In this warranty case decided by the district court pursuant to admiralty and diversity jurisdiction and applying Kentucky law, plaintiff towboat-owner appeals the grant of summary judgment for defendant towboat-builder and the judgment, following a bench trial, in favor of a second defendant, the manufacturer of the engines used in plaintiff's towboats. We affirm the district court's action with respect to both defendants.

In 1981, plaintiff contracted with defendant Marine Builders to construct two towboats for use on the Ohio and Green rivers. Their contract stated that plaintiff would be in charge of selecting the gears and engines to be installed in the boats. The agreed price of the boats did not include the cost of the engines and gears; Marine Builders was to bill plaintiff for these items at Marine Builders' cost.

Although plaintiff was contractually responsible for choosing the gears and engines, there is evidence in the record indicating that Marine Builders participated in the selection. Plaintiff chose Caterpillar 3512 engines and was initially inclined to select Twin Disc 540 gears, but Marine Builders steered plaintiff toward Twin Disc 530 gears as an adequate and more economical selection.

Since the Caterpillar 3512 engine had never been used in conjunction with the Twin Disc 530 gears, plaintiff arranged to have defendant Caterpillar perform a certain test: a "torsional vibration analysis" (TVA). Such a test predicts the amount of stress, in pounds per square inch, generated by the engine vibrations and exerted upon other parts (here, the Twin Disc gearbox). Caterpillar's TVA referred to (+/-)3000 psi as the "maximum allowable" stress to which the gearbox could be subjected. The TVA also found that, with the Caterpillar engine, the gearbox might be subjected to as much as 3030 and 3168 psi in some situations. The TVA concluded, however, that "this unit is free of serious torsional vibration." (App. at 372).

The parties dispute both the purpose of the test and the meaning of its reported results. Plaintiff essentially contends that the test was intended to predict whether any problems would be caused by using the Caterpillar engines with the Twin Disc gears. Caterpillar claims, and the district court agreed, that the test was intended only to determine if the combination would result in the failure of engaged parts (in the power train) caused by excessive stress.

Based on the TVA's conclusion that the combination would be "free of serious torsional vibration," plaintiff went ahead and had Marine Builders assemble the two towboats with Caterpillar 3512 engines and Twin Disc 530 gears. Shortly after delivery, both towboats suffered repeated failures in the part of the gearbox known as the clutch assembly, particularly in the reverse clutch packs. The repaired gearboxes would operate for only a month or two before the clutch assembly would again break. Plaintiff then had Twin Disc, the gearbox manufacturer, perform its own TVA. That company's test showed stresses in the gearbox well beyond 3000 psi.

Plaintiff's expert testified that the repeated gear failures were caused by excessive torsional vibrations. Defendant Caterpillar proffered, and the trial judge accepted, several other alternative causes of the gear problems. These findings, which plaintiff now contests as clearly erroneous, will be addressed below.

The gear problems were permanently cured when a special rubber piece was installed between the engine and the gearbox. The purpose of this rubber piece, known as a torsional coupling, was to reduce the vibrations experienced by the gearbox.

Plaintiff sued Marine Builders, Caterpillar, and various suppliers on breach of contract, breach of warranty, strict liability, and negligence theories. All such claims were dismissed except for the warranty claim against Caterpillar. Plaintiff now appeals only the dismissal of its warranty claim against Marine Builders, and the judgment for Caterpillar following a bench trial. We turn first to the grant of summary judgment in favor of Marine Builders.

Plaintiff's warranty claim against Marine Builders alleged that this defendant's express warranty covering "all assembled parts" extended to the "match" of the Caterpillar 3512 engines and the Twin Disc 530 gears. Marine Builders' warranty stated:

Marine Builders, Inc. guarantees all assembled products and parts, (except component products or parts on which written warranties issued by the respective manufacturers thereof are furnished to the original purchaser, as to which Marine Builders, Inc. assumes no liability) against defective materials, or workmanship.... No other warranties expressed or implied are intended or given by Marine Builders, Inc.

(App. at 85). Due to the explicit exclusion of parts warranted by their respective manufacturers, plaintiff does not contend that the gearbox and engine--separately warranted by Twin Disc and Caterpillar--are covered by Marine Builders' warranty. Rather, plaintiff's theory is that the "match" of the engine and gearbox qualifies as an "assembled part" not excluded by the parenthetical text governing other manufacturers' parts.

The district judge rejected plaintiff's "match" argument on the ground that plaintiff was contractually charged with selecting the engines and gears--an undisputed fact more relevant to a claim of implied warranty than to the express warranty claim advanced by plaintiff.

We review de novo a district court's grant of summary judgment. EEOC v. University of Detroit, 904 F.2d 331, 334 (6th Cir.1990). We examine the grant of summary judgment to determine "whether the evidence presents a sufficient disagreement to require submission to a [finder of fact] or whether it is so one-sided that one party must prevail as a matter of law." Booker v. Brown & Williamson Tobacco Co., 879 F.2d 1304, 1310 (6th Cir.1989) (citation omitted).

We find that the trial court properly granted summary judgment to Marine Builders. The analytical basis of our conclusion differs, however, from that of the trial court.

Plaintiff claims that Marine Builders breached an express warranty, not an implied warranty.1 Despite this insistence on an express warranty theory, the district judge and the parties in their appeal briefs have treated plaintiff's claim as if it were one of implied warranty. This discrepancy accounts for the importance attributed to the plaintiff's contractual obligation to select the engines and gears. Plaintiff concedes that it originally agreed to control the selection of engines and gears, but it claims that this contractual provision was subsequently modified when Marine Builders suggested the Twin Disc 530 over the other gear model (Twin Disc 540) which plaintiff had been considering. To fortify its claim that plaintiff wielded control, Marine Builders points not only to the contractual provision but also plaintiff's subsequent arrangement with Caterpillar to investigate the compatibility of the engines and gears.

With these facts, both parties seek refuge in the principle that no implied warranty arises when a buyer "exercises his own judgment in the selection of the goods and does not rely on the judgment of the seller regarding its suitability for his purposes." Holterman v. Louisville Bridge & Iron Co., 254 S.W.2d 493, 494 (Ky.App.1953). Plaintiff says it relied on Marine Builders' judgment; Marine Builders counters that plaintiff exercised its own judgment.

The district judge, while acknowledging plaintiff's contention that it relied on Marine Builders' advice, held that the contract irrevocably placed the burden of choosing the parts on plaintiff. He granted summary judgment to Marine Builders, holding that "[w]hile Marine Builders' warranty applied to 'all assembled products,' it did not apply to the match of the engines and gears, which were to be selected by plaintiff." (App. at 33).

We agree that the warranty covering "all assembled parts" does not cover the match of engines and gears. Our decision, however, is not grounded in the identity of the party controlling the selection of the engines and gears, but in a common sense interpretation of the warranty language. "[W]here an adjudication of parties' rights and obligations under a contract rests solely upon a reading of the contract itself, and involves no findings of fact or questions of credibility, we review the district court's conclusions of law free of the clearly erroneous standard of Fed.R.Civ.P. 52(a)." In re Stevenson Assoc., Inc., 777 F.2d 415, 418 (8th Cir.1985); Dugan & Meyers Constr. Co. v. Worthington Pump Co., 746 F.2d 1166, 1168 (6th Cir.1984), cert. denied, 471 U.S. 1135 (1985).

The import of Marine Builders' warranty is clear. The boat assembler assumed no liability for defects in products supplied and warranted by their original manufacturers, but only for defective workmanship in installing those parts. Plaintiff does not attribute the gear failures to faulty installation. It is only by a tortured interpretation of the warranty language--and a misguided foray into implied warranty principles--that plaintiff calls the "match" of the engines and gears an "assembled part" covered by Marine Builders' express warranty.2

We thus affirm the summary judgment in favor of Marine Builders.

We now turn to the more technical aspects of this case: the dispute over what Caterpillar's TVA was intended to predict, what it did in fact predict, and whether the conditions it predicted, or should have predicted, were among the actual causes of the repeated gear failures. The trial judge's findings of fact with respect to these issues we review under the "clearly erroneous" standard. Fed.R.Civ.P. 52(a). We are not bound by this standard, however, in reviewing the judge's mixed findings of fact and law. K & M Joint Venture v. Smith Int'l, 669 F.2d 1106, 1111-12 (6th Cir.1982); Cordovan Assoc. v. Dayton Rubber Co., 290 F.2d 858, 860 (6th Cir.1961). We may, in general, review such findings de novo.

The parties stipulated at trial, and agree on appeal, that the TVA was accompanied by an implied warranty of fitness. It is this warranty which forms the basis of plaintiff's claim against Caterpillar. Plaintiff argues that the parties intended the TVA to determine the overall compatibility of the engines and gears. Caterpillar counters that the TVA's purpose was indeed to predict the components' "compatibility," but Caterpillar's use of that term is more limited; it says the TVA was intended to predict only whether there would be failures in the power train caused by excessive torsional vibrations.

The trial judge found that plaintiff's president, "an astute, knowledgeable and experienced riverman," bargained with Caterpillar for an analysis, "the narrow scope of which was stated on the face of the report." (App. at 40). The judge did not at that point indicate the text of the report to which he was referring. Included on "the face of the report" is the all-important statement, "[T]his unit is free of serious torsional vibration." The judge's views on the TVA's narrow scope are later clarified, however, in his finding that "the test did not attempt to predict failures in non-engaged parts." (App. at 42). This finding is crucial in light of the judge's additional finding that the gear failures occurred in the non-engaged parts of the gearbox, i.e., not the drive shaft.

Plaintiff refers this court to the testimony of the Caterpillar engineer who performed the TVA. That engineer testified that the sole purpose of the TVA was to predict the compatibility of the engines and gears, and that by compatibility he meant whether there would be excessive levels of stress in the system. (App. at 258-59). In further testimony by that same Caterpillar engineer, however, we see just how elusive is the meaning of the term "compatibility." The Caterpillar engineer later testified that the TVA was intended only to analyze the "torsional compatibility" of the engines and gearboxes, and that such an analysis would not attempt to predict whether any non-engaged parts would fail. (App. at 253-54). In contrast, plaintiff's engineering expert indicated that the purpose of a TVA is to test whether "the system" as a whole will fail, not just the shafts. (App. at 113). Plaintiff also emphasizes the absence of any text in the TVA explicitly limiting its findings to engaged parts.

Although we are generally free to interpret independently the terms of a contract or other legal document, the meaning of this TVA (which concluded, "[T]his unit is free of serious torsional vibration") necessarily turns on the purely factual determination of the parties' intent in ordering and performing the TVA. This factual finding was made by the trial judge; he found that the parties bargained for and received a report, "narrow in scope," which "did not attempt to predict failures in non-engaged parts." Given the conflicting evidence of the parties' intent, we cannot say that the judge erred in his findings, nor can we say that we would reach a different conclusion reviewing the TVA de novo.

The same is true of the trial judge's finding that the TVA's figure of (+/-)3000 psi (for the "maximum allowable" level of stress) was only "an arbitrary bench mark as opposed to a standard which should not be exceeded." (App. at 40). Arguing that the TVA was incorrect even in its narrow scope, plaintiff claims that the stresses on the gearbox did in fact exceed 3000 psi and that such high vibrations caused the gear failures. While a typical engineer might be loathe to predict smooth sailing in a system subjected to stresses beyond its components' established standards, the judge based his finding on evidence that the (+/-)3000 psi figure was considered so conservative as to be safely exceeded. Without such a factual finding we could not, on a de novo reading of the TVA, divine the significance of the (+/-)3000 psi figure--and again, we cannot say that the judge erred in his factual findings with respect to this figure.

Plaintiff also contests the trial judge's finding that the gear failures were not caused by torsional vibrations that the TVA should have, but did not, predict. Plaintiff's expert testified that the TVA, having predicted stresses in the drive shaft, should also have predicted stresses in the smaller, weaker, non-engaged parts. These are the parts that failed. Therefore, plaintiff argues, even if the TVA did not explicitly address the prospect of failure in the non-engaged parts, it should have done so if the TVA essentially warranted that there would be no gear failures caused by excessive stresses. Here then enters the hotly-contested issue of causation.

The trial judge found that most of the breakage occurred in the reverse clutch packs. He found this significant "because, according to [plaintiff's port engineer], most of the time the boat is run in forward gear while the reverse packs are disengaged. Substantial damage[ ]," the judge further found, was "sustained by other parts while they were disengaged and in stressless positions." (App. at 42). Plaintiff presents no evidence, and we can find none, which demonstrates that these findings are clearly erroneous.

Even if we were to agree with plaintiff that the judge clearly erred in finding that these non-engaged parts failed for reasons other than vibrational stresses,3 it would not alter the result we reach in this case. If the judge did not err in finding that the TVA was not intended to, and did not, predict the behavior of the non-engaged parts, his findings as to the origin of those parts' failures are immaterial.

AFFIRMED.