square curve over two torsional cycles. VT at 15:20. In this regard, the agency provided to Goodyear a printout, dated March 4, 1977, from its own test machine that showed the application of a square radial load curve over two torsional cycles and in-phase cycle loading; this printout identified MIL-T-11891B as the applicable military standard. Revision D was rescinded in February 1988 because the government found that no contractors, including those previously qualified under revision B, could satisfy the more stringent road test requirement contained therein-a requirement that was not related to the bushing qualification test requirements. VT at 14:59-15:00. After 1988, MIL-T-11891B was the applicable military standard, and was in effect at all relevant times concerning Varec's qualification tests in 1983 and source approval in 1991. VT at 12:07, 15:00.

As noted above, the Army accepted, without witnessing, Varec's bushing qualification test results from Varec's own test machine, which had not been approved by the Army, despite the specific requirements of MIL-T-11891B that tests be conducted by the government on an approved machine. VT at 11:30. For the reasons stated below, we find, from our review of the record, that the qualification test results of Varec's bushings demonstrate that the test was not conducted as contemplated by MIL-T-11891B.

First, we conclude that Varec bushing test was not conducted with in-phase loading. The certified test results provided to the agency show that for six of the required eight bushings the radial and torsional cycles were not maintained in an exact 4 to 1 ratio.11 This demonstrates that the cycles were not applied inphase, since if the respective loads were applied in-phase, the 4 to 1 ratio would necessarily have been maintained because of the testing machine's mechanical linkage. VT at 15:40. Varec, although responding to a number of challenges concerning its testing machine and bushing qualification tests (i.e., the age and temperature of the test bushings), does not assert that the test machine that it used applied the radial and torsional loads in-phase; instead, Varec asserts that in-phase loading is not a requirement of MIL-T-11891B.

We also are unpersuaded that Varec's bushing test machine applied the radial load as a square curve over two torsional cycles, even though Varec, in comments to the agency after our hearing, asserted, for the first time, that its test machine applied a square radial load curve similar to that of the governmentowned test machine, and provided a hand-drawn illustration of its radial load curve, dated July 10, 1992, that it asserts was applied in qualifying its bushings. Varec repeated this assertion to us in an affidavit that included the hand-drawn

curve.

Varec has produced no contemporaneous evidence of the radial load curve actually applied by its test machine in 1983. In contrast, the evidence produced to substantiate the radial load curves applied by Goodyear's and the government's

11 While the Army asserts that the Varec bushing test was sufficiently close to a 4 to 1 ratio to be acceptable to the Army, the primary point is, as detailed below, that this evidences that the Varec bushing qualification test was materially less stringent than that contemplated by MIL-T-11891B.

test machines are printouts generated by the test machines themselves. As noted by Goodyear, it would be impossible to accurately plot, without an electronic recording device, pressure versus time for a load that completes its cycle in less than 1 second, as is the case for the cycles on the bushing test machines. In addition, the test data provided by Varec for its bushing qualification tests call into question whether a square radial load curve was applied. Unlike the results of testing on Goodyear's bushings, Varec's bushings, although tested over more cycles than required, showed minimal signs of deflection (or fatigue). VT at 15:50-57. Varec and the agency argue that Varec's bushings must have evidenced only nominal deflection because the bushings are so much better than required by the agency's qualification test standards. The test results show that Varec's bushings showed nominal deflection over the life of the test, while the test results for Goodyear's qualified bushings showed a steady deflection over the life of the test until eventual failure well beyond the required test limits. We concur with the testimony of Goodyear's engineer that the most likely explanation for this is not that Varec produced "super" bushings, but that the Varec bushings were not tested in accordance with MIL-T-11891B. VT at 15:56-57, 16:13-14.

Finally, Varec does not state how in 1983 it knew the shape of the required radial load curve to be applied or that it was to be applied over two torsional cycles, since Varec's qualification tests were conducted prior to the release of the C and D versions of MIL-T-11891, which for the first time specifically informed bidders of the required shape of the radial load curve. Also, there is no representation from the Belgian ministry of defense that a square radial load curve was applied by Varec's test machine during the qualification tests of Varec's bushings, although an affidavit has been provided from a member of the Belgian military who witnessed the qualification tests, which generally states that Varec passed the bushing tests required by MIL-T-11891B. Given the lack of supporting evidence that the tests were conducted in accordance with the agency's minimum requirements, we find that Varec's qualification tests were less stringent than those required of other bidders.12

The Army and Varec argue that MIL-T-11891B does not expressly require that the radial and torsional loads in the bushing qualification test be applied inphase or that the radial load be applied as a square curve. However, as noted above, MIL-T-11891B contemplated that all bushing qualification tests would be conducted by the government on its own test machine. Thus, it must be kept in mind that in interpreting the requirements of MIL-T-11891B, these requirements are based on testing on a particular government machine. Since the government-owned test machine applies the radial and torsional loads in-phase and the radial load on a square curve, and the Army does not contend that its own bushing test machine tests for characteristics beyond the requirements of MIL-T-11891B, we find that MIL-T-11891B contemplates that the bushing

12 As noted above, FOC was unable to pass the bushing qualification test on the government-owned test machine in time to receive award under the IFB.

qualification test be applied in-phase and that the radial load be applied as a square curve.

Consistent with the foregoing interpretation of the MIL- T-11891B qualification testing requirements, Goodyear contends that the Defense Standardization Manual DOD 4120.3-M (August 1978), "Defense Standardization and Specification Program Policies, Procedures and Instructions," and the Defense Standardization and Specification Program Manual SD-6 (June 1986), "Provisions Governing Qualification," provide that qualification testing of foreign-made products be done in the United States at an approved facility and that:

Test data generated as a result of qualification testing in accordance with SD-6, Provisions Governing Qualification, shall be the only test data evaluated during the qualification process. Test data generated prior to authorization of, or outside the purview of, qualification tests (e.g., from first article tests unless generated under the cognizance of the qualifying activity) may not be used in lieu of qualification test data.

DOD 4120.3-M, which controls the QPL process, states that it is "mandatory for use by all DOD activities," and FAR § 9.203(c) provides that "[i]nstructions concerning qualification procedures" are included in DOD 4120.3-M. As Goodyear points out, if Varec's bushings were tested in accordance with DOD 4120.3-M and MIL-T-11891B, they would have been tested on the KRC test machine and subjected to radial and torsional cycle loading that was in-phase and to a square radial load curve.

The record shows that the Army, in accepting the Belgian certified test results, did not comply with the above provisions of DOD 4120.3-M.13 The Army contends, however, that it was required to accept Varec's certified test results, despite the failure of Varec to obtain the agency's prior approval to conduct qualification testing, because of a memorandum of understanding (MOU) between the United States and Belgian governments that required full consideration of Varec's application for qualification14 and because DFARS § 225.872-3(f)(1) required that the agency:

Consider the adequacy of qualifying country service testing on a case-by-case basis. Departments or agencies that must limit solicitations to sources whose items have been service tested and evaluated by the department/agency shall consider supplies from qualifying country sources that have been tested and accepted by the qualifying country for service use.

In our view, the Belgian MOU and DFARS § 225.872-3 permit the agency to accept certified test results from the Belgian ministry of defense in some circumstances. Specifically, DFARS § 225.872–3(f) instructs the agency to consider the adequacy of qualifying country testing on a case-by-case basis and to consider the acceptability of supplies from sources that have been tested and accepted by the qualifying country. The MOU also requires that the agency give full consideration to applications for qualification.

13 The agency states that the intent of DOD 4120.3-M and SD-6 is to ensure the reliability of qualification test data.

14 The MOU between the United States and Belgium is set out as appendix T:139 to the DOD FAR Supplement (DFARS).

However, this regulation and MOU do not provide, or imply, automatic acceptance of qualification test results certified by Belgian authorities. See T.G.L. Rubber Co., Ltd., supra (TACOM reasonably refused to qualify the protester's product despite test results certified by the Israeli ministry of defense). Rather, DFARS § 225.872-3(f) and the MOU only allow the Army to accept, for the purpose of evaluating a product for potential qualification for listing on a QPL, the certified test results of the Belgian ministry of defense if the Army can be reasonably assured that the tests were performed in accordance with its requirements.15 In this case, since Varec has been permitted to test its own bushings on a nongovernment machine, contrary to the express requirements of DOD 4120.3-M and MIL-T-11891B, the Army was required to ensure that Varec and the Belgian ministry of defense were aware of the operational requirements that the bushing qualification test must satisfy and that the tests were actually performed in accordance with those requirements before the Army qualified the product for listing on the QPL.

As discussed above, the record here shows that Varec and the Belgian ministry of defense were not aware of the requirement that in-phase cycle loading be performed or that they were aware of the requirement that the radial load curve be applied as a square curve over two torsional cycles. The record also shows that these deviations from the qualification tests contemplated by MIL-T-11891B, to which all other qualified products have been subjected, are material. In this regard, the Army admits that if the torsional and radial load cycles are not maintained in-phase, the bushing qualification test is less severe than if the loads are applied in-phase. 16

Nevertheless, the agency argues that even if Varec's bushings were not tested with the torsional and radial load cycles being maintained in a 4 to 1 ratio and in-phase, the data presented was sufficient for the agency to conclude that Varec's bushings would satisfy its requirements. Specifically, the agency argues that Varec's bushings passed qualification testing over approximately 175,000 cycles while MIL-T-11891B only requires testing over 150,000 cycles. In the agency's view, the longer test ensures that, even though the loads were not applied in-phase, the maximum pressure on the bushing was applied often enough to be equivalent to in-phase testing over 150,000 cycles.

The problem with the agency's analysis is that it assumes that the radial load pressure applied on Varec's bushings was by a square curve over two torsional cycles.17 If such a square radial load curve were not applied during the test, then the maximum pressure on the bushing would occur much less often than assumed by the agency. Moreover, the agency's chief engineer testified that the shape of the radial load curve was probably more important than the in-phase requirement, and that there would be a difference between test results of bush

15 We note in this regard that DFARS § 225.872-4(a) requires that qualifying country sources be responsive to the terms and conditions of solicitations, such as QPL requirements.

16 We note that another bidder's (FOC's) bushings were tested to these stringent requirements on the governmentowned test machine and failed to qualify in time to receive award under this IFB. See Florida Ordnance Corp.,

supra.

17 The agency specifically qualifies this argument by this assumption.

ings that received the radial load pressure over a triangle curve rather than a square curve. VT at 17:25, 17:35. As noted above, we are unpersuaded that a square radial load curve was used to test Varec's bushings. Under the circumstances, although Varec's bushings passed tests that occurred over 175,000 cycles, rather than the required 150,000 cycles, we do not find that this provides a reasonable basis to conclude that Varec's bushings would similarly pass proper testing on a test machine that used the much more stringent testing procedure of applying the torsional and radial load cycles on a 4 to 1 ratio, inphase, and applying the radial load as a square curve. In this regard, the agency's witnesses testified that they had no way of translating the Varec test results into results that could be expected from the KRC test machine without more evidence comparing the two test machines. 18 VT at 13:52, 17:27.

The Army argues that even if it waived certain QPL requirements for Varec, Goodyear was not prejudiced thereby. We will not disturb a procuring agency's award absent a prejudicial violation of procurement laws or regulations. MTS Sys. Corp., B-238137, Apr. 27, 1990, 90-1 CPD ¶| 434. As noted above, the use of a QPL is inherently restrictive of competition, but is properly permitted where a QPL is necessary to satisfy an agency's minimum needs. Here, the Army consistently represents that the acquisition of qualified tank components, as included on the QPL, is a firm agency requirement. Nevertheless, as we found above, the Army has waived certain QPL requirements for only Varec. Other firms may have sought to compete if the same requirements were to be waived for all bidders;19 for that matter, FOC may well have been able to qualify its product in time to receive award if the same bushing qualification requirements were waived for it. Furthermore, Goodyear contends that it would have offered a much lower bid price had it known that the Army was going to waive certain bushing qualification test requirements "to account for the fact that it would be competing with companies that had not been required to meet or maintain the QPL requirement, which involves substantial expense." Given the fact that FOC took more than 6 months to find a rubber compound that would satisfy the bushing qualification tests performed at KRC, see Florida Ordnance Corp., supra, we have no basis to question Goodyear's assertion. Under these circumstances, we find that Goodyear was prejudiced by the Army's waiver of certain qualification requirements without amending the solicitation or QPL.20 See MTS Sys. Corp., supra.

In sum, we find that the agency had no reasonable basis for qualifying Varec's product and including it on the QPL. Since Varec's product was not properly on the QPL, the agency's acceptance of Varec's bid under the IFB was improper. See 40 Comp. Gen., supra (award may not be made to a bidder listed on a QPL

18 There also was testimony that the bushing test results would vary from machine to machine, even if the tested bushings are of the same rubber compound. VT at 16:59-17:00.

19 In this regard, the Army has informed us that it currently has applications for qualification of tank track components from other MOU signatory countries and that it intends to qualify these countries' products in the same fashion that it qualified Varec's.

20 In any case, as noted above, Goodyear has withdrawn several other timely, pre-bid-opening protests of the agency's inclusion of Varec's product on the QPL based upon the parties' agreement to await our decision in this protest.