This appeal requires us to consider whether the district court's rulings, that restricted direct examination of plaintiffs' expert as to the probable cause of a malfunction in a refrigerator that allegedly caused a fire that destroyed plaintiffs' home, constituted harmful error. Because we find that the court's evidentiary rulings were erroneous and harmful within Fed.R.Evid. 103(a), we will reverse and remand for a new trial.

On April 10, 1979, plaintiffs Francis and Barbara Breidor purchased from defendant Sears, Roebuck and Company a refrigerator manufactured by defendant Whirlpool Corporation. From that date until July 24, 1980, the refrigerator was in continuous use, and plaintiffs experienced no problems with the unit. On the evening of July 24, 1980, plaintiffs' daughter used the refrigerator sometime between 10:00 and 10:15 p.m., and it was functioning normally. She then left the house. Shortly thereafter a fire broke out that destroyed much of the family's home and its contents. No member of the Breidor household was home when the fire was reported at around 11:00 p.m.

Alleging that the fire was caused by a defect in the refrigerator, plaintiffs brought suit in Philadelphia County Court of Common Pleas against Sears and Whirlpool. The suit was predicated on strict liability under Restatement (Second) of Torts Sec. 402A (1965). Defendants removed the case to the District Court for the Eastern District of Pennsylvania pursuant to 28 U.S.C. Sec. 1441 (1976).1 The case proceeded to a bifurcated trial. The liability portion of the trial focused on the origin and cause of the fire. Defendants contended and offered expert testimony to show that the fire began at a low point alongside the refrigerator, that the unit was not defective, and that there is not enough oxygen inside a refrigerator to sustain a fire. Plaintiffs contended that the fire originated inside the upper part of the refrigerator, and was caused by a defect in the refrigerator. Plaintiffs' expert witness, William Emory, testified that the fire originated in the upper part of the refrigerator but stated that he could not find a specific defect. Because of this concession, and despite repeated attempts by plaintiffs' counsel, Emory was not allowed to render an opinion as to the cause of the fire.

The jury returned a verdict in favor of Sears and Whirlpool, and this appeal followed. The plaintiffs contend that the district court abused its discretion in refusing to permit Emory to give his opinion as to the cause of the fire, and that the exclusion of this evidence affected plaintiffs' "substantial rights" within the meaning of Fed.R.Evid. 103(a).

Because this appeal turns on the admissibility of Emory's testimony, it is necessary to recount it in some detail. Emory, a professional engineer and fire investigator, was qualified as an expert on the origins and causes of fires. He first testified that the origin of the fire was in the "upper part of the refrigerator."2 Next, he turned to the cause of the fire. Emory testified that, as a result of further investigation, he was able to eliminate several possible causes of the fire, including a smoldering cigarette or other object on the floor next to the refrigerator, an electrical breakdown in the wall outlets or the wires running to the refrigerator, or a malfunction in the refrigerator's compressor. Elimination of these possible causes reinforced Emory's initial conclusion that the fire originated in the upper part of the refrigerator.

Emory then discussed the probable cause of a fire that originates in the upper part of a refrigerator. He explained that he had gone to Sears to get information about the electrical components that were in the upper part of the refrigerator, but that he could not get this information. He also testified that he observed that the thermostat was located in the top of the refrigerator along with some other component. Using this information, he decided to try to cut into the refrigerator casing to see if he could find some problem with the thermostat, or some other source of ignition that could have caused the fire. Emory was then asked:

Q. When did you do that?

A. October 10th. The cut out portion in the back [of the refrigerator] was something I hacked out with a hack saw blade or whatever we had. We were not too successful. I could not find a specific defect. I gave it considerable thought and what I finally came up with is there had been a breakdown in the electrical ...MR. SHORT (defendant's counsel): I object.

The gentlemen said that he could not find a defect, and, therefore, I object to any expression of opinion.

THE COURT: What did you say?

Repeat your answer.

THE WITNESS: I could not find a specific defect due to the location of the component in the inner part and due to the fire cone.

What I did do is make an analysis as is common in engineering and determine the source of fire what [sic] would have had to be ignition by electrical objects.

MR. SHORT: Objection, Your Honor.

THE COURT: You can't tell us what might have been but you can state your opinion.

A. In my opinion, there was an electrical breakdown which creates an electrical arcing which is over two thousand degrees farenheit.

The process of electrical arcing produces a pressure wave due to the temperature and the pressure conditions there would have been sufficient to have pushed open the upper door.

MR. SHORT: Objection.

THE COURT: It will be stricken.

Emory next testified that, because the fire often destroys all direct evidence of its cause, fire experts often cannot find any specific cause of a fire. Emory explained that this lack of direct evidence frequently requires fire experts to rely on circumstantial evidence to establish the probable cause of a fire.3 Emory then explained that in this case he could not find any direct evidence (i.e., a "specific defect"), because of the fire damage to the refrigerator and because the location of the electrical components inside the metal casing of the refrigerator made it impossible to access them.4

Emory also testified that the basis for his proffered expert opinion was: (1) the construction of the refrigerator; (2) the damage he observed; (3) the parts list provided by Whirlpool; (4) his background and experience in investigating other fires; (5) the source of the fuel (cabinet insulation); and (6) his knowledge of electricity and his experience with electrical objects.5

Plaintiffs' counsel also attempted--unsuccessfully--to have Emory testify that he had eliminated all other possible causes of the fire except for an electrical malfunction in the thermostat unit.

Q. Based on that opinion, and also based on your review and investigation, would you tell this jury all of the possible ways a fire could have originated inside that refrigerator or freezer?

MR. SHORT: Objection.

THE COURT: I will sustain the objection.

....

Q. Mr. Emory, is there any other possible explanation for how the fire would have originated inside the refrigerator other than to be originated by electricity?MR. SHORT: Objection, sir.

THE COURT: I will sustain the objection.

Q. Does your investigation point to any possible origin outside the refrigerator that could have started the fire?

MR. SHORT: Objection.

THE COURT: I will sustain the objection.

Despite this testimony establishing that fire experts frequently rely on circumstantial evidence in forming an expert opinion as to the cause of a fire, and that, because of the fire damage and the location of the electrical components inside the refrigerator casing, there was no direct evidence available in this case, and despite the fact that Emory was apparently prepared to testify that the evidence pointed to an electrical malfunction in the thermostat as the only possible cause of the fire, the district court repeatedly refused to permit Emory to state his opinion as to the probable cause of the fire.

Expert opinion evidence is admissible if it "will assist the trier of fact to understand the evidence or to determine a fact in issue." Fed.R.Evid. 702. "[T]he trial judge has broad discretion in the matter of the admission or exclusion of expert evidence, and his action is to be sustained unless manifestly erroneous." Salem v. U.S. Lines Co., 370 U.S. 31, 35, 82 S. Ct. 1119, 1122, 8 L. Ed. 2d 313 (1962); Seese v. Volkswagenwerk A.G., 648 F.2d 833, 844 (3d Cir.); cert. denied, 454 U.S. 867, 102 S. Ct. 330, 70 L. Ed. 2d 168 (1981); see Fed.R.Evid. 103(a) ("Error may not be predicated upon a ruling which admits or excludes evidence unless a substantial right of the party is affected ....")

The grounds for the district court's refusal to admit Emory's expert testimony are not clear from the record. Based upon a careful reading of the trial transcript we have gleaned two possible reasons for the court's decision. First, some of the district court's statements to plaintiffs' counsel appear to reflect the court's concern that, because Emory could not find a specific defect, his opinion as to the probable cause of the fire was mere guess work. Second, evidence in the record suggests that the court may have considered Emory's testimony unnecessary because the plaintiffs had adduced enough evidence of a malfunction in the upper portion of the refrigerator to permit the case to go to the jury under the theory of MacDougall v. Ford Motor Co., 214 Pa.Super. 384, 257 A.2d 676 (1969). We consider these possible bases for the court's disputed evidence rulings in turn.

To the extent that the allegedly speculative nature of Emory's testimony was the basis for the district court's evidentiary rulings, the court abused its discretion. Where a fire investigator identifies the cause of fire in terms of probabilities (as opposed to mere possibilities) by eliminating all but one reasonable potential cause, such testimony is highly probative. Gichner v. Antonio Troiano Tile & Marble Co., 410 F.2d 238, 247 (D.C.Cir.1969); Lanza v. Poretti, 537 F. Supp. 777, 785-786 (E.D.Pa.1982). See also Annot., 88 A.L.R. 2d 230, Secs. 5, 8, & 9 (1963).

As we read the record, Emory was prepared to testify that he had eliminated all possible causes of the fire except for an electrical malfunction in the thermostat. It was Emory's theory that an electrical arcing in the thermostat caused a huge build-up of heat inside the refrigerator, creating a heat wave of sufficient force to push open the refrigerator door and allow the fire to escape into the kitchen. There was some evidence in the record to support this theory.6 Where there is a logical basis for an expert's opinion testimony, the credibility and weight of that testimony is to be determined by the jury, not the trial judge. See, e.g., Annot., 88 A.L.R. 2d 230, Sec. 9 (1963).

Gichner v. Antonio Troiano Tile & Marble Co., 410 F.2d 238, 247 (D.C.Cir.1969), is closely analogous to this case.7 In Gichner, the owner of a building sued his lessee for damages resulting from a fire.8 At trial, plaintiffs' expert testified that he was unable to find any direct evidence of the cause of the fire due to the fire damage, but that it was his expert opinion that the fire was caused by careless smoking. Plaintiffs' expert explained that he reached this conclusion by eliminating all other reasonably possible causes of the fire. Id. at 244. At the close of plaintiffs' case, the district court granted judgment for the defendants on the ground that, "although the fire investigator's conclusions were good enough for fire department purposes, the evidence was not sufficient to 'justify the Court in inferring by a fair preponderance of the evidence it has been established that careless smoking was the cause of the fire.' " Id. at 240-241. The court of appeals reversed,9 holding that "where a fire investigator identifies the cause of a fire in terms of probabilities, as opposed to mere possibilities, by eliminating all potential causes of the fire but one, that testimony is not only relevant, but in some circumstances may be a basis for decision." Id. at 247.10

The proffered expert testimony of Mr. Emory is like that of the expert in Gichner and is admissible for the same reasons. The testimony falls well within the ambit of Fed.R.Evid. 702 because it is helpful to the trier of fact in determining the truth in a difficult factual context--the origin of a fire. Helpfulness is the touchstone of Rule 702, and Emory's testimony is helpful because of his special knowledge and skill concerning the origins and causes of fires. Contrary to appellees' contentions, the testimony is not speculative or lacking in foundation.11 Emory's thorough investigation described above indicates a solid foundation for his testimony; the mere fact that Emory could not identify a specific defect does not mean that he was speculating when he offered his expert opinion as to the cause of the fire. Moreover, testimony by Emory that the probable cause of the fire was an electrical malfunction in the thermostat would be helpful because it would afford to the jury an explanation of how a fire could have started in the upper part of the Breidor's refrigerator; the defendants' expert had testified that the fire started outside the refrigerator.B.

A second possible explanation for the district court's decision to exclude Emory's testimony on the cause of the fire is that the court thought that the testimony was cumulative and hence excludable under Fed.R.Evid. 403. To understand this possible basis for exclusion, it is necessary first to understand the theory of liability upon which plaintiffs proceeded.

The court charged the jury under the "malfunction" theory of MacDougall v. Ford Motor Co., 214 Pa.Super. 384, 257 A.2d 676 (1969). Under MacDougall, section 402A actions are governed by the evidentiary standards of warranty law rather than negligence. Under these standards, the occurrence of a mechanical malfunction evidences a "defective condition" without proof of the specific defect in design or assembly causing the malfunction. To recover under MacDougall, the plaintiffs did not have to prove the existence of a specific defect; they only had to show "the occurrence of a malfunction of machinery in the absence of abnormal use and reasonable secondary causes. Id. at 391, 257 A.2d at 680.12 The plaintiffs produced evidence that the refrigerator received only normal use, and Emory's and the fire investigator's testimony that the fire started inside the refrigerator provided probative evidence of the absence of reasonable secondary causes. Thus, the court may have believed that the plaintiffs already had sufficient evidence in the record to recover under the MacDougall theory, and therefore that Emory's testimony as to the actual cause of the fire was cumulative.13

Upon a review of the entire record, we conclude that the cumulative-evidence rationale does not support the decision to exclude Emory's testimony, even under the MacDougall theory of liability,14 for two reasons. First, his testimony was necessary to rebut the defendants' contention that a fire could not have been sustained in, and escaped from, a closed, air-tight refrigerator. If permitted, Emory would have testified that a "pressure wave" caused by an electrical arcing in the thermostat unit situated in the upper part of the refrigerator "would have been sufficient to have pushed open the upper door."15 Because the defendants subsequently offered testimony from their own expert that tests done by starting various types of fires inside refrigerators demonstrated that the fires would burn themselves out due to lack of oxygen, Emory's testimony was crucial to the plaintiffs' case. Without it, plaintiffs could not establish causation; the jury therefore could not have found in their favor under any theory. Accordingly, Emory's testimony as to the cause of the fire was not cumulative of other testimony already adduced in the record, and the district court's exclusion of it was an abuse of discretion.

Emory's testimony was also not cumulative because, if he had been permitted, he would have testified as to all the possible causes of a fire in the upper part of a Whirlpool refrigerator. Without evidence that it was possible for a fire to start in the upper part of a refrigerator, plaintiffs were placed at a substantial disadvantage in countering defendants' expert testimony that the fire started outside the refrigerator.

We have concluded that the district court abused its discretion when it excluded Emory's expert opinion testimony. This exclusion affected plaintiffs' substantial rights in at least two ways. First, the absence of Emory's testimony seriously impaired plaintiffs' ability to counter the defendants' evidence that the fire started outside the refrigerator. But for the court's restrictive evidentiary rulings, the jury would have received evidence explaining the mechanism of the starting of a fire in the upper portion of the refrigerator and evidence that an electrical malfunction in the thermostat was the only possible cause. Second, Emory's testimony would have assisted the plaintiffs in proving that a fire that started inside the refrigerator could have escaped and caused the Breidor's house to catch fire; without Emory's testimony plaintiffs were at a substantial disadvantage in countering the defendant's experts' testimony that the fire could not have been sustained in and escaped from a closed, air-tight refrigerator.

Accordingly, we will vacate the district court's judgment and remand for a new trial.