Albert Whatley brought this products liability action to trial for injuries arising from his exposure to the asbestos products of Raymark Industries' predecessor, Raybestos-Manhattan. The jury found in Whatley's favor but he contends the jury's findings of 90.01% contributory causation by ten settling defendants is without sufficient evidence. Raymark's cross-appeal challenges the award for future medical expenses. We affirm in part but modify the recovery against Raymark.

Albert Whatley was employed for over thirty years as a chipper at an Alabama shipyard. Whatley's work, smoothing the finish of iron surfaces onboard ships, often placed him in close unventilated spaces. Although he did not personally use asbestos-containing products, Whatley worked alongside other tradesman, including insulators, who used such products. Over the course of his employment Whatley was exposed to asbestos dust consequential to the use of asbestos-containing products. Whatley retired in 1982 and later developed lung cancer, which was diagnosed and operated on in 1986.

Whatley filed this product liability suit against Raymark Industries and thirteen other defendants, all of whom he alleged caused his lung ailments by the asbestos from their products. One defendant was severed after filing insolvency proceedings. Whatley reached a settlement with twelve of the defendants and proceeded to trial against Raymark alone. The jury returned a verdict in Whatley's favor. Raymark and ten of the settling defendants were found to have manufactured defectively designed asbestos products that were producing causes of Whatley's injuries. The jury assigned a percentage of contributory causation to each defendant, 9.09% to Raymark and nine of the settling defendants and 9.10% to another settling defendant. The damages award was correspondingly reduced by 90.01% to represent Raymark's 9.09% responsibility.

Whatley filed this appeal contending that there is insufficient evidence to support the percentage of causation allocated by the jury to the settling defendants. Raymark responds that there is sufficient evidence in support of the contributory causation by the settling defendants or, in the alternative, if the evidence is insufficient as to the settling defendants it is also insufficient as to Raymark. Raymark also presents a cross-appeal challenging the award of future medical expenses.

A. Comparative Causation of Settling Defendants

In Duncan v. Cessna Aircraft Co., 665 S.W.2d 414 (Tex.1984), the Texas Supreme Court established a system of comparative causation to be applied in strict products liability cases.1 Under this system the jury apportions responsibility among all--including non-settling defendants, settling defendants, and the plaintiff--whose acts or products combined to cause the plaintiff's injuries. The non-settling defendants' liability and the plaintiff's recovery are to be reduced by the total percentage share assigned to the settling defendants. Id. at 429. The non-settling defendant must establish the liability of settling defendants in order to benefit from Duncan 's system of comparative causation. Dartez v. Fibreboard Corp., 765 F.2d 456, 474 (5th Cir.1985); see Shipp v. General Motors Corp., 750 F.2d 418, 425 (5th Cir.1985) (Duncan placed the allocative burden of comparative causation on the defendant).

Whatley contends that Raymark failed to provide sufficient evidence of his exposure to the asbestos containing products of the settling defendants and is thus not entitled to a reduction in the damages award. The federal standard for review of sufficiency of the evidence applies in diversity cases. This court must review all the evidence in the light most favorable to the jury's verdict and must affirm the verdict unless the evidence points "so strongly and overwhelmingly in favor of one party that the Court believes that reasonable men could not arrive at a contrary [conclusion]." Boeing Co. v. Shipman, 411 F.2d 365, 374 (5th Cir.1969).

As an initial matter, Raymark argues that Whatley's "judicial admissions" provide more than sufficient proof of the liability of all ten settling defendants. Specifically, Raymark contends that statements made by Whatley in his complaint, pretrial order, answers to interrogatories as well as statements by counsel and witnesses, are judicial admissions and as such are binding upon Whatley. Whatley responds that allegations in his complaint claiming that the settling defendants made defective products that caused his harm constitute inconsistent pleadings and as such are not binding. Liberal pleadings and flexible joinder are favored as methods of promoting efficient and fair judicial administration. See, 5 C. Wright & A. Miller, Federal Practice and Procedure Secs. 1282, 1283 (1969); 7 C. Wright & A. Miller, M. Kane, Federal Practice and Procedure Secs. 1601, 1602, 1604 (1986 & Supp.1988). Although Whatley's pleadings are not technically inconsistent, we find the same policies to be applicable here. See Continental Ins. Co. of New York v. Sherman, 439 F.2d 1294, 1298 (5th Cir.1971).

Duncan specifically allows the non-settling defendant to reduce his liability to the plaintiff only upon providing evidence to the jury of the percentage of causation attributable to the settling defendants. Not only do the statements to which Raymark points provide little real evidence of the liability of the settling defendants, but they provide no evidence upon which a jury could determine the percentage or extent of liability as required by Duncan. To make Whatley's pleadings conclusive on this issue would render the Duncan requirement meaningless. We decline to adopt Raymark's view and look to the totality of the evidence in reaching our conclusions.2

The Texas Supreme Court in Duncan did not quantify the evidence required to support a finding of contributory causation. That decision and Fifth Circuit decisions in its wake, however, refer to the non-settling defendant's burden as that of establishing liability. See Duncan, 665 S.W.2d at 434; Dartez, 765 F.2d at 474. In light of this language and in the absence of language in support of a different burden under Duncan, we look to the burden of proof that has been imposed on litigants in other asbestos product cases to determine whether sufficient evidence of liability for contributory causation was provided in this case.

Issues of insufficiency of the evidence as to causation require consideration of the nature of the product and the extent of the plaintiff's exposure. Gideon v. Johns-Manville Sales Corp., 761 F.2d 1129, 1143 (5th Cir.1985). A survey of Fifth Circuit challenges to the sufficiency of the evidence in asbestos litigation reveals a varying requirement of proof. In Gideon, evidence of exposure, challenged as insufficient by a non-settling defendant, consistent of the plaintiff's testimony that he worked around the defendant's products for 25 years. There was no testimony or other evidence, however, on the plaintiff's contact with the products or any dust created from their use. Id. at 1144. The court held that the evidence provided insufficient proof of causation. In Martin v. American Petrofina, Inc., 779 F.2d 250, 251-52 (5th Cir.1985), the court held sufficient evidence was offered in support of the jury's finding that, "more probably than not," the plaintiff had been exposed to the defendant's products. The plaintiff could not remember working with the defendant's product in his job removing insulation. Evidence was provided, however, from a supply officer who testified that the defendant's products were used by the plaintiff's employer during the relevant years and from insulators who testified that they had used the defendant's products. The plaintiff in Dartez v. Fibreboard Corp., 765 F.2d 456, 470 (5th Cir.1985) was found to have offered insufficient evidence of his exposure and of the nature of the defendant's product. The identification of defendant's product was made by a co-worker and the evidence showed the product to have a low asbestos fiber emittance and the plaintiff's exposure to be short (16 to 18 weeks). In Migues v. Fibreboard Corp., 662 F.2d 1182, 1185 (5th Cir. Unit A Dec. 1981), uncontradicted testimony of former co-workers that decedent-plaintiff had worked with defendant's products, and that use of the products created asbestos dust which all the workers inhaled, provided sufficient evidence to support the jury's conclusion that plaintiff had been exposed to defendant's products. Similarly, although unable to prove with any specificity that he came into contact with the defendants' products, the plaintiff in Jackson v. Johns-Manville Sales Corp., 727 F.2d 506 (5th Cir.1984) provided sufficient circumstantial evidence of causation to support a jury verdict. The evidence showed that defendants' products were ordered for use on a number of ship hulls on which the plaintiff worked and co-workers testified that they worked with the products under very dusty circumstances. Id. at 523. Finally, in Halphen v. Johns-Manville Sales Corp., 737 F.2d 462 (5th Cir.1984), the court rejected an insufficiency of the evidence challenge to a jury's finding that the defendant's products played a substantial part in the plaintiff's illness and death. The plaintiff suffered from a rare cancer linked to asbestos exposure. The evidence showed that plaintiff's employer used the defendant's products during the relevant time period and that it was likely that the plaintiff had been exposed to those products. Id. at 466.

The evidence presented in support of Whatley's exposure to the products of the settling defendants is not dissimilar to that presented in other asbestos cases. The testimony of two insulators, Benny Gill and Donald Cook, who worked for Whatley's employer provided the bulk of the evidence in support of the settling defendants' liability. Both insulators were identified as co-workers by Whatley. The parties also stipulated what products were manufactured by nine of the defendants and the asbestos content of all but two of those products.

Benny Gill identified products made by nine of the settling defendants as products "used on all of the ships." Gill, who had worked as an insulator for Whatley's employer since 1965, gave summary descriptions of the appearance and use of the products. Donald Cook testified that he worked on the same jobs as Whatley for approximately six months out of each year during the 1950s, '60s and '70s. Cook provided lengthy testimony on the use of Raymark's product, asbestos cloth (manufactured by Raymark's predecessor, Raybestos Manhattan), the dusty conditions created by the product, and Whatley's exposure to those conditions. Cook also identified products of six of the settling defendants as products used by Whatley's co-workers.

Whatley's trial testimony and his answers to interrogatories show that he worked alongside tradesmen using asbestos-containing products including cloth, pipe covering, cement, and block. He gave extensive testimony on the use, nature, and prevalence of asbestos cloth on the ships, and said that asbestos dust was created by all forms of asbestos products as well as cloth.

The specific evidence offered in support of the liability of each settling defendant was as follows:

Owens-Corning Fiberglas, makers of Kaylo block and pipe covering. Gill testified that the product was used on the ships. He described the block as dusty and explained how workmen cut the block with a handsaw before installation. Cook testified that there was some though little use of the blocks when Whatley was around. The parties stipulated a product content of 15% amacite chrysotile asbestos fiber.

Fibreboard Corporation, makers of Pabco block and pipe covering. Gill described how the products were used on the ships and what the products looked like. Cook verified that the products were used on the ships. The parties stipulated a product content of 12% to 15% asbestos fiber.

Pittsburgh Corning Corporation, successor to Uranco Industries, maker of unibestos block and pipe covering. Gill recalled the use of unibestos but could not recall the manufacturer. He described the pipe covering as hard crusted and prone to giving off splinters after it was sawed. Cook testified that unibestos was used very little. He identified the packaging of Pittsburgh's unibestos product. The parties stipulated a product content of 6.5% amacite.

Eagle-Picher Industries, maker of installation cement. Gill testified that the product was used on the ships and gave a one sentence description of its use. Cook testified that very little of the product was used but that it was used around Whatley. The parties stipulated a product content of 8.8% and 12% chrysotile asbestos fibers.

Celotex Corporation, makers of Carytemp block and pipe covering. Gill provided no evidence on this product. Cook identified the product packaging and testified that the products, mostly pipe covering, were used on the ships. The parties stipulated a product content of 6% to 7% asbestos fiber.

H.K. Porter, maker of asbestos cloth. Gill testified that a "bunch" of asbestos cloth made by both H.K. Porter and Raybestos was used on the ships and that it was often hard to differentiate between the products of the two manufacturers. Cook testified that he did not see the name of any manufacturer other than Raybestos on asbestos cloth used in the shipyard. The parties' stipulation as to H.K. Porter did not include a fiber content.

GAF, successor to Ruberoid pipe covering, maker of calsilite. Gill testified that the product was used on the ships. He described the color and size of the pipe covering. Cook could not remember using the product. The parties stipulated the product content of 13.5% asbestos.

Combustion Engineering, maker of No. 12 block. Gill testified that the product was used on the ships. He described the color and heat capacity of the block and recognized the product container. Cook could not remember using the product. The parties stipulated a product content of 4.5% asbestos.

Owens-Illinois, Inc. No evidence--in the form of testimony, stipulation, or otherwise--on the products of this manufacturer was offered to the jury.

Armstrong, maker of Armatemp cement. Gill testified that the product was used on the ships. He described how the product was used without adding details of its nature or characteristics. Cook could not remember using the product. The parties' stipulation provided the asbestos content and fiber type as unknown.

The liability and contributory causation of settling defendants Owens-Corning Fiberglas, Fibreboard Corporation, Pittsburgh Corning Corporation, Eagle-Picher Industries, and Celotex Corporation is supported by sufficient evidence. Donald Cook recalls using, albeit to varying degrees, the products of all five of these defendants. Both Cook's and Whatley's testimony emphasized the relative frequency with which the two men worked together. Gill testified on the use at the shipyard of all but Celotex products. The asbestos content of each was also stipulated and presented to the jury. The proof provided against these five defendants is commensurate with that deemed sufficient in other cases. The jury's findings are affirmed as to these five defendants.

Nor can we say that there was insufficient evidence in support of the jury's finding of liability of and contributory causation by the settling defendants H.K. Porter, GAF, and Combustion Engineering. Donald Cook was unable to recall the use of these defendants' products at the shipyard. Benny Gill, however, attested to their use and the evidence established that Gill worked with Whatley. The jury was entitled to make the inference that Whatley was exposed to products used by Gill in his job as an insulator. The jury also had before it the stipulated asbestos content of products made by GAF and Combustion Engineering. Although no fiber content was stipulated for H.K. Porter's asbestos cloth, both Gill and Whatley provided lengthy testimony on the frequent use of asbestos cloth, the frequent practice of tearing the cloth, and the great amounts of dust it created when torn. In addition, Gill testified that a "bunch" of H.K. Porter's asbestos cloth was used on the ships. There was sufficient evidence to support the jury's findings that the products of these three defendants were producing causes of Whatley's injuries.

The evidence is insufficient, however, to establish the liability of the last two settling defendants, Owens-Illinois and Armstrong. Owens-Illinois was not mentioned at trial. No evidence was submitted to the jury in support of this defendant's liability. As was done with two of the settling defendants, Owens-Illinois should have been excluded from the jury's verdict form for lack of proof. The only evidence in support of Armstrong's liability was Gill's testimony in which he stated that Armatemp cement was used on the ships and described its use to provide a hard, smooth finish. The jury received no evidence of the fiber content of the cement, whether the cement was dusty, or whether other aspects of the product or its use made it likely that asbestos would be inhaled by those working on the ships. The jury's finding of contributory causation as to Owens-Illinois and Armstrong is without the support of sufficient evidence of exposure.

Raymark contends that a rejection of the sufficiency of the evidence against a settling defendant necessitates the same conclusion on the evidence presented against it. That does not follow. Whatley provided far more evidence toward Raymark's liability than was provided toward the liability of any other manufacturer. Whatley, Gill, and Cook testified on the use, frequency of use, characteristics, and fiber content of Raybestos asbestos cloth. The evidence is clearly sufficient to support the jury's finding of Raymark's liability.

2. Reallocation of Damages Awarded by the Jury

Whatley provided sufficient evidence upon which the jury could find that he sustained injuries and that Raymark was a cause of those injuries. Under Duncan, Raymark was then allowed to "reduce" its liability to Whatley by the percentage of causation allocated by the jury to the settling tortfeasors. 665 S.W.2d at 429-30; see Dartez, 765 F.2d at 474; Acord v. General Motors Corp., 669 S.W.2d 111, 117 (Tex.1984); O'Neal v. Sherck Equipment Co., 751 S.W.2d 559, 560 (Tex.App.--Texarkana 1988, no writ). The plaintiff's burden is not increased; the allocative burden is placed on the defendant. Shipp v. General Motors Corp., 750 F.2d 418, 425 (5th Cir.1985); see Moore v. Johns-Manville Sales Corp., 781 F.2d 1061, 1063 (5th Cir.1986).

Duncan established joint and several liability of each tortfeasor subject to the jury's assessment of comparative causation which may limit liability of any or all defendants. Duncan, 665 S.W.2d at 434; see Dartez, 765 F.2d at 474 (the two controlling principles of Duncan are contributory causation and joint and several liability). In order to reduce its liability, Raymark was required to provide proof of liability of the settling defendants. Raymark provided insufficient evidence to support a finding of contributory causation of settling defendants Owens-Illinois and Armstrong. Raymark's liability to Whatley and Whatley's recovery from Raymark, thus, should not be reduced by any allocation to those two companies. Raymark is responsible for 27.27% of the total award, which represents a reduction for the percentage of causation attributed to the eight remaining settling defendants.

The dissent insists that, despite Raymark's failure to sustain its burden of proving the causation of settling defendants, the court may not increase the plaintiff's recovery above 9.09% of the total damages without usurping the jury function. This view confuses defendant's liability to plaintiff with defendant's ability to reduce plaintiff's recovery because of plaintiff's settlement with other tortfeasors. Raymark has been found liable for plaintiff's injuries. Had there been other non-settling defendants, an allocation of causation attributed by the jury to the non-settling defendants would bear only upon contribution between them; each of them would be liable to the free-from-fault plaintiff for his full damages. In the first hypothetical put by the dissent, A (plaintiff) could indeed stand on B's (non-settling defendant) liability for the damages less the 60% attributed to D (settling defendant).3 This is what joint and several liability means. In the second hypothetical, again, the allocation between non-settling defendants would be immaterial to plaintiff; the defendants are jointly and severally liable for full damages. Despite the dissent's remonstration, the result reached today is faithful to the Texas law; the result does not reallocate causation but simply reinstates that part of Raymark's liability it failed to reduce.

B. Sufficiency of the Evidence for Award of Future Medical Damages

In its cross-appeal, Raymark claims that no evidence was presented to show that Whatley will need future medical care or that the damages awarded for such were reasonable or necessary. A plaintiff must show a reasonable probability that medical expenses will be incurred in the future. Keeler v. Richards Mfg. Co., 817 F.2d 1197, 1202 (5th Cir.1987). Under Texas state law, such expenses must be reasonable and necessary. Pan Am. Ins. Co. v. Hi-Plains Haulers, Inc., 163 Tex. 1, 350 S.W.2d 644, 646 (1961). Our inquiry is whether the evidence points so strongly and overwhelmingly in favor of Raymark's contention that a reasonable jury could not have reached an opposing verdict. Boeing, 411 F.2d at 374-75.

The district court granted Raymark's original motion for judgment notwithstanding the verdict and struck the jury's award for future medical expenses of $150,000. Following the entry of that judgment, Whatley responded with his own motion for judgment or new trial. Evidently reconsidering, the district court granted Whatley's motion and entered a reformed judgment of $120,000 for future medical expenses.

In support of his claim for future medical damages, Whatley offered the testimony of Dr. Gary Friedman, a specialist in pulmonary diseases. Dr. Friedman testified that 95 percent of all lung cancer patients die within five years of diagnosis. This testimony was uncontradicted. Although Whatley had recovered from his prior surgery, we cannot say the jury had no basis upon which to find a reasonable medical probability that Whatley would require future medical care for a recurrence of his lung cancer.

Raymark further challenges Dr. Friedman's testimony on the cost of any future medical care as insufficient to support an award of damages. When asked by Whatley's counsel about the expenses for terminal care of cancer victims, Dr. Friedman gave his opinion that between $80,000 and $120,000 would be required for care at a nationally known Houston hospital. The "necessary" prong of the Texas standard for medical expenses is satisfied by our determination that there was sufficient evidence of a reasonable medical probability that future care would be required. Challenging the "reasonableness" of this estimate of future medical expenses, however, Raymark contends that the prices charged by a renowned Houston hospital are not representative of the Alabama hospitals that Mr. Whatley will be more likely to attend.

Although Whatley may not have provided a wealth of evidence in support of an amount for his future medical expenses, he provided evidence sufficient for affirmance on appeal. See, e.g., Simeon v. T. Smith & Sons, Inc., 852 F.2d 1421, 1427 (5th Cir.1988) (per curiam) (testimony of doctor that plaintiff had a fifty-fifty chance of needing another operation and that cost would be about $10,000 even though he claimed not to be an expert on costs; testimony was sufficient to support remittitur to $10,000); City of Pasadena v. Freeman, 731 S.W.2d 590, 595-96 (Tex.Ct.App.--Houston [14th Dist.] 1987), aff'd, 744 S.W.2d 923 (Tex.1988) (award of future medical costs was supported by testimony of treating doctor who gave an educated guess on cost and duration of future treatment). The $120,000 award entered by the district court reflects the high side of Dr. Friedman's estimate. The district court did not abuse its discretion in its resolution of this issue. Keeler v. Richards Mfg. Co., 817 F.2d 1197, 1202 (5th Cir.1987). We affirm the award of $120,000 for future medical expenses.

We modify the judgment to award Whatley the recovery of 27.27% instead of 9.09% of his damages against Raymark. The cause is remanded for entry of the appropriate judgment.

With due respect to my colleagues, I dissent on one point in this case. In allocating the liability shares of two settling co-defendants back to Raymark, thus raising its proportionate liability to the plaintiff from 9% to about 27%, we have invaded the province of the jury. Duncan v. Cessna Aircraft Co., 665 S.W.2d 414 (Tex.1984), bold and far-reaching as it was, did not ordain this result, and I know of no other principle that justifies it.

Despite Whatley's best efforts, the jurors obviously disbelieved his contention that Raymark, the lone, non-settling defendant at trial, was somehow more liable than other asbestos manufacturers to whose products Whatley was exposed. As a result, the jury found Raymark equally responsible with ten other settling asbestos manufacturers for causing Whatley's lung cancer. It assessed each of them with approximately 9% fault, and judgment was rendered accordingly.

On appeal, we have correctly found insufficient evidence to hold that Owens-Illinois or Armstrong Products exposed Whatley to harmful asbestos fibers. Thus, as to nine manufacturers, including Raymark, there was sufficient evidence to connect their products to the plaintiff's injury. The question then becomes how to treat the jury verdict with two fewer manufacturers involved. Three possibilities exist. First, we could remand for a new trial. Second, we could assume that the jury, having clearly determined that Raymark was no more responsible proportionately for Whatley's cancer than any other manufacturer, should bear 1/9th rather than 1/11th of the liability, and we could award 11% of the verdict against Raymark. Third, purporting to apply those cases which have held, in the wake of Duncan, that a non-settling defendant bears the burden of establishing the percent causation of settling tortfeasors, we could reach the majority's result here.

As I see it, the majority's result is a fact-bound wolf masquerading in the sheep's clothing of the Duncan line of cases. I would demonstrate this by showing two closely related, plausible post-Duncan scenarios involving settling tortfeasors. First, suppose A sued defendants B, C and D and settled with D. According to Duncan, B and C are entitled to point the finger at D as the absent tortfeasor, and the jury is responsible for allocating the share of comparative causation among the parties. The maximum exposure of B and C, absent contributory negligence, will be reduced by the extent of D's percentage contribution to the tort. Suppose the jury finds B 10% liable, C 30% liable and D, the settler, 60% liable. Suppose further that on appeal, C is legally exonerated. The appellate court obviously cannot continue to hold B liable for 10% of the tort, but should it simply add C's 30% to the judgment previously entered against B? Such a result, analogous to that of the majority here, would surely not make sense, because we cannot speculate how a jury might have viewed B's negligence if compared only with that of the absent D rather than the personally participating defendant C. The allocation of causation among the three participating tortfeasors could only be redone before a new trier of fact.

Likewise, consider the possibility that in the present case, two defendants had failed to settle, and they both impleaded the eight other settling asbestos manufacturers. Suppose further that the jury, unlike this one, had awarded different shares of causation among the eleven potential defendants. If we had reversed the judgment because of insufficient evidence of liability as to two settling tortfeasors, would we have the license to apply the percentages applicable to the exonerated tortfeasors in some manner against the two non-settling appellants? I doubt it. What non-arbitrary formula could we employ? The relative shares of liability, when some potential tortfeasors have been excluded from the jury's consideration, might change dramatically. We would have to order a new trial.

This case differs from the first hypothetical, but not from the second, in only one respect: that cases following Duncan have held that the non-settling defendant has the burden of proof regarding the liability of settling tortfeasors. Shipp v. General Motors Corp., 750 F.2d 418, 425 (5th Cir.1985). That difference is not dispositive, however, because it is still the province of the trier of fact to determine the percent share of causation assigned among the defendants and the settling tortfeasors.

The Texas Supreme Court in Duncan characterized the allocation of comparative causation without qualification as a jury issue:

We hold that in multiple defendant cases in which grounds of recovery other than negligence are established, the non-settling defendants' liability and the plaintiff's recovery shall be reduced by the percent share of causation assigned to the settling tortfeasor by the trier of fact.

665 S.W.2d at 429 (emphasis added).

In this case, the entire calculus of comparative causation has been changed by eliminating two settling tortfeasors. Unpleasant as it might be for the parties involved, I must conclude that a retrial is the only way to revise the factual determination of Raymark's percent causation. This situation is no different than if some particularly offensive piece of evidence had entered the trial in regard to one tortfeasor, and we were required to order a new trial devoid of its infectious influence. The majority's usurpation of the fact-finding function of comparative causation becomes clear in light of the consequences of their re-formulated judgment. The jury found that Raymark was no more responsible for Whatley's lung cancer than any other asbestos manufacturer; the majority have held Raymark 27% responsible, as against the 73% residual causation by eight other tortfeasors.

The posture of this case is unusual, because there is one plaintiff, one non-settling defendant, and virtually identical issues of liability pertaining to the settling tortfeasors and Raymark. In almost any other postulated Duncan-type case, the majority could not justify the expedient of avoiding a new trial by creating their own allocation of comparative causation. That the case is unusual, however, does not prevent it from being incorrectly decided, as I believe it has been. I would reverse and remand for a new trial.