General Motors Corporation (GM) appeals from a judgment on a jury verdict awarding damages to the plaintiffs in this products liability case. GM's contention on appeal is that it is not liable to the Chases on the brake defect theory that plaintiffs relied upon at trial because the collision in which they were involved was unavoidable. Further, GM argues that it was prejudiced by the district court allowing into evidence the plaintiffs' videotaped brake tests and in not allowing GM's own videotaped tests to rebut that of the plaintiffs. GM also claims that it was prejudiced at trial by the introduction of evidence of vehicle recall. Because the plaintiffs' videotaped tests fail to comply with the standard for admissibility in this circuit and because the evidence of recall was improperly admitted, we vacate and remand for a new trial.

At the outset, we should say we are of opinion that GM adequately preserved its claims of error. We also think that, with the exclusion of the substantial segments of evidence that this opinion requires, it is more appropriate not to express an opinion on the sufficiency of the evidence, which we do not.

On January 23, 1982, the Chases were involved in a collision with another vehicle on a two-lane highway near their home in West Virginia. Plaintiffs were returning from Williamsport, Maryland about midnight after a tennis party and headed south on U.S. Route 11. There had been snow accumulation and cold temperatures in the area for several days. The road was covered with snow and ice.

A northbound vehicle slipped on the ice and crossed the center line out of control into the path of the plaintiffs' car. The Chase automobile, a 1980 Chevrolet Citation, was proceeding downhill and on a curve to the left. Chase, driving the car, applied the brakes to avoid hitting the oncoming vehicle. The Chase car turned somewhat to its left and the cars collided, with the front right corner of the Chase car, striking the front of the other auto. Mrs. Chase, who had been riding on the passenger side, received a severe blow to the head, resulting in brain damage and physical impairment. Chase and the other driver were not seriously injured.

Plaintiffs brought this products liability action against GM and B & M Chevrolet-Cadillac Corporation (B & M) on January 10, 1984. They alleged negligence in design, strict liability and breach of implied warranty concerning the brake design on the Citation. The complaint charges that the allegedly defective brake design was the direct and proximate cause of the injuries incurred by Mrs. Chase. During the trial, B & M was voluntarily dismissed. The case was sent to the jury on theories of strict liability and negligence in design, and the jury returned a verdict in favor of the plaintiffs for $2,211,309.45 for Mrs. Chase and $50,000.00 for Chase.

The plaintiffs were allowed, over the objection of GM, to introduce into evidence videotaped braking tests, one performed by the National Highway Traffic Safety Administration (NHTSA) and another by Dr. Rudolf Limpert. Accompanying the NHTSA video was a written report describing the test using an M-3 Citation1 and the conclusions drawn therefrom. The report was also admitted into evidence. Dr. Limpert testified as an expert witness for the appellees explaining his videotape of the test which used other M-3 Citations. Dr. Limpert further testified to his conclusions about the accident based on the test results.

GM attempted to admit a videotape made by its expert, William Gillespie, performing the NHTSA/Limpert test protocol on five other vehicles. The trial court sustained plaintiffs' objection to the admissibility of Gillespie's test videotape and explanatory testimony. The court stated that the ground for sustaining the motion was that the vehicles in GM's test were not 1980 Citations with brake components identical to that of the Chases' car.

The NHTSA and Limpert tests, alleged to demonstrate the physical principle in dispute, were performed in daylight, on flat, straight surfaces under controlled protocols at test facilities or on an airport runway. The tests used experienced drivers at regulated vehicle speeds and who steered straight ahead. Brake pedal force in the NHTSA tests was mechanically applied to produce single axle lock. The NHTSA tests were performed on three different surface types under wet conditions. Dr. Limpert, applying the brakes himself, similarly tested the vehicle on both wet and dry surfaces.

In contrast, the conditions appertaining the night of the wreck were quite different. Chase was driving downhill on a slight curve to the left and was so turning his car at the time of the accident. While we must, for present purposes, credit his trial testimony that he did not further turn his car to the left, Chase admits it was already slightly so turning at the time of the wreck. The night was quite cold and the surface of the road was covered with snow and ice. Instead of applying measured or pre-ascertained brake pressure, Chase instinctively applied the brakes. "It was an instinct, to apply the brakes," Chase testified. He testified that he was trying to stop the car as soon as he could. Thus, the conditions obtaining at the scene of the accident were vastly different from any test performed either by NHTSA or the plaintiffs' expert, Dr. Limpert.

In the case of Gladhill v. General Motors Corp., 743 F.2d 1049 (4th Cir.1984), we were presented with a fact situation indistinguishable from that of the case at hand. The Gladhills, like the Chases, had purchased a 1980 Chevrolet Citation. The automobile was purchased in November 1979. Their claim was that the brakes would occasionally lock with only slight pressure on the pedal, causing the car to skid. In September 1980, Gladhill, while driving alone, left the road and collided with a utility pole. His claim was that the faulty brakes on the car had caused him to leave the road. His accident occurred at night on a hill sloping downwards at a sharp curve between the two elbows of the curve. GM was permitted to introduce a videotape demonstration of a brake test of a Citation which had been conducted at a GM test facility on a flat straight asphalt surface in daylight by an experienced test driver. GM introduced the test to show that, while a rear brake might have locked, the car would continue to travel in a straight line. The jury verdict and the judgment were for GM, and the plaintiffs appealed. We reversed and stated, at p. 1051 of the opinion, that: "Indeed, the circumstances of the accident, as alleged, are so different from this test as to make the results largely irrelevant, if not misleading." We continued, on p. 1052: "It is possible to call almost any evidence of this type 'a demonstration to illustrate a principle,' but when the demonstration is a physical representation of how an automobile behaves under given conditions, those conditions must be sufficiently close to those involved in the accident at issue to make the probative value of the demonstration outweigh its prejudicial effect." We continued that the conditions of the demonstration in Gladhill were dissimilar in such fundamental and important respects that the risk of prejudice to the plaintiffs outweighed the probative value of the evidence. 743 F.2d at 1052.

In the case before us, the Chase car was going downhill, the test cars were on a flat surface; the Chase car was on a road covered with ice and snow, the test cars were on roads either dry or merely wet or damp; the Chase car was turning slightly to the left at the time the brakes were applied, the test cars were steered in a straight line; the Chase accident occurred in subfreezing temperature, the temperature at the tests was not subfreezing; and Chase applied the brakes instinctively to stop as quick as he could, the drivers of the test cars applied measured force to the brakes. Thus, the conditions of the tests performed both by the plaintiffs' expert and by NHTSA, as in Gladhill, were so dissimilar to the conditions existing at the time of the accident "... in such fundamental and important respects that the risk of prejudice to ... [GM] outweighed the probative value of the evidence." 743 F.2d at 1052. We further note that Gladhill, on retrial, required the exclusion of "... all evidence generated by this experiment, including testimony pertaining thereto." 743 F.2d at 1052 and Fn. So, on retrial, not only will the videotapes be excluded but also the NHTSA test results and testimony obtaining thereto.2

GM also complains that the trial court excluded a videotape and test results produced by it under similar conditions as those offered by the plaintiffs except that the GM tests were run on different makes of automobiles with brakes similar to those of the Citation in order to prove that the other vehicle's behavior was similar to that of the Citation, with the sought-for conclusion that the GM vehicle was designed with a standard of care consistent with the state of the art at the time of the sale of the Citation in question. The district court gave as its reason for excluding the GM tests that they were not conducted using a Citation, rather different vehicles. And, we should say that such reasoning is ordinarily correct if one can consider, as the district court apparently did, that the GM tests and tape were offered to contradict the tests and tapes offered by the plaintiffs. In this case, however, on retrial, the GM tests and tapes should be excluded for the same reasons we have given to exclude the plaintiffs' tests and tapes.

If the holding of the district court is taken to mean that evidence as to the state of the art in a products liability case, where a standard of care, or whether the product is susceptible to a strict liability analysis, is in question, is inadmissible, it was incorrect. The question in products cases such as this one is the standard of care "at the time the product was marketed." Werner Co. v. Upjohn, Inc., 628 F.2d 848, 859 (4th Cir.1980). There is no doubt that almost every product which is marketed can be manufactured to be more safe than it is. The question is: did the manufacturer use reasonable care in designing and manufacturing the product at the time it was marketed, not whether it could possibly have been made better or more safe, or later has been made better or more safe. In this respect, the state of the art at the time the product was marketed, while not conclusive, is evidence of due care. See 1A Products Liability, Frumer and Friedman, Sec. 2.26[d][c] (1987). It is also evidence of whether or not a product is "unsafe" under the theory of recovery of strict liability in tort which has been adopted in West Virginia. Morningstar v. Black and Decker Mfg. Co., 162 W.Va. 857, 253 S.E.2d 666, 682-83 (1979).

In this respect, we do not mean to imply that on retrial expert testimony is to be excluded. In products liability cases, expert witnesses necessarily may play a significant part, and this case is no exception. When their testimony, however, is based upon the result of tests which were conducted under such different circumstances than those obtaining at the time of the accident complained of so as to make the results "largely irrelevant, if not misleading," Gladhill p. 1051, that testimony must be excluded. On retrial, we do not intend to exclude expert testimony.

To recall, this accident occurred in January 1982; the car had been purchased by the Chases in April 1980. In July 1980, GM made a design change in the braking system of Citations of the type purchased by the plaintiffs. This change involved the lining material of the rear brakes. As a result of complaints of rear brake lock involving 1980 model Citations of the type owned by the plaintiffs, GM, in February 1983 in response to the NHTSA, recalled plaintiffs' vehicle to make the change it had made in July 1980 on Citations assembled subsequent to that time.

The recall letter recited that the Chases' car under certain operating conditions and braking situations might experience premature rear wheel skid, most likely on a wet or slick surface, and authorized the Chases to bring their car in to their dealer for the July 1980 modification to be made to the braking system. Only a few days after the wreck, however, Chase had sold his car for scrap and the same had been crushed, so the Chase vehicle was not available either for modification or inspection.

The district court excluded the February 1983 recall letter on the ground that it had been required by the government and was not any voluntary act on the part of GM, but it allowed testimony during the trial of the fact of recall. This fact was used powerfully by plaintiffs' attorneys in closing argument, as excerpts show:

There is further confirmation that this was a bad, dangerous, defective design by GM itself. Following this collision, in 1983, the Chase car and all the M-3 cars were recalled voluntarily by GM, recalled because they wanted to change, and did change, the Bendix lining to the less aggressive 40-50 lining, and that occurred per GM's chart.

* * *

* * *

[GM] says that they, that the brakes weren't bad and that the recall was a publicity move on the part of the businessmen at General Motors, the fellows that sit in Detroit. Well, if that's right, why in the world would they spend ten millions of dollars bringing every Citation in off the highway to change the brake parts? Does that make sense to you? Does that sound like a sound business decision by brilliant business executives, who are paid hundreds of thousands of dollars every year to make decisions like that?

* * *

* * *

They replaced every Bendix lining on every Chevrolet X-Car in America, and they have had no trouble since that point in time, no problem, but they come in here and want you, the jury, to believe there was nothing wrong with those brakes; great braking system.

Exclusion of the letter is not the issue on appeal; it is admitting the evidence of recall. GM argues that it should be excluded under F.R. of Evid. 407, which is, in pertinent part:

When, after an event, measures are taken which, if taken previously, would have made the event less likely to occur, evidence of subsequent measures is not admissible to prove negligence or culpable conduct in connection with the event.

The "event" spoken of in that rule in our case is the date of the accident, which was in January 1982. Wright and Graham Federal Practice and Procedure, Evidence, 1980, Sec. 5283. That the evidence of the recall was used "to prove negligence or culpable conduct," is undoubted as shown by the argument recited above. That the February 1983 recall was taken "after ... [the] event" is also undoubted. That the plaintiffs take the position that the changes occasioning the recall "would have made the event less likely to occur" is also beyond question. Therefore, we are of opinion that evidence of the fact of recall was improperly admitted under Rule 407. This is not a new principle and has been the rule in the courts of the United States at least since 1892, when the Court stated in Columbia & Puget Sound R. Co. v. Hawthorne, 144 U.S. 202, 12 S. Ct. 591, 36 L. Ed. 405 (1892):

But it is now settled, upon much consideration, by the decisions of the highest courts of most of the states in which the question has arisen, that the evidence is incompetent, because the taking of such precautions against the future is not to be construed as an admission of responsibility for the past, has no legitimate tendency to prove that the defendant had been negligent before the accident happened, and is calculated to distract the minds of the jury from the real issue, and to create prejudice against the defendant. (Citations omitted) 144 U.S. at 207, 12 S.Ct. at 593.

Our decision here is in accord with our opinion in Werner v. Upjohn Co., 628 F.2d 848 (4th Cir.1980), cert. den. 449 U.S. 1080, 101 S. Ct. 862, 66 L. Ed. 2d 804 (1981), in which we held that Rule 407 applied not only to a theory of negligence but also to a theory of strict liability in products liability cases, and we adhere to that opinion.

But that is not the end of the matter. The change in the brake design of the Citation occurred in July 1980, some two or three months after the plaintiffs' car was manufactured and sold to them. That was a "measure" which was "taken" before the "event" mentioned in Rule 407. So Rule 407 does not exclude evidence with respect to that change. On retrial, the plaintiffs are perfectly free to explore the July 1980 change in the braking system and the reasons for it if such change might tend to be evidence of negligence, or if such evidence is otherwise admissible under the law of strict liability as it exists in West Virginia.

Two further things deserve brief mention.

The district court's instruction on proximate cause which might not be reversible error, a question we do not reach, would be more accurately stated if it contained the word "collision" or some such language instead of the word "damage" or "damages." There is no need that we see to walk so close to the enhanced theory of recovery, which we have held does not apply in West Virginia in McClung v. Ford Motor Co., 472 F.2d 240 (4th Cir.1973).

While we do not think the district court abused its discretion in admitting evidence of the cost of attendant care for Mrs. Chase, if the figures submitted to us by GM are correct, and we are unable to ascertain this because of the general verdict, it would seem that the amount so attributed by GM is several times as much as attendant care would reasonably cost. That, however, is a question we do not reach, for the case must be tried again.

The judgment of the district court is vacated and the case is remanded for a new trial not inconsistent with this opinion.

VACATED AND REMANDED.