was frequently necessary. Gibbons testified the belly dumps did have trouble unloading, though he would not characterize the occurrence as typical. He testified he never saw a scraper have trouble unloading on the Servidone project.
McCreery testified he was surprised to see material not coming out of the belly dumps on the videotape. Prof. Sowers stated he had never seen scrapers have such trouble and noted the placement process was “much more difficult than ordinary.” Barth testified he anticipated some trouble getting material out of hauling units when he bid this job. However, he also stated that “[t]he main reason was because we had observed that problem on Lane’s job.” He also testified he had experienced problems getting material out of scrapers on projects at Granger Dam, Farm to Market Road 1382, and in Houston on at least one highway job. Farm to Market Road 1382 was a road project at the Joe Pool site. Barth did not testify such occurrences were ordinarily encountered in dam construction.
Defendant also offered testimony by Coleman and Gibbons to the effect that some of plaintiff’s unloading problems were attributable to the depth of cut of Servidone’s scrapers and Holland loaders. Defendant did not contend that Servidone took particularly large cuts; rather, the tenor of the testimony was that smaller cuts would have assisted unloading. For example, Gibbons testified that while Servidone’s scrapers were taking predominantly 8-inch cuts, he had seen the contractor at Aquilla dam taking 4 to 6-inch cuts. The court finds that without some showing that the cuts Servidone was taking were significantly or unusually large relative to general experience, the above evidence is simply a suggestion of a remedial measure that Servidone might have taken, and thus is not relevant to the issue of material difference. Further, Coleman noted that even with the use of smaller cuts at Aquilla the curling and cohesion problems remained, though to a lesser extent. And, he noted reducing cut thickness may not increase efficiency where smaller cuts cause a greater loading time.
The court finds that the scrapers and the belly dumps had unusual problems unloading a significant portion of the time, and that such problems are not ordinarily encountered in dam construction.
Servidone also showed it encountered problems in processing the material which significantly deviated from those ordinarily encountered in such work. Fantozzi testified that processing of the fill commonly required 10 to 15 passes of the disc plow. His previous experience in similar work had been that 1 to 4 passes was sufficient. Lem Hughes, Servidone’s fill foreman, testified “some of it I can get up there in 5 or 6 passes and get it. Some of it takes 10, 12, and some of it out there I’ve had to run anywhere from 18 to 25 passes on it.” At Granger dam it had taken roughly 6 passes *365to process the material. Decker testified that on his later visits he observed the disc plows making 6 to 12 passes. He also noted Servidone was having trouble getting the proper moisture. Coleman saw Servidone making a “large” number of disc passes “throughout” the project on “various occasions.” The specifications do not require a particular number of disc passes to control moisture; rather, they simply require discing until the proper moisture is reached. The court finds Servidone was making significantly more disc passes than would ordinarily be expected.
Coleman suggested that Servidone over-plowed the fill “in some instances” due to a lack of supervision. Coleman did not describe the standard , by which he estimated the fill was overplowed. As noted earlier, determinations of proper moisture were in the first instance performed on a qualitative, educated guess basis by Servidone’s personnel on site. The court finds that Servidone’s judgment was reasonable. It certainly had little incentive to overwork lifts. Consequently, the court interprets Coleman’s testimony, in part, as support for the proposition that Servidone was required to make, in Coleman’s view, an unusually large number of disc passes to achieve moisture requirements.
Servidone also showed it was required to bring additional disc plows on site and more powerful equipment to pull them. Fantozzi testified that he had originally planned to use 2 disc plows running 1 shift per day to process the material at Joe Pool. The Government estimator, Sears, estimated using 3 disc plows running 1 shift per day. McCreery projected the use of 2 disc plows. Barth did not indicate how many disc plows Abrams estimated using on the Servidone project. Fantozzi testified further that as the contract progressed the 2 plows proved grossly insufficient to maintain an appropriate rate of production. Citing a foreman’s report dated August 20, 1982, he testified Servidone had 5 disc plows by then and as many as 7 later in the contract. Servidone started a night shift on August 23, 1982, exclusively to process additional fill. This testimony was generally confirmed by other evidence. Gibbons saw hauling units waiting for disc plows to finish processing previous lifts “which means [Servidone] was hauling more material than could be processed.” Coleman noted the night shift was used solely to process material. Decker noted that Servidone was using considerably more disc plows on his second visit and was having trouble manipulating and processing the material.
D-9 bulldozers were used to pull the plows because the D-8’s dragged through the material, and suffered more frequent mechanical breakdowns related to overloading, such as failure of torque converters and transmissions. The evidence showed it was very unusual to use a D-9 to pull a 36-inch Rome disc plow of the type used by Servidone. James Ray, a marketing manager for Rome Industries with experience using this type of equipment, stated that he had never seen disc plows of the type used by Servidone pulled with a D-9 bulldozer. Rome recommends either a D-6 or D-7 for the TRCH-16 plow, and a D-7 or D-8 for use with the TRCW-16 plow. Barth had also never seen a disc plow pulled by a D-9. He speculated it might not have been absolutely necessary, but may have simply increased efficiency by increasing speed. However, no direct or circumstantial evidence was offered to support this proposition. Coleman testified that farm tractors were used at Aquilla dam, D-7’s and D-8’s at North Fork dam, and D-8’s at Granger. McCreery and Sears both used D-8’s for disc plows in their estimates for the project. Defendant noted that Rome’s “Harrow Selection Guide” did recommend a D-8 or D-9 bulldozer to pull Rome’s TYH-16 disc, a heavier model than Servidone used. However, Ray testified that the TYH-16 is designed to plow stumps out of freshly cleared areas. It is not recommended for construction jobs.
Servidone was required by the Government to add weights to the disc plows it used for processing. Hudson testified that he ordered Servidone to weight the plows because it appeared to him that the lift thickness being placed was not being pen*366etrated by the plow. The relevant contract specifications provide: “Discs which are too light to achieve the desired cutting depth shall be ballasted or replaced with heavier discs as directed by the Contracting officer.” Contract Section 2D, Technical Provision (“TP”) 9.1.6. Hudson’s direction was not improper, but it is additional evidence of the unusual nature of this material. Joe Servidone and Fantozzi indicated they had never before weighted disc plows. Ray testified that he has never seen a TRCW-16 plow (the largest plow used by Servidone) weighted, and noted there is no provision in the design of the plow for adding weights. He stated he would not recommend weighting it since that particular model has an overhead frame which already adds significant weight. The smaller model used by Servidone, the TRCH-16, has a provision for a weight box. However, he testified that weight is usually added when the ground is “very, very, extremely hard and dry.” He had never seen such a plow weighted on a freshly dumped earth fill. Decker considered that it was “unusual” to ballast disc plows. McCreery of Johnson Brothers stated he planned using the Rome plow with “a small ballast tray” on back. It was not clear the model he was referring to; however, he stated: “It’s unusual when we had to use the ballast boxes.”
The court finds that the weighting of disc harrows, particularly the heavier ones used by Servidone, is an unusual occurrence in such work and is some indication that the conditions differed materially from those ordinarily encountered. In addition to the cost of adding weights to the plows, the increased weight placed additional strain on the towing bulldozers.
Servidone also offered evidence that it took unusual measures to try to deal with the material. It experimented with lining the interiors of the belly dumps with a plastic material to facilitate discharge. This failed, however, to alleviate the problem, as the material adhered even to the plastic. Servidone also used the plastic material on the Holland loader to prevent the collection of dry chunks of material which had been causing excessive belt wear. This modification proved more successful. Fantozzi had never previously taken such measures.
While defendant made an effort to show that some of Servidone’s production problems were due to inexperience or improper methods, such evidence was not sufficient to rebut the evidence that the soil was unusually difficult to work. Although not necessary to that finding, the court notes the testimony of McKee, senior Vice President for Engineering with Contract Surety Consultants. He has a degree in civil engineering, and 30 years’ experience in construction contracting. At the time his company rendered its services to St. Paul Fire & Casualty, he had no interest in keeping an inefficient contractor on the job. After 5 inspections of the Joe Pool site, as well as visits to company headquarters, McKee concluded that Servidone was a capable, efficient contractor, which would do the work as economically as some, and more economically than most.
The above evidence, considered in total, compels a finding that the soil was unusually tough and differed materially from conditions ordinarily encountered or generally recognized as inhering in earthen dam construction.7 This was reflected in Servidone’s inability to excavate and process the soil in a way that could be anticipated,
b. Were the soils unusual by comparison with North Texas soils?
Defendant was not able to contest the uniqueness of the soil. Indeed, it states in its own pretrial brief: “these soils were among the most plastic and, hence, toughest in existence.” Defendant’s Pretrial Submission at 50. Defendant’s real argument with respect to uniqueness is that the *367materials were not “unusual” or materially different when compared with certain other dams in the Ft. Worth District and North Texas generally. The court accepts the premise behind defendant’s argument— that the soils in question have to be unusual for the region in which they are located.
In Charles T. Parker Construction Co., 193 Ct.Cl. 320, 433 F.2d 771, the plaintiff contended it had encountered a differing site condition in the form of hard and abrasive rock. In denying plaintiff’s claim, the court, applying the differing site condition claim, relied on the following:
The agency decision found that hard, abrasive rock was generally recognized and usual in this geographical area; and that hard and abrasive rock is precisely what the plaintiffs encountered. It further found that [plaintiff’s] claimed costs were not causally related to the discovery of unknown and unusual physical conditions “differing materially from those ordinarily encountered and generally recognized as inhering in work of the character provided for in this contract.”
Id. at 334, 433 F.2d at 778.
Professors Nash and Cibinic in their treatise, Administration of Government Contracts (2d ed. 1985), suggest a similar test: “Unusual conditions are judged by the ‘normal conditions for the area’, Layne Texas Co., IBCA [No.] 362, 65-1 B.C.A. [(CCH)] 114,658 (1965). The condition must significantly deviate from the norm for the area and the type of work, Unitech, Inc., ASBCA [No.] 22,025, 79-2 B.C.A. [(CCH)] 1113,923 (1979).” Nash & Cibinic, supra, 386. To similar effect is Paccon, Inc., ASBCA No. 7,643, 62 B.C.A. (CCH) 113,546 (1962). While the decisions and commentary do not indicate how narrowly the geographic area in question should be drawn, the court is persuaded that Servidone cannot rely on its lack of experience with very heavy clays in general, or its lack of experience in North Texas in particular, as a reason to use a very broad comparison of soils.
Defendant suggests that the scope of reference be North Texas, and particularly the Dallas/Fort Worth area, which it asserts is known for having soils with high liquid limits. Even assuming, however, that the degree of unusualness of Joe Pool soils is measured against other soils in North Texas, Servidone has met its burden. Defendant attempted to show that the material at Joe Pool was the same as that used in other dams in the Ft. Worth district, relying principally on an assessment of the various liquid and plastic limits taken from borings at 7 dams in the Ft. Worth district. The court finds first that the character of the soil at Joe Pool varied significantly from the other dams on a qualitative basis not fully reflected in the Atterberg limits. Much of this evidence has been discussed above. Thus, what follows is a summary discussion.
Lem Hughes, who had extensive experience with earthmoving work, stated the materials were different than any he had experienced in Texas. Further, he stated that the material at Granger dam, which defendant uses to compare with Joe Pool, was much easier to process. Richard Pierce testified he has never seen soils this tough in the Dallas/Ft. Worth area. He noted also that he had performed several jobs in the Eagle Ford Shale formation in which Joe Pool is located and has never seen soils this tough.
Claude Wise, Servidone’s quality control chief, had experience as testing engineer on a project at the Dallas/Ft. Worth Airport, “[m]oving a massive amount of fill material, construction of approximately five miles of runway, another ten miles of taxi ways, various aprons, three terminal buildings, and everything that goes with a modern airport.” He testified that while the project had liquid limits up to the “mid 70’s,” he had never seen excavation or processing problems like those at Joe Pool.
Jay Taylor testified regarding the experience of Downing Brothers, a firm which was a subcontractor to Servidone on the lime stabilized subgrade. He was general superintendent of the company during the contract period. He has performed lime *368stabilization work on more than 100 projects, including highways, airports, subdivisions, dams, and roads. Downing Brothers primarily does work within a 200-mile radius around Waco, Texas, which includes the Dallas area. Lime stabilization requires using a pulver mixer to grind up the clods and break down the soil so that a certain amount of it will pass through a quarter inch screen. Pulverization is similar to discing soil. Taylor testified that in his experience this normally took 3 to 5 passes. On the Servidone job, however, this took 25 passes. He believed this was a result of the high plasticity of the soil. Although he had worked with high PI soils before, including Waco dam (1 of those designated by defendant for comparison on DX 190), he had not encountered the same problems with those soils that he had at Joe Pool.
Gibbons, testifying for defendant, noted that he had seen soils behaving in a manner similar to those at Joe Pool at Carswell Air Force Base near Ft. Worth. But he had also seen material “drift” into the pan of a scraper at Segoville, near Dallas. Coleman had seen curling problems in scraper loading, similar to those at Joe Pool at Granger, North Fork, and Aquilla dams; however, when asked if he had seen similar problems with belly dumps unloading he had seen this only on the Lane work at Joe Pool. Further, he testified that he had seen farm tractors towing disc plows at Aquilla, D-7’s and D-8’s at North Fork, and D8’s at Granger. As noted, no one indicated they had ever seen D-9’s used to pull disc plows anywhere in the Dallas/Ft. Worth area. Coleman had never seen water trucks get stuck to the extent they did on the Servidone contract. While he testified this occurred because the spray bars “sprayed water directly on the drive wheels,” a photograph dated June 28, 1985, (DX 24) shows a truck with a spraybar that clearly directs the water well away from the wheels. The same appears on the video tape.
Abrams had similar difficulties at Granger dam in certain high LL clays it encountered, but Barth stated that his company was dealing primarily with low plasticity CL clays on that project. Both Decker and McKee of Contract Surety Consultants had seen similar scraper problems at Aquilla. However, McKee attributed some of the problem to prewetting the borrows. Further, each testified that they had never encountered material like that at Joe Pool, even considering Aquilla.
In sum, the anecdotal evidence strongly suggests a finding that Joe Pool was very unusual even compared with other North Texas Corps projects. Such a finding is even more strongly supported, however, by defendant’s evidence based on the Atterberg data.
Both Prof. Sowers and Dr. Poulos agreed that LL and PI are some indication of the toughness of a particular soil. Dr. Poulos stated that, in general, the higher the LL of a soil, the more difficult it is to work. He said the same thing concerning the PI. Prof. Sowers, for example, testified that “as the plasticity index goes up [] soils become more resistant to [ ] moisture change,” and are “nastier ... to work with.”
The court is persuaded by the expert testimony, as well as that of the contractors who were familiar with the use of LL data, that there is a close, but not direct, correlation between LL and PI, on the one hand, and toughness, as that term was used throughout trial, on the other. As those 2 measurements increase, soil will be more cohesive and slippery.
The court is not convinced, however, that the degree of toughness of a particular soil is as predictable as defendant suggests. The testimony of Prof. Sowers indicates that while a soil as a general matter may become tougher as PI or LL increases, it is a qualitative, rather than quantitative, analysis and is not the type of information which would have allowed a typical contractor to predict the nature of materials encountered on this project. He also doubted that a contractor, given the information as it appeared in the Servidone documents, would go through the exercise of calculating and plotting the data as suggested by defendant.
*369Prof. Sowers and Pierce both testified that mineralogy plays a significant role in the behavior of material as well. Pierce was asked if it is generally true that the more highly plastic soils are always tougher: “No. That’s not necessarily true, because it depends a lot on the clay minerals present; and, the Atterberg limits is really not a direct measurement of that.” He noted that x-ray diffraction showed that a large percentage of the clay in the soil at Joe Pool was a smectite clay, which is a type of montmorillonite clay made up of very, very small plates. He noted the effect of such plates on the ability of soil to hold moisture and stated, “[t]his soil just seems to have unusual properties in its ability to hold water.”
It became apparent that a more precise way of predicting the toughness of a soil is through the application of LL and PI data to the A-line or U-line analysis. The Aline/U-line graph was devised by Prof. Casagrande at Harvard University in 1947. Prof. Casagrande is widely recognized as one of the pioneers in soil classification research. The vertical axis of the graph is the PI and the horizontal is the LL. Prof. Casagrande plotted various soils on the graph and, based on previous experience of his own and others in the field, developed dividing lines to differentiate relatively broad classes of materials according to their PI and LL values. The primary dividing line is the “A-line,” which, as shown on DX 187R, begins at LL = 20, PI = 0 and continues across the chart with a positive slope of approximately %. All materials falling above the A-line are considered clays and all those below the A-line are considered silts. The clays are further subdivided by a line described by LL = 50. Those clays which have a LL of 50 of greater are described as CH clays. Those clays having a LL less than 50 are described as CL clays.
A second dividing line on the graph is the U-line. This line is a purely empirical one which, when originally plotted by Casagrande, approximated the uppermost limit of all data he had collected. The line begins at the points LL = 8 and PI = 0 and rises toward the right upper comer of the chart with a positive slope of 9/io. This line is noted in the Corps of Engineers manual, Laboratory Soils Testing, EM1110-2-1906 (Nov. 30, 1970), DX 14, which states at III — 13, 14, “A plot of liquid limit versus plasticity index for natural soils has never been known to fall above the upper limit line.” The manual suggests that the line may be used to spot check lab tests, since presumably any data point falling above the line would be an error. Both Prof. Sowers and Dr. Poulos testified that soils subsequently have been encountered which plot above the line.
Dr. Poulos testified that a soil with a high LL, yet which is not close to the U-line, would be less difficult to work than a LL soil which is on the U-line. While an elevated LL or PI can independently contribute to the toughness of a soil, proximity to the U-line is a more accurate predictor.
Dr. Poulos testified that when a soil is very close to the U-line, it very quickly gets soft and slippery when wet. When dried, it gets extremely hard. He also noted that while a material generally becomes more difficult to work as the PI and LL increase, a soil with a significantly lower LL and PI may still be more difficult to work than another with higher values if the first is closer to the upper limit line. In other words, a soil with a low LL may have relatively high toughness if it is near the upper limit, although generally materials are harder to work as they go up parallel to the U and A-lines — that is, when both PI and LL increase simultaneously.
The U-line on Prof. Casagrande’s original chart, DX 188, represents the upper limit of all data through approximately 1960. The data were not limited to the United States, but included material from other countries. DX 187R is a U-line analysis of borings on the Servidone project and those of 7 other dams in the Ft. Worth district that have significant CH clay content. These 7 dams were selected by defendant from DX 190, a list of 21 dams in the Ft. Worth district, and they appear to have involved the most difficult soil of these 21 projects. The yellow zone marks the Joe Pool data. This exhibit demonstrates *370graphically that Joe Pool soils represent the worst of the worst. Not only do most of the data points appear to rest at the higher LL values, but the lower LL data points tend to fall on or above the U-line. By comparison, the data for the other 7 dams included in the comparison show that the bulk of data points for those dams falls below a LL of approximately 65 percent. Perhaps as much as half of the data points at Joe Pool falls above that number. Further, even at the lower LL, the materials for the 7 other dams range well down below the U-line, whereas the materials at Joe Pool hug the upper limit line.
Dr. Poulos cautioned that in determining how the material will behave in the field, one should be careful about distinguishing a few points. He also testified that in his view the soils at the other dams were not significantly different from the CH material found at Joe Pool. However, while that may be true on a dam by dam comparison, in comparing the dam to the norm created by the others in the area, Joe Pool is by far at the worst end of the scale, particularly considering that defendant has chosen the 7 worst of the 21 North Texas dams for comparison.
Dr. Poulos further testified that the other dams would each have a significant portion of their data points in the yellow zone, possibly as much as 70 percent. It is clear, however, that the bulk of these points would fall at the lower LL levels, suggesting they would still be less tough than Joe Pool. For example, Dr. Poulos cited Granger dam as being the “upper limit”. Although reaching a LL of 80 percent rather than 90 percent, “it covers the entire range of Joe Pool otherwise.” However, Barth, who worked at Granger for Abrams, testified that the contractor used mostly CL clay materials. Consistent with previous testimony that CL clays are easier to work than CH, Lem Hughes testified that the Granger materials processed reasonably well.
The court notes finally that Dr. Poulos acknowledged that the materials which he observed at Joe Pool were as tough as any he had encountered. He further explained that the only clays he had experienced which were more difficult were sodium bentonite clays, otherwise known as montmorillonite. The record reflects that there is montmorillonite in the soil at Joe Pool.
Conditions have been considered unusual where “anticipated materials have appeared in higher proportions or degrees than reasonably anticipated,. . . .” Nash & Cibinic, supra, 386 (citations omitted); accord Paccon, Inc., 62 B.C.A. 113546; Ballenger Corp., DOT CAB Nos. 74-32, 74-32A, 74-32H, 84-1 B.C.A. (CCH) If 16,973 (board found that significant amounts of montmorillonite in soil contributed to differing site condition). Both the qualitative assessments made by individuals experienced in the Dallas/Ft. Worth area and the Atterberg limits analysis of Dr. Poulos show that the materials at Joe Pool were unusual and differed materially not only from other soils generally, but also with respect to soils generally encountered in the Dallas/Ft. Worth area.
c. Should Servidone have anticipated the soil conditions?
Defendant never effectively countered the assertion that the materials used at Joe Pool were extremely unusual in their toughness. Rather, it tried to show that the behavior of the soil was not remarkable considering its high liquid and plastic limits. This position became evident while Dr. Poulos watched the videotape showing the equipment operating on the project. When viewing a bulldozer pushing a belly dump he was asked whether that was unusual and replied, “It’s not unusual, in this kind of material____” He noted further he would have expected this “based on the plans and specifications.” He explained that the reason he did not find these high liquid limits and high plasticity indices unusual is because he was familiar with the kinds of soils in the area from his other work. The work referred to was the soil classification work Dr. Poulos performed with Prof. Casagrande. When asked whether it was unusual to use clays of this high degree of plasticity in dam construction, he stated:
*371It is not unusual in this sense: A dam, an earth dam, is built out of local materials. One doesn’t tend to haul in materials from a long distance. Therefore, when one designs a dam in Texas where the soils are highly plastic like these soils are, one uses the soils that are present. Therefore, when you keep asking about how unusual it is, it’s not within the experiences which I have designed a dam. I’ve given you the description of my experience____ [0]ne simply can’t do anything else in any economical fashion but to use these soils.
This testimony is pertinent because even if the soil was very unusual, Servidone cannot recover unless it demonstrates that conditions at Joe Pool were unknown — that is, that a reasonably prudent contractor would not have anticipated the conditions from a review of contract data, see P.J. Maffei Building Wrecking Corp. v. United States, 732 F.2d 913, 917 (Fed.Cir.1984), or readily discovered them upon investigation or inquiry. See Stock & Grove, Inc. v. United States, 204 Ct.Cl. 103,119, 493 F.2d 629 (1974); Kaiser Ind. v. United States, 169 Ct.Cl. 310, 340 F.2d 322 (1965). The contractor is assessed with the knowledge that a reasonable investigation would have produced. See Vann v. United States, 190 Ct.Cl. 546, 571, 420 F.2d 968 (1970); Nash & Cibinic, supra, 377. This review does not necessarily require a contractor to hire a geologist to interpret data. Stock & Grove, 204 Ct.Cl. at 119, 493 F.2d at 637; PHL Contractors, IBCA No. 874-11-70, 73-2 B.C.A. (CCH) If 10,293 (1973).
The boring logs contained information about the LL and PL of the soil, which, in turn, could have been used to generate the PI. Defendant argues that a reasonably prudent contractor would have known that the high LL and PI values in the specifications were a warning of very difficult soil. Although in its post-trial brief it does not also contend that Servidone should have performed the A-line U-line analysis, that is implicit from the extensive testimony defendant adduced on that issue.
The court has found that there is some correlation between the Atterberg data and toughness, and that this correlation is made most apparent by an A-line U-line analysis. The latter information was not furnished. The inquiry thus shifts to whether a reasonable, prudent contractor would have been aware of the significance of the data that was furnished, namely LL and PL information.
Defendant points out that Barth was familiar with the terms LL, PL, and PI. He testified that he viewed elevated LL and elevated PI to mean virtually the same thing, i.e., the net effect is to make soils more difficult to work. Soils in the range of a PI of 60 or higher he considered would be “very difficult”. He had heard of the U-line, although he did not understand its operation. Even plaintiff’s witness, Jay Taylor, superintendent for the lime subcontractor, testified that the relationship between a high PI and toughness of the soil is well known in the excavation industry.
Sears, the government estimator, testified that he was aware that PI is descriptive of what might be expected of a material. He indicated he was aware that PI was an indicator of how hard it is to get moisture into and out of a material and so might affect cost. However, he testified that in looking at boring information like that on the Servidone project, he would not determine the PI for any of the borings. Rather, he takes “the test results which are presented and tr[ies] to get an overall understanding, a comprehension of the job and the problems or the ease or the difficulty that might be encountered.” He would not do any mathematical manipulation with the data. Further, he did not know what an A-line U-line graph was.
John McCreery, another contractor with significant experience in earth moving, had great difficulty locating the LL information on the specifications, although that is not fully reflected on the transcript. This was consistent with his testimony. It was obvious that he places very little weight on LL or PI in evaluating a project. He showed no ability to determine PI from the boring data. Nonetheless, he did indicate an awareness that higher LL materials tend to curl and stick.
*372Fantozzi testified that, at least at the time of the bid, he did not understand that there was any correlation between PI and the difficulty of breaking apart soil. In his view, the proximity of a soil’s moisture content to ideal would be of more importance to him than would LL.
The court is presented, therefore, with a range of reactions to the Atterberg data. At one end is Barth, who understood both how to calculate PI and applied PI and LL in evaluating a bid. At the other end is Servidone, whose witnesses indicated no knowledge of a connection between high LL and toughness. None of the lay witnesses expressed any meaningful knowledge of the A-line U-line analysis. With respect to the more basic Atterberg data, LL and PI, the court notes that the 3 individuals whose experience is primarily with Texas soils — Barth, Taylor, and Sears — expressed the greatest knowledge about the Atterberg limits, and expressed the greatest concern about the effect of high liquid limits. As witnesses, the court was most impressed with Barth and Taylor. They were both forthright, and gave no indication of trying to color their testimony to a particular result. They were both generally familiar with work in the area. More particularly, the court accepts what Barth or Abrams would do when going to an entirely new area for excavation work as strongly indicative of what a reasonable, prudent contractor would do.
The court concludes that Barth and Taylor can be used as a measure of what the reasonable contractor, bidding on a project in North Texas, would have known about the general correlation between high LL and difficulty in working the soils. Even McCreery, whose work was primarily outside of Texas, had some knowledge about the connection between plasticity and the phenomenon of which Servidone complained.
In addition, plaintiff should have known about the general difficulty of working in soils in North Texas, or else should have informed itself about local conditions and thus would have discovered that soils in the area are tough to work. The court bases this conclusion in large part on the testimony of Barth. He described the process Abrams goes through when it bids a job in a new geographical area. For example, he considered the Abrams bid at Granger to be in a new area in that Abrams was a contractor from “West Texas and New Mexico, five, six hundred miles away from Granger Dam.” He noted weather conditions and soil conditions were very different. He stated that when faced with such a project “we would want to get familiar with just the general working area. We would go around and observe other construction projects under way.” Further, he noted that “ideally” Abrams would bid a few jobs in the area very conservatively, to get a feel for the competition, or bid some small jobs. Abrams had bid jobs in the Central and North Texas area including dams over several years prior to the Granger bid. He stated he would be very reluctant to jump into a new area on a job the size of Granger, which he noted was comparable in size and scope to Joe Pool.
Barth testified that Abrams had done a site visit when it bid the Lane project. He testified that Abrams’ normal procedure was to dig holes for visual observation of the material and “[to see] how it digs, and how the holes stand up and whether there’s water there____ We would take samples, occasionally, probably not as extensively as the Corps [had] done, but we would take samples and take ’em to a lab and have some liquid limit tests run.” He stated “our visual observation of the material would give us confidence or concern, as the case may be, about how it would excavate and how the scrapers could excavate it and what kind of loads they could carry.” J.D. Abrams’ Lane tests essentially confirmed the conditions shown in the drawings, namely, that they were dealing with high LL clays. Barth did not testify, however, that the test holes gave him cause for concern. When asked if, according to his subsequent observations of Lane’s performance, the conditions on the Lane job appeared to be unusual, he stated: “No, they didn’t. I think they were representative of what we had at Granger, and what Holloway had at Aquilla.” Barth also stat*373ed that as a general matter when inquiries are made to Corps officials during bidding, “the usual answer is do whatever it says in the plans and specs.”
With respect to Servidone’s site visit, Fantozzi testified that he visited the site with Anthony Giordano and Tom Spagnoletti, also of Servidone. Giordano is a professional engineer and contractor who was then Vice President of Servidone. Fantozzi testified that Spagnoletti was a “two-year college graduate with about fifteen years’ experience.” They toured the entire project site along with Gary Hames, an employee of Servidone, and then remained on the job site for the remainder of the day. “We went to the borrow areas ... looked at the material that was there. We didn’t bring a backhoe or we didn’t excavate anything. We didn’t really feel there was a need to; the borings were quite, you know, were pretty extensive on this job. And we basically — what we saw in our investigation was what we had previously learned from the plans and specs.” Fantozzi noted: “[W]e didn’t... try to verify a boring log because we have worked on many Corps of Engineers’ projects in the past, and the borings that are given by the Corps have always been ... accurate and good borings____” He did remember picking up and handling material in the borrows. They observed Lane’s project, and viewed an embankment which Lane had constructed, but there was no embankment work going on. They did not speak with any Lane employees. Fantozzi noted that Lane had also taken out bid documents for the completion phase of the Joe Pool project and stated Servidone “wouldn’t really talk to Lane ... a competitor about how they’re going to bid the job or anything like that.”
Despite the fact that Servidone had never performed any jobs in Texas before Joe Pool, Fantozzi did not recall asking anyone in the Dallas/Ft. Worth area about working with the soil there. Servidone did not cut a test trench or otherwise attempt to assess the reaction of the soil with equipment. The adequacy of Servidone’s site investigation must be measured against what a reasonable, intelligent contractor, experienced in the particular field of work, would discover. See Stock & Grove, Inc., 204 Ct.Cl. at 119, 493 F.2d at 637; Foster Constr., 193 Ct.Cl. at 614r-15, 435 F.2d at 887-88; McCormack Contr. Co. v. United States, 18 Cl.Ct. 259, 263 (1989). By comparison with what Barth and others did, the court is left with the distinct impression that Servidone’s 1-day, somewhat desultory site visit was inadequate, particularly given the vast majority of the contract work required excavation and processing of soil.
Servidone had no prior experience in Texas, and no significant experience with predominantly CH soils. The soils it was faced with indicated very high liquid limits and by far the bulk of the job involved working with dirt. Servidone is charged, based on the analysis above, with knowledge that a high LL is an indicator of toughness. It was faced with an acceptance curve it had never heard of before— the LLC.C. Under these circumstances, a reasonable, prudent contractor would have further informed itself about local anomalies in excavation and construction. In this case, even a minimal inquiry to others in the Dallas/Ft. Worth area who were involved in earthmoving projects would have put Servidone on notice of the unusual character of North Texas soils generally. Servidone was not limited to talking with Lane employees about conditions in the area. The court heard much evidence regarding the nature of soils in the North Texas, and it was apparent that virtually anyone who had performed work in that part of Texas recognized that the area has a great deal of difficult CH clay material. Gibbons, for example, testified that he understood all clay in North Texas is slippery when it is wet. Dr. Poulos stated that even by picking up and handling the soil, one would notice it was sticky.
Servidone was also imprudent in not inquiring as to the significance of the LLCC and in not informing itself about the effect of high liquid limits. Fantozzi testified that if he sees something in specifications which he does not understand, he asks about it. The court agrees that prudence *374dictates that course of action. Servidone employees, however, deviated from that practice in this case. The contract clearly warns that acceptability of in place moisture content will be “determined by the ‘LIQUID LIMIT CORRELATION METHOD’ ”. Section 2D, TP 10.3.2. See also id. at TP 8.1 & 10.1. The contract contains materials descriptions which are explicit in terms of the range of LL levels. Section 2C ¶ 5.
The court finds that Servidone did not act reasonably in preparing its bid. The relevant inquiry, however, is how greater caution should have been reflected in Servidone’s bid. The court has found that Servidone should have anticipated some of the difficulty of the soils at Joe Pool, either because of the elevated LL levels, or because a reasonable contractor would have exercised more diligence and caution if it had no experience in North Texas. The standard of reasonable anticipation does not require a contractor to expect the worst, however. See Redman Service, Inc., ASBCA No. 8,853, 63 B.C.A. (CCH) 113,897 at 19,338. Indeed this would be contrary to the purpose of the differing site condition clause. Cf Foster Constr., 193 CtCl. at 613-14, 435 F.2d at 887. As discussed above, a reasonable contractor generally would have been aware that soils with high LL levels, and soils in North Central Texas generally, would be relatively tough to work.
Despite concluding that Servidone was unreasonable in failing to appreciate the warning offered by the elevated LL’s, it does not follow that there was not a differing site condition. The court has also found, and reiterates here, that the connection between LL and toughness is not a direct one. Moreover, the mineralogy of the soil affects toughness in ways not fully anticipated by the LL and PI readings. There was evidence, unrebutted by defendant, that montmorillonite was present in the Joe Pool soils and that, as Dr. Poulos expressed it, montmorillonitic clays are the worst he is acquainted with. Although the boring logs reflected the presence of that material, there was no evidence or argument that a typical contractor would appreciate the attendant risk.
The court has also found that proximity to the U-line is a better predictor of toughness than simple LL and PL data. The court is puzzled, however, by defendant’s offer of evidence concerning how Joe Pool clays plot against the U-line. As discussed above, that evidence supported Servidone’s argument about the uniqueness of the soils. Even though such charts had been developed by the Government prior to the solicitation, see Design Memorandum No. 9, at 88, the Corps’ charting of the Joe Pool Proctor data against the A-line and U-line was not provided to bidders. Even if it had been, it is far from clear that a reasonable contractor would have appreciated the significance of the graph without some accompanying explanation of the Corps’ analysis of its significance as expressed at trial through Dr. Poulos — namely, that the graph predicts extremely tough soils.
Nor has defendant contended, much less established, that a reasonable, prudent contractor would have developed its own plot of the Atterberg data against the U-line. .. None of the non-expert witnesses expressed any understanding of the A-line/U-line analysis. The court notes that the Corps’ Laboratory Soils Testing manual, DX 14, discusses the upper limit line only as a method for determining test errors. The manual discusses the relationship between toughness and plasticity in a footnote warning that when soils fall on the chart “well above the A-line” (not near the U-line) they will be tougher to roll out for purposes of doing a plastic limit test. DX 14 at III-ll.
In sum, even a reasonable contractor, having done an adequate site inspection, and having general knowledge about the difficulty of high LL soils, would not have been warned about the actual condition of these soils. The court thus holds that Servidone is entitled to recover damages to the extent the condition it encountered could not have been reasonably anticipated. See Nash & Cibinic, supra, 397 (citing Bernard McMenamy Contractor, Inc., ENGBCA No. 3413, 77-1 B.C.A. (CCH) 1112,335 *375(1976); Continental Drilling Co., ENGBCA No. 3455, 75-2 B.C.A. (CCH) ¶ 11,541 (1975)). While the plaintiff acted unreasonably in preparing its bid, it would constitute a windfall to defendant to hold that fact precluded a finding of a differing site condition in these circumstances. The court has found that no bidder, no matter how experienced, would have anticipated the conditions actually found. The differing site condition thus consists of that difference between the conditions a reasonable contractor, represented by Barth, could have anticipated, and conditions actually encountered. In its damages calculations, treated separately, the court quantifies the extent of the differing site condition.
These findings implicate Servidone’s separate argument that the Government breached an implied duty by not disclosing to bidders relevant, superior knowledge of the soil conditions which the Government possessed. In particular Servidone contends that A-line, U-line plots compiled by defendant prior to the contract should have been disclosed.8 Defendant counters that it was under no obligation to disclose information other than that provided. The Government’s failure to disclose to the contractor vital information that the Government possesses and that the contractor cannot get from a source other than the Government may give rise to a claim for breach of contract. American Shipbuilding Co. v. United States, 228 Ct.Cl. 220, 225, 654 F.2d 75, 79 (1981); Hardeman-Monier-Hutcherson v. United States, 198 Ct.Cl. 472, 458 F.2d 1364 (1972); Helene Curtis Ind., Inc. v. United States, 160 Ct.Cl. 437, 312 F.2d 774 (1963).
In American Shipbuilding, the court specified 4 elements of a breach of contract claim based on failure to disclose:
(1) [A] contractor undertakes to perform without vital knowledge of a fact that affects performance costs or duration, (2) the Government was aware the contractor had no knowledge of and had no reason to obtain such information, (3) any contract specification supplied misled the contractor, or did not put it on notice to inquire, and (4) the Government failed to provide the relevant information.
228 Ct.Cl. at 225, 654 F.2d at 79 (citation omitted). An element implicit in the above formulation is that the Government indeed possessed the vital information. See Johns-Manville Corp. v. United States, 13 Cl.Ct. 72, 132-33 (1987) (issue framed as whether the Government possessed knowledge superior to what plaintiff knew or should have known), vacated on jurisdictional grounds, 855 F.2d 1571 (1988).
It is this latter element that is missing from Servidone’s proof of superior knowledge. There was no testimony to suggest that the Corps recognized that the unusually difficult nature of these soils was reflected in the A-line/U-line analysis. For example, while Schmidt testified on cross-examination that soils at the U-line are unusually tough, the question asked his opinion based on his current knowledge, as he “[sat] here today.” The court is left with the definite impression after trial that it was only the experts who fully appreciated the uniqueness of how the Joe Pool data plotted on the graph. The testimony suggests that Corps employees felt this project would be similar to other dam projects in the area. In addition, as the court has held, there is an element of subjective judgment in evaluation of Atterberg data, as well as a degree of general unpredictability in how soils with the same Atterberg characteristics will react.
As a result, Servidone failed to show the Government possessed superior knowledge vital to performance. Cf Helene Curtis Ind., Inc., 160 Ct.Cl. at 444, 312 F.2d at 778 (noting “[i]n 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”.)
Such cases as Hardeman-Monier-Hutcherson, 198 Ct.Cl. 472, 458 F.2d 1364, and Ragonese, 128 Ct.Cl. 156, 120 F.Supp. 768, are inapposite. In Hardeman, plain*376tiff entered into a contract to build a pier in a remote section of Australia. The Government was in possession of studies indicating that the site was subject to extremely severe weather and sea conditions during certain periods of the year. The solicitation contained information relating to winds, rainfall, and temperature, but the severe conditions noted in the reports were not furnished. The Government was held liable when it refused to provide the reports to the contractor upon request prior to bidding. In Ragonese the Government was held liable because it failed to effectively communicate the extent of groundwater at the site. It was shown that the water level in 5 of 16 test borings by the Government was above the level of the sewer to be constructed by the plaintiff. It was further shown that significant water was found in many of the other 11 test borings but that the water level could not be taken because the holes had collapsed. In both these cases, the Government was in possession of specific information which it was aware would have put the contractor on notice of the conditions to be expected.
While the court is unwilling to find that failure to furnish the A-line/U-line analysis was a breach of contract, for the same reasons the defendant cannot rely on that analysis to claim that Servidone should have been aware of the likelihood of extreme toughness of the soils. Either the U-line data are highly useful in predicting toughness or they are not. If they are, then if the Corps appreciated that fact, it should have furnished the A-line/U-line analysis, along with some explanation. If the Corps did not appreciate the significance of the graph, then it can hardly be argued that Servidone should have developed that information on its own and taken it into account.
The following summarizes the court’s findings on the issue of a differing site condition as to the embankment work:
1. The drawings and specifications contained LL and PL information which a reasonable and prudent contractor would have taken into account as general indicators of toughness of the soil.
2. The LL readings on the Joe Pool borings were sufficiently elevated that a reasonable and prudent contractor would have anticipated difficulty in working with the soils.
3. A reasonable and prudent contractor would have taken into account general soil conditions in the North Texas area or would have informed itself about those conditions. If Servidone had done that in this instance, it would have discovered that soils in the North Central Texas area are generally difficult to work, something Servidone did not anticipate in its bid.
4. Servidone’s bid was imprudent in that it did not at all take into account the implications for toughness in the Atterberg data.
5. Although the Atterberg data furnished in the contract was some indication of toughness, the actual degree of toughness of the soil was sufficiently in excess of that indicated in the contract or that ordinarily encountered or generally recognized as inhering in work of the character provided for in the contracts so as to constitute a materially different site condition.9
B. Defective Specifications
In Count IV of the complaint Servidone alleges the Government breached an implied warranty by providing specifica*377tions that were inadequate to achieve the Government’s desired result. Servidone contends in particular that the LLCC was an ineffective means of controlling moisture in the fill.
The underlying question is whether the specifications supplied were suitable for their intended purpose. See John McShain, Inc. v. United States, 188 Ct.Cl. 830, 833, 412 F.2d 1281, 1283 (1969) (“[Ajlthough Government-furnished plans need not be perfect, they must be adequate for the task or ‘reasonably accurate.’ ”). See generally United States v. Spearin, 248 U.S. 132, 136-38, 39 S.Ct. 59, 61, 63 L.Ed. 166 (1918) (discussing breach of warranty); Hoi-Gar Mfg. Corp. v. United States, 175 Ct.Cl. 518, 360 F.2d 634, 638 (1966) (same). Servidone has no right to complain if the project it ultimately constructed was essentially the same as the one it contracted to construct. In this case it was. The dam was built according to specifications. Although the contractor had extensive unanticipated difficulties, that fact does not establish that the specifications were unsuitable. Indeed, for the reasons discussed above, the court concludes that the LLCC was not unreliable, and that the specifications were not defective.
II. HARASSMENT AND OVERTESTING CLAIMS
In Count VI of the amended complaint Servidone avers that the quality assurance testing program carried out by the Government was either a change from the testing procedures specified in the contract or a breach of the Government’s duty to cooperate. Servidone specifically alleges that the quality assurance (“QA”) testers tested at inconvenient times, took non-representative samples, immobilized unnecessarily large areas, and generally performed in a way that disrupted work. It also alleges that the overall number of tests was excessive.
Most of the testimony to support these assertions came from Fantozzi. The court does not discount the accuracy of his overall assessment that the testing appeared irrational on occasion. But the court is persuaded from the testimony of the QA personnel, as well as the fact that the vast majority of soil samples passed the relevant tests, that the instances of taking non-representative samples were isolated by comparison with the bulk of the testing. The same is true of claims that the size of the area retested was too large. Fantozzi was asked, for example, how often the Government tested to a depth of 5 feet, a figure Fantozzi had used. His answer was that it was not a daily occurrence, but it happened on numerous occasions. Testimony of others was that the cuts were not that deep, but that to the extent several lifts were disturbed, it was necessary to test deep because several lifts might have been placed in the interval between rework and retest.
Servidone also attempted to demonstrate that the amount of testing done by the Government was far in excess of that initially contemplated by either party and beyond what was necessary or reasonable. The question raised by this aspect of the complaint is whether the Corps abused its discretion under the contract to assure quality, or failed in its duty to cooperate with the contractor and not hinder, delay, or increase the costs of its performance. See Adams v. United States, 175 Ct.Cl. 288, 296-300, 358 F.2d 986 (1966), WRB Corp. v. U.S. 183 Ct.Cl. 409, 469, 509 (1968).
The evidence does not support these claims. The contract permitted testing as deemed necessary in the discretion of the Corps. Servidone was unable to establish that the actual amount of testing was inconsistent with the Corps of Engineers’ Manual, or with Schmidt’s suggestions to the technicians, or with good judgment. Dr. Poulos testified that the testing frequency was “perfectly reasonable”. The over-testing claim is rejected.
A related issue is Servidone’s claim that certain Corps personnel held animosity towards it and demonstrated that by petty harassment, overtesting, and too strict an enforcement of acceptance requirements. There was clear evidence of unprofessional *378conduct by some Corps personnel or technicians hired on the Corps’ behalf. That conduct was isolated, however, and the particular effects were never tied to any quantifiable injury. See Roberts v. United States, 174 Ct.Cl. 940, 946, 357 F.2d 938, 943 (1966). The Corps has substantial discretion in its decisions related to acceptance. The evidence was insufficient to show that discretion was exercised in an arbitrary way. See Adams, 175 Ct.Cl. at 295, 358 F.2d at 990-91.
In Counts VII and IX, Servidone alleges that defendant required it to employ more QA and contractor quality control personnel than called for in the contract. The court rejects this claim. Contract provisions outlining both QA and quality control state that the personnel described are a minimum number. The CO had discretion under the contract to approve technicians and presumably to waive any deficiencies in a resume. Moreover, Servidone’s complaint that the technicians were not qualified is not linked to any specific damage claim or right to relief. With the exception of Mickey Davis, Servidone never attempted to demonstrate incompetence on the part of the technicians.
III. THE REWORK CLAIMS
In Count VIII of the complaint, Servidone claims that it was required to remove and tear out material which was later found to meet the requirements of the contract in every respect. The complaint avers that the incident occurred on September 27, 1983, and that it submitted a claim for $6,085.40 on October 6,1983, which was denied by the CO on March 26, 1985. (Defendant avers in its answer that the claim was not submitted until November 29, 1983, and that the CO’s final decision was issued on March 21, 1985.)
Servidone offered no evidence at trial relating to this incident. Further, neither the original claim nor the CO’s decision was offered into evidence. As a result, Servidone has failed to carry its burden of proof as to the claim described in Count VIII of its complaint.
A related issue, which apparently is part of the overall embankment claim, is Servidone’s contention that it unnecessarily reworked approximately 150,000 cubic yards of fill. It points to the testimony of Dr. Poulos that he would expect rework in the range of 1 percent to 2 percent of the total embankment, which consisted of approximately 10,000,000 cubic yards of fill. The evidence suggests that the actual amount of rework was in the range of 250.000 cubic yards. Using Dr. Poulos’ low estimate, Servidone concludes that 150,000 cubic yards of rework were unnecessary, and therefore should be separately compensated as a differing site condition. There is no explanation for why the amount should not be based on the high end of Dr. Poulos’ estimate. In any event, the court rejects that claim for 2 reasons. First, there was no attempt made to tie the additional rework to a differing site condition. Servidone merely assumes that the additional work was due to that character of the soil that the court has found to constitute a differing site condition. Based on the plaintiff’s inexperience with heavy clay soils and lack of knowledge of the LLCC, it would appear highly plausible that some of the excess rework was due to something other than a differing site condition.
The second reason the claim is rejected has to do with the way plaintiff has presented its entire case. As discussed below, plaintiff fundamentally has made a total cost claim. In its post-trial brief, Servidone explains that costs associated with overtesting and rework damages are “included in its direct costs under work items 7, 8, 9, 12, and 80.” Plaintiff’s Post-Trial Brief at 72. Servidone explains that the damages associated with this claim can be isolated by utilizing its calculation of the actual costs per cubic yard of fill. Based on $4 per cubic yard, and an estimate of 150.000 excess yards, Servidone seeks $600,000.00 for this claim. It is clear that this amount duplicates part of the costs outlined in its overall damages claim, which makes no mention of this $600,000.00. Because the court is allowing the bulk of the differing site condition claim as to embankment work, it would appear to be inappro*379priate to separately allow a claim for rework.
IV. THE INUNDATION CLAIMS
On 4 occasions during plaintiff’s performance of the contract, portions of the jobsite were flooded. In Counts V and X of the complaint, Servidone claims that the flooding at the site was caused by the Government’s flawed design for drainage in general and the diversion channel in particular, and that groundwater data in the borings was incorrect, further aggravating the problem. Servidone suggests that if site elevations and proper groundwater levels were provided in the solicitation, it would have been put on notice of the potential problem.
In its post-trial brief, Servidone makes clear that it is not making a claim based on excess rainfall. The amount of rainfall actually falling on the site was within the volumes reflected in the specifications. Instead, Servidone argues that the government-designed diversion channel, through which Mountain Creek was temporarily channelized during construction, was inadequate to handle heavy rainfall, given the unusual soil conditions. Servidone’s argument can be summarized without serious prejudice as consisting of its observation of 2 phenomena: 1) The site was seriously flooded on occasion; and 2) the flooding appeared to be above, rather than below, the diversion channel. The assumption behind Servidone’s claim is that the diversion channel with its culverts was warranted to discharge all rainwaters from the Mountain Creek/Walnut Creek watershed upstream of the dam site, without causing backup on the construction site. The court holds that nothing in the contract documents warrants such an interpretation.
The contract contains a clause entitled “Physical Data,” which indicates the historical precipitation at the project site. The contract reflects an average number of 80 days of precipitation per year and an average experience of 35.95 inches of precipitation per year. The anticipated number of weather delay days was 250 calendar days during the contract term. The drawings also contain hydrographs for Mountain Creek over a 17-year period showing that discharge is typically well under 2000 cubic feet per second (“CFS”). There are numerous spikes, however, which reflect that volume frequently reached over 10,000 CFS, and sometimes as high as 35,000 CFS.
Servidone contends that “Defendant offered no proof of the adequacy of the design for its intended purpose.” Plaintiff’s Post-Trial Brief at 58. Defendant is not required to offer such proof, however, since the plaintiff never established a prima facie case of inadequate design. Nothing in the contract documents constitutes a representation that the diversion channel with its culverts would always immediately discharge all rainwater, no matter how high the flow in the creek. Rather, Servidone had to be aware that the very process of dam construction has the effect of impounding water and constricting its flow. Nor did Servidone establish a link between the unusual soils and the flooding. That line of argument fails for lack of proof.
In Count II of the complaint, Servidone raised a related inundation claim. It contends that it encountered a Type I differing site condition in borrow pit Cl closest to the Mountain Creek end of the embankment. In its post-trial brief, Servidone claims that costs associated with the lack of access to borrow pit Cl added $100,-000.00 to its costs to complete.
A portion of designated borrow area Cl had been used by Lane. The pit Lane excavated became inundated with water, and was maintained in that condition and stocked with fish. The contract warned bidders that a portion of Cl was flooded; that the flooded portion had to be maintained as a fish pond; and that any further use of Cl required maintenance of a 50-foot buffer, absent written approval from the CO. Servidone developed its own Cl borrow pit. It alleges that due to flooding, the fish pond overtopped its banks and flooded Servidone’s Cl borrow pit.
The court declines to find, as requested by Servidone, that the flooding of the Cl pit was due to the “condition of the material and in part to the location of the Borrow *380Area Cl.” Plaintiff’s Proposed Findings of Fact No. 62. There was no testimony that the nature of the soil had anything to do with that problem, and the “location” of the pit is nothing more than a designation by the Corps of an area available for excavation. The more important variable in determining whether the pit would flood was Servidone’s own choice about the precise location of the Cl pit it developed, the amount of buffer it left between its excavation and the pond, and the proximity of the pit to Mountain Creek. Those were variables plaintiff controlled. It could have chosen not to use Cl at all. There was certainly no warranty, implied or otherwise, that overtopping would not occur.
The court accepts Coleman’s testimony that at least part of the inundation of Servidone’s Cl pit resulted when waters over-topping Walnut Creek and Mountain Creek were directed artificially into the pit by a drainage ditch dug by Servidone. That testimony was never effectively rebutted. Even assuming that the fish pond was the source of some of the excess water in the new Cl pit, plaintiff has not established that this was caused by a differing site condition.
V. LIME STABILIZED SUBGRADE CLAIM
The contract required Servidone to construct a 5-mile long roadway on the top of the embankment and its approaches. The roadway consists in part of 6 inches of subgrade stabilized by the introduction of lime, which is then compacted. In Count XII of the complaint, Servidone seeks an equitable adjustment for work it performed to achieve compaction of the lime stabilized subgrade. Servidone contends the work was beyond the contract requirements. It offers 3 alternative theories of recovery: defective specifications, constructive change, and differing site condition. Defendant argues that the contract contains performance, rather than design, specifications, and that in any event Servidone did not follow the specifications. It also contends that problems arose because of the chemical properties of the water used to help in curing the lime.
At the outset, the court notes that the specifications are both design and performance based. To an extent not found in the rest of the embankment work, the lime stabilization had to be conducted according to relatively detailed procedures, yet the final result had to be the achievement of the following: 95 percent of maximum compaction (using a modified Proctor test), a certain specified soil particle size, and a 4 percent lime content.
The process of obtaining compaction according to contract specifications involved a number of steps. The surface first had to be cleaned of debris, and the soil opened up. Lime was then to be introduced using either a slurry or dry method. The soil was then mixed, cured for a period of time, and the result tested against required particle size and lime content. If the soil met contract requirements, it was then compacted with rollers to achieve the required 95 percent compaction.
Servidone subcontracted the subgrade work to Downing Bros., a Texas firm with significant experience in lime stabilization. From the outset, Downing had significant problems. The soil did not open up easily; it took extra applications to get the lime content up; the soil did not pass the particle size test until an extraordinary number of passes of the mixer; and the soil would not achieve the specified compaction and moisture content. Downing gave up and eventually Servidone finished the work.
Richard Pierce, a geotechnical engineer with Law Engineering, was asked by Prof. Sowers to consult with respect to Servidone’s problems in achieving lime stabilization. -He prepared a report which expressed his views as to the reasons for the problems and made recommendations. Pierce also testified at trial. The court is persuaded by this testimony, as well as the evidence concerning efforts made by Servidone and Downing to achieve the contract requirements, that while the specifications were not improper, they were extremely difficult to achieve, in primary measure because of the peculiar characteristics of the soil and the high moisture content and *381inadequate compaction of the sub-base. For reasons discussed in connection with the Type II differing site claim on the embankment work, the court finds that the soil differed from that normally encountered due to its toughness, plasticity, and moisture sensitivity, and that these characteristics would not have been anticipated by a reasonable contractor.
There were 2 additional causal factors raised in connection with lime stabilization problems. Dr. Poulos testified that the lime subgrade was too wet. Water makes clay soil, but particularly this clay soil, less susceptible to compaction. Based on the testimony of Pierce and Taylor, however, the court finds that some, though not all, of this excess moisture was due to migration from oversaturated soil within the embankment. Defendant has not contended that Servidone should bear the responsibility for the moisture content of the embankment itself, and in any event the court would decline to so hold.
The second factor was the character of the water Servidone used to make the lime slurry mixture. Technical Provision 2.4 of the contract states that water used in the lime application process must be free of substances deleterious to the lime-water mix. See Section 2L, TP 2.4. In the report of plaintiff’s own expert, Pierce, the inference is drawn that the water used had a high sulfate content which might impede the mixing of lime and water. At trial, Pierce testified without contradiction that the water used was having a “deleterious” effect on the lime slurry operation.
In view of these findings, the court concludes that the plaintiff has demonstrated a Type II differing site condition with respect to the lime stabilization work. A portion of Servidone’s additional expense in this connection can be attributed to over-saturation of the lime subgrade and to the quality of the water it used, however. The Government should not pay for damages to the extent of the latter contributing factor. Defendant never attempted to quantify the effect of the mineral content of the water, yet it was clear that plaintiff had significant problems because of the nature of the soil. The court concludes, by way of a jury verdict, that the plaintiff may recover 90 percent of its proven additional costs associated with lime stabilization work. See Specialty Assembling & Packing Co. v. United States, 174 Ct.Cl. 153, 184, 355 F.2d 554, 572-73 (1966) (not essential that damages be ascertainable with absolute certainty); Locke v. United States, 151 Ct.Cl. 262, 267, 283 F.2d 521, 524 (1960) (if reasonable probability of damages can be clearly established, uncertainty as to amount will not preclude recovery).
VI. MOUNTAIN CREEK II CLAIM
The contract required the filling of the bed of Mountain Creek at points at which it would be covered by the dam. This work was done in 2 phases. In doing the first stretch of the creek bed, it was apparently determined that substantial muck and soft soil had to be excavated from the creek bed. Plaintiff submitted a claim in connection with that extra excavation, and the Corps eventually agreed to pay a reduced sum. In Count XI of its complaint, Servidone claims it was directed by the Government to excavate and place material in a similar manner in connection with the second phase of Mountain Creek. A claim was submitted on September 27, 1985, which was denied by inaction (“Mountain Creek II claim”). In its damages presentation, Servidone seeks to recover $349,-447.00.
Although this claim was not treated extensively during trial, based on Fantozzi’s testimony and the court’s examination of the relevant profile drawings of the excavation, it is uncontested that Servidone was required by contract to excavate to a depth of 460 feet at Mountain Creek. In fact, because it encountered material that was too soft and mucky, Servidone was directed to continue excavating an additional 20 feet, and then backfill. Servidone was paid for the extra excavation and backfill at the contract rate. What it seeks is the difference between the contract rate and its actual expense. Fantozzi testified that costs related to excavating in the creek bed were much higher. The pit created by the additional excavating was steep, narrow, wet, *382and soft. Machinery had a difficult time maneuvering. Processing in the pit during backfill was also more difficult. This is borne out by the photographs and video. Defendant put on no evidence to counter this claim. The court concludes that Servidone carried its burden of proof on the issue of liability.
Servidone’s accountant isolated direct costs, in excess of amounts already paid, associated with Mountain Creek II. Defendant did not successfully rebut that testimony. The court finds that the claim is proven to the extent of the costs identified by Servidone’s accountant.
VII. THE STANDBY CLAIM
As part of its Type II differing site condition claim, plaintiff separately seeks compensation for what it claims are increased costs attributable to equipment standby time which resulted from the unusual soil conditions. The total amount of this claim as of December 31, 1989 was $2,710,067.00. Very little trial time was devoted to this issue. In part, that is because it is inextricably linked, in terms of liability, with the larger Type II differing site condition claim as to embankment work. Support for the claim came primarily from Servidone’s expert accounting witness, Craig Jacobsen. Jacobsen merely compiled hours during which equipment was or was not used. He admitted that he had “no way to determine standby as attributable to a specific event on the job____” He utilized the company’s books and records with respect to daily equipment use and from them could determine whether equipment was idle. He extracted a total of 160,230 hours of total standby time. After eliminating standby time separately claimed elsewhere, he multiplied the total hours during which Servidone-owned equipment was in standby status, times one half the equipment ownership rate, and incorporated leased equipment at its actual hourly rental cost.
The court has very little confidence in these calculations, not because the records are inaccurate or because Jacobsen, who is obviously a skilled accountant, erred. Rather, the weakness in this entire claim is the absence of any meaningful proof of a link between down time and the condition of the soil.
It is not accurate to conclude, as plaintiff does in its post-trial brief, that “Coopers & Lybrand and Fantozzi reviewed plaintiff’s equipment use records to determine the amount of increased standby equipment hours resulting from the nature of the soil at the site____ Coopers & Lybrand identified 160,230 hours when plaintiff was not able to use its equipment 40 hours/week due to the nature of the soil at the site.” Plaintiff’s Post-Trial Brief at 61 (emphasis added). The equipment reports, attached to the contractor’s daily quality control reports, merely reflect if a piece of equipment was used or idle for a given length of time. There is no way Jacobson, from these hundreds of daily reports, could limit the standby hours claimed to the effects of the soil, and he made no effort to do so. Although Fantozzi said he “assisted” in preparation of the claim, he merely furnished standby hours to Jacobsen on individual pieces of equipment. Jacobsen admitted that he made no distinction between equipment that was standby, i.e., waiting to be used, but useable, and equipment that might have been surplus, i.e., not needed at all.
In sum, there is no affirmative evidence that the equipment standby costs were directly attributable to soil conditions. While some of the equipment idle time was no doubt attributable to standby, and while some of that standby may have been attributable to soil conditions, Servidone has provided no clear proof of that, and in any event, has not given the court the means to isolate the relevant standby hours.
VIII. MOBILIZATION CLAIM
Servidone claims $335,577.00 as the cost of unnecessarily mobilizing and demobilizing equipment. The court is unable to find any explicit articulation of this claim in the record. While Jacobsen extracted certain equipment costs which comprise the present damages claim, he testified that he was instructed by Fantozzi which costs to *383look for. Specifically, Jacobsen explained that these were costs for “substantial additional equipment [that] was brought onto the project.” Fantozzi, however, never presented to the court an explanation for why liability should be assessed at all, much less beyond that already allowed in connection with the embankment claim. The claim fails for lack of proof.
IX. INTEREST
Plaintiff is entitled to collect interest at the statutory rate on any recovery as of the date of receipt of the relevant claim. 41 U.S.C. § 611. The principal embankment claim was received on March 1, 1984. The lime stabilization claim was received on December 16, 1985. The Mountain Creek II claim was received on September 30, 1985.
As an additional element of the underlying cost of the changed condition, however, Servidone also seeks recovery of interest it paid on borrowings during the course of the contract. At trial, plaintiff established that it borrowed over $30,000,-000 during the period of performance of the contract. Although defendant contests the assertion that all of this borrowing can be traced to the Texas project, for the court’s present purposes, it is sufficient to state that, at a minimum, a substantial portion of these borrowings is attributable to the contract in question. (The court recognizes that the question of whether these borrowings can be traced to the differing site condition is a different one.) At the time of closing argument, cumulative interest on the total amount borrowed was over $13,000,000. Servidone seeks recovery of that entire amount.
Two questions are presented by this element of plaintiff’s claimed damages. First, does the law recognize interest on borrowed funds as an element of costs? If it does, has Servidone proven that its interest costs are attributable to the differing site condition?
Servidone has not persuaded the court that the law accommodates the claim. As defendant correctly points out, as a general matter, interest is not allowed on monetary claims against the Government unless “plainly authorize[d]” by statute or by contract. Brookfield Constr. Co. v. United States, 228 Ct.Cl. 551, 559, 661 F.2d 159, 165 (1981). See also 28 U.S.C. § 2516(a) (1982) (“[i]nterest on a claim against the United States shall be allowed in a judgment of the United States Claims Court only under ... a[n] Act of Congress expressly providing for payment thereof”.) General Provision 19 of the contract controls the pricing of change orders. In turn, it incorporates section 15 of the Defense Acquisition Regulations (“DAR”). The applicable DAR is found at 32 C.F.R. § 15-205.17 (1982).10 That section provides that “[i]nterest on borrowed funds (however represented) [is] unallowable.” The applicability of such a provision in the present case conclusively precludes recovery of interest. Nor is the court willing, as plaintiff suggests, to review the continuing validity of cases rejecting an “equity capital” approach to compensating contractors under these circumstances. See Getty Oil Co. v. United States, 767 F.2d 886, 888-89 (Fed.Cir.1985); Framlau Corp. v. United States, 215 Ct.Cl. 185, 197-98, 568 F.2d 687, 693 (1977).
Servidone asks in the alternative that the court consider the heavy cost of its borrowings when it reviews the percentage of profit to apply to costs. This also the court declines to do. The applicable regulation disallows interest “however represented.” If the court were to amplify profit because the cost of borrowing is otherwise not reflected, it would amount to an award of part of the interest claimed. Having disallowed the claim for these reasons, it is not necessary to address the second issue.
X. DAMAGES
To recapitulate, the court has allowed the following claims: 1) part of the embankment claim; 2) part of the lime stabilization *384claim; and 3) part of the Mountain Creek II claim. The following claims are not sustained: 1) the harassment and over-testing claims; 2) the rework claims; 3) the inundation claim; 4) the standby claim; and 5) the interest claim.
A. The Embankment Claim
Servidone’s damages presentation was made primarily by Craig Jacobsen. He prepared an extensive report and testified at length. Plaintiff's claim was the subject of an audit by the Government’s accounting witness, Richard Houlihan, with the Defense Contract Auditing Agency (“DCAA”). While both had expertise in accounting, Jacobsen had considerably more experience in auditing construction companies and in general impressed the court as very knowledgeable and competent.
In preparing Servidone’s damages claim, Jacobsen completely reconstructed plaintiff’s accounting records by going back to original documents. It was a massive undertaking and thoroughly explained at trial. Defendant never successfully challenged the accuracy of Jacobsen’s methodology in any meaningful way. There was nothing sloppy or superficial in his approach. For that reason, the court is persuaded that the figures Servidone presents are correct.
Defendant’s primary challenges to Jacobsen’s methodology relate to fundamental choices about how to derive costs. The first is the assertion that Jacobsen improperly used a “total cost” approach, which is generally disfavored in the law. See WRB Corp. v. United States, 183 Ct.Cl. at 426; J.D. Hedin Constr. Co. v. United States, 171 Ct.Cl. 70, 86-87, 347 F.2d 235, 246-47 (1965). While Servidone disputes this characterization of its damages presentation, it is obvious that defendant is correct. Jacobsen identified all direct and indirect costs associated with the embankment items, and then treated as damages those costs in excess of the amounts bid for embankment items. While this approach is a modified total cost claim in the sense that it isolates embankment items, it suffers from the same presumptions of any total cost claim — namely, that the original bid was an accurate reflection of a reasonable cost, and that all costs in excess of that had to be attributable to the differing site condition.
The fact that the embankment claim is presented as a total cost claim is not fatal to Servidone’s recovery, however. The Court of Claims recognized that in some types of work, and for certain types of changed conditions, the total cost approach may be the only viable means to identify costs. WRB Corp., 183 Ct.Cl. at 426; Turnbull, Inc. v. United States, 180 Ct.Cl. 1010, 1025-26, 389 F.2d 1007 (1967). Acceptability of the total cost approach depends on proof of 4 conditions: 1) The nature of the losses makes it impossible or highly impracticable to determine the actual losses directly with a reasonable degree of accuracy; 2) the plaintiff’s bid was reasonable; 3) its actual costs were reasonable; and 4) it was not responsible for the added costs. WRB Corp., 183 Ct.Cl. at 426.
The court has no misgivings in concluding that the total cost approach is justified here. The first element is easily satisfied. It would be impossible to isolate the incremental impact of the nature of the changed condition. The total volume of dirt excavated and placed was not affected. Rather, the change was reflected in the additional amount of effort expended to process that dirt. The only meaningful way to express that increase is in the difference between the reasonably anticipated costs and the reasonable actual costs. This is precisely a claim in which the total cost approach is the most appropriate means of evaluating the impact of the changed condition.
The second element, reasonableness of plaintiff’s bid, has already been addressed in part. The court has concluded that the differing site condition consists only in the difference between the conditions plaintiff should have anticipated and the conditions actually encountered. Although conditions were worse than anyone reasonably would have anticipated, plaintiff should not get *385the benefit of its own failure to anticipate that level of difficulty that a reasonable contractor should have expected. The court therefore will not utilize plaintiff’s own bid in doing a modified total cost analysis because that bid was not reasonable within the meaning of WRB Corp. Instead, a determination of what was a reasonable bid must be made from the bids of others. See Ingalls Ship Building Division, Litton Systems, Inc., ASBCA No. 17597, 78-1 B.C.A. (CCH) ¶ 13,038 at 63,-668-72.
The court has already determined that Abrams acted as a reasonable contractor. Its total bid was $33,961,358.77. This is $8,180,020.00 more than Servidone’s total bid. During closing argument, the court asked plaintiff’s counsel whether total bid or unit estimates are a better reflection of costs. Counsel argued that the total bid is more accurate because contractors may have reasons not directly related to actual cost to make imbalanced bids. Two factors militate against giving weight to that observation in this case. First, plaintiff’s modified total cost approach begins with the assumption that its own estimated costs for the embankment items were completely accurate. Those costs formed the foundation of Jacobsen’s analysis. Second, the court counts it unlikely that a bidder would materially imbalance a bid with respect to embankment costs, when those constituted the majority of the total costs. Since Servidone’s claim isolates the embankment items, the court will base its analysis on bids for that particular work.
Servidone’s estimate of the total cost of the four contested embankment items, $14,-767.424.00, was $9,262,459.00 less than Abrams’ estimate of $24,029,883.00, for the same items. The bid for embankment items of Resource Construction Co., the third lowest bidder, was $16,343,696.00. This is $1,576,272.00 more than Servidone’s estimate for the same work. The difference between Servidone’s embankment bid and the Government estimate of $19,326,-845.00, with a profit of 8.5 percent, is $6,202,202.00. The record does not reflect the embankment bid of those contractors which bid more than the 3 qualified low bidders, although the court notes that the high bid on all items of work was over $46,000,000, $21,000,000 more than Servidone’s bid.
It is plain that Servidone underbid. The figures extracted above suggest an underbid on embankment items in the range of $1% million to over $9 million. As discussed above, the court is persuaded that Barth as an individual and Abrams as a company were the best model presented of the reasonable contractor. The court is similarly persuaded that Abrams’ bid is the most realistic, reasonable, and objective assessment available to the court of the work as of the time of the bids. While Abrams may have had a slight advantage because of its work on a roadbed near the dam, the court will not discount the difference between Servidone’s and Abrams’ bids on that basis for 3 reasons. First, Servidone should have been aware of the limitations created by its own inexperience. As discussed above, there were sufficient indicators in terms of the Atterberg data, the use of a LLCC and its lack of experience in Texas that should have warned Servidone to become better informed. Second, the record is clear that while Servidone is a very competent contractor, it was initially at a disadvantage on this project because of its unfamiliarity with work in highly plastic soils and with the use of a LLCC. The evidence suggests that there was, indeed, a learning curve for Servidone on this project. Some of that unfamiliarity should have been anticipated in the bid. Third, the court does not elsewhere attempt to assess the effect of Servidone’s error in projecting the variance between actual and optimum water content. See discussion supra p. 357. Rather than attempt in the form of a jury verdict to assess the full impact of Servidone’s underbid, the court concludes that the most accurate means of reflecting both Servidone’s failure to anticipate those problems reasonably apparent in the bid information and its own lack of experience is to utilize Abrams’ bid amount for embankment work. There exists, therefore, a realistic assessment for what Servidone should have bid.
*386This approach is consistent also with Servidone’s contentions in its post-trial brief. There it states, in reference to Proctor data, that “[i]f the Government had furnished this information to bidders, the bids would have been many, many millions of dollars more than the Fort Worth District wanted to spend.” Plaintiff’s Post-Trial Brief at 32 (emphasis added). The court rejected Servidone’s contention that liability should attach to the Government’s failure to provide Proctor data. Rather, it has held that plaintiff’s problems, if any, with the LLCC were due to its own lack of caution or experience. A fair inference from Servidone’s brief is that, if it had understood the operation of the LLCC, it would have bid substantially higher.
The final element of the test set out in WRB Corp. is whether the contractor was responsible for the additional expense. Defendant, though never contesting Servidone’s general competence as a contractor, made some effort to show that Servidone’s operations were flawed. The court has already rejected most of these contentions as a possible explanation for the differing site condition. See discussion supra pp. 362, 366. The court is persuaded that there is insufficient evidence to reject a modified total cost approach because of evidence that plaintiff was responsible for the added costs. Although there are elements of plaintiff’s costs that the court eliminates, those can be sufficiently isolated to be treated separately.
The element of damages that has not yet been discussed is whether plaintiff's actual costs were reasonable. See WRB Corp. 183 Ct.Cl. at 426. The number of hours expended and costs of labor are not in issue. Defendant has not challenged the reasonableness of the claimed costs. Nor has it questioned the accuracy of Servidone’s bookkeeping or Jacobsen’s analysis. What remains, instead, are disputes concerning the proper method to assess overhead, profit, and equipment rates.
Profit
Servidone is entitled to a reasonable profit on additional costs allowed in connection with the equitable adjustment. See Macke Co. v. United States, 199 Ct.Cl. 552, 568-69, 467 F.2d 1323, 1332-33 (1972). It claims profit at a rate of 10 percent on direct and indirect costs. Defendant challenges that percentage, asserting that the appropriate rate is 8.5 percent, the rate Servidone used in calculating its bid. The 10 percent rate is based on nothing other than Servidone’s request. While that rate is not per se unreasonable, plaintiff has not justified an increase in the rate they originally contemplated. It points to DAR § 3-808.3, 32 C.F.R. § 3-808.3 (1982), which concerns negotiated price adjustments. That provision suggests that a higher rate than the original bid rate can be appropriate if the change work is significantly different from the contracted work, and if a “detailed analysis” is done. In this case, the change work is arguably identical in nature to the original work. In any event, plaintiff has not provided an analysis to warrant an increase in the profit rate. Accordingly, profit on allowed costs will be based on a rate of 8.5 percent.
Equipment rates
Other than the Government’s challenge to the claim for interest on borrowings, the bulk of the costs questioned in the DCAA audit relates to rates charged for use and ownership of equipment. In the claim audited by Houlihan, Servidone was seeking recovery of approximately $34,-000,000. Houlihan challenged nearly $13,-000,000 of the claimed $22,500,000 in direct costs. Of that $13,000,000, virtually all the challenged costs relate to equipment. There are 2 primary disputes. The first and largest relates to equipment ownership expense.
Servidone’s claim for equipment ownership costs is based on multiplying an uncontested number of hours each piece of equipment was operated by an hourly ownership or rental cost drawn from rates adopted by the Associated General Contractors (“AGC”). This method is called for by *387the contract.11 The government audit rejects the use of AGC rates, apparently because the audit led to the conclusion that the plaintiffs claimed costs were substantially above the costs charged internally on the company’s books with respect to the Lakeview project.
The court rejects defendant’s position with respect to AGC rates for two reasons. First, it is inconsistent with the contract. The contract specifies use of AGC rates, and there is no reason not to honor that provision. See Nolan Bros., Inc. v. United States, 194 Ct.Cl. 1, 13-15, 437 F.2d 1371, 1377-78 (1971); Holloway Constr. Co. v. United States, 18 Cl.Ct. 326, 332-33 (1989).
Second, the court does not accept Houlihan’s explanation for why Servidone’s internal accounting for costs demonstrates the inapplicability of AGC rates. Servidone’s books and records reflect an internal accounting of project costs. The court is persuaded, however, after Jacobsen’s explanation, that Houlihan’s extraction of internally booked cost is inappropriate as a device to assess damages. Jacobsen testified, without contradiction, that the internal assignment of equipment chargeout rates is fundamentally a fiction. In some respects it is no more than a prediction of what costs will be. It can be made more accurate only by subsequent, continuous reconciliations. In any event, the court has no confidence that using the types of costs isolated by Houlihan, or the annual data, is a fair method, after the fact, for limiting costs in this case. It was obvious that use of internal chargeout rates could only be an accurate reflection of costs if Servidone’s internal recordkeeping was adjusted to segregate all elements of costs allowed under the AGC analysis, and if, over an adequate number of years, there were accurate reconciliations. Even if AGC rates could be determined by this type of analysis, defendant did not persuade the court its proposal was fairer or more accurate. The court notes that this issue was not raised by defendant in its post-trial brief.
Defendant also argued that if AGC rates were used for equipment ownership, those rates would have to be adjusted downward by reflecting a greater than average number of months of use on the Lakeview project. This argument raises a legitimate inquiry. The foreword to the AGC rate manual provides that the rates are to be used as a guide, and that they can be adjusted based on actual experience. The factor which defendant points to as dictating an adjustment is length of work season. The AGC rates presume a work season varying from 6 to 9 months for the equipment in issue, based on “central states.” There is no question that Servidone was able to work the equipment, at least during the Lakeview project, virtually year-round. Moreover, the project took 4 years to complete, although the embankment work was finished in less time. Jacobsen testified, however, that the rate has to reflect the lifetime experience of the equipment, taking into account all the jobs on which it has been used. He testified that Servidone has done jobs in northern tier states, which presumably would balance out the additional use Servidone got in Texas. Although it is not clear that the testimony presented anything like a comprehensive picture of Servidone’s construction projects, a substantial number of those discussed was not in the South. On balance, the court is left with an impression that the audit could have made a case for adjusting the rates, but that it did not do so.
With respect to operating costs, the DAR does not specify that AGC rates must be used. Instead, it reflects that these must be separately allowed. Jacobsen testified, and the court finds, that it was impossible to pull these figures from Servidone’s *388books and records. He also explained that this phenomenon was not unusual. Jacob-sen relied instead on rates recommended by the 1985 Dataquest Cost Reference Guide. This reference is, according to Jacobsen, who has wide experience in construction cost accounting, commonly used in the industry. For each piece of equipment, Jacobsen was able to pull from Dataquest a projected typical cost for use expenses. These were totaled and, along with the AGC ownership rates, multiplied by the hours of use, pulled from the daily reports. Jacobsen was unable to compare the results of using Dataquest rates with Servidone’s actual equipment operating costs.
Houlihan did not directly address the reliability of Dataquest rates, and did not offer a clear alternative. Once again, the court is left with the impression that the DCAA audit could have challenged more effectively the necessity of relying on Dataquest rates. It did not do so, however, and the court has no basis for concluding that the Dataquest approach is unreasonable. Although it does so reluctantly, the court feels compelled, simply because of the number of issues and the volume of testimony on the accounting of the claim, to record that the audit and associated testimony lacked persuasiveness and coherence. It was seldom clear precisely what was done or why it was done. Moreover, the DCAA audit, based on sampling, was considerably less detailed and comprehensive than Jacobsen’s reconstruction.
In conclusion, with respect to direct costs for equipment within the embankment claim, the court finds that the costs claimed were reasonably incurred.
Overhead
Servidone claims jobsite overhead at a rate of 9 percent, and home office overhead at 3.5 percent. The latter figure is not contested. At trial, the Government questioned the jobsite overhead rate, although the issue does not appear in its post-trial brief. In its original claim, Servidone calculated a jobsite overhead rate of 11.2 percent. The claim was revised after Coopers and Lybrand’s analysis to 9 percent. Jacobsen testified that several items that Servidone treated on its books and records and in its claim as direct costs were categorized by Houlihan as indirect costs, and vice versa. Houlihan admitted that his attempt to locate and categorize costs as direct or indirect was limited. Although there may be outer parameters as to how costs can be categorized, the court is left with a difference of opinion by accountants as to where, given the company’s own practice, and given generally accepted accounting principles, costs should be placed. Jacobsen’s assessment is accepted for two reasons. First, the court finds that both he and his approach had more credibility. Second, it is not inconsistent with the court’s adoption, elsewhere, of Servidone’s approach to direct costs. Damages, therefore, will be calculated using a home office overhead rate of 3.5 percent, and a field office overhead of 9 percent.
The defendant has not contested bond premium and insurance costs.
Summary of Embankment Recovery
Servidone claims a total of $24,-537,098.00 as direct and indirect costs on its embankment claim. As discussed above, in order to utilize a modified total cost approach as to the embankment claim, the court must have confidence in the validity of the actual cost figures. See Skip Kirchdorfer, Inc. v. United States, 14 Cl.Ct. 594, 610-14 (1988). The plaintiff bears the burden of proving the amount of its damages. Roberts v. United States, 174 Ct.Cl. 940, 956, 357 F.2d 938, 949 (1966); Neal & Co. v. United States, 17 Cl.Ct. 511, 513 (1989).
As discussed above, the plaintiff’s embankment claim is presented as a total cost claim. The court rejected two claims made by Servidone in connection with embankment work, the claim that costs were increased due to a differing site condition with respect to borrow pit Cl, and the contention that its costs were aggravated by unwarranted testing and rework. For the reasons that follow, the embankment claim should be adjusted down to reflect the impact of the court’s findings with respect to those arguments. Cf Skip Kirchdorfer, Inc., 14 Cl.Ct. at 612 (where court *389reduced plaintiff’s cost claim to reflect poor quality of proof).
In its proposed finding number 64, Servidone contends that its “costs and time of performance were increased due to the unavailability of Borrow Area Cl.” In its post-trial brief, Servidone estimated the impact of that additional cost at $100,000.00. It is apparent from the way Jacobsen prepared his analysis, however, that this additional cost, as well as that for the asserted overtesting and rework, is built into the $24,537,098.00 claimed in additional embankment costs. Having found that any additional costs in connection with borrow area Cl should be borne by plaintiff, the court is compelled to make an adjustment. Using Servidone’s suggested figures, the court disallows $100,000.00.
Plaintiff’s contention that excessive rework and overtesting cost it an additional $600,000.00 was also rejected. This was Servidone’s estimate of the amount of rework it did in excess of that which should reasonably have been expected. It failed, however, to demonstrate that the entire amount of rework was due to a differing site condition, or to a de facto change order by the Government. What remains are the possibilities that this was routine work that should have been expected, or that is was due to other causes for which no liability would attach to the Government, such as Servidone’s misjudging of the moisture content, or that it was due to the unusual nature of the soil. It would be unfair to require the Government to pay for rework associated with the first or second possibilities. Yet it is apparent from Servidone’s approach to damages that is precisely the result of its present claim. Jacobsen did not isolate any amount for overtesting. It is only in its post-trial brief that plaintiff estimates that the cost associated with all the claimed overtesting is $600,000.00. Since this amount is already built into Jacobsen’s calculations, at least a portion must be removed. While the court is persuaded that a primary cause of rework was the unusual nature of the soil, Servidone had the burden of proof on the issue, and failed. The court therefore makes an estimate in the nature of a jury verdict that half of those claimed costs should be disallowed.
Accordingly, Servidone’s allowed direct and indirect costs on embankment work, after appropriate changes in profit, bond and insurance charges, are $23,703,-582.00. As discussed previously, Servidone’s bid was inadequate. It should have bid $24,029,883.00 for embankment work. Instead it bid $14,767,424.00. It should not recover to the extent of the $9,262,459.00 difference between these sums. Consequentially, Servidone has demonstrated damages with respect to the embankment claim in the amount of $14,441,123.00.
B. Lime Stabilization Claim
The court has already concluded that 90 percent of Servidone’s proven lime stabilization costs is allowable. Plaintiff claims $116,641.00. This consists of direct and indirect costs. Jacobsen certified the accuracy of $58,333.00 as direct labor and equipment charges in excess of the amount paid by the Government. He was unable to certify the calculation of additional standby costs of $33,907.00. These figures were provided without explanation by Fantozzi, apparently as his best estimate. The court elsewhere has disallowed other standby cost claims. It perceives no additional reason to allow these. Consequently, Servidone is allowed direct costs in the sum of $52,499.70 (90 percent of $58,333.00). After adjustment for profit of 8.5 percent and other indirect costs, the allowed amount is $65,482.85.
C. Mountain Creek II
Servidone’s present claim for Mountain Creek II is $349,447.00, inclusive of profit and all direct and indirect costs. Of the direct labor and equipment charges, Jacob-sen was only able to verify an amount of $157,625.00. He could not trace in Servidone’s books and records the additional claimed labor and equipment and equipment charges of $71,493.00. As in the case of the lime stabilization work, this is apparently an amount claimed for standby equipment costs. This amount is disallowed as *390unproven. Similarly, Servidone’s additional claim for $94,848.00 in “extended overhead, including profit,” was unsubstantiated. Accordingly, direct costs are allowed in the amount of $157,625.00. After adjustment for profit and indirect charges, the total amount allowed is $196,605.58.
CONCLUSION
The Clerk is directed to enter judgment for plaintiff in the amount of $14,703,-211.43, plus interest pursuant to 41 U.S.C. § 611. Interest on $14,441,123.00 will run from March 1, 1984; interest on $196,-605.58 will run from September 30, 1985; and interest on $65,482.85 will run from December 16, 1985.
. All three figures are taken from exhibits. Figure 3 has been modified by the court to superimpose the final LLCC.
. Defendant actually permitted a slight deviation.
. During the period of August 9, 1982 through August 19, 1982 materials which were placed using the initial curve or the first adjusted curve as control were not rejected if they complied with the moisture requirements based upon either of the curves.
. A second reason for the scatter, and one that Poulos considered less significant, is that soils with the same LL can have different mineral characteristics, and thus different points of opti*359mum compaction. The averaging imposed by the curve is not helpful in that regard, since the results are reflective of genuine differences. The court discounts much of the apparent deviation from the LLCC as attributable to errors in the Proctor data. It is also noteworthy in this respect that generating LL results (necessary for use of the LLCC) does not have the same potential for error.
. Material for the lime stabilized subgrade was also taken from the borrow areas. Thus, while the focus of Pierce’s testimony was the lime stabilized subgrade, the court considers his testimony and reports clearly relevant to much of the soil used in the embankment.
. Defendant suggests that the slickness was caused in part by Servidone’s practice of applying a downshot of water on each new lift as it was spread. Hudson testified it would be appropriate to apply a downshot whenever a dry crust had formed on the fill, and conceded that he often suggested such a course of action. Fantozzi went further, testifying that the procedure was effectively required between lifts. The court is persuaded that Fantozzi’s impression was an accurate one. Given the difficulty Servidone was already experiencing moving equipment through the fill, and given its general resistance to adding water, it is more likely that the addition of water, if indeed it was unnecessary, was at the Corps’ direction.
. Servidone also introduced evidence that the sheepsfoot tamping rollers used to compact the fill did not "walk out." Walking out refers to the phenomenon of a sheeps foot roller sinking less and less into the soil with each succeeding pass over the same soil. On the last pass, the drum is riding a few inches above the soil surface. This apparently was intended as additional evidence either that the material was unusual or was overly wet. The evidence on this point is inconclusive and the court does not rely on it.
. The court rejected, above, Servidone’s related argument that the proctor data underlying creation of the LLCC should have been supplied to bidders.
. In Count VII of the complaint, Servidone claims that the cumulative effect of all additional and changed work caused by the Government — based on either a differing site condition or defective specification theory — was so great as to amount to a cardinal change in the contract. "The basic standard [to show a cardinal change] is whether the modified job ‘was essentially the same work as the parties bargained for when the contract was awarded.” Air-A-Plane v. United States, 187 Ct.Cl. 269, 275, 408 F.2d 1030, 1033 (1969) (citations omitted). The court finds that no claim can be supported under this standard. The contract required plaintiff to complete the Joe Pool dam by excavating, processing, and placing approximately 10,000,000 cubic yards of fill material taken from borrow areas on site. Although it was considerably more difficult than could reasonably have been anticipated, Servidone was called on to do "essentially the same work." Plaintiff’s performance was within the scope of the contract.
. While Bell v. United States, 186 Ct.Cl. 189, 205-06, 404 F.2d 975, 984 (1968), permitted a recovery of interest for a claim brought under the changes clause, the relevant clause apparently did not explicitly exclude interest, as does the one at bar.
. The applicable DAR provision, incorporated into the contract by the pricing of adjustments clause, provides in pertinent part, as follows:
Evaluation of costs for the use of construction plant and equipment, in sound and workable condition, which are owned or controlled by a contractor or subcontractor and are furnished or [sic] the proper and economical performance of a fixed-price type contract shall be based upon the "Contractor’s Equipment Ownership Expense Schedule" Sixth Edition, 1966, published by the Associated General Contractors of America, Inc.
32 C.F.R. § 15-402.1(c) (1982).