OPINION OF THE COURT
SEITZ, Chief Judge.
Defendant (“Ford”) in this diversity action under Pennsylvania’s Wrongful Death and Survival Statutes appeals a judgment based on a jury’s verdict in plaintiff’s favor. Vizzini v. Ford Motor Co., 72 F.R.D. 132 (E.D.Pa.1976).
I.
FACTUAL BACKGROUND
Salvatore Vizzini (Vizzini), late husband of plaintiff, died in a one-vehicle accident at approximately 12:30 p. m. on Saturday, September 1, 1972. Vizzini was killed on impact when his 1972 Ford F-100 pickup truck skidded through the “T”-type intersection of two rural roads near his vacation home in Seabreeze, New Jersey, and struck a tree. The skid marks on the macadam road were measured at 363 feet in length, and were in a nearly straight line. The marks were left by the tires on only one set of wheels. The marks were almost entirely on the wrong side of the road. When Vizzi-ni was found, he was protruding from the waist up out of the driver’s side door window. Although the truck was equipped with lap-type safety belts, when found, Viz-zini was not wearing a safety belt.
Vizzini had worked at his job as a trouble-shooter for the Philadelphia Electric Co. the entire day and night preceding the accident, i. e., from 8:00 a. m. Friday, August 31, straight through until 8:00 a. m. Saturday, September 1. His son then drove him from his suburban Philadelphia home to Seabreeze, New Jersey, where the family had a vacation cottage. During the 28 or so hours preceding his death Vizzini only had short naps which were taken during the drive to Seabreeze, a trip that took approximately 214 hours.
Vizzini remained at this vacation home for about IV2 hours before he left on his fatal ride. During that time he drank three bottles of beer. The autopsy revealed a blood alcohol concentration of .168, in excess of the percentage required for a presumption of intoxication under both Pennsylvania and New Jersey law. After drinking the beer, Vizzini left the cottage driving the F-100 pickup, and shortly thereafter was involved in the accident in which he died. There was no witness. The day of the accident was sunny, dry, and clear, and the road was free of hazards and obstructions.
Vizzini’s wife, administratrix of his estate, then brought this action against Ford. Her complaint alleged negligence in the manufacture of the truck or, in the alternative, that the truck was defective and unreasonably dangerous under a strict liability theory. She based her claim upon a defect in the left front brake assembly that was discovered during post-accident inspections of the truck. She alleged that a self-adjuster cable in that brake assembly was negligently or defectively assembled in that it was not properly connected. As a result, she claimed, the left front brake shoe was unable to expand to compensate for wear, resulting eventually in failure of the brake which in turn caused the accident. Plaintiff also alleged that there was leakage of brake fluid from the master cylinder, and that such leakage contributed to the failure of the truck’s brake. She asserted that the failure of the left front brake of her husband’s truck was the proximate cause of her husband’s death.
The case first was tried to a jury on the issue of liability. Plaintiff’s expert gave his opinion that the accident had been caused by the failure of the truck’s front brakes. Her experts further testified that [757]*757such failure was caused by a defectively manufactured self-adjuster system, aggravated by the defective master cylinder. Ford’s experts testified that the alleged defects could not have caused the accident, and Ford argued that the sole proximate cause of the accident was the negligence of Vizzini himself.
The jury’s special verdict was in the form of answers to six interrogatories. The jury found Ford to have been negligent in the manufacture of the truck, and that such negligence was a proximate cause of the accident. It found Vizzini to have been negligent as well, and that his negligence was also a proximate cause of the accident. And it found that the truck was defective to the point of being unreasonably dangerous, and that such defect was a proximate cause of the accident.
A trial on the issue of damages was then held before the same jury. After hearing extensive evidence on plaintiff’s damages claim, and after two full days of deliberation, the jury was unable to agree on damages. In a note to the court the jury explained that one juror was unable to agree with the others on a damage figure. The note indicated that the one holdout juror had compromised his or her earlier vote on liability in order to achieve unanimity on the first verdict, but had done so in the belief that the liability verdict as rendered would result in a “draw” with no damages assessed against either party. Accordingly, that juror was unable to agree with the others on the amount to be awarded plaintiff.
After ascertaining that the deadlock was indeed insoluble, the court declared a mistrial. It denied Ford’s motions for judgment n. o. v. and for a new trial on liability as well as damages, and ordered a new trial limited to the issue of damages under the strict liability claim.
At that second trial on damages, plaintiff produced testimony, inter alia, concerning the deceased’s income, expenses and employment history. Plaintiff also put an economist on the stand who testified that an employee in Vizzini’s position at the time of his death could have expected to receive yearly pay increases of 3% per year, a figure the economist based on projected increases in national productivity. An actuary testified for plaintiff to the value of the various fringe benefits Vizzini was receiving as part of his compensation package, to his life expectancy, and to the value of his services as a husband and father. The actuary computed the present value of Vizzi-ni’s total lost future earnings using a 6% simple interest factor.
Ford objected to various aspects of plaintiff’s evidence on damages. In addition, Ford’s offer of proof concerning non-usage of safety belts by Vizzini was rejected by the court. The court also denied Ford’s request that it be allowed to demonstrate the impact of taxes on Vizzini’s future earnings. Ford presented no further evidence in rebuttal. The jury awarded plaintiff $421,000.00 under the Pennsylvania Survival Act, and $52,298.00 under the Pennsylvania Wrongful Death Act.
II.
FORD’S MOTION FOR JUDGMENT NON OBSTANTE VEREDICTO
Ford on appeal seeks a reversal of the district court’s refusal to grant its j.n.o.v. motion. The motion was based on the ground that the plaintiff failed to introduce adequate evidence that the allegedly defective self-adjuster cable caused Vizzini’s accident. The trial judge also had denied Ford’s earlier motion for a directed verdict. Though he characterized plaintiff’s causation evidence as “very thin,” 72 F.R.D. at 135, the trial judge did believe that sufficient evidence had been introduced to warrant submission of the question of causation to the jury.
Before deciding the merits of Ford’s j. n.o.v. motion, we first must determine what standards are to guide this court in gauging the sufficiency of the evidence. The federal courts, as well as the commentators, have split over whether federal or state standards should control such determinations in diversity cases, see 9 C. Wright & A. Miller, [758]*758Federal Practice & Procedure § 2525 (1971). This circuit, however, in cases applying Pennsylvania law, has found there to be “no difference in the standard followed in Pennsylvania courts and in the federal courts” on the question of “the quantum of evidence necessary to establish a jury question on the issue of causation.” Kridler v. Ford Motor Co., 422 F.2d 1182, 1183-84 & 1183 (3d Cir. 1970). This court, therefore, when applying Pennsylvania law, has proceeded without distinguishing between the federal and Pennsylvania approaches. See Neville Chemical Co. v. Union Carbide Corp., 422 F.2d 1205, 1211-12 (3d Cir. 1970).
In Neville Chemical Co. v. Union Carbide Corp., 422 F.2d 1205 (3d Cir. 1970), this court characterized the standard by which j. n.o.v. motions are evaluated under federal law as requiring “that we ‘determine whether, as a matter of law, the record is critically deficient of that minimum quantum of evidence from which a jury might reasonably afford relief.’ ” 422 F.2d at 1211-12, (footnote omitted), quoting Denne-ny v. Siegel, 407 F.2d 433, 439 (3d Cir. 1969). The court in Neville also stated the Pennsylvania standard:
The Pennsylvania test, as stated in Smith v. Bell Telephone Co. of Pa., 397 Pa. 134, 153 A.2d 477 (1959) was summarized in Denneny as requiring “the trial judge [to] determine as a minimum requirement of a prima facie case whether the plaintiff has produced ‘substantial evidence’ of ‘sufficient facts’ having the capacity of supporting a logical conclusion.” [footnotes omitted]
Neville Chemical Co. v. Union Carbide Corp., supra, 422 F.2d 1205, 1211, quoting Denneny v. Siegel, 407 F.2d 433, 440.
Our task, then, in evaluating Ford’s motion for judgment n.o.v. is the same whether Pennsylvania or federal standards are applied. And since plaintiff has the benefit of the jury’s verdict in her favor, in reviewing the district court’s denial of Ford’s motions “we must consider the evidence in the light most favorable to plaintiff, drawing all reasonable inferences in favor of the plaintiff and against the defendant.” Kademenos v. Equitable Life Assurance Society of the United States, 513 F.2d 1073, 1074 (3d Cir. 1975); accord, Moyer v. Ford Motor Co., 205 Pa.Super. 384, 386, 209 A.2d 43, 44 (1965).
Ford urges us to reverse the district court’s denial of its j.n.o.v. motion on the ground that plaintiff failed to show sufficient evidence to prove any causal connection between the alleged defect in the self-adjuster cable of the truck’s left front brake and the accident that killed Vizzini. Plaintiff, of course, even under a strict liability theory, must prove not only the existence of the defect but also that such defect was a proximate cause of the injury at issue. Cornell Drilling Co. v. Ford Motor Co., 241 Pa.Super. 129, 359 A.2d 822, 825 (1976). Ford argues that the expert upon whom plaintiff chiefly relied, Mr. Pruyn, offered only a conclusory assertion on causation. That assertion, Ford argues, was buttressed by no objective data supporting his theory that the weakness in the self-adjuster cable was capable of causing brake failure. No tests of similarly faulty cable, or of the cable in Vizzini’s truck itself, were run by Mr. Pruyn. No treatises were cited to support his theory, no authorities offered for the assertion that an unconnected cable similar to that in Vizzini’s truck could lead to brake failure.
Moreover, Ford argues that Mr. Pruyn’s testimony was often inconsistent with important points of plaintiff’s theory of the case. The brake pedal reserve and the evidence of residual braking action in the left front tire, both noted by Mr. Pruyn in his post-accident inspection of the truck, were inconsistent with the brake failure that plaintiff alleged occurred. Further, Ford’s own test movies showed that the skid marks found at the scene of the accident were not consistent with those that would be expected to occur after front brake failure. Ford also points to inconsistencies in the testimony of plaintiff’s witnesses concerning the number of wheels that left the skid marks, and concerning the alleged leakage from the master cylinder. And Ford argues that its own experts and tests demonstrated conclusively that the alleged defect did not cause the accident in which Vizzini died.
[759]*759Ford relies on Moyer v. Ford Motor Co., 205 Pa.Super. 384, 209 A.2d 43 (1965), and Denneny v. Siegel, 407 F.2d 433 (3d Cir. 1969) to support its assertion that plaintiffs expert testimony was insufficiently supported by the facts to create a jury question on the issue of causation. In Moyer, however, plaintiff’s only evidence in support of his theory that the accident at issue resulted from a “freezing up” of the front wheels was what the court called the “rather confused” testimony of an automobile mechanic who “had neither seen nor examined the damaged automobile” and who admitted there were several explanations for the accident. 205 Pa.Super. 386-87, 209 A.2d at 44. The defendant in Moyer had produced two experts, one of whom had examined the car, and both of whom testified that the wheel had not frozen up.
Similarly, in Denneny the plaintiff offered the testimony of two experts to prove that her post-operative infection had been caused by the hospital’s negligence in allowing into the operating room persons wearing unsterilized clothing. But one of plaintiff’s experts had been unable to determine any cause for plaintiff’s infection, while the other acknowledged that, had the infection been caused by non-sterile conditions, any number of factors other than the one alleged by plaintiff might have created those conditions. 407 F.2d at 436.
This case is different. Here plaintiff’s experts were adequately qualified to testify as experts. They testified to a defect in the adjuster cable of the left front brake. There was credible evidence of leakage shortly before the accident from that master cylinder controlling the left front brake. Plaintiff introduced testimony that the skid marks at the scene of the accident were left by only the rear tires, thus supporting her theory that the front brake had not worked properly. She offered explanations for the presence of some braking power in the left front wheel after the accident. Moreover, plaintiff used some of the tests conducted by Ford itself, as well as admissions elicited on cross-examination of Ford’s witnesses concerning the relationship of self-adjuster cable failure to brake failure, to bolster her case.
Viewing the record as a whole, and in the light most favorable to plaintiff’s case, we cannot say as a matter of law that her case is “critically deficient of that minimum quantum of evidence from which a jury might reasonably afford relief.” Neville Chemical Co., supra, 422 F.2d at 1211-12. Much of plaintiff’s evidence is conflicting. Some of her witnesses were confused. But enough evidence was introduced to support plaintiff’s causation claim, provided the jury chose to believe that evidence. The credibility of plaintiff’s witnesses was for the jury. “It is not necessary . that every fact or circumstance point unerringly to liability;' it is enough that there be sufficient facts for the jury to say reasonably that the preponderance favors liability.” Smith v. Bell Telephone Co. of Pa., 397 Pa. 134, 138, 153 A.2d 477, 480 (1959). It was therefore not error to deny Ford’s j. n.o.v. motion.1
III.
FORD’S MOTION FOR A NEW TRIAL AS TO ALL ISSUES
Ford seeks a reversal of the district court’s denial of its motion for a new trial on the ground, inter alia, that it was inconsistent with the sound exercise of discretion for the trial judge to limit the second trial below to the issue of damages. Under Fed.R.Civ.Pro. 59(a) the trial court, in its sound discretion, may limit the grant of a new trial to only a portion of those issues litigated in the original proceeding, including the issue of damages only. 6A Moore’s Federal Practice 159.06 at 59-81 (1973). The trial court’s decision to grant a limited new trial is reviewable by this court when that decision comes before us, as it does in this case, as part of an appeal from ' a final judgment. Springfield Crusher, Inc. v. Transcontinental Ins. Co., 372 F.2d 125 (3d Cir. 1967).
[760]*760The standard of review of an order under Fed.R.Civ.Pro. 59(a) “is a matter of federal procedure and is, in no wise, subject to state practice.” Silverii v. Kramer, 314 F.2d 407, 413 (3d Cir. 1963). Accord, 11 C. Wright & A. Milier, Federal Practice & Procedure § 2802 (1973). And the standard we apply on review is whether, viewing the circumstances as a whole, it was consistent with the sound exercise of discretion for the trial judge to limit the new trial to the issue of damages. Though the discretion of the trial judge is broad in such cases, it is not boundless.
We recognize, as we must, that a motion for a new trial ... is addressed to the sound discretion of the trial judge and its denial is not ordinarily reviewable in the absence of a showing of exceptional circumstances such as an abuse of discretion. . . . There is an abuse of discretion within the meaning of the rule, when the action of the trial judge is clearly contrary to reason and not justified by the evidence.
Springfield Crusher, Inc. v. Transcontinental Insurance Co., 372 F.2d 125,126 (3d Cir. 1967) (citations omitted).
Ford argues that the issues of liability and damages were so intertwined that it was improper to allow the second jury to try the damage issue only. The circumstances under which it is proper to limit a new trial to certain issues were set out by the Supreme Court in the venerable case of Gasoline Products Co., Inc. v. Champlin Refining Co., 283 U.S. 494, 51 S.Ct. 513, 75 L.Ed. 1188 (1931). That case resolved a split among the circuits by establishing the propriety of partial new trials; but it also qualified the power to do so:
Where the practice permits a partial new trial, it may not properly be resorted to unless it clearly appears that the issue to be retried is so distinct and separable from the others that a trial of it alone may be had without injustice. . Here the question of damages on the counterclaim is so interwoven with that of liability that the former cannot be submitted to the jury independently of the latter without confusion and uncertainty, which would amount to a denial of a fair trial.
Id. at 500, 51 S.Ct. at 515 (citations omitted).
In the forty-six years since the decision in Gasoline Products, the judicial limitation on the power to grant partial new trials has retained its force. It has been read to prevent limited new trials where a tangled or complex fact situation would make it unfair to one party to determine damages apart from liability, 6A Moore’s Federal Practice 159.06 at 59-89 (1973), or where “there is reason to think that the verdict may represent a compromise among jurors with different views on whether defendant was liable.” 11 C. Wright & A. Miller, Federal Practice and Procedure § 2814 at 96 (1973) (footnotes omitted).
This circuit has consistently observed the stricture of Gasoline Products. For example, this court has required that partial new trials should be granted “ ‘only in those cases where it is plain that the error which has crept into one element of the verdict did not in any way affect the determination of any other issue.’ ” Romer v. Baldwin, 317 F.2d 919, 922-23 (3d Cir. 1963), quoting Thompson v. Camp, 167 F.2d 733, 734 (6th Cir.), cert. denied, 335 U.S. 824, 69 S.Ct. 48, 93 L.Ed. 378 (1948). Accord, e. g., Darbrow v. McDade, 255 F.2d 610, 611 (3d Cir. 1958).
The district court opinion in Feinberg v. Mathai, 60 F.R.D. 69 (E.D.Pa.1973) represents what we believe to be a sound approach in circumstances similar to those involved in this appeal. In Feinberg, plaintiff’s decedent had died in a crash between the defendant’s automobile and the motorcycle upon which plaintiff’s decedent was a passenger. The trial was bifurcated, with damages tried separately from liability. The trial judge stated that his original decision to bifurcate the trial was based on “the close question of liability” and the hope that a verdict for plaintiff on liability might induce a settlement on damages. Id. at 70. The jury, in answers to several interrogatories, found both the defendant and the driver of the motorcycle negligent. It also found that plaintiff’s decedent had [761]*761not assumed the risk of injury. The court accordingly molded a liability verdict in plaintiff’s favor.
After the damages phase of the trial, however, the jury returned what the court termed a “patently insufficient verdict as to damages.” Id. at 71. Plaintiff then moved for a new trial limited to damages only, while defendant argued that any new trial granted should extend to all issues.
The trial court began its analysis from the premise that it could limit the grant of a new trial only if “the- error requiring a new trial [had] not affected the determination of any other issue.” Id. at 70. The court noted the requirement of Romer v. Baldwin, supra, that partial new trials should not be granted “unless the issue to be retried is distinct and separate and the error which has crept into one element of the verdict did not in any way affect the determination of any other verdict.” Id. 60 F.R.D. at 70. Given the close question of liability and the jury’s failure to find adequate damages in the face of strong evidence of substantial damages, the trial court found it “possible that the jury interwove testimony concerning liability over into damages, tainting the verdict with elements of compromise.” Id. at 71. It therefore required a new trial on all issues.
But even had the court been satisfied that the jury did not take into consideration an element of compromise, it nevertheless felt that a new trial was dictated. For if there were no compromise
then there is little left but to conclude that the jury misunderstood or disobeyed the instructions given it. If that was the case, then there is no reason to believe that the jury actually followed the court’s charge on liability.
Id.
In the circumstances of the case now before us, we conclude that the issue of damages is not “so distinct and separable from the [issue of liability] that a trial of it alone may be had without injustice.” Gasoline Products Co., Inc. v. Champlin Refining Co., 283 U.S. at 500, 51 S.Ct. at 515. We are most reluctant to disturb the trial judge’s ruling on the limited new trial. But it is difficult to say that allowing a second jury to determine the issue of damages in isolation from the whole of the circumstances surrounding the case was not an injustice to Ford.
The question is a close one. On this record we think the original jury’s failure to agree on damages indicates that its deliberations on damages were affected by the liability issues in the ease. Here the liability evidence presented by plaintiff, though sufficient to warrant submission of the issue to the jury, was “very thin.” Plaintiff’s expert witnesses were self-contradictory on many points, and were strongly contradicted by Ford’s experts on many others. Several facts inconsistent with plaintiff’s theory of the case were not entirely explained away. Nevertheless, after nine hours of deliberation the jury returned a verdict for plaintiff on the strict liability count.
On the other hand, the evidence on damages was sufficiently convincing to support a substantial award. But after nearly two full days of deliberation on damages — substantially longer than it took to decide the liability issue — the jury could not agree on an award. As in Feinberg, supra, this indicates that the jury’s deliberations on damages may have been influenced by its views on liability. Or it may indicate that the jury confused its responsibilities in one or both phases of the trial. In either event, we believe that the problems that led to a mistrial in the damages portion of the trial also affected the liability phase.
The Fifth Circuit reached a similar result on analogous facts, overruling the trial judge’s determination in Hatfield v. Seaboard Airline RR Co., 396 F.2d 721 (5th Cir. 1968). In that case the evidence of liability was close, and the evidence of damages was extensive. The record reflected some confusion among the members of the jury on the effect of a finding of contributory negligence. In any event, the jury’s answers to special interrogatories found the railroad negligent, and the plaintiff free from any contributory negligence. The jury, however, awarded only $1.00 in damages. The trial judge refused plaintiff’s request for a limited new trial on damages, and from that decision the plaintiff appealed.
[762]*762The Court of Appeals for the Fifth Circuit recognized that the scope of review of a refusal by a trial judge to grant a new trial was “necessarily narrow.” Id. at 723. Nevertheless, it found it an abuse of discretion for the judge to have denied a new trial in the circumstances of the case: the issue of liability was close and hotly contested; there was evidence of jury confusion on an important issue; and after lengthy deliberations the jury had awarded only “nominal damages in the teeth of an uncontroverted showing of substantial damages.” Id.
The Court of Appeals directed a new trial on liability as well as damages. The verdict on damages, it held, pointed “irrefutably to some kind of jury impropriety which could not help but affect the verdict as a whole.” Id. at 724.
It is highly unlikely that the jury, upon proper deliberation, found the first four interrogatories in accordance with the court’s instructions, and then, in contravention of those same instructions, refused, either through misunderstanding or through willfulness, to assess the damages ensuing. On the contrary, the nominal damage award can be seen only as the result of either a compromise on one of the liability issues or as an attempt to render a verdict for Seaboard with Seaboard paying the costs. In either event the misconduct necessarily contaminated the entire verdict, and precludes a new trial on the damage issue alone.
Id. (footnote omitted).
Our decision that there must be a complete retrial is also consistent with the reasoning of those cases that have allowed a new trial limited to damages only. In those cases, the error requiring a new trial was an error relating peculiarly to the damages portion of the trial only, one that did not implicate the jury’s liability verdict. E. g., Wagner v. Reading Co., 428 F.2d 289 (3d Cir. 1970).
We therefore reverse the decision of the lower court, and order a new trial on the issue of Ford’s liability under the strict liability claim as well as on the issue of damages.2
IV.
FORD’S ASSIGNMENTS OF ERROR ON THE ISSUE OF DAMAGES
Ford asserts that the trial court committed several errors concerning the issue of damages. To provide guidance for the trial court we will discuss those issues which the district court decided and which probably will recur in a new trial.
A.
Ford argues that plaintiff did not meet her burden of proving the cost of her husband’s maintenance. Under the Wrongful Death and Survival Acts plaintiff bore the burden of showing the amount Vizzini spent in support of himself. Such amounts did not, of course, constitute part of her losses after his death. Blackburn v. Aetna Freight Lines, Inc., 368 F.2d 345 (3d Cir. 1966).
Controlling Pennsylvania law “only requires that a reasonable quantity of information must be supplied by plaintiff so that the jury may fairly estimate the amount of damages from the evidence.” Ashcraft v. C. G. Hussey & Co., 359 Pa. 129, 132-33, 58 A.2d 170, 172 (1948). And as the court noted in Smail v. Flock, 407 Pa. 148, 180 A.2d 59 (1962), the Pennsylvania Supreme Court has consistently held that the law
“does not require that proof in support of claims for damages or in support of claims for compensation must conform to the standard of mathematical exactness. In Western Show Co., Inc., v. Mix, 308 [763]*763Pa. 215, 162 A. 667, 668, we held that evidence in support of a claim for damages was sufficient ‘if it afforded a reasonably fair basis’ for calculating the plaintiffs loss.”
407 Pa. at 154, 180 A.2d at 62, quoting Getz v. Freed, 377 Pa. 480, 485, 105 A.2d l02, 104-05 (1954) (citations omitted) (emphasis in the original).
In this case, plaintiff testified that her husband habitually kept for his own expenses approximately $50.00 out of his weekly salary of $400.00 and gave the rest to her for family use. She adduced testimony, inter alia, concerning her husband’s employment situation and his personal habits. Counsel for Ford had ample opportunity to cross-examine plaintiff and other witnesses. He was able to bring out on such examination that some of the $350.00 set aside for family expenses may have gone for items used by Vizzini. The evidence thus was sufficient to allow the jury to estimate fairly the amount of damages. “Certainly it lacked mathematical exactness and was far from precise or fully detailed, but under the law of Pennsylvania it is not essential that it should be.” Blackburn v. Aetna Freight Lines, Inc., 368 F.2d 345, 347 (3d Cir. 1966).
B.
Ford objects also to the testimony of plaintiff’s actuary on the value of certain fringe benefits due Vizzini from his employer. The plaintiff introduced testimony by her actuary, Mr. Goodfarb, concerning the value of certain sick pay and insurance benefits. Ford argues that those benefits had already been accounted for by Goodfarb in his forecast of Vizzini’s future income, since a worker in Vizzini’s position would be expected to receive either income for days worked or insurance benefits for days missed, but not both.
We agree with the trial judge that such testimony was properly admitted. Ford had ample opportunity to examine Mr. Goodfarb and to present rebuttal testimony on the question of what elements are properly includable in the calculation of lost income. Fringe benefits such as those testified to by Mr. Goodfarb are properly a part of the damages portion of plaintiff’s case, see Curry v. United States, 338 F.Supp. 1219 (N.D.Cal.1971), and we agree it was for the jury to evaluate the value of such benefits, including whether any duplication of benefits occurred. The trial judge, however, must take care to instruct the jury that duplication of damages is not allowable and that the jury should scrutinize the components of plaintiff’s damage claim to insure no double recoveries occur.
C.
We next consider Ford’s objections to the exclusion of evidence concerning the impact of income taxes on the lost future income of plaintiff’s decedent. Pennsylvania law controls this issue, and it is clear. The rule in Pennsylvania is that “in fixing damages for the determination of decedent’s earning capacity, the income tax consequences of the matter should not be taken into consideration.” Girard Trust Corn Exchange Bank v. Philadelphia Transportation Co., 410 Pa. 530, 538, 190 A.2d 293, 298 (1963). See Frankel v. United States, 321 F.Supp. 1331, 1348-49 (E.D.Pa.1970), aff’d without comment on this issue, sub nom. Frankel v. Heym, 466 F.2d 1226 (3d Cir. 1972). We agree with the trial court that the brief comment of the Superior Court in Gallagher v. Four Winds Motel-Hotel, 233 Pa.Super. 1, 335 A.2d 394 (1975), is not to be read as altering, even if it could, the well-established Pennsylvania rule on this matter.
D.
Ford argues that the formula used by plaintiff for computing the present value of her damages is outdated. Plaintiff’s actuary used a 6% simple interest formula, while Ford argues higher interest rates and compound interest are the norm.
Pennsylvania law is clear in requiring that “in reducing damages to their present worth, the interest must be computed simply and at the lawful rate of six per [764]*764cent only.” Brodie v. Philadelphia Transportation Co., 415 Pa. 296, 300, 203 A.2d 657, 659 (1964). Plaintiff’s actuary was thus correct in his method of computing the present worth of plaintiff’s future losses, and Ford’s objection on this point is without merit.
E.
Ford contends that the trial court erred in rejecting its offer of evidence concerning the consequences of non-usage of seat belts by Vizzini. During the damages phase of the trials below, Ford asked that it be allowed to prove by way of expert testimony that had Vizzini been wearing a lap-type safety belt his injuries, would have been far less severe than they actually were. There was evidence already on the record that the truck was equipped with such belts, and that Vizzini had not been wearing a belt at the time his body was discovered.
Ford proffered this evidence on non-usage solely in an attempt to reduce damages. The trial judge, noting that there seemed to be no Pennsylvania precedents on point, rejected the offer on the grounds that such evidence made the determination of damages too conjectural, since the jury would have to determine what the extent and nature of Vizzini’s injuries would have been had he suffered some lesser harm as a result of wearing a belt. The court also noted that evidence of non-use was “akin to a claim of contributory negligence, which is not a defense in a strict liability case,” 72 F.R.D. at 139, and rejected the evidence for this reason as well.
Pennsylvania’s appellate courts have not yet considered the propriety of the “seat-belt defense,” either in strict liability cases or in more traditional negligence actions. The new trial here will be limited to the question and extent of Ford’s liability under the strict liability claim only. Therefore we must predict how Pennsylvania’s appellate courts would view evidence concerning non-usage in a case governed by strict liability principles.
Pennsylvania bases its strict liability doctrine on the Restatement (Second) of Torts, § 402A. Following the Restatement’s reasoning, Pennsylvania does not recognize the defenses of contributory negligence or comparative negligence in cases tried under § 402A. For example, in McCown v. International Harvester Co., 463 Pa. 13, 342 A.2d 381 (1975), the defendant sought to introduce evidence of plaintiff’s negligent driving as a defense to plaintiff’s strict liability claim. The evidence of negligence was offered both to defeat liability under the strict liability claim and to reduce damages in an amount equal to the plaintiff’s lack of due care.
Chief Justice Jones, writing for a unanimous Pennsylvania Supreme Court, rejected defendant’s evidence on both grounds. To admit such evidence of contributory negligence as a defense to the strict liability claim, the Court reasoned, would be to “defeat one theoretical basis for our acceptance of Section 402A” by contradicting the consumer’s “normal expectation of product safety.” Id. at 16-17, 342 A.2d at 382. The Court said that the defendant’s argument that the evidence of negligence should serve to reduce damages “would create a system of comparative assessment of damages for 402A actions” unknown to Pennsylvania law. Such a doctrine of comparative negligence, the Court continued, would be “unwise to embrace ... in the context of an appeal involving Section 402A.” Id. at 16, 342 A.2d at 382 (footnote omitted). In a footnote, the Court added that “[t]o initially apply a theory of comparative negligence to an area of the law in which liability is not premised on negligence seems particularly inappropriate.” Id. at n. 3, 342 A.2d at 382 n. 3.3
[765]*765We therefore believe that evidence of non-usage cannot be admitted in Pennsylvania under any general theory of comparative negligence. Ford also argues, however, that the evidence is admissible under the doctrine of avoidable consequences to prove plaintiff’s failure to mitigate damages. The general rule in Pennsylvania as elsewhere is that a plaintiff may not recover for harm resulting from the defendant’s wrongful act or omission that the plaintiff reasonably could have avoided. Ford argues that Vizzini reasonably could have avoided severe injuries by buckling his seat belt, notwithstanding the fact that the cable defect originally may have caused his truck’s brakes to fail. Ford contends that it should not be liable for damages caused only by that failure to wear the belt.
The rule of avoidable consequences traditionally has been applied only where plaintiff has failed to mitigate damages by some act or omission after the original injury had already occurred. “[0]ne injured by the tort of another is not entitled to recover damages for such harm as he could have avoided by the use of reasonable effort or expenditure after the commission of the tort." Restatement (Second) of Torts, § 918(1) (Tent. Draft No. 19, 1973) (emphasis added). See Restatement of Torts § 918(1) (1939). Plaintiff’s negligence prior to the injury that contributes to the injury generally has been addressed in terms of contributory and comparative negligence. C. McCormick, Handbook of the Law of Damages § 33 (1935) (hereinafter McCormick on Damages); W. Prosser, Handbook of the Law of Torts § 65 (4th ed. 1971) (hereinafter Prosser on Torts).
In this case, the negligence that Ford asserts against the plaintiff occurred before the injury. We thus do not face the traditional situation in which the rule of avoidable consequences has been applied as we would, for example, had plaintiff failed to seek medical aid after being injured by the tort of the defendant. Some few courts have extended the rule of avoidable consequences to encompass antecedent negligence of the plaintiff, though many others have refused to do so.4 See Prosser on Torts § 65 at 423-24; Comment, The Seat Belt Defense, 19 N.Y.L.F. 672 (1974). And the Restatement (Second) of Torts, though it confines the rule of avoidable consequences per se to negligence after the commission of the injury, recognizes that in some cases “apportionment may also be made where the antecedent negligence of the plaintiff is found not to contribute in any way to the original accident or injury, but to be a substantial contributing factor in increasing the harm which ensues.” Restatement Second of Torts § 465, comment c at 511 (1965). Cf. id. § 433A, comment f at 438.
Pennsylvania’s appellate courts, however, have not considered whether to extend plaintiff’s duty to mitigate damages to encompass negligent acts prior to the occurrence of the injury at issue. Nor have they considered the applicability vel non of the rule of avoidable consequences in strict liability cases. Rather, the Pennsylvania decisions have applied the rule in tort cases only where the suit was premised on negligence principles and only where conduct after the occurrence of the injury was involved. E. g., Chamberlin v. Morgan, 68 Pa. 168 (1871) (failure to seek medical aid after injury); Gervis v. Kay, 294 Pa. 518, 144 A. 529 (1928) (failure to sell stock at favorable price after breach by defendant of brokerage contract to sell); James v. Ferguson, 401 Pa. 92, 162 A.2d 690 (1960) (failure to seek medical aid after injury).
We therefore first must predict whether, in a case based on the strict liability principles of § 402A, the Pennsylvania courts would extend the rule of avoidable consequences to encompass plaintiff’s antecedent negligence.5
[766]*766The rule of avoidable consequences is closely akin to the doctrine of contributory negligence. Both describe the legal implications of plaintiff’s negligent acts. Both are based on the policy of conservation of resources. And “[b]oth rest upon the same fundamental policy of making recovery depend upon plaintiff’s proper care for the protection of his own interests.” Prosser on Torts § 65 at 423.
The two concepts differ in that, as traditionally applied, contributory negligence comes into play before the injury occurs. Avoidable consequences “comes into play at a later stage [after] the defendant has already committed an actionable wrong.” McCormick on Damages § 33 at 128-29. Even so, McCormick went on to say that “this distinction may be a mere matter of using different labels for two sides of the same bottle.” Id. at 129. Similarly, Pros-ser suggested “that the doctrines of contributory negligence and avoidable consequences are in reality the same.” Prosser on Torts § 65 at 424.
The two doctrines do differ in effect. A finding of contributory negligence is a complete bar to recovery. Failure to mitigate damages, on the other hand, results only in a proportionate reduction in plaintiff’s damages. This distinction is grounded more in the different times that each doctrine comes into play than it is in any difference between the types of negligent acts involved. Contributory negligence focuses on plaintiff’s negligence before the injury, and so on the cause of the injury. Where plaintiff’s negligence is held to be a contributing cause of the injury recovery is denied entirely, at least in part because of the traditional common law reluctance to apportion fault. Prosser on Torts § 65 at 416-418. But the avoidable consequences rule does not focus on the cause of the injury. It focuses on the negligence that aggravates an already existing injury. Since that negligence does not involve fault for causing the original injury, no corresponding bar to total recovery developed.
In the “avoidance consequence” cases, the initial damage cannot logically be charged to the plaintiff’s own negligence as a cause, while the later damages may be. If no such division can be made, the plaintiff’s negligence will bar all recovery, notwithstanding that it is subsequent in point of time to that of the defendant.
Id. § 65 at 423 (footnote omitted.) See McCormick on Damages § 33 at 127-130; Ford Motor Co. v. Dallas Power & Light Co., 499 F.2d 400, 414-15 (5th Cir. 1974); Alcoa Steamship Co. v. Charles Ferran & Co., 383 F.2d 46, 54 (5th Cir. 1967).
To extend the avoidable consequences rule to bar proportionate recovery for plaintiff’s prior acts of negligence shifts the focus away from the question of plaintiff’s aggravation of an existing injury to the question of plaintiff’s contribution to the original cause of that injury. Thus, the inquiry becomes essentially one into the contributory negligence of the plaintiff. Whether labelled an extension of “avoidable consequences” or an “apportionment of damages,” to reduce recovery in an amount equal to plaintiff’s antecedent negligence is to apply comparative negligence standards in the assessment of damages.
We believe that McCown v. International Harvester Co., supra, is evidence that Pennsylvania’s courts would not allow the introduction of such comparative negligence principles into a § 402A case in this manner. In McCown the defendant-manufacturer admitted the defect, but argued that the damages ought to be reduced by an amount equal to plaintiff’s negligence in driving recklessly prior to the accident. In this case Ford similarly seeks to reduce plaintiff’s damages for Vizzini’s antecedent negligence. We believe the Pennsylvania courts, as they did in McCown, would reject [767]*767such evidence as improper in cases premised on § 402A principles.6
We do not believe McCown distinguishable on the ground that the negligence there asserted against the plaintiff concerned the cause of the accident, while that here asserted by Ford goes to the cause of the injury. “Proximate cause” is the legal description of acts to which the law assigns responsibility for the harm suffered by plaintiff. Restatement (Second) of Torts §§ 430 et seq. It is not a rigid term confined only to the determination of mechanical cause and effect. Rather it is a flexible concept designed to effectuate the varied policies that determine whom the law holds liable for harm suffered by the plaintiff.
McCown makes it clear that the policy of § 402A is best served, once the alleged defect is determined to have been the proximate cause of the injury, by foreclosing inquiry into the contributing negligence of plaintiff prior to the injury. Of course, the plaintiff in a strict liability case must prove the existence of the defect, and that the defect was a proximate cause of the injury. It is possible that the alleged acts of negligence on the part of plaintiff prior to the injury should be considered in that determination of proximate cause, McCown v. International Harvester, supra, 463 Pa. at 17, 342 A.2d at 383 (Pomeroy, J., concurring), but that question is not before us.
The evidence of the effects of Vizzini’s failure to wear his safety belt was thus properly excluded below. We emphasize that in so holding we do not determine •the applicability vel non of the traditional rule of avoidable consequences to § 402A actions. Nor need we reach the difficult and controversial question, not yet decided by Pennsylvania’s courts, of whether and under what circumstances the failure to wear a seat belt may be considered negligence. Thus, we also do not decide whether the admission of evidence of non-usage of seat belts makes the determination of damages too speculative an undertaking. Finally, we express no opinion on the propriety of extending Pennsylvania’s avoidable consequences rules to encompass plaintiff’s prior negligence in cases not premised on § 402A principles.7
The only other reported decision we have been able to find concerning evidence of non-usage of seat belts in a strict liability case is in accord with our position. In Horn v. General Motors Corp., 17 Cal.3d 359, 131 Cal.Rptr. 78, 551 P.2d 398 (1976), the California Supreme Court rejected evidence of non-usage on the grounds that to allow it would introduce elements of contributory and comparative negligence that had no place in strict liability cases. The Court specifically rejected the defendant’s attempt to have the evidence admitted for the limited purpose of reducing damages only:
Defendant’s cognate claim that the proffered evidence was relevant to prove that some of plaintiff’s damages could have been avoided by reasonable action on her part is again merely another word formulation of the same inadmissible claim of contributory negligence.
Id. at 370, 131 Cal.Rptr. at 84, 551 P.2d at 404.
We believe the Pennsylvania courts would reject evidence concerning non-usage [768]*768of seat belts by plaintiff in a case tried under the principles of the Restatement (Second) of Torts § 402A, even for the limited purpose of reducing damages. Such evidence therefore was properly excluded below.
F.
Ford next contends that the district court erred in allowing plaintiff to introduce evidence of a 3% yearly growth in national productivity for the purpose of showing that her husband’s wages could have been expected to have increased by approximately 3% per year in the years after his death. Dr. F. G. Adams, an economist, testified that his studies led him to project a yearly growth in national productivity of 3%, and that he believed a worker in Vizzini’s position could be expected to have shared in that growth rate. Ford objected, arguing the testimony was improper because it was simply a method of accounting for inflation. The trial judge admitted the testimony, citing decisions of this court which the trial judge believed approved such evidence where a proper foundation had been laid. He believed such a foundation appeared in this case. 72 F.R.D. at 137-38.
We believe Pennsylvania law controls on this issue. The Supreme Court of Pennsylvania has not yet considered whether or in what manner inflation or productivity increases may be accounted for in damage awards. This is a difficult and changing area of the law, and the various state and federal courts are not agreed on the proper approach to such issues.8 But we believe that a recent decision of the Pennsylvania Superior Court, Havens v. Tonner, 243 Pa. Super. 371, 365 A.2d 1271 (1976), pet. for allocatur denied, No. 692 (Pa.Supreme Court, July 11, 1977), is evidence of the thinking of the Pennsylvania courts on this point.
In Havens, the Pennsylvania Superior Court held that evidence of increased productivity, was “simply a substitute for inflation and equally speculative and inadmissible in a calculation of future earnings,” id. at 378, 365 A.2d at 1274. The Court went on to say that
[a]ny estimate of future earnings over a substantial period of years based upon economic predictions is necessarily extremely speculative in nature. Much more satisfactory is evidence of the earning potential of the individual in question.
243 Pa.Super. at 380, 365 A.2d at 1275. Thus, evidence of increased productivity should not be admitted at a new trial.9
CONCLUSION