ON PLAINTIFF’S MOTION AND DEFENDANT’S CROSS-MOTION FOR SUMMARY JUDGMENT

PER CURIAM:

This case was referred to Trial Commissioner Louis Speetor with directions to prepare and file his opinion on the issues of plaintiff’s motion and defendant’s cross-motion for summary judgment under the order of reference and Rule 166(c). The commissioner has done so in an opinion and report filed on June 25, 1970, wherein such facts as are necessary to the opinion are set forth. Defendant filed a request for review by the court and the case has been submitted to the court on oral argument of counsel and the briefs of the parties.

In this case the burden was on the defendant to prove that the plaintiff was responsible for the misfunctioning of the turbines. The court agrees with the trial commissioner, for the reasons he gives, that the Corps of Engineers Board of Contract Appeals could not properly find that the defendant has sustained that burden. On several grounds, taken together, the Board’s contrary determination is not entitled to acceptance. First, the Board’s reasoning is summary, conclusionary and inadequate, and lumps conclusions of law with purported factual findings in such a way that it is impos *1372 sible to tell how much the Board relied on legal conclusions which were erroneous and how much on true findings of fact unaffected by legal error. See Loral Electronics Corp. v. United States, 387 F.2d 975, 980, 181 Ct.Cl. 822, 831-833 (1967); Sundstrand Turbo v. United States, 389 F.2d 406, 418, 182 Ct.Cl. 31, 52-53 (1968). In this connection, it is important that the Board appears to rely very heavily on its erroneous legal conclusion that the “Cooperation” clause of the contract required plaintiff to investigate the consequences of, and then make the necessary adjustments for, the switch to single bearing generators. As Commissioner Spector points out, plaintiff had no such responsibility under that clause or under any other part of its contract, but the error nevertheless seems central to, and necessarily infects, the Board’s determination. * As for the few factual findings which are untainted by legal error, they are insufficient in themselves to sustain the administrative result. The court is satisfied that on the whole record the Board could not reasonably and correctly come to the conclusion that the Government adequately proved that the problems with the turbines came about because of any failure by plaintiff in its responsibilities.

Since the court agrees with the opinion and recommended conclusion of the trial commissioner, as hereinafter set forth, it hereby adopts the same, together with the foregoing discussion in the paragraph above, as the basis for its judgment in this case. Therefore, plaintiff’s motion for summary judgment is granted, defendant’s cross-motion is denied and judgment is entered for plaintiff in the sum of $112,053.86.

OPINION OF COMMISSIONER

SPECTOR,

Commissioner: Plaintiff’s claim, in the amount of $112,053.86, grows out of a contract dated October 19, 1956, for the design, manufacture, testing and supervision of erection of two 90,000 horsepower hydraulic turbines for the fourth and fifth power plant units at the Garrison Dam Project, Riverdale, North Dakota. The total estimated contract price was $2,569,200. It was one of four independent contracts with defendant contributing to the completion of these power units, there being separate contracts for the manufacture of the generator, the manufacture of the governor, and for the erection and installation of the power units. Hydraulic turbines such as these are equipped with upper and lower sets of stationary and rotating “seal rings” which increase the efficiency of the turbine by preventing water from by-passing it. 1 The claim has heretofore been the subject of decisions adverse to plaintiff by defendant’s contracting officer, and, on appeal, by the Army Corps of Engineers Board of Contract Appeals. Describing the claim, the board opinion states:

As will appear in greater detail hereinafter, soon after the installation and early operation of the Unit 4 and 5 turbines damage to seal rings was discovered. [In testimony and briefs the parties have usually referred to the rings in question as “seal” rings. In the specifications they are called “wearing” rings.] The dispute focuses principally on this damage and contentions of the parties as to the cause of such damage. * * *

The claim consists generally of costs incurred thereafter by plaintiff at de *1373 fendant’s direction, plus costs incurred by defendant and withheld from sums otherwise due plaintiff. Under the “Disputes” provision contained in the contract, which is worded in accordance with the so-called Wunderlich Act, 2 such decisions by the contracting agency, to the extent that they concern a dispute on a question of fact, are final “unless determined by a court of competent jurisdiction to have been fraudulent or capricious or arbitrary or so grossly erroneous as necessarily to imply bad faith, or not be [sic] supported by substantial evidence.” Decisions concerning a question of law are not mentioned in this provision. 3

General Statement of Facts

The facts hereinafter related generally parallel those recited in the board’s opinion, occasionally supplemented by uncontested facts, and facts not mentioned in the agency opinion although mandated by the record. 4

Baldwin Locomotive Works, plaintiff’s predecessor, had previously furnished the turbines for the first, second, and third power plant units at the Garrison Dam. The bidding documents leading to the present contract for the fourth and fifth turbines, provided that a number of the bid items could be omitted upon award if “ * * * a bid is based on furnishing turbines sufficiently similar to the installed turbines * * 5 Be *1374 cause plaintiff’s bid proposed to furnish turbines identical to those successfully employed ' in units 1, 2 and 3, it was awarded a contract omitting items 7 through 9, and 12 through 20 of the 20 bid items, thereby saving defendant approximately $108,000. Eliminated were first model and additional model tests; tools, wrenches and lifting devices; and spare parts. Also, the original shop drawings for units 1, 2 and 3 were permitted to be reused.

Plaintiff commenced performance and by letter dated January 10, 1957, submitted shop drawings for units 4 and 5 to defendant for approval. All but five of these drawings were identical to those previously submitted and approved for units 1, 2 and 3. Among these identical drawings previously submitted and approved, were those two which contained the dimensions for the upper and lower stationary and rotating seal rings and the manufacturing goals for the seal ring clearances. On February 5, 1957, defendant approved and returned these two drawings which recited a manufacturing goal of between .050" and .053" for the upper seal ring clearance and between .062" and .066" for the lower seal ring clearance.

During February 1959, plaintiff completed delivery of the components for turbine units 4 and 5, and the turbines were installed in the project under a separate contract, as previously mentioned, between defendant and P. S. Lord Company. The board opinion acknowledges that:

Turbine units 1, 2 and 3 at Garrison Dam were identical in all material respects to units 4 and 5 furnished under the instant contract. Units 1, 2 and 3 have operated without difficulty ever since their original installation with seal ring clearances nearly identical to the original field clearances of units U and 5. The original field clearances for units 4 and 5 were substantially less than the minimum clearances shown on the approved shop drawings. The minimum clearance measured for the upper seal of unit 4 was .039 in. and for the lower seal .041. Original minimum field seal clearances for unit No. 5 were .043 in. for the upper seal and .040 in. for the lower. [Emphasis supplied.]

The contract specifications provide in TP 2-15 that:

* * * The wearing rings on the runner and stationary wearing rings shall be made of suitable steel, having sufficiently dissimilar characteristics from those of the runner to prevent seizing in case of accidental rubbing. The rings shall be securely fastened in place by approved means. The radial clearances between the moving and stationary wearing rings shall be as small as possible, consistent with safe operation and with the clearances required in the turbine and generator guide bearings. * * * [Emphasis supplied.] 6

*1375 Specification SC-4 states that “[a]pproval by the Contracting Officer of the Contractor’s drawings shall not be held to relieve the Contractor of any part of the Contractor’s obligation to meet all of the requirements of these specifications or of the responsibility for the correctness of the Contractor’s drawings.”

Meanwhile, defendant prepared specifications for a separate contract for the generators which were to be coupled to the turbines in power units 4 and 5. The bidding documents gave prospective bidders the option of furnishing either a less expensive one-bearing generator, or the more expensive two-bearing generator previously installed in units 1, 2 and 3. This option was provided upon the recommendation of the Interior Department which was responsible for purchasing and marketing the power generated at Garrison Dam, and therefore interested in cost-saving. General Electric Company had manufactured the two-bearing generators installed in units 1, 2 and 3, under a contract containing the same option. General Electric expressed a preference for the two-bearing generator in units 4 and 5 also, and was opposed to the option permitting one bearing both before and after award to it of the unit 4 and 5 contract. Its preference was stated to be based on the greater mechanical stability and additional margin for unusual or abnormal electrical or mechanical disturbances that might affect operation.

General Electric was awarded the contract for the unit 4 and 5 generators on July 2, 1957, approximately 9 months after the award of plaintiff’s contract for the turbines, and 6 months after approval of plaintiff’s shop drawings showing seal ring clearances as above detailed. The contract as awarded to General Electric stated in TP 2-10d. that the “generator shall be provided with one guide bearing located below the rotor. An upper guide bearing shall be provided if it is deemed necessary by the Contractor for satisfactory operation * * [Emphasis supplied.]

General Electric elected to supply the less expensive single bearing generator. Its reason for doing so was explained by its representative at a meeting held after the damage to plaintiff’s turbines had occurred, as hereinafter described. The General Electric representative then stated:

Now, mechanically, a mechanical system with an upper guide bearing is always better. There is no question about it. In general, it is a more stable machine — it’s indicated by the difference in the critical speed and so on. Now, this has been brought to the attention of the Corps that is, if they had chosen to insure getting duplicate units, you could have so required in the specifications. In the course, we’re feeling our way under the stress of a lot of things, some of which you are familiar with I am sure, and we have had to do everything we can to beat competition, if you will, and this construction is generally offered and where we’re satisfied that it is satisfactory construction we use it. * * *

Both plaintiff’s contract for the turbines, and General Electric’s contract for the generators contain similar clauses entitled “Cooperation with Other' Manufacturers.” In plaintiff’s contract, it reads as follows:

SC-7. Cooperation with Other Manufacturers. The Contractor for the turbines shall exchange with the governor and generator manufacturers, all necessary drawings, dimensions, templates, gauges, and other information required to insure the complete and proper design and manufacture of all connecting or related parts of the turbines, governors, and generators. No extra compensation shall be claimed because of any modifications required to accomodate [sic] equipment of another manufacturer, except as is otherwise specifically provided herein. All adjustments shall be made between the respective manufacturers without involving extra cost to the Government. Two copies of all drawings and all correspondence relating to drawings and *1376 specifications interchanged between manufacturers shall be sent to the Contracting Officer.

The board opinion acknowledges that “[t]here was apparently a normal interchange of information between General Electric and Baldwin. According to letters in evidence, General Electric, after receiving its notice of award, requested and received of Baldwin turbine data necessary to design of the generator. In turn it furnished the turbine manufacturer with pertinent details concerning the generator and, specifically, the information that no upper guide bearing was to be used.” The board opinion continues :

There is no evidence that the Government ever pointed out to appellant that the single-bearing generator might necessitate changes in the turbine design.

Defendant provided neither General Electric nor plaintiff with copies of the other’s specifications. There is no evidence that General Electric provided plaintiff with detailed specifications, stability tests or any other information that would forewarn plaintiff of the possibility that the single bearing generator General Electric had selected might conceivably require changes in the turbine design.

While the turbines were being installed (following their delivery in February 1959.), large quantities of rock, gravel and debris were observed at both the penstock intake areas and the draft tube discharge areas, relating to power units 4 and 5. In order to operate power units 1, 2 and 3, it had been necessary for defendant to seal off the installation sites for units 4 and 5 with a concrete bulkhead and, as a result, eddy currents stirred by the flow through adjacent units 1, 2 and 3, had apparently caused the accumulation in the units 4 and 5 area. In August 1959, P. S. Lord Company, the installation contractor, attempted to install steel gates in the discharge area serving the unit 4 and 5 draft tubes, so as to permit removal of the aforementioned concrete bulkhead, and the pumping out of the water between the steel gates and the bulkhead. The accumulation of debris was so heavy that it was necessary for P. S. Lord Company to employ a siphon pump to move it approximately 8 feet downstream before the steel gates could be seated. There was no other attempt to remove the rock, gravel and debris located at either the intake or discharge points serving power units 4 and 5, prior to commencement of operation of those units.

Unit 4 was completed and ready for test in early December 1959. As quoted above from the board opinion, the as-built seal ring clearances for units 4 and 5 were “nearly identical to the original field clearances” of units 1, 2 and 3 which had been operating without difficulty since their original installation. These upper seal ring clearances for unit 4, as measured on September 18, 1959, ranged from .039" to .049". The lower seal ring clearances as of that date ranged from .041" to .055". By way of comparison with existing unit 3, for example, the latter’s average upper seal ring clearance in 1954 was .047" with a minimum clearance of .042". By 1961, a chemical buildup on the seal rings had caused the average upper seal ring clearance on unit 3 to decrease to .026" with a minimum measurement of .021". The as-built lower seal ring clearance of unit 3 in 1954 averaged .050" with a minimum measurement of .043", but by 1961, chemical buildup had reduced the average clearance to .038" with a minimum of .030".

As previously indicated, these as-built or field clearances were generally less than those set forth in the approved shop drawings. A reasonable variation between as-built and shop-drawing clearances is normal when dealing with a large runner. The variation usually results from a slight out-of-roundness, compounded by the difficulty in maintaining design tolerances on a runner 18 feet in diameter. The installed clearances were, in any event, found suffi *1377 cient by both plaintiff’s engineering department and Government’s on-site inspector.

On December 9, 1959, the mechanical test run on unit 4 was commenced. The unit was operated for about 4 hours at a wicket gate opening of about 17 percent. During this period the unit was operated without power load and therefore without fluctuations of electrical load. The turbine was run at gradually increasing speed until full speed (90 r. p. m.) was reached about 1% hours after startup. As the board opinion states:

* * * The unit ran smoothly without excessive noise. The dial indicators located above the turbine bearing and in contact with the shaft showed a .0015 inch skate. The dial indicators attached to the indicators for the thrust bearing showed a smooth gentle movement.

Bearing temperatures were normal and the turbine unit was operating freely and smoothly. The board opinion continues:

At about 1 P.M., after four hours of smooth operation, the turbine began to make a loud, thumping noise in time with the turbine rotation. After this noise commenced it was noted that the turbine bearing indicators were showing a fluctuation of .007 inch in exact time with the rotation. The thrust bearing movement was erratic and more than twice normal. After approximately 7 minutes the thumping noise subsided. The turbine bearing movement returned to .003 inch and the thrust bearing movement to near normal. The unit was operated for another 15 minutes at which time the Government’s mechanical engineer, hearing a dull thump, ordered the unit shut down. 7

*1378 After some difficulty in closing the aforementioned steel draft tube gates, unit 4 was unwatered and examined. The board’s opinion describes the results of that examination as follows:

* * * Portions of the stationary rings had embedded themselves in the rotating rings. Apparently, some of these embedded particles acted as tools to cut into the stationary rings. Most of the damage was confined to the center portions of the rings. Top and bottom edges were almost undamaged. Many chips of metal were found inside the head cover and between the runner band and the discharge ring. The rotating rings had only a few cuts and gouges but had many areas where the surface had been built up with addition of metal apparently from the stationary rings. [Emphasis supplied.]

Inspection of unit 4 following shutdown also revealed quantities of debris of all types in the penstock, turbine unit and draft tube. This debris ranged from nails, wire and bits of metal to a four foot length of 2x4 lumber.

Defendant paid P. S. Lord $26,382.73 for dismantling, repairing and reassembling unit 4, and this was accomplished between December 9, 1959 and March 1, 1960. That amount has been withheld from monies otherwise due plaintiff under its contract, and represents part of this claim. By letter dated December 30, 1959, defendant had written plaintiff seeking assurance that plaintiff would agree to continue the contract guarantees in force after defendant’s repairs to unit 4, so long as defendant obtained final seal ring clearances after repair no smaller than those measured on unit 4 before initial startup.

Spare upper and lower stationary rings, which had been furnished by plaintiff under its prior contract to supply the turbines for units 1, 2 and 3, were used to replace the damaged rings. The spare rings were replaced by plaintiff at a cost of $7,980 and that is also part of plaintiff’s claim herein.

The board opinion observes that:

After the repairs made by P. S. Lord, unit No. 4 had lower seal ring clearances ranging from Wiooo to 58Aooo of an inch and upper seal ring clearances ranging from 3%ooo to 594ooo of an inch. These clearances were approximately the same as the original field clearances on units Nos. 1, 2 and 3. [Emphasis supplied.]

Before starting up unit 4 for the second time, defendant added to its list of check-out procedures the requirement that the dial indicators for measuring skate and determining injurious throw or oscillation be continuously observed for the first 12 hours of operation. On February 29, 1960, unit 4 was put back in operation. After iy2 hours of operation, a knocking similar to that heard on December 9, 1959, was heard. Prior to this noise there had been no observable evidence of unusual or abnormal shaft fluctuations, under defendant’s new check-out procedure requiring continuous observation for the first 12 hours. The unit was immediately shut down and dewatered. It was again necessary to send divers down to remove debris from the draft tube threshold in order to seal the steel gates. Although no signs of damage could be detected with the feeler gauges, it is quite possible that the type of gouging later discovered and hereinafter described, would not be apparent without dismantling the unit. After some minor work on the generator, the unit was placed in commercial operation on March 21,1960, and continued in operation until June 6, 1960, when it was again closed down for reasons other than a malfunction. During this shutdown, further damage to the seal rings was observed.

Turning now to power unit 5, the record shows that it too was troubled with problems and, initially at least, these problems were related to debris. On April 1,1960, a leak developed in the concrete stop logs which P. S. Lord Company had been using, in place of steel gates, to temporarily seal the unit 5 draft tube *1379 discharge gates. 8 These stop logs had also been placed with the help of a diver who moved accumulated debris. The sudden inflow from the leak caused water to rush up through a 24-inch pipe which normally functioned as a drain from the penstock to the draft tube. Thus sudden backflow caused steel shot blasting material 9 to scatter throughout the turbine, and a significant amount entered the lower seal ring clearance. Plaintiff’s erection engineer and the Government inspector worked for a number of hours cleaning the turbine.

By April 15,1960, unit 5 was ready for test operations. The as-built upper seal ring clearances ranged from .043" to .054" and the lower seal ring clearances from .040" to .068". As in the case of unit 4, these were approximately the same as original field clearances in units 1, 2 and 3, which have operated without difficulty since their original installation. With respect to its initial operation, the board opinion states:

The mechanical run on Unit No. 5 was started on 15 April 1960 and during preliminary tests it apparently functioned smoothly without indication of any mechanical difficulties. On 4 May 1960, the unit was placed on the line generating 20,000 KW power. Load rejection tests were made at loads of 20,000, 40,000 and 60,000 KW. The shaft throw was checked during load rejections and found to range from .0015 to .002 inch with up to .005 skate at the rate of one skate to ten RPM of the shaft. Throughout 4 May and on through the following day, the unit ran smoothly. At 4:00 P.M. on 5 May the unit was put on condensing for the first time. 10 The unit continued to run smoothly with a shaft throw of .0015 inch and no unusual noise. The unit continued to perform well on condensing throughout the night of 5 May. At 9:00 A.M. on 6 May the unit was again placed on generating at a 20.000 KW load. The unit experienced a load increase of approximately 12.000 KW for several moments at 10:30 A.M. It settled down again to an approximate 22,000 KW load without unusual noises or fluctuations. At about 10:45, a light knock was heard in the turbine. The noise suddenly became louder and the Baldwin erection supervisor instructed the operator to shut the unit down.

It was again necessary to send a diver down to clear debris at the sill of the discharge area of the draft tubes in order to seal the steel gates and dewater the unit. After dewatering, examination of the interior of unit 5 revealed the presence of debris. Gravel was found inside the head cover, an area that had been carefully cleaned out prior to operation. 11 A slight metal buildup had occurred in several locations on the center portion of the lower rotating ring and there were corresponding minor gouges on the lower stationary ring. A piece of metal chip was jammed between the lower rotating and stationary seal rings. The latter was *1380 removed by P. S. Lord Company with a hacksaw. The sum of $2,018.28 paid by defendant for this inspection and repair has been withheld from sums otherwise due plaintiff, and is included in its claim.

Since, as later detailed, defendant seeks to hold plaintiff responsible for these problems on the theory of inadequate seal ring clearances, it should also be observed that plaintiff on May 17, 1960, applied a 50-ton jack in eight different directions to the unit 5 turbine shaft, just below the turbine bearing, to determine if it was physically possible for the rotating seal rings to contact the stationary seal rings through shaft deflection, without any intervening foreign material. Although the shaft was jacked over until it made contact with the turbine bearing and could travel no further, contact between the seal rings was not achieved. The smallest clearances thus obtained were .033" between the lower seal rings, and .029" between the upper seal rings.

The board was unimpressed with this test. Its opinion observes:

* * * Mr. Kurzawa did not concern himself with action at the generator bearing. Although a 50-ton jack was used, it didn’t exert anything like 50-tons of force. * * * Appellant has not shown that the test made with the jack involved forces comparable to those which existed during actual operation.

Although defendant had initially indicated that it would agree to operate unit 5 without further repair or adjustment, it advised plaintiff by letter of May 17, 1960, that it was rejecting both units 4 and 5. Unit 4 was then operating satisfactorily. By telegram June 20, 1960, defendant announced the release of Mr. Kurzawa (plaintiff’s erection supervisor), effective June 2, 1960.

On June 7 and 8, 1960, defendant called a conference at Garrison Dam to discuss the difficulties which had been encountered. As earlier mentioned, the power units were the product of four independent contracts with defendant, and therefore representatives of plaintiff, the generator manufacturer, the installation contractor and the governor manufacturer were invited to the conference. The defendant’s position was stated by the Area Engineer, Colonel Hogrefe, as follows : * '

Perhaps I should state at this time that the installation as a whole is considered unsatisfactory until we have units that we feel that we can rely on and will give us the service here which we anticipated when we put these contracts out. Basically, there are a number of differences here. We feel that inasmuch as the units in toto cannot be considered acceptable that we would like the contractors to come up with what they propose or consider should be done in order to give us satisfactory units and installation. Now, unfortunately we do not have a single contractor to work with; we have all four, but the units themselves are a product of the work of four companies rather than of one and until we have a satisfactory unit, we are attempting to deal with you each individually as far as your contracts go, but as a group in order to find out what can be done to make these units acceptable.

Also representing the Government at the meeting was Mr. John Parmakian of the Bureau of Reclamation of the Interior Department, user and marketing agency for power produced at Garrison Dam. The board opinion states in this respect:

Although single guide bearing generators were approximately $24,000 cheaper than the two-bearing type and had been considered by the Bureau of Reclamation as a means of saving money, Mr. Parmakian said that he was *1381 personally opposed to their use because of experiences outside his Bureau. 12

Although he later stated at the hearing before the appeals board that specific suggestions for action without the test data were purely “conjecture and guesswork,” at this June 1960 conference he stated:

Isn’t the purpose of all this that we have a very sensitive primadonna machine on units 4 and 5. We want to get it less sensitive. Now, the contention of Baldwin is that foreign material and the contention of the user is that this is sensitive and doesn’t appear to be foreign material. Now, what do we do next to make it of less of a primadonna. That is what we are talking about.

Mr. Crawford [Baldwin Representative] : Increase clearance to 100 thousandths like you want and let it go at that.

Mr. Parmakian: Now you’re talking. Let’s pursue that. 13

It is fair to state that four different theories have been advanced to explain the problems encountered on power units 4 and 5, as follows:

1. Inadequate seal ring clearances.

2. Use of a single bearing generator. (This is combined with 1 in the board’s opinion, but they must be considered separately since the single bearing generator was the product of an independent contract between defendant and General Electric, under which the parties had certain responsibilities, as above outlined.)

3. Excessive debris.

4. Insufficiently dissimilar metals in the rotating and stationary seal rings. (This theory has apparently long since been abandoned by defendant.)

The above analysis is supported by the following excerpt from the board’s opinion:

From the outset of the conference [of June 1 and 8, 1960] the Government took the position that the damage to the seal rings was due either (1) to inadequate clearances considering the relative instability of the single bearing generator or (2) to insufficiently dissimilar metals in the rotating and stationary seal rings. Baldwin argued that it felt the clearances, being nearly identical with those in the successfully operating units 1, 2 and 3, were adequate, that the metal in the seal rings was not the cause of the trouble and that in its opinion the damage was caused by foreign materials entering the seal ring clearance. (Long in advance of the hearing before this Board, the Government ceased pursuing the metals theory.)

On 25 June 1960 the Government wrote Baldwin directing that it submit a proposal for increased seal ring clearance; that new seal rings for units 4 and 5 be furnished with material complying with the “nonseizing” provisions of the specifications; and that the turbine guide bearing clearances be reduced to those shown on approved shop drawings. At a later date measurements resulting from the *1382 tests taken at Mr. Parmakian’s suggestion were furnished to Baldwin together with additional data prepared by General Electric as to the mechanical stability of the generators. On 22 July 1960 the Government repeated its request for the proposals outlined in its 25 June letter.

That letter of June 25, 1960, also provides an explanation of how plaintiff happened to be selected, from those present at the conference, to effectuate these' changes. It reads in part:

During this conference it was stressed that the government desired a joint proposal from the contractors for repair of the referenced units. Representatives of the contractors either would not or could not singly or jointly furnish such proposals. The above cited corrective measures are the minimum considered essential at this time. * * * It is obvious that the power-on-the-line schedules for units 4 and 5 cannot be met and that every effort must be made to preclude any avoidable delays.

To assist you in determining the proper clearances for the seal rings, the General Electric Company has been directed to furnish you all pertinent information concerning their generator and to include their computations on its stability under operating loads.

Plaintiff continued to protest that the cause of the difficulties was excessive debris, and it suggested that if the units were operated at less than full head, the heavy accumulation of debris would flush through the turbine without harm. This suggestion was based on plaintiff’s belief that the sudden intake, which characterized the prior operations, caused large amounts of accumulated debris to rush into the turbine head, jamming the seal ring clearances. Plaintiff also suggested that the snorkel on unit 4 be cut to the same length as that on unit 5 and the units then be restored to operation.

In late July or early August, defendant proceeded to clean accumulated debris and foreign material from in front of the penstock intake gates and draft tube discharge gates. A mobile crane with an electric magnet was used to remove steel wire, boulders and gravel and load it into a barge. These cleaning operations by the Government continued for 4 months, that is, until approximately November 1, 1960.

During this same period, in early August 1960, Mr. Parmakian of the Reclamation Bureau consulted with other representatives of defendant and persuaded them to increase the seal ring clearances of units 4 and 5 to about twice their originally installed measurements. The board opinion relates what followed:

By telegram dated 6 August 1960, letter dated 8 August 1960, and a telephone call of 8 August 1960, the Government ordered Baldwin to make “repairs” consisting of increasing the unit 4 and 5 five lower seal ring clearances to approximately 9 %ooo inch and upper seal ring clearances to approximately 7 %ooo. The Government also ordered Baldwin to furnish a new set of spare stationary rings for the upper and lower seals to replace the spare parts used by the Government in the initial repair of unit No. 4.

Under protest appellant proceeded to comply with these directives. The work commenced on or about 1 September 1960. Unit 4 was placed back in operation on 28 September 1960 and unit 5 on 18 October 1960.

It appears that the Government’s decision to order opening of the seal ring clearances was based primarily on the recommendations of Mr. Parmakian which were set forth in a document entitled “Travel Report” dated 9 August 1960. Based on what he termed “current field test data” he made an analysis for determining required minimum seal ring clearances for units 4 *1383 and 5. This analysis, taking into consideration seven so-called “conditions,” was set forth in Mr. Parmakian’s report as follows:

Condition 1 — Closure of seal ring clearances due to application of reservoir pressure to the casing as read from the scratch gages.

Condition 2 — Out of roundness of runner after repairs are made (estimated).

Condition 3 — Radial expansion of runner due to centrifugal forces.

Condition 4 — Closure of seal ring clearances due to guide bearing clearances with shaft touching opposite sides of the guide bearing.

Condition 5 — Increase in Condition 4 due to movement of turbine bearing resulting from full servomotor reaction.

Condition 6 — Closure of seal ring clearances due to unequal deflection of generator bridge.

Condition 7- — One-half of minimum design clearances shown on turbine manufacturer’s drawings.

Total 1 to 7, inclusive — Suggested minimum seal ring clearances.

CA339&]

It is noted that Mr. Parmakian’s suggested minimum clearances were about double the mínimums measured after the seal ring troubles occurred in Units 4 and 5. They were arrived at by adding together the allowances made for all such conditions. [Emphasis supplied.]

Mr. Parmakian testified that each of the deflections considered in the seven conditions could occur under normal operating conditions and that it would be possible for a large percentage of them to occur at the same time in the same direction. With respect to the large deflection introduced by Mr. Parmakian in Condition 7, he testified as follows:

I picked out a safety — not a safety, no, I added an additional amount of clearance, approximately equal to about half of what the turbine manufacturer had in its original design drawings.

I don’t consider these as being a factor of safety, I consider these as being all-inclusive to include a lot of other conditions that could not be *1384 proved out at the spot, because of reservoir conditions, and because of many other difficulties.

Also, the expensive — the cost of actually doing this over and over again, if this machine could be repaired in a day — completely cleaned out with no damage whatsoever—

Parmakian arrived at his conclusions with respect to inadequate clearances on units 4 and 5 without any consideration of the successful operation of Units 1, 2 and 3.

Mr. Parmakian’s “Travel Report” (see footnote 13 swpra), was prepared without observation of the turbines while they were either dismantled or in operation. Of the seven factors considered as conditions affecting design of minimum seal ring clearances, only two were based on actual measurements. Furthermore, the measurements actually taken were of doubtful accuracy as they did not reflect operational values, nor was any consideration given, as the board opinion found, to the prior successful operation of units 1, 2 and 3.

At an August 9, 1960 meeting held by defendant at plaintiff’s facility, the generator contractor refused to take any corrective action unless issued an amendment to its existing contract or a new contract. The generator contractor was subsequently paid $54,400.85 for the costs it incurred in the course of effecting changes in units 4 and 5, as hereinafter described. 14

Plaintiff agreed at that meeting to proceed under protest. Mr. Kurzawa, plaintiff's erection engineer (who had been previously removed from the project by defendant), returned to Garrison Dam to effect the required changes. Employing defendant’s installation contractor, P. S. Lord Company, plaintiff increased the upper seal ring clearances on both units to .070", and the lower seal ring clearances on both units to .090" whereupon the units were returned to service in September and October 1960.

On December 24, 1960, the Government notified plaintiff by telegram that it had scheduled final inspection and testing of units 4 and 5. These tests were arranged for the period between January 9, 1961 through January 20, 1961, in order to complete them before expiration of the 2-year guarantee period. The telegram also stated that the Corps of Engineers had complied with plaintiff’s suggestion of August 9, 1960, by operating the turbines for 90 days without condensing operations. This recommendation had been intended to permit the turbines sufficient time to flush through excess debris.

By letter dated January 17, 1961, defendant requested Baldwin’s concurrence in extending the contract guarantee period for 1 year from the date of repair on the ground that malfunctioning of the turbine units had rendered it impossible to conduct its tests within 2 years from the date of delivery. Plaintiff refused to extend the guarantee period, stating that the difficulties with the seals were not its responsibility.

In rejecting this request of defendant, plaintiff pointed out that comparable turbine units, with comparable single bearing generators, had been operating without difficulty at the Niagara Power Station in New York, with minimum seal ring clearances less than those originally installed at Garrison on units 4 and 5, and approximately one-third of those ordered by defendant in August 1960.

Plaintiff requested and received seal ring clearance data in July 1961 which showed that units 1, 2 and 3 had been operating for years with clearances equal to or smaller than those originally provided for units 4 and 5; that the clearances of all five units were decreasing due to the accumulation of calcium-like deposits on the seal rings; and that in a matter of months, the enlarged unit 4 *1385 and 5 clearances had been substantially decreased.

In August 1961, plaintiff furnished defendant with further details concerning trouble-free operation at Niagara, and invited defendant’s representatives to visit and inspect the one-bearing Niagara units which were operating with small clearances. In November 1961, General Electric furnished plaintiff with additional data on the stability of one-bearing generator units, demonstrating that the difference between the one-bearing and two-bearing generators was insignificant with respect to possible lateral deflection at the turbine seal rings caused by any side thrust forces exerted upon the turbine runner.

Plaintiff’s claim was filed in November 1961. As later stipulated by the parties and filed with the board, it totals $112,053.86, made up as follows:

1. Withheld from amounts otherwise due plaintiff ................ $28,401.01

2. Cost of replacing stationary rings .. 7,980.00

3. Erection engineering service ...... 7,149.46

4. Field engineering service ........ 9,199.24

5. Cost of increasing seal ring clearances ............_ _........ 59,324.15

$112,053.86

The claim was denied August 9, 1963, by the contracting officer for the following reasons:

a. By reason of Paragraphs SC-14 and SC-15 of the contract specifications, Contractor has undertaken to provide turbines that will be designed to give satisfactory performance. This undertaking was not an “impossibility” and, therefore, Contractor cannot recover any extra costs incurred by providing a satisfactory installation. * * *

b. It is obvious that the turbines were defectively designed because after greater clearances were obtained, the turbines operated satisfactorily. [Emphasis supplied.]

The subsequent board decision dated July 1, 1966, characterizes that contracting officer’s decision as follows:

On the ground that after the clearances were increased there was no further trouble, the Contracting Officer concluded that the proximate cause of the trouble was insufficient clearance perhaps coupled with differences in the generators (between those furnished for units 4 and 5, and those furnished for units 1, 2 and 3).

Paragraph SC-14 of the specifications cited in the above quoted portion of the contracting officer’s decision provides as follows:

SC-1 k. Guarantees. If within one year after any one turbine unit and its related equipment is placed in commercial operation, but not later than two years after the completion of delivery of the turbine unit and its related equipment f. o. b. delivery point, any parts of the turbine or appurtenances are found defective because of design, workmanship, or material, the Contractor shall, at his own expense, furnish necessary replacement parts of design, workmanship, and material approved by the Contracting Officer. The pitting of a turbine runner beyond the limits specified in paragraph 2-05 will be considered a defect within the meaning of this paragraph. [Emphasis supplied.]

The contracting officer apparently construed this clause as constituting plaintiff an insurer (save for “impossibility”) of the satisfactory operation of the turbines. On appeal, while affirming the contracting officer’s decision, the board construes the provision somewhat differently. It states:

The Government relies mainly on the guarantee clauses of the contract in justification of its withholding moneys from Baldwin-Lima-Hamilton and its refusal to pay for increasing the seal ring clearances. In proceeding under the guarantee clauses, the Government has the burden of showing that damage to the seal rings occurred as the result of the Contractor’s defective design, material or workmanship. See Appeal of The Heil Co., ASBCA No. 10047, 65-2 BCA Par. 4924, and cases cited therein. Re *1386 gardless of the burden which a claim under the guarantee clauses imposes, the Government, of course, has the burden of proof as to affirmative claims against Baldwin; in this case, the items for which the Government is withholding money otherwise due to the contractor. Appeal of Mecham Construction Company, Inc., ENG BCA No. 2112. [Emphasis supplied.]

After setting forth these standards, the board observes that “neither party has been able to prove with absolute certainty the cause of the seal ring damage. The extent of the Government’s burden, however, is not to establish with such certainty the cause of the damage which it has alleged. It must show by a preponderance of the evidence that the most probable cause of the damage was defective design, materials or workmanship when considered with other possible causes. Appeal of Jefferson Construction Co., ASBCA No. 7008, 1962 BCA 3409.”

The board recites the main facts not in dispute as follows:

* * * (1) [T]hat damage occurred to the seal rings three times — • twice to unit 4 and once to unit 5, (2) since the seal ring clearances were increased in September and October, 1960, both units have operated without further serious seal ring damage so far as anyone knows and (3) prior to operating the units with increased clearances, the Government made some degree of clean-up of the areas in front of the penstock intake and the draft tube discharge areas.

After discussing the contentions of the parties, it concludes:

We think the Government has sustained its burden of proof and hold that the most probable cause of the damage in question was appellant’s failure to install units 4 and 5 with adequate seal ring clearances.

In response to plaintiff’s citation of successful operation with the same or lesser clearances at Garrison (units 1, 2 and 3) and at Niagara, the board observed that there were “significant differences” in those turbine installations so that “the Board can hardly conclude that seal ring clearances which have proved adequate at one must necessarily be adequate for another.”

The opinion flatly rejects the plaintiff’s theory that seal ring problems were attributable to excessive amounts of foreign material in these words:

Lacking convincing evidence that the amounts of foreign material involved here were unusually large or unusually hard we shall not consider at length the mechanism by which appellant contends the material entered the clearance and caused damage. Turbines are expected to handle normal amounts of foreign material without serious damage. If they cannot do so, there is something wrong with the design. In this case Mr. Kurzawa testified that the damage could have been caused by as few as six tiny particles of quartz having a diameter no greater than the thickness of a dime. We would have reservations about the adequacy of the design of a large turbine which could be so severely damaged by six tiny particles of quartz. Additionally, appellant has not persuaded us of the reasonableness of saddling the Government with the burden of making the area around the penstock intake so immaculate that such minute particles could not enter the turbines.

Lacking proof of the existence of excessive quantities of foreign material in and around the turbines at the time they were first operated, we attach no significance to the fact that there was no further damage to the seal rings after the Government made efforts to clean debris away from the penstock intakes and the draft tube gates. This cleaning simply happened to coincide with the enlargement of the seal ring clearances which, in our opinion, obviated further damage.

The board’s analysis of the issue in this case is quite correct. Its conclusions, however, to the extent that they depend on determinations of fact, are not *1387 supported by substantial evidence; nor are those conclusions correct as to the crucial issues of law upon which they necessarily depend. 15

This case is a classical “whodunit,” reminiscent of recently publicized efforts to reconstruct and to determine the causes of accidents in space or flight, under circumstances where theory must ultimately substitute for fact, because the facts have largely been obliterated by the incident. The facts from which inferences and theories must be drawn in this case, are typically meager. As the board admits, all that one can know for sure is that damage occurred, and that there was no further trouble after both of the causes urged upon the court by the parties, were simultaneously eliminated. Because the drastic enlargement of the seal ring clearances, and the extensive cleanup of the accumulation of debris, coincided in point of time, there is no longer available the “controlled experiment” cherished by our scientific brethren. One can now but conjecture as to which of the two causes, if either, was the culprit in this case.

As outlined above, four possible theories have been advanced at one time or another to explain what happened here. The board correctly stated that “the Government has the burden of showing that damage to the seal rings occurred as the result of the Contractor’s defective design, material or workmanship * * * [and] * * * the burden of proof as to affirmative claims against Baldwin; in this case the items for which the Government is withholding money otherwise due to the contractor.” It concluded that the Government had sustained this burden of proof when, out of these four alternatives, it selected as the “probable” cause of the damage, a design deficiency predicated upon insufficient seal ring clearances. If that, indeed, does constitute a determination “concerning a question of fact,” it is not supported by evidence in the record which can be characterized as substantial. 16 And this is aside from the question of whether the board erred as a matter of law in finding that defendant had sustained its previously acknowledged “burden of proof,” in the face of these alternatives.

Moreover, it cannot as a matter of law be deemed a correct interpretation of the “Guarantees” clause quoted above, to characterize as “defective design” seal ring clearances virtually identical to those successfully employed on power units 1, 2 and 3; and virtually identical to those employed by defendant when it repaired unit 4 after the first mishap. In view of the provisions of TP 2-15, “Wearing Rings and Facing Plates,” to the effect that “radial clearances between the moving and stationary wearing rings shall be as small as possible, consistent with safe operation” to insure maximum power generating efficiency, plaintiff might more readily have been charged with defective design had it substantially enlarged the clearances over its prior successful experience.

Returning to the available “evidence” relating to the cause of the damage, Mr. Swain, the Government’s erection engineer, and the only witness for defendant who actually observed the seal ring damage, testified that he was unable to form an opinion as to what initiated the damage. This is because, once initiated either by debris or by a contact, the damage tended to propagate itself, in the course of which it destroyed any clues as to its origination. It is of little proba *1388 tive value, therefore, that after initial damage there was contact between the rings, with resultant damage consistent with rubbing. In fact, specification TP 2-15, “Wearing Rings and Facing Plates,” shows that the design contemplated some transient contact since it prescribed steel in the rotating and stationary rings of sufficiently dissimilar characteristics “to prevent seizing in case of accidental rubbing." [Emphasis supplied.]

Plaintiff argues, from the fact that gouging initially discovered on unit 4 occurred at the center of the lower seal ring rather than at its edges, that the damage .is more consistent with debris than deflection of the runner. Here, once again, one can only engage in surmise. The seal ring clearances were so relatively small that a slight deflection could conceivably produce damage at the center of the ring rather than at its top or bottom edge.

There is considerable evidence identifying debris as the culprit. This was the testimony of all of plaintiff’s witnesses, supported by photographs, and it was not rebutted by defendant’s witnesses. One of the latter, Mr. Soucek, conceded that the turbines were carrying a heavy load of material prior to the Government’s extensive cleanup operations. The absence of fluctuation in the dial indicators prior to damage is also more consistent with debris than it is with insufficient clearance as an initiating cause. Finally, there are the actions of the Government in devoting 4 months to cleaning and flushing the intake and discharge areas, while plaintiff was complying with orders to widen the seal ring clearances. In summary, there is little if any credible “evidence” linking the damage to defective design of the seal ring clearances. It is little more than a guess, elevated to a finding of fact, and that which is offered to support it is not substantial evidence within the definition of that term. The later doubling of the clearances has more of the characteristics of a change than of a “repair,” and appears more concerned with future safety than with the cause of the damage.

However, the board opinion charges the plaintiff herein with “inadequate clearances considering the relative instability of the single bearing generator." (Emphasis supplied.) This is a determination necessarily involving an interpretation of plaintiff’s contract, plus an interpretation of defendant’s contract with General Electric, plus an evaluation of defendant’s own residual responsibilities after it elected to award four separate prime contracts for the construction of these power units. It is therefore a determination on a question of law, not final, and subject to plenary consideration herein. 17

At the time this contract was awarded to plaintiff and its contract' price was fixed, it was. encouraged to provide turbines identical to those which it had previously and successfully provided for units 1, 2 and 3. As a result, a saving of $108,000 was realized by defendant. Plaintiff’s shop drawings, identical to those submitted for units 1, 2 and. 3, were approved. All this occurred long before defendant awarded a contract for a different type of generator to General Electric, against the latter’s advice. The board opinion acknowledges that the two contractors had the normal interchange of information contemplated by the “Cooperation” clause contained in both contracts. 18 Field clearances were identical *1389 with those successfully employed on units 1, 2 and 3, and were approved by defendant for all five units. Defendant employed the same clearances in its initial repair of unit 4. General Electric’s contract contained a provision requiring it to provide an upper guide bearing “if it is deemed necessary by the Contractor for satisfactory operation.” General Electric now urges that use of a single bearing generator would not in fact cause any instability in the turbine. The experience at the Niagara Power Plant, utilizing single bearing generators in conjunction with small seal ring clearances, is offered to absolve the single bearing generator of any responsibility for the damage at Garrison.

All the foregoing notwithstanding, defendant urges that plaintiff had a last clear chance to adjust for the single bearing generators, and should have known that they called for larger seal ring clearances than those successfully employed on the earlier power units.

This conclusion inexplicably absolves General Electric and defendant of any responsibility for the exercise of similar foresight or perception. 19 It ignores the fact that the Government, when it awarded four separate prime contracts for the construction of the power units, itself became in effect the prime contractor, with some residual responsibility for coordinating the efforts of its contractors. 20 Nowhere in its contract did plaintiff undertake to become the insurer or systems analyst or manager of the entire power unit in which its turbine was installed. To hold plaintiff responsible because its product sustained the damage is equivalent to holding that the owner of the vehicle damaged in an accident is for that reason responsible for causing the accident.

Even assuming, therefore, that use of the single bearing generator was in fact the cause of this damage, it is an error of law under the terms of these contracts to hold plaintiff responsible therefor. 21

In view of the foregoing, it is unnecessary to determine whether, as urged by plaintiff, the board erred as a matter of law in applying an erroneous burden of proof standard. 22

CONCLUSION

Plaintiff’s motion for summary judgment is granted, defendant’s cross-motion is denied, and judgment is entered for plaintiff in the sum of one hundred twelve thousand, fifty-three dollars and eighty-six cents ($112,053.86). 23

Notes

Another legal error infecting the determination is the conclusion that plaintiff was required to design the turbines so as to avoid occasional, transient contacts.

. A turbine consists essentially of a series of vanes, known as the “runner,” attached to a central rotating shaft within a stationary shell. Water flowing against the runner, causes it and the shaft to rotate. The shaft is coupled to a generator which converts this mechanical energy into electricity. The seal rings mentioned in the text circle both the upper and lower borders of the runner and turn with it. There are stationary seal rings on the turbine shell, and the distance between these and the rotating seal rings on the runner, is known as the “seal ring clearance.”

. 68 Stat. 81, 41 U.S.C. §§ 321-22 (1964 ed.), as follows:

“§ 321. Limitation on pleading contract-provisions relating to finality; standards of review.

“No provision of any contract entered into by the United States, relating to the finality or conclusiveness of any decision of the head of any department or agency or his duly authorized representative or board in a dispute involving a question arising under such contract, shall be pleaded in any suit now filed or to be filed as limiting judicial review of any such decision to cases where fraud by such official or his said representative or board is alleged: Provided, however, That any such decision shall be final and conclusive unless the same is fraudulent or capricious or arbitrary or so grossly erroneous as necessarily to imply bad faith, or is not supported by substantial evidence. * * *

“§ 322. Contract-provisions making decisions final on questions of law.

“No Government contract shall contain a provision making final on a question of law the decision of any administrative official, representative, or board. * * ”

See also United States v. Utah Constr. & Mining Co., 384 U.S. 394, 86 S.Ct. 1545, 16 L.Ed.2d 642 (1966) ; United States v. Anthony Grace & Sons, Inc. 384 U.S. 424, 86 S.Ct. 1539, 16 L.Ed. 2d 662 (1966); and United States v. Carlo Bianchi & Co., 373 U.S. 709, 83 S.Ct. 1409, 10 L.Ed.2d 652 (1963).

. Cf. “Disputes” article in current use which “does not preclude consideration of law questions in connection with decisions provided for in paragraph (a) above [on questions of fact] ; provided, that nothing in this contract shall be construed as making final the decision of any administrative official, representative, or board on a question of law.”

. Ray D. Bolander Co. v. United States, 186 Ct.Cl. 398, 409 n. 11 (1968); Maxwell Dynamometer Co. v. United States, 386 F.2d 855, 870, 181 Ct.Cl. 607, 631, (1967); and Loral Electronics Corp. v. United States, 387 F.2d 975, 980, 181 Ct.Cl. 822, 832 (1967).

. Specifically, the Invitation for Bids provided in pertinent part:

“11. Evaluation and Award.

“d. * * * However, where the bidder heretofore has made model tests or field tests on the turbine he proposes to furnish, the Contracting Officer, at his discretion, if sufficient data are furnished with the bid to enable him to determine the acceptability of the tests, may deduct from the bid the unit price cost of such tests as stated in the bid. The total bid of any such bidder, less the costs stated for said tests, will be considered to be the net amount of the bid submitted for the purpose of determining the lowest acceptable bid and in making the award.

“e. If a bid is based on furnishing turbines sufficiently similar to the installed turbines that the tools, wrenches, lifting devices and spare parts on hand will fit the turbines to be furnished, the total bid less the bid price for Items Nos. 9 and 12 to 20 inclusive will be considered to be the net amount of the bid submitted for *1374 the purpose of determining the lowest acceptable bid and in making award.

“f. If a bid is based on furnishing turbines sufficiently similar to the installed turbines to permit extensive reuse of shop drawings for the installed turbines, $6,000 will be deducted from the net amount of the bid for the purpose of determining the lowest acceptable bid and in making award. This amount of $6,000 represents the lessened cost to the Government for processing the shop drawings and other engineering work.

“i. Notwithstanding any other provisions of these specifications concerning the method of award, award will be made on all items listed in the schedule as a whole to one bidder, except that the Government reserves the right to not make award on Item Nos. 7, 8, 9 and 12 to 20 inclusive.”

. The incentive for maintaining seal ring clearances “as small as possible” consistent with safety is that water bypassing the runner represents a substantial loss of power, and therefore a corresponding loss of power revenue.

. This important incident is described in somewhat greater detail by the Government engineer in charge of installation, as follows:

“At 3 minutes before 1:00 p. m. a loud thumping noise was heard in the turbine. Mr. Kurzawa, Baldwin-Lima-Hamilton Representative, and Mr. Pearson, mechanical inspector, were close by and upon entering the turbine pit they cheeked the dial indicators at the turbine bearing. The indicators were fluctuating .007 of an inch in exact time with the rotation. They also noticed that the seal water supply pipe fastened to the wall of the turbine pit was moving in jerks and sounded like it was being struck by a hammer. The striking noise was in exact time with the rotation (90 times per min.). Mr. Harris, General Electric erecting engineer, checked the dial indicators at the thrust bearing and found the movement erratic and more than twice as much as normal. At approximately 1:04 p. m. the noise subsided and the dial indicators at turbine bearing showed a reduced movement to .003 of an inch. At this time the dial indicators at the thrust bearing settled down to near normal movement. Mr. Swain was called and arrived at the Powerhouse at 1:10 p. m. After being informed of what Had happened and hearing a dull thump, when listening with his ear at the seal water supply pipe, he ordered the unit shut down, of which action Mr. Kurzawa concurred. The unit stopped rotating at 1:20 p. m.

“Mr. Hutchins was immediately notified of what had happened and the action taken. During the disturbance, the turbine bearing temperature showed a sharp rise from 40° C to 44° C. When Mr. Kurzawa was queried concerning the advisability of him having the unit shut down immediately upon hearing the noise in the turbine, he stated that the conditions were identical with those experienced when large pieces of timber would go through the unit and in most of such cases, serious damage did not result. He further stated that if large pieces of timber or metal should get caught in the runner, the method of shut down should be by using the head gates.

“The draft tube stop logs were placed in Unit No. 4 draft tube opening during the afternoon of 9 December. Excessive leakage at the stop logs made unwatering impossible until divers succeeded in sealing off the leak's. The unit was unwatered at 1:00 p. m. on the 10th of December. Examinations made during the afternoon of 10 December disclosed damage (galling) at four locations on the lower seals and at three locations on the upper seals. A piece of 2" x 4" timber approximately 4 feet long was found lying across two runner buckets.”

. The only available set of steel gates was being used on unit 4.

. The steel shot blasting material had been left there by defendant’s penstock construction contractor, Southwest Welding Company, and was apparently plugging the drain line.

. The condensing mode of operation consists of closing the wicket gates and stopping the water input to the turbine runner, depressing the draft tube water level below the turbine runner level by the application of air pressure, and then operating the generator as a motor, with the turbine spinning free.

. Later, in August 1960, the Corps of Engineers, Missouri River Division Laboratory, did a sieve analysis of “sand” taken from the turbine head of unit 5 on August 12, 1960. . “Sand” was defined as any material that would pass through a quarter-inch sieve. The test indicated that among the debris, there were pieces of rock and metallic material small enough to enter the seal ring clearances.' Some of this debris consisted of siliceous rock, igneous rock and bits of concrete, all of which were found to be harder than steel on the Mohs scale. The latter is a standard measurement employed by geologists for identifying and classifying minerals in terms of hardness.

. It will be recalled that it was the Bureau of Reclamation that had decided that the reactive capability provided by the double bearing generator was unnecessary in units 4 and 5, and had suggested that the Army Corps of Engineers (the agency awarding and administering these contracts), use the cheaper single bearing, thus effecting a saving in the cost of the power which Reclamation was to purchase and market.

. Somewhat later, on August 9, 1960, Mr. Parmakian issued a “Travel Report,” so-called, setting forth his recommended clearances for units 4 and 5. In that report, reference is made to a June 15, 1960 report (not in the board record), in which he concluded “ * * * that with one guide bearing at the generator the turbine seal ring clearances which were provided for units 4 and 5 were inadequate.”

. Notwithstanding a provision in the generator contract, quoted earlier in the text, requiring General Electric to furnisli a two-bearing generator should the one-bearing design interfere with satisfactory operation.

. Notes 2 and 3 supra. See also Ray D. Bolander Co. v. United States, 186 Ct.Cl. 398, 405 (1968).

. “The substantiality of evidence must take into account whatever in the record fairly detracts from its weight. * * * Congress has merely made it clear that a reviewing court is not barred from setting aside a Board decision when it cannot conscientiously find that the evidence supporting that decision is substantial, when viewed in the light that the record in its entirety furnishes, including the body of evidence opposed to the Board’s view.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 488, 71 S.Ct. 456, 464, 465, 95 L.Ed. 456 (1951).

. Paccon, Inc. v. United States, 399 F.2d 162, 185 Ct.Cl. 24 (1968).

. The interpretation of a contract by the parties before it becomes the subject of controversy, is deemed by the courts to be of great, if not controlling, weight. General Warehouse Two, Inc. v. United States, 389 F.2d 1016, 181 Ct.Cl. 180 (1967); Micrecord Corp. v. United States, 361 F.2d 1000, 176 Ct.Cl. 46 (1966); Houston Ready-Cut House Co. v. United States, 96 F.Supp. 629, 119 Ct.Cl. 120 (1951). In this ease, the conduct of the plaintiff, the Government, and the generator contractor indicated they were satisfied with the information exchanged between plaintiff and General Electric.

*1389 Also, the Government appeared initially to be assuming responsibility for the damage, and it did not invoke the provision requiring the contractor to effect necessary repairs where it was charged with responsibility.

. Despite General Electric’s special contractual responsibility quoted above. As noted, General Electric has been reimbursed for its costs in connection with enlargement of the seal ring clearances. Also, Cf. Helene Curtis Indus., Inc. v. United States, 312 F.2d 774, 160 Ct.Cl. 437 (1963), on the Government’s responsibility to disclose knowledge it possesses.

. The use of separate prime contracts, in lieu of a single contractor who then subcontracts the separate elements of the power unit, undoubtedly resulted in savings by avoiding the pyramiding of profit and overhead. But it correspondingly shifts to the Government the management responsibility ordinarily vested in a single prime contractor.

. Cf. C. W. Regan, Inc., 67-2 BCA ¶ 6454; Hardeman-Monier-Hutcherson, 67-1 BCA ¶ 6210; Drake America Corp., 60-2 BCA ¶ 2810, and cases cited therein, with the board decision in this case.

. Cf. Spartan Aircraft Co. v. United States, 100 F.Supp. 171, 120 Ct.Cl. 327 (1951).

. That amount having been stipulated by the parties and filed with the board. Cf. United States v. Anthony Grace & Sons, Inc., 384 U.S. 424, 86 S.Ct. 1539, 16 L.Ed.2d 662 (1966).