Herman Miller, Inc., appeals from a judgment awarding Celina Goode and Thomas Goode damages for injuries suffered by Celina when a coffee urn resting on a shelf unit manufactured and sold by Miller collapsed, spilling hot coffee onto Celina Goode's lap. Miller claims, among other things, that there was no evidence at the bench trial that the design defect found by the trial court caused the coffee spill. The Goodes appeal the amount of damages awarded for loss of consortium. We are persuaded that plaintiffs failed to prove that any defect in the shelf unit was more likely than not a proximate cause of the accident. We reverse.

On December 11, 1982, Celina Goode was severely burned when a large coffee urn fell from a wall-mounted shelf manufactured and sold by Miller. At the time of the accident, Goode and three co-employees were eating lunch in the office of the nursing supervisor in the neo-natal intensive care unit at Louisiana State University Medical Center. Goode sat on a loveseat next to the shelf unit on which a coffee maker with a fifty-five cup capacity sat. The cord to the coffee maker plugged onto an electric socket located next to the loveseat.

Elsie Lou Curlin, a ward clerk at the medical center, placed the coffee urn on the shelf at approximately 7:30 a.m. No one in the office saw any bumping of the shelf or movement near it. While the three co-employees testified that their immediate attention was focused on Goode after the incident, Karen Curry and Barbara Jackson noticed that the mounted shelf was ajar.1 According to Jackson, the "table" on which the coffee urn sat collapsed, spilling the coffee onto Goode's lap.

The shelf unit is part of a large number of shelves, frames, cabinets, and lockers described as a coherent structive system, named Co-Struc. We are concerned with two parts of the shelf unit. The largest piece, a CST Frame, is a one-piece injection-molded plastic shelf frame consisting of three sides, with the center piece or backside attaching to a wall-rail. The frame is sixteen and three-fourths inches high, twenty-two and seven-sixteenths inches wide, and thirteen and one-eighth inches deep. The entire CST Frame is removable so that it can be sterilized, and is interchangeable with any other Miller component that mounts on the wall-rail.

The frame has a plastic lid, known as a CS Lid, which is also removable so that it can be washed and sterilized. The lid is twenty-two and seven-sixteenths inches wide and twelve and three-fourths inches deep. The top surface of the lid has a ridge running along each of its two short sides, which is designed to fit into the recessed bottom of similar lids in order to allow them to be safely stacked when they are transported for washing. The bottom surface of the lid has a lip around its outer edges to secure it on top of the frame. In addition, the frame has a ridge along each of its two short sides that fits into the recessed bottom of the lid in order to hold the lid on the frame. Elimination of the ridges on top of the lid would allow the surfaces of the lid to be used interchangeably.

The evidence showed that if the lid is placed on the frame correctly, the outer lip provides stability to its positioning and reduces the movement or instability of the frame itself. However, when the lid is inverted, there exists an uneven degree of stability to the lid. The ridges provided on the top of the lid as a stacking feature provide some stability in side-to-side movement, and, as a result, the inverted lid would appear to be fixed in its position when one pushes it in one direction, but is unstable if moved in the other. No warning appears on the shelf unit, either to inform the user of the false impression of stability upon inversion or to label the top and bottom of the lid.

Although the shelf unit similarly gives a false impression of stability if the lid is placed right-side up on the frame, but off-center, the Goodes did not argue, nor did the court find, that this constituted a design defect. The sole theory of recovery is that inadvertent inversion of the lid leads to an unstable and therefore dangerous product.

The shelf unit involved in this accident was one of hundreds used at the medical center. Charles Stevens, a project manager for Miller, testified that the lid was designed to sustain a sixty-six pound load with less than one-fourth inch deflection. According to the technical fact sheet for Miller products, the lid would sustain a one hundred thirteen pound load with no more than one-fourth inch deflection. These units are used to support patient monitoring equipment in the Intensive Care Unit, and to store other monitors which exceed the weight of the coffee urn which fell and injured Goode. Because the units are interchangeable and moved frequently throughout the medical center, the actual shelf unit involved in the accident was not presented at trial, although an identical model was produced for the court.

The Goodes sued Miller in a Louisiana district court, but the case was removed to federal district court on the basis of diversity jurisdiction. The court, sitting without a jury, issued a memorandum opinion, which found that the shelf was defective under Louisiana law, and awarded the Goodes $64,208.64. Miller appeals, claiming that there was no finding that the lid was inverted at the time of the accident and thus, the defect found by the court did not cause the accident. It also claims that there was no evidence that the shelf was in normal use, even if inverted, at the time of the accident and that the shelf unit was not defective. The Goodes appeal, claiming that the $1,000.00 awarded to Thomas Goode for loss of consortium was insufficient.

Recovery for products liability under Louisiana law requires proof that: (1) the product was defective; (2) the product was in normal use at the time the injury occurred; (3) the defect proximately caused the injury; and (4) plaintiff's injury might reasonably have been anticipated by the manufacturer. Hunt v. City Stores, Inc., 387 So. 2d 585, 589 (La.1980). Each element must be proved by a preponderance of the evidence. See Alexander v. St. Paul Fire & Marine Ins. Co., 312 So. 2d 139, 144 (La.App.1975). Miller argues that the Goodes failed to meet their burden of proving that the design of the shelf unit was defective, that the product was in normal use at the time of the accident, and that the defect caused the injury.

A product is defective under Louisiana law if it is "unreasonably dangerous" for normal use, meaning that the product is dangerous to an extent beyond that which would be contemplated by an ordinary user. DeBattista v. Argonaut-Southwest Insurance Company, 403 So. 2d 26, 30 (La.1981), cert. denied, 459 U.S. 836, 103 S. Ct. 82, 74 L. Ed. 2d 78 (1982); see also 402A, Comment i, Restatement (Second) of Torts. If the product has reasonably foreseeable dangers, not obvious to the ultimate consumer, Louisiana law requires the manufacturer to either adequately warn or guard against them. Straley v. Calongne Drayage & Storage, Inc., 346 So. 2d 171, 176 (La.1977). Ordinarily, whether a product's design renders a product defective is a fact question under Louisiana law. Cobb v. Insured Lloyds, 387 So. 2d 13, 19 (La.App.1980).

In determining whether a product is defective, the Louisiana Supreme Court has at times applied a "risk/utility test," weighing the likelihood and gravity of harm against the benefits and utility of the manufactured product. Hunt, 387 So. 2d at 589. This test is utilized when "the product as marketed ha[s] something wrong with it that allow[s] it to function in a way that was not intended." Perkins v. F.I.E. Corp., 762 F.2d 1250, 1272 (5th Cir.1985). This analysis requires consideration of:

the gravity of the danger posed by the challenged design, the likelihood that such danger would occur, the mechanical feasibility of a safer alternative design, the financial cost of an improved design, and the adverse consequences to the product and consumer that would result from an alternative design.

Id. at 1273 n. 62 (quoting Barker v. Lull Engineering Co., 20 Cal. 3d 413, 429, 573 P.2d 443, 453, 143 Cal. Rptr. 225, 234 (1978)).

The trial court found the product to be defective under both the consumer expectation and the risk/utility tests, because consumers did not comprehend the dangers presented by an inverted lid and because simple steps could be taken to alleviate the dangers presented by inadvertent inversion of the lid. The court found that a warning or label could be placed on the lid, or that the ridge on the top shelf could be removed so that the surfaces of the lid could be used interchangeably. The court did not view the utility of the ridge, used to permit easy stacking while cleaning, as worth the risks of harm it creates.

Miller argues that the evidence was insufficient to support a finding of design defect, because the difference between the two surfaces of the lid was obvious to the ultimate users of the product and because a warning or label on the lid would defeat its intended use in sterilized environments. However, the trial court found, after hearing the testimony and observing two ward clerks handle the CS Lid, that the ward clerks were hesitant when asked to indicate the top and bottom of the lid. The court stated:

[T]he daily user[s] of the CST frame and lid were unsure about the proper use of the product. Doris Turk and Elsie Lou Curlin, ward clerks who removed and cleaned the CST lid daily, were hesitant and appeared uncertain when questioned about the CST frame and lid. They were slow to designate the "top" side of the lid without first placing the lid on the frame themselves.

We are not prepared to hold that the able trial court erred in that conclusion. The court had the unique opportunity to observe the witnesses and to view the ward clerks handle and manipulate the lid. Doris Turk, one of the ward clerks responsible for cleaning and replacing the lids daily, incorrectly identified the top and bottom of the lid and maintained that identification upon examination until allowed to actually place the lid on the frame. While it is true that each witness eventually identified the top and bottom of the lid correctly, the hesitancy noted by the court supports its finding that the difference between the top and bottom of the lid was not obvious, especially when these shelves are replaced repeatedly throughout the day by busy ward clerks.

Furthermore, Turk also testified that she had experienced trouble with the lid before the incident. On several occasions, Turk noticed that the lid slipped off after she had placed it on the frame. Since the experts agreed that slippage could not occur if the lid were placed on the frame correctly, her testimony indicates that either the top and bottom of the lid were not as obvious as Miller urges or that she placed the lid right-side up but failed to engage the lid properly on the frame.

The evidence also showed that the distinction was not obvious even after placing the lid on the frame, since movement of an inverted lid in two directions (back and to one side) gave a false appearance of stability. Only if pushed in every direction would the instability of the inverted shelf be apparent. Stating that "the slippage of an inverted shelf ... was unexpected, presenting a danger not considered, not contemplated by the ordinary user," the court found that the ward clerks were not aware of the risks of inattentive use of the shelf unit.

The court also found that the risk of harm resulting from inadvertent inversion of the lid outweighed the utility of the lid's stacking ridges.2 Considering the likelihood of inverted use of the lid, the inexpensive design features feasibly preventing such use, and the degree of harm posed by the current design, the trial court found the unit to be defective. Because the ultimate users of the lid could not identify without difficulty the top and bottom of the lid, the court concluded that a warning was required or that precautions should be taken to make the surfaces of the lid interchangeable. We cannot say that this decision was clearly erroneous.

Normal use is a term of art marking the scope of the manufacturer's duty. LeBoeuf v. Goodyear Tire & Rubber Co., 623 F.2d 985, 989 (5th Cir.1980). Under Louisiana law, normal use includes reasonably foreseeable misuse, and a manufacturer is obligated to take into account such misuse in designing its product. Johnson v. Tenneco, Inc., 752 F.2d 160, 162 (5th Cir.1985). A product can therefore be unreasonably dangerous under Louisiana law for failure to warn or guard against unintended uses, if such uses are reasonably foreseeable. Bridges v. Chemrex Specialty Coatings, 704 F.2d 175, 179 (5th Cir.1983).

Miller challenges the court's finding that the shelf was in normal use at the time of the accident on two grounds: (1) that the shelf unit was not intended to support a 55-cup urn of coffee, and (2) that it was an obvious misuse of the unit to place the lid on the shelf upside-down. The court, however, noted that both Miller's project manager and Miller's technical fact sheet indicated that the frame and lid could support weights in excess of sixty-six pounds. Moreover, the hospital put hundreds of these units to a variety of uses, many of which involve supporting monitors weighing more than the coffee urn. While the court did not expressly state that placing a coffee urn on the unit was a foreseeable use of the product, it did find that the unit was advertised as capable of holding loads of similar weight.

A trial court's findings are to be read liberally. See 9C. Wright & A. Miller, Federal Practice and Procedure Sec. 2580, at 719 (1971). "If a trial judge fails to make a specific finding on a particular fact, the reviewing court may assume that the court impliedly made a finding consistent with his general holding so long as the implied finding is supported by the evidence." In re Texas Mortg. Services Corp., 761 F.2d 1068, 1075 n. 12 (5th Cir.1985). We are persuaded that the court impliedly found that placement of the coffee urn was a normal use of the product and cannot say that it was clearly erroneous in doing so.

The court also found inversion of the lid to be a foreseeable misuse of the product. Noting the ease of inversion, the uses of the inverted lid,3 and the inability of the witnesses to identify the bottom of the lid without difficulty, the court found inversion of the lid reasonably foreseeable and therefore within the rubric "normal use." Again, we cannot say that this finding was clearly erroneous.

Consequently, Miller cannot, on the basis of abnormal use, escape its duty to either provide an adequate warning of the danger of lid inversion or to guard against the danger in some other way. We must now try to decide whether Miller's failure to take such precautions caused Goode's injury.

Miller complains that the district court made no finding, and there was no evidence, that the shelf lid was inverted at the time of the accident. Miller points to several other hypothetical causes assertedly equally plausible from the record. Hence, Miller argues that the presence or absence of a warning or label on the shelf lid is immaterial, as this "defect" was not found to be the cause of the accident.

In its memorandum opinion, the court stated, "In this case, the issue is whether the inverted use of the lid of the component structure frame constituted a foreseeable hazard that made the product unreasonably dangerous in the absence of a special warning of the potential peril." This statement necessarily implies that the court found the lid was inverted at the time of the accident because foreseeability would simply not be the issue of the case otherwise. We turn then to whether the finding enjoys sufficient record support.

Under Louisiana law,

[m]ere proof that something is possible is of little probative value as to an ultimate issue of fact, unless it be established with reasonable certainty that all other alternatives are impossible. Where several possible causes in fact are established with equal probability but no one is shown by the preponderance which the law demands, the plaintiff has not discharged the burden of proof.

Nailor v. International Harvester Co., 430 So. 2d 784, 790 (La.App.1983) (emphasis added). Miller posits numerous other possible causes of the accident. It notes that plaintiff's tendered expert, James Jackson of Haag Engineering,4 testified that the coffee urn could have fallen off the shelf because it was placed too far forward on the shelf, or that the lid could have been placed off-center, but right-side up, and therefore fallen. Miller also suggests that the coffee urn, which was not produced at trial, could have itself broken or collapsed. Finally, since the Goodes did not produce the actual shelf unit involved in the accident, Miller suggests that the shelf might have been broken in the years of use before the accident.

The court explained that because the evidence showed that the frame and lid were unstable, it was persuaded that the defective shelf caused the plaintiff's injuries. Like the court, we are persuaded that several of the posited causes are not as likely as inversion of the shelf. Samuel Shomer, the assistant dean for academic administration at the medical center, testified that the shelf unit involved in the incident was being used in his office at the time of the trial. This would suggest that the lid was not itself broken, as suggested by Miller. Furthermore, because the coffee urn sat safely on the shelf for more than four hours, we are persuaded that the court had grounds to exclude improper placement of the coffee urn as a possible cause. We are further informed in this inquiry by the fact that the lid itself was ajar after the accident. The evidence showed that the lid would not dislodge if placed properly on the frame.

But we find no record basis for concluding that it is more likely than not that the lid was inverted, rather than placed improperly, but right-side up, on the frame. The evidence showed that with both inversion of the lid and improper placement of the lid, movement of the lid in two directions gives a false impression of security. The evidence also showed that with both inversion and improper placement of the lid, little effort would be necessary to dislodge the lid no matter what was placed upon it. Moreover, plaintiff's tendered expert testified that the accident could have occurred either way. Upon review of the record, including the product itself, we are persuaded that the court erred in concluding, as it must have done, that the instability of the frame and lid could be caused only by inversion of the lid.

Thus, because two equally plausible proximate causes exist for this accident, we are persuaded that the Goodes failed to carry their burden of proving that the design defects found by the court caused the injury to Celina Goode.

Goode cross-appeals seeking an increased damages award. Because we have ruled that the Goodes failed to carry their burden of proof on the issue of causation, we need not pursue this contention.

The judgment of the trial court is REVERSED.