In this products liability diversity case, plaintiffs appeal from a final judgment entered on an adverse jury verdict, assigning reversible error in several of the district court's evidentiary rulings, principally the exclusion of the proffered testimony of plaintiffs' expert witness, Keith Austin. Finding no abuse of discretion by the district court, we affirm.

On November 23, 1973, a fire engulfed the apartment of plaintiffs Roy and Elsie Horton, destroying the contents of the apartment, severely burning them, and killing Jerry Horton, Roy's brother, who had been visiting Roy and Elsie. Plaintiffs suspected the origin of the fire to be a television set sold by defendant W. T. Grant and manufactured by the Matsushita defendants. In order for plaintiffs to recover, it was necessary for them to prove that the television was defective when it was manufactured and sold, and that the defect was a proximate cause of the fire. Plaintiffs offered the testimony of one Keith Austin, who was apparently conceded to be a qualified expert in both television design and manufacture and the investigation of television-related fires. In the latter capacity, Austin was allowed to testify that the probable origin of the fire was the television set, and this testimony was corroborated by other evidence and expert testimony. However, plaintiffs were also required to show that the television was defective.

Unfortunately, plaintiffs' television set was totally destroyed in the fire, and hence was unavailable for Austin's inspection. Defendants had discontinued manufacturing the model purchased by plaintiffs, so an equivalent new set was unavailable. Instead, plaintiffs obtained a used set of unknown prior history, but which was conceded to be identical in original design and manufacture to plaintiffs' set, and submitted it to Austin for study. Based on his observations of this set, Austin was prepared to testify substantially as follows:

At two points in the set's chassis, wires were routed in close proximity to capacitors, which are electronic components for storing electricity. While capacitors, in operation, do not become hot, they may absorb heat from the air inside the set's cabinet given off by other components. The surface of the capacitors may thus become sufficiently warm to melt the insulation on any wires which touch the components. The insulation on the two wires in the set under inspection was in fact charred. If the set continued to operate in this condition, the insulation could burn through, allowing the bare conductor in the wires to touch other components and to result in a short circuit. A short circuit could produce enough heat to ignite various materials in the set's cabinet, starting a fire of sufficient intensity to have done the damage observed in the plaintiffs' apartment.

Austin could not determine from the schematic circuit diagrams supplied by the manufacturer whether the wire routing which he considered defective was mandated by the set's design. However, he testified that because of the neatness with which the wires were soldered and the manner in which they were bundled into a harness, it was unlikely that the defective condition had been introduced by a repairman. Rather, it was Austin's testimony that based on his familiarity with the manner in which television sets are usually assembled, the questionable wire routing was present when the set left the factory.

The district judge was of the view that Austin's proffered testimony concerned a test or experiment, and that it was admissible only if the test was performed under conditions substantially similar to those which would have been present in the plaintiffs' television. Applying this reasoning, he refused to admit the evidence, even assuming that the alleged defect the wires' being routed near the capacitors had originated at the factory. Austin could not exclude the possibility that the set he had examined had been abused before he inspected it, and that this abuse had produced the charring of the capacitors which Austin found indicative of projected failure. And Austin had not performed any tests to show that under normal operating conditions the wires would have been exposed to enough heat by virtue of their proximity to the capacitors to char the insulation. Alternatively, the judge indicated that he was skeptical of Austin's ability to testify that the set had not been altered since it had left the factory, but it is not clear to us whether he would have excluded the evidence on this ground alone.

Austin was conceded to be a qualified expert, and we believe that the admissibility of his testimony should have been determined under the standards applicable to expert testimony generally, rather than under the rules used in passing on the admissibility of the results of experiments and tests conducted by non-experts. It is generally held that relevant testimony from a qualified expert may be received if and only if he is in possession of such facts as would enable him to express a reasonably accurate conclusion as distinguished from mere conjecture. Although all of the facts and observations relied upon by an expert need not be independently admissible, there still must be an adequate basis for his testimony, and it is within the discretion of the district court to decide whether such a basis has been shown. See Gilbert v. Gulf Oil Corp., 175 F.2d 705, 709 (4 Cir. 1949).

Applying this standard to the instant case, we do not believe the district judge abused his discretion in refusing to admit Austin's testimony. We are not unmindful of the difficulty of proving a manufacturing or design defect under the circumstances of this case. However, the possibility which troubled the district judge that the set which Austin tested had at some point been subjected to unreasonable abuse is a serious concern. It would seem that Austin or another expert could have conducted further tests which would have established the dangerous potential of the wire placement, rather than relying exclusively on the charring of the insulation in a test set of unknown prior history.

We have considered the possibility that Austin's proffered testimony could be interpreted to suggest that the routing of the wires in the set was inherently dangerous, without regard to whether any charring of the insulation had already occurred. If such an interpretation were permissible, and if Austin's testimony that the wiring had not been altered is credited, the absence of a prior history of the test set should not have precluded Austin from expressing his opinion that the set was defective, without discussing the charring. However, as we read Austin's proffer, he relied heavily upon the fact that the wires were damaged, and we believe the district judge was within his discretion in concluding that this fact was crucial to Austin's opinion.

This case was tried from March 25 to April 1, 1975, prior to the effective date of the Federal Rules of Evidence. We therefore express no opinion as to whether the district court's ruling would have been correct under Rule 703, which governs the admissibility of expert testimony.

Plaintiffs object to several other evidentiary rulings by the district court. They claim that one Thornton, a defense expert, was permitted to state an opinion based on observations of a used set, while, as we have discussed, plaintiffs' expert Austin was denied such an opportunity. Our reading of the record is that Thornton was permitted only to state that the circuit design of the model in question, as revealed by its schematic diagram, was a good one. He was not allowed to base an opinion on any observation of an actual set. The quality of the circuit design was not an issue in the case; plaintiffs' theory was one of improper component and wire placement. Thus we find no merit in this objection.

Plaintiffs complain that they were not permitted to introduce evidence showing that defendants were given notice by Underwriters' Laboratories of certain defects in the design of the set. Defendants had previously shown that the set ultimately did win UL approval, and it was defendants' contention that the defects noted by UL had been rectified prior to the beginning of production. The district court indicated it would permit plaintiffs to introduce the correspondence from UL if plaintiffs would assure the court that they intended to show that the conditions which UL had complained of were in fact not corrected. Plaintiffs declined to give such an assurance, and the court refused to admit the evidence. Clearly, if the defects involved were eliminated before the plaintiffs' set was produced, the existence of the defects at one point in the pre-production design was irrelevant to any issue being tried. Since plaintiffs effectively conceded that such was the case, the district court was correct in excluding the evidence.

Plaintiffs object to the admission of testimony concerning their smoking and drinking on the night of the fire, claiming that this evidence was irrelevant and prejudicial. Plaintiffs Roy and Elsie Horton testified at trial as to their observations of the fire as they fled from their apartment. Even moderate intoxication might have impaired their powers of perception and retention, and thus was a proper subject of impeachment. Testimony concerning the intoxication of Jerry Horton was properly excluded, since Jerry died in the fire and did not testify at trial. While the consumption of tobacco ordinarily does not impair one's ability to perceive and retain information, evidence of the plaintiffs' smoking was admissible on another ground. One of defendants' theories was that the fire was started by careless smoking. While this theory would normally be thought to raise a question of contributory negligence, an issue which the court declined to submit to the jury, it also tends to negate plaintiffs' theory that the television set was the cause of the fire by setting up an alternative cause. Thus testimony showing plaintiffs' smoking was admissible to disprove liability.

Finally, plaintiffs argue that the district court should have granted a direct verdict in their favor. Even assuming that the uncontradicted evidence established that the fire emanated from the television set, this fact was not enough to establish liability under any theory. At a minimum, plaintiffs were required to show that the set was defective when it left the factory, and that the defect caused the fire. Such proof was lacking because of the exclusion of testimony by plaintiffs' expert Austin.

AFFIRMED.