Appellants challenge the judgment for Dartez on his claim that he was injured by exposure to the asbestos-containing products they manufactured. We vacate the judgment and remand for a new trial against all appellants except Raymark Industries. The judgment against Raymark is reversed and that corporation is dismissed.
I
James Dartez began working as an insulator using asbestos-containing products in 1957. Except for the year 1959, he continued in that occupation until 1966. He then followed other pursuits until 1980 when he returned to the insulating trade. In 1982 Dartez filed the present action against fourteen corporations alleging that he had been injured by exposure to their asbestos products between 1957 and 1966. His complaint, which was based on the theories of negligence and product liability, did not allege any injury after his return to work in 1980. Prior to trial Dartez settled with three of the defendants. Two others, Johns-Manville Sales Corp. and Unarco Industries, Inc., were severed from this action when they commenced voluntary bankruptcy proceedings. One defendant was dismissed upon Dartez’s own motion.
After a three day trial, the jury returned a verdict in favor of three defendants, but found the remaining five corporations to be liable to Dartez under both theories in the amount of $200,000. The five defendants held liable were Celotex Corp., Fibreboard Corp., Owens-Illinois, Inc., Pittsburgh Corning Corp., and Raymark Industries, Inc. (formerly Raybestos-Manhattan). Defendants filed motions to amend and correct the judgment, for judgment notwithstanding the verdict, for a mistrial, and for a new trial, but all motions were denied.
II
On appeal the defendants challenge the district court’s decision to admit the following evidence: (1) the deposition of Dr. Kenneth Smith, a physician formerly employed by Johns-Manville; (2) the deposition of Edward Ames, a public relations director for Owens-Corning Fiberglas; (3) the minutes of a committee meeting of the Asbestos Textile Institute; (4) several medical articles discussing the dangers of asbestos; and (5) evidence pertaining to the relationship between asbestos exposure and the diseases of cancer and mesothelioma. They also argue that the record does not contain sufficient evidence to support the jury’s verdict.. In addition, they assert that they were denied a fair trial because of remarks made to defense counsel by the trial judge in the jury’s presence. Finally, they contend that the judge used the wrong formula to adjust the jury’s award to account for the payment Dartez had received from the settling defendants.
III. THE EVIDENTIARY RULINGS
A. Dr. Smith’s Deposition
Dr. Smith, who is now deceased, was employed by Johns-Manville between 1944 and 1966. During most of this period he was the only physician in the company’s full-time employ. He served as Medical Director for approximately thirteen years. In 1976 Smith was deposed in another asbestos case, De Rocco v. Forty-Eight Insu-lations, Inc., No. 7880 (Allegheny Cty. PA Ct.C.P.1974). He testified about the knowledge of the relationship between asbestos exposure and disease which he had obtained while working for Johns-Manville, the scien-*461tifie studies on the subject, and the extent to which this information was disseminated throughout the industry. Portions of this deposition were read into evidence at Dar-tez’s trial.
Defendants assert that Dr. Smith’s testimony is inadmissible because it is irrelevant, unduly prejudicial, and hearsay. Their assertion is not correct.
Under the Federal Rules, evidence is relevant if it has “any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.” Fed.R. Evid. 401. Defendants contend that Smith’s testimony is irrelevant because it relates only to Johns-Manville. Their contention reflects a misunderstanding of a critical issue in any product liability action: the state of the art pertaining to any possible risks associated with the product. Dartez was required to establish that the dangers of asbestos were reasonably foreseeable or scientifically discoverable at the time of his exposure before these defendants could be found liable. Borel v. Fibreboard Paper Products Corp., 493 F.2d 1076,1088 (5th Cir.1973), cert. denied, 419 U.S. 869, 95 S.Ct. 127, 42 L.Ed.2d 107 (1974).
Borel holds all manufacturers to the knowledge and skill of an expert. They are obliged to keep abreast of any scientific discoveries and are presumed to know the results of all such advances. Moreover, they each bear the duty to fully test their products to uncover all scientifically discoverable dangers before the products are sold. Id. at 1089-90. The actual knowledge of an individual manufacturer is not the issue. If the dangers of asbestos were known to Johns-Manville at the time of Dartez’s exposure, then the same risks were scientifically discoverable by other asbestos corporations. Therefore, the testimony of the medical director of the industry’s largest member is relevant to plaintiff’s attempt to meet the evidentiary burden defined by Borel.
Even relevant evidence may be excluded if its probative value is outweighed by the danger that it will unfairly prejudice or confuse the jury. Fed.R.Evid. 403. However, because Rule 403 permits the exclusion of probative evidence, it is an extraordinary remedy that must be used sparingly. United States v. Thevis, 665 F.2d 616, 633 (5th Cir.), cert. denied, 459 U.S. 825, 103 S.Ct. 57, 74 L.Ed.2d 61 (1982).
Smith’s testimony has a significant probative value in the jury’s assessment of the state of the art. Dartez used only those portions of the deposition that were relevant to this issue. Defendants assert that this testimony was prejudicial because it allowed the jury to find these defendants liable for the knowledge and conduct of Johns-Manville. But this argument misses the point of Borel —the knowledge of one manufacturer can be a proper basis for concluding that another manufacturer should have warned of a specific danger. Rule 403 is designed to exclude evidence that has an “undue tendency to suggest decision on an improper basis____” Fed.R. Evid. 403 advisory committee note. Smith’s testimony, while contrary to defendants’ interest, cannot be construed as unfairly prejudicial under this standard.
Finally, defendants object to the Smith deposition on the grounds that it is inadmissible hearsay. Plaintiff asserts that the deposition is admissible under’ Fed.R.Evid. 804(b)(1) which provides:
The following are not excluded by the hearsay rule if the declarant is unavailable as a witness:
Former Testimony
Testimony given as a witness at another hearing of the same or a different proceeding, or in a deposition taken in compliance with law in the course of the same or another proceeding, if the party against whom the testimony is now offered, or, in a civil action or proceeding, a predecessor in interest, had an opportunity and similar motive to develop the testimony by direct, cross, or redirect examination.
*462The application of the former testimony exception is problematical in this case. Although Celotex and Pittsburgh Corning were represented at Smith’s deposition, the other defendants were not. Therefore a decision to admit the deposition under this rule would require us to determine whether the asbestos companies present when Smith was deposed constitute “predecessors in interest” for all other asbestos producers, despite the absence of any direct legal connection between them.
The Sixth Circuit, when confronted with the identical question, answered in the affirmative and permitted the use of this same deposition against asbestos companies not represented in De Rocco. Clay v. Johns-Manville Sales Corp., 722 F.2d 1289, 1294-95 (6th Cir.1983). The court reasoned that the De Rocco defendants satisfied Rule 804(b)(l)’s predecessor in interest requirement because they had a full opportunity to cross-examine Smith and to register any objections to his testimony and because their motive to confront Smith was similar to that of the Clay defendants. Id. at 1295. See Lloyd v. American Export Lines, Inc., 580 F.2d 1179, 1185 (3d Cir.), cert. denied, 439 U.S. 969, 99 S.Ct. 461, 58 L.Ed.2d 428 (1978).
It is not necessary to decide in today’s case whether we agree with the Sixth Circuit’s broad definition of predecessor in interest. Even if the Smith deposition is not admissible under the former testimony exception, it is encompassed within the residual exception to the rule against hearsay, Fed.R.Evid. 804(b)(5) which admits:
A statement not specifically covered by any of the foregoing exceptions but having equivalent circumstantial guarantees of trustworthiness, if the court determines that (A) the statement is offered as evidence of a material fact; (B) the statement is more probative on the point for which it is offered than any other evidence which the proponent can procure through reasonable efforts; and (C) the general purposes of these rules and the interests of justice will best be served by admission of the statement into evidence. However, a statement may not be admitted under this exception unless the proponent of it makes known to the adverse party sufficiently in advance of the trial or hearing to provide the adverse party with a fair opportunity to prepare to meet it, his intention to offer the statement and the particulars of it, including the name and address of the declarant.
This provision was added because Congress recognized that the specifically defined exceptions would not encompass every situation wherein a particular piece of hearsay demonstrated such reliability and appropriateness that it should be considered by the finder of fact. Congress believed that the residual exception was necessary to avoid the distortion of the specific exceptions beyond the reasonable circumstances they were intended to include. This exception was designed to protect the integrity of the specifically enumerated exceptions by providing the courts with the flexibility necessary to address unanticipated situations and to facilitate the basic purpose of the Rules: ascertainment of the truth and fair adjudication of controversies. Sen. Comm, on the Judiciary, S.Rep. No. 1277, 93d Cong., 2d Sess. (1974), reprinted in 1974 U.S.Code Cong. & Ad.News 7051, 7065-66; 11 J. Moore, Moore’s Federal Practice, § 803(24)[7] (2d ed. 1982). While we recognize that the residual exception must be used sparingly, United States v. Thevis, 665 F.2d at 629, we conclude that its use is justified here.
Although we do not decide whether the De Rocco defendants are predecessors in interest to the present defendants for the purposes of Rule 804(b)(1), we can agree with the Sixth Circuit’s substantive conclusion that the De Rocco defendants had the same motive as the present defendants to subject Smith to searching cross-examination and to raise any appropriate objections. Cf. United States v. Feldman, 761 F.2d 380 (7th Cir.1985) (Wisdom, J.). If these defendants had been present at the deposition, it is unlikely that any additional material information would have been elic*463ited. Assuming these defendants had been able to elicit from Smith testimony that they had no actual or constructive knowledge of the risks associated with asbestos exposure, the jury would still be entitled to find them liable under Borel on the basis that Smith’s testimony established that these risks were scientifically discoverable. Given these considerations, we conclude that this sworn testimony taken at an adversary proceeding presents the “equivalent guarantees of trustworthiness” required by Rule 804(b)(5) when proffered to establish the state of the art.
Smith’s testimony has a unique probative value in regard to the state of the art because his employment as Medical Director for Johns-Manville placed him in the most logical position to observe the development of scientific knowledge about the dangers of asbestos and the diffusion of such knowledge throughout the industry. His testimony on the state of the art has the same relevance in every asbestos case, regardless of which corporations are the defendants. In view of the commonality of the state of the art issue, it is unlikely that these appellants could have strengthened their defense if they had been represented at the deposition. To limit the use of this evidence to those defendants who happened to have been sued by De Rocco would be an overly technical and arbitrary ruling, contrary to the spirit of the Federal Rules of Evidence which were designed to admit all trustworthy proof.
Plaintiff notified defendants of his intent to introduce testimony from this deposition on July 2, 1982, more than a year before trial. Accordingly, the notice requirement of Rule 804(b)(5) was satisfied and this evidence was properly admitted.
B. Edward Ames’s Deposition and Related Documents
1.
Ames served as public relations director for Owens-Corning Fiberglas during the 1930’s and 1940’s. The record does not indicate if he is still employed by that corporation. Like Smith, Ames was also deposed in an earlier asbestos suit. Ames testified about Owens-Corning’s knowledge of the dangers of asbestos exposure during the 1930’s and 1940’s. Dartez introduced portions of this deposition to establish the state of the art.
Defendants objected on the grounds that the deposition is irrelevant and inadmissible hearsay. They claim that Ames’s testimony is irrelevant because he did not testify as to the knowledge of the present defendants and because most of his testimony was related to Owens-Corning’s attempts to establish the relative safety of fiberglas as compared to asbestos. However, as discussed above, the state of the art is defined in terms of what the industry as a whole knew or could have discovered by properly fulfilling their duty to test these products. Hence the fact that Ames’s testimony describes only the knowledge of a nonparty does not affect its relevance on this material point. For the same reason, the fact that he discussed Owens-Corning’s awareness of these dangers only in relationship to the corporation’s efforts to prove that working with fiberglas was safer than working with asbestos does not make it irrelevant.
Although this evidence is relevant, it should have been excluded as hearsay. The former testimony and residual exceptions to the rule against hearsay, Rule 804(b)(1) and (b)(5), both require that the declarant be unavailable at trial. Fed.R. Evid. 804(a). Plaintiff made no attempt to establish that Ames was unavailable. Admission of his former testimony without such a showing was erroneous. Perricone v. Kansas City Southern Railway Co., 630 F.2d 317, 321 (5th Cir.1980).
Plaintiff asserts that Ames’s testimony is admissible under Rule 801(d)(2) which specifies that a statement by the agent of a party opponent concerning a matter within the scope of his employment is not hearsay if offered against that opponent. However, Ames was not employed by any of the defendants at trial, but by Owens-Corning Fiberglas. Owens-Corning was dismissed *464from this action two months before trial after entering into a settlement agreement with Dartez. Hence Rule 801(d)(2) is inap-posite because Ames’s employer was not party to the trial. Even if Owens-Corning had remained a party, this rule would not have authorized the use of Ames’s testimony against the other defendants.
Defendants also contend that the portions of Ames’s testimony in which he refers to certain bibliographies of articles related to the dangers of asbestos should have been excluded because the best evidence rule, Fed.R.Evid. 1002, requires the introduction of the bibliographies themselves. This objection was not raised at trial and we will not address it here for the first time. If the evidence is introduced on retrial, this argument may then be presented to the district court.
2.
In conjunction with the Ames deposition, plaintiff introduced a series of memoranda exchanged between Ames, the president of Owens-Corning, and the corporation’s attorney, along with copies of correspondence between Ames and a Dr. Can-field. All the documents are dated between 1940 and 1945. The memoranda discussed a contemplated plan to collect information regarding the health risks connected to asbestos exposure and circulate it to the insulators’ union to counteract the asbestos producers’ efforts to encourage the unions to charge a premium for work on projects using Owens-Corning’s fiberglas products instead of asbestos. Ames’s correspondence with Canfield was intended to gather information as to whether a patient of the doctor had died of exposure to fiberglas or to asbestos.
Defendants contend that these documents are inadmissible because they are irrelevant, hearsay, and not authenticated. The evidence is relevant to the state of the art issue for the same reasons that the Ames and Smith depositions are relevant. See Jackson v. Johns-Manville Sales Corp., 750 F.2d 1314, 1317-19 (5th Cir.1985) (en banc).
Although these documents are hearsay, they could be admitted under the exception provided by Fed.R.Evid. 803(16) for “[statements in a document in existence twenty years or more the authenticity of which is established.” The authentication requirement is governed by the standards set forth in Fed.R.Evid. 901(a): “The requirement of authentication or identification as a condition precedent to admissibility is satisfied by evidence sufficient to support a finding that the matter in question is what its proponent claims.” See United States v. Koziy, 728 F.2d 1314, 1321-22 (11th Cir.), cert. denied, — U.S. -, 105 S.Ct. 130, 83 L.Ed.2d 70 (1984). Ancient documents are most frequently authenticated under the provisions of Rule 901(b)(8) which states that authenticity can be established by
[ejvidence that a document or data compilation, in any form, (A) is in such condition as to create no suspicion concerning its authenticity, (B) was in a place where it, if authentic, would likely be, and (C) has been in existence 20 years or more at the time it is offered.
11 J. Moore, Moore’s Federal Practice § 803(16)[4] (2d ed. 1981).
Defendants have not challenged the dates on these documents which indicate the papers were about 40 years old at the time of trial. Nor have they raised any arguments concerning their condition. The record shows, however, that plaintiff failed to introduce evidence to establish that the papers were found in the expected place— here the corporate records of Owens-Corning. The Ames deposition might have authenticated the memoranda, but it was improperly admitted itself.
. At trial, plaintiff contended that the Owens-Corning documents were authenticated by a letter from the attorney representing Owens-Corning in this litigation. This letter, which was introduced into evidence for this one purpose, was a response to plaintiff’s request that certain documents be produced. Production obviously was not the purpose of the request because copies of the documents were attached to the let*465ter. Counsel for Owens-Corning treated it as a request for authentication. In his response Owens-Corning’s counsel authenticated several documents and declined to authenticate others. We need not decide whether such evidence would be sufficient to establish authenticity as a general rule. The letter in evidence refers to documents only by number, and the record does not provide the means to determine which number refers to which document. Therefore, the letter does not establish the authenticity of any particular document.
C. The Minutes of Committee Meetings of the Asbestos Textile Institute
The Asbestos Textile Institute was a trade organization for the manufacturers of asbestos textile products. Dartez introduced into evidence the minutes of several of the Institute’s committee meetings held between 1954 and 1956. These documents tend to show that the Institute’s members were aware of the risks associated with asbestos products and that they were concerned about possible unfavorable publicity if these risks became widely known.
Fibreboard and Owens-Illinois objected to this evidence on the grounds of irrelevancy because there was no showing that either defendant was a member of the Institute or had any knowledge of its proceedings. As discussed above, these factors do not affect the relevance of this evidence to the state of the art issue.
All defendants argue that the exhibits are excludable hearsay and were not authenticated. Because these documents appear to be more than 20 years old, they could be admitted under the ancient document exception to the rule against hearsay, Fed.R.Evid. 803(16), as set forth in the preceding section of this opinion. However, plaintiff’s introduction of these documents suffers from the same fatal defect as his introduction of the documents associated with the Ames deposition: he has not satisfied the authentication requirements of Rule 901. There is nothing in the record to indicate the source of the minutes.
Defendants also contend that these documents are too remote in time to be probative of any point. However, these meetings occurred only a few years before Dartez began working as an insulator in 1957. Moreover, under Jackson v. Johns-Manville this remoteness would not affect their relevance to the state of the art issue. 750 F.2d at 1318-19.
D. Medical Articles
To establish the relevant medical knowledge plaintiff introduced several articles from various publications. Defendants objected to this procedure on the grounds that the articles were inadmissible hearsay because they were not introduced in compliance with the restrictions Fed.R. Evid. 803(18) places on the use of such material. Rule 803(18) provides a limited exception to the rule against hearsay for learned treatises such as these articles
[t]o the extent called to the attention of an expert witness upon cross-examination or relied upon by him in direct examination, statements contained in published treatises, periodicals, or pamphlets on a subject of history, medicine, or other science or art, established as a reliable authority by the testimony or admission of the witness or by other expert testimony or by judicial notice. If admitted, the statements may be read into evidence but may not be received as exhibits.
Plaintiff’s medical expert did not testify about the disputed articles. Moreover, these articles were admitted as exhibits and apparently given to the jury. Such action was improper. See Johnson v. William C. Ellis & Sons Iron Works, 609 F.2d 820, 822-23 (5th Cir.1980). The reason for the Rule’s restrictions on the use of learned treatises is to avoid the possibility that the jury will misunderstand and misapply the technical language within such an article if they are allowed to consider the publication itself instead of receiving the information through the testimony of an expert in the field. Fed.R.Evid. 803(18) advisory committee note.
*466E. Evidence Pertaining to Cancer and Mesothelioma
Dartez does not claim to be suffering from cancer or mesothelioma. At trial he introduced testimony that his exposure to asbestos had increased his risk of developing those diseases in the future.1 For our purposes, this evidence will be divided into four categories: (1) Dartez’s own testimony as to his fear of these diseases; (2) testimony by Dr. Comstock, Dartez’s physician and expert witness, as to what he told Dartez about his chances of developing either disease and the consequences; (3) Comstock’s testimony about the costs of treating cancer and mesothelioma; and (4) Comstock’s expert testimony on the causative relationship between these diseases and asbestos exposure. Defendants objected to this testimony because it did not establish a reasonable medical probability that Dartez would develop either disease and because it was unduly prejudicial. All objections were overruled.
On appeal defendants reassert their same objections. Dartez maintains that the evidence is admissible because the testimony of his medical expert, Dr. Comstock, satisfies the standard of reasonable medical probability. He also asserts that the testimony is necessary to prove his claim for mental anguish stemming from his increased risk of developing these diseases. Finally, he maintains that the evidence is admissible under Borel v. Fibreboard Paper Products Corp., 493 F.2d 1076 (5th Cir.1973), cert. denied, 419 U.S. 869, 95 S.Ct. 127, 42 L.Ed.2d 107 (1974) to establish the extent of the defendants’ duty to warn.
1.
We recently addressed the admissibility of evidence related to cancer and mesothelioma in Gideon v. Johns-Manville Sales Corp., 761 F.2d 1129 (5th Cir.1985), another Texas asbestos suit wherein the plaintiff conceded he was not suffering from either disease, but alleged that his chances of developing these illnesses had been increased greatly by his exposure to defendants’ products. In Gideon we ruled that Texas would permit a plaintiff to recover damages for a disease that may develop in future years only if he introduces expert testimony establishing that there is a reasonable medical probability that the disease will appear. “Possibility alone cannot serve as the basis for recovery, for mere possibility does not meet the preponderance of the evidence standard. Certainty, however, is not required: the plaintiff need demonstrate only that the event is more likely to occur than not.” Id. at 1137 (footnote omitted). If this standard is met, plaintiff is entitled to recover all damages associated with the disease at issue, including compensation for mental anguish.
The testimony adduced by Dartez did not rise to the level of a reasonable medical probability. The Comstock testimony was based upon studies of the proportional mortality rate — the percentage of insulators diagnosed as having asbestosis who die of each particular disease — and Comstock’s personal observations at five post-mortems on individuals who had been exposed to asbestos. In relevant part he stated:
you are looking at insulators who have been diagnosed as having asbestosis. The proportion of those insulators who die of lung cancer, depending on the individual study and there is some variation, will range from around thirty-five to as high as fifty or sixty percent. Looking at the same population, the risk of the proportion of the deaths that would be due to mesothelioma will range from as low as about seven percent to a high also approaching forty percent.
On cross-examination he characterized these figures as the “extremes” in the range of possibilities. He also stated that the risk of lung cancer for an insulator exposed to asbestos who was a moderate smoker was only “approaching fifty percent.” Dartez testified that he smokes one *467pack a day. Finally, Comstock testified that if Dartez followed instructions and quit smoking, his current risk of cancer would be reduced by approximately one-third.
This proof does not establish a reasonable medical probability that Dartez will develop either cancer or mesothelioma because of his exposure to defendants’ products. Accordingly, under Texas law he is not entitled to recover damages for any increased risk of those diseases. For the reasons set forth in parts 2 and 3 of this section, most of the disputed testimony is admissible to establish plaintiff's mental anguish and defendants’ liability. However, the testimony on the expense of treating cancer and mesothelioma should have been excluded as irrelevant because Dartez is not entitled to recover such costs.
2.
Gideon specifically pretermitted the question of whether Texas law precluded recovery for mental anguish resulting from fear of developing a disease where there is no showing that the disease is medically probable. Id. at 1138. The Texas courts have not directly addressed this question in the context of latent diseases allegedly caused by defective or unreasonably dangerous products. From the Texas decisions described below, we conclude that State’s courts would permit recovery for mental anguish which arises from exposure to a latent disease even though that disease is not medically probable.
Texas has long recognized plaintiff’s right to compensation where such suffering is the natural result of the injury at issue. Gulf, C. and S.H. Railway v. Dickens, 118 S.W. 612, 616 (Tex.Civ.App. 1909, no writ). See Sutton Motor Co. v. Crysel, 289 S.W.2d 631, 634 (Tex.Civ.App.1956, no writ). In Kimbell v. Noel, 228 S.W.2d 980 (Tex.Civ.App.1950, ref. n.r.e.), plaintiff’s physician testified that he had diagnosed her as having a precancerous condition caused by an injury to her breast during an automobile accident. Despite the fact that the trial court sustained defendants’ objection after the testimony was given, defendants asserted on appeal that reversal was required. The appellate court found that the trial court had cured any possible defect by sustaining defendants’ objection, but also went on to determine that the testimony was admissible, in part because it was adduced to establish the mental anguish element of plaintiff’s case. Id. at 982-83.
Similarly, in Dulaney Investment Co. v. Wood, 142 S.W.2d 379 (Tex.Civ.App.1940, writ dism’d judgmnt cor.) the court upheld a decision permitting a plaintiff who had injured his arm in an accident in defendant’s elevator to testify about his fear of paralysis, despite defendant’s objection that the testimony was conclusory, inflammatory, and prejudicial.
[I]t appears to us that where the plaintiff has testified to a physical condition tending to show a partial loss of use of one arm and one leg, since the accident, it was not error to permit him to testify to his fear of paralysis. If that is not a part of his “mental anguish” occasioned by the injury received, then we are at a loss to understand what would constitute mental anguish.
Id. at 384. Plaintiffs have been permitted to present evidence of fear of hydrophobia, lockjaw, and blood poisoning after being bitten by a dog, Gamer v. Winchester, 110 S.W.2d 1190, 1193 (Tex.Civ.App.1937, writ dism’d); Trinity & Sabine Railway Co. v. O’Brien, 18 Tex.Civ.App. 690, 46 S.W. 389, 391 (1898, no writ) and of fear of blood poisoning after a foot was mangled in a railroad accident. Southern Kansas Railway Co. v. McSwain, 55 Tex.Civ.App. 317, 118 S.W. 874, 875 (Tex.Civ.App.1909, no writ). In one dog-bite case the court noted that the possible diseases
are matters that would quite naturally disturb a person, living in a day in which we are all made germ-conscious by scientists and physicians. If actual bodily injury can be shown by appellee, as a result of the attack, these mental fears may be considered by the jury in arriving at the issue of appellee’s mental anguish.
Gamer, 110 S.W.2d at 1193.
With the exception of Kimbell v. Noel, none of these cases even discuss the con*468cept of medical probability. In Kimbell the issue of probability formed the basis of defense counsel’s objection at trial, but it was not expressed as a factor in the appellate court’s decision.
Here Dartez has been injured by the accumulation of asbestos fibers in his lungs. Gideon, 761F.2d atll37. According to Comstock, medical studies have established a connection between asbestos exposure and both mesothelioma and cancer. People today are sensitive to the dangers of cancer, just as they were “germ-conscious” when Gamer was decided. Mental anguish would reasonably follow after Dartez was informed by a reliable source — his physician — that his exposure to defendants’ products had heightened his risk of developing these deadly diseases. Under Texas law he is entitled to compensation for mental anguish proximately caused by his asbestos exposure, even if such distress arises from fear of diseases that are a substantial concern, but not medically probable. The fact that the evidence does not establish a reasonable medical probability that Dartez will develop either disease is relevant to the jury’s determination of the reasonableness of the anguish he feels and the amount of compensation to which he is entitled. However, it does not suffice to bar the proof for purposes of establishing mental anguish.
Dartez’s own testimony regarding his emotional reaction to the realization of the risks he now faces and Comstock’s testimony as to what he told Dartez about these risks were properly admitted to establish the mental anguish element of plaintiff’s case.
3.
The final type of cancer and mesothelioma evidence — Comstock’s expert testimony on the causative relationship between asbestos exposure and those diseases — is relevant to the issue of liability. Evaluation of the scope of defendant’s duty to warn requires full disclosure of risks. Borel established that a manufacturer is obliged to warn of all scientifically discoverable dangers associated with its product so that the user may decide whether to expose himself to the risks involved. 493 F.2d at 1089-1090. See Jackson v. Johns-Manville Sales Corp., 750 F.2d 1314, 1320 (5th Cir.1985) (en banc). In Borel we criticized the warnings then utilized by three asbestos companies because the cautions did not “intimate[ ] the gravity of the risk: the danger of a fatal illness caused by asbestosis and mesothelioma or other cancers.” 493 F.2d at 1104 (on petition for rehearing). Evidence that asbestos may be a carcinogen is relevant to defining the scope of defendant’s duty to warn and the extent of their breach of that duty. Jackson, 750 F.2d at 1320.
Our conclusion is not affected by the fact that Dartez has not developed cancer or mesothelioma and, according to the evidence, does not face a medical probability that he will do so. Whether a specific disease has been diagnosed in an individual plaintiff does not determine the scope of defendants’ duty to warn. What is significant is whether warning of the nondis-closed risks could have averted plaintiff’s injury, or afforded him the opportunity to make a knowing choice. Here the proof did tend to establish that cancer, mesotheli-oma, and asbestosis stem from the same source — the inhalation of the asbestos fibers. Had defendants warned of these risks, plaintiff could have quit his job or worked with a respirator to avoid the damage, or he could have made a conscious decision to accept the health risks involved.
4.
In Jackson v. Johns-Manville we recognized that in some asbestos cases the quantum of evidence adduced about cancer and mesothelioma may reach a level where further proof should be excluded as unduly prejudicial under Fed.R.Evid. 403. 750 F.2d at 1320-21. In the case before us, however, the evidence related to these two diseases was not so overwhelming as to cross the line recognized in Jackson. See Gideon, 761 F.2d at 1137.
F. The Effect of The Errors
Erroneous evidentiary rulings by a trial judge can be treated as harmless only if *469the error “does not affect the substantial rights of the parties.” Fed.R.Civ.P. 61. The errors identified above did affect substantial rights of the defendants. The Ames deposition and the accompanying documents, the minutes from the Asbestos Textile Institute, and the scientific articles could easily have tipped the balance in favor of finding liability. Because we cannot say with conviction that this evidence did not affect the jury’s determination, the result of this trial cannot stand. Brown v. Miller, 631 F.2d 408, 412-13 (5th Cir.1980).
An additional substantial problem exists. When Comstock’s inadmissible testimony on the costs of treating cancer or mesothe-lioma is considered in conjunction with the district court’s instructions on how to calculate damages; it is clear that the jury was allowed to include in their award future damages associated with the actual occurrence of such diseases, including medical expenses and lost income. Dartez is not legally entitled to recover for these damages.
For these reasons, we vacate the judgment and remand the cause for a new trial.
IV. SUFFICIENCY OF THE EVIDENCE
The jury determined that defendants were liable under both negligence and product liability theories. Defendants contend that the evidence does not support either finding. Because Raymark’s arguments are significantly different from those of the other defendants, Raymark’s liability will be discussed separately in Part B of this section.
A.
Borel v. Fibreboard Paper Products, 493 F.2d 1076 (5th Cir.1973), cert. denied, 419 U.S. 869, 95 S.Ct. 127, 42 L.Ed.2d 107 (1974) defined the evidentiary burdens in proof of strict liability against a manufacturer. To establish that a product is unreasonably dangerous plaintiff must prove that the magnitude of the danger it presents outweighs its utility. Id. at 1087. Even where this balance favors the product’s utility, adequate warnings must be given or the product still will be classified as unreasonably dangerous. Id. at 1089. Such warnings must advise the product’s users of any risk associated with the product that is “scientifically discoverable”, id. at 1088, by one with the “knowledge and skill of an expert.” Id. at 1089. See Gideon, 761 F.2d at 1143. An affirmative duty, commensurate with the potential dangers, is placed on the manufacturer to test its products. Borel, 493 F.2d at 1089-90.
Failure to warn was one of the theories of liability submitted to the jury in the present case. Plaintiff introduced evidence that asbestos was dangerous and that defendants failed to warn of these dangers despite the fact that the risks had been identified by experts many years before Dartez’s exposure occurred. Defendants countered with evidence that the danger to insulators was not discoverable until the late 1960’s, after Dartez had ceased working as an insulator. The parties also introduced conflicting evidence as to whether the level of exposure experienced by Dar-tez was sufficient to be dangerous and whether Dartez actually had sustained any damage to his lungs.
Weighing conflicting evidence and inferences and determining the relative credibility of witnesses in order to resolve a factual dispute is the province of the jury. Boeing Co. v. Shipman, 411 F.2d 365, 375 (5th Cir.1969) (en banc). Their decision must be accepted if the record contains any “competent and substantial evidence fairly tending to support the verdict ... even if different inferences and conclusions also might be supported by the evidence.” Stewart v. Thigpen, 730 F.2d 1002, 1007 (5th Cir.1984).
Their finding that asbestos products are dangerous is supported by the testimony of Drs. Comstock and Smith. Smith’s testimony, along with many letters and other documents introduced by plaintiff, supports the conclusion that the asbestos companies knew of the risks associated with their products during the time of Dar-*470tez’s exposure, or at least that they would have discovered these dangers if they had fulfilled their Borel obligation to test their products. 493 F.2d at 1089-90.
The evidence was adequate to support a verdict of liability against defendants Celo-tex, Fibreboard, Owens-Illinois, and Pittsburgh Corning.
B.
Raymark’s arguments pertaining to the sufficiency of evidence must be addressed separately for two reasons: (1) the only Raymark product at issue, asbestos cloth, is substantively distinguishable from the asbestos insulation products of the remaining defendants and (2) Dartez’s exposure to the Raymark cloth was quantitatively different from his exposure to the other products.
According to Dartez’s testimony he used asbestos cloth on two jobs, one in 1962 with Bill Howard that lasted for a week and a half and another in the early 1960’s at the Monsanto Chocolate Bayou plant during which he used asbestos cloth on four different occasions with a total exposure of sixteen to eighteen weeks. The brand of cloth used at the Howard job was never identified. Dartez testified that he used two brands of cloth at the Chocolate Bayou project. He identified one as a Johns-Man-ville product, but did not know which company had manufactured the other brand. One of Dartez’s co-workers on that job, Thomas Hartwell, testified that the second brand was made by Raymark. Dartez testified that he wrapped asbestos cloth around fittings and joints, but never stated that he cut asbestos cloth or that his use of this cloth produced dust. The absence of testimony regarding dust produced from asbestos cloth contrasts sharply with his testimony regarding the large amounts of dust generated by working with the other asbestos products. Hartwell testified that the cloth produced dust when it was cut, but that it was not as dusty as the pipe covering.
The asbestos cloth does not emit as many fibers as the other insulation products because the textile manufacturing process minimizes the potential for fiber release.
We described that process in Gideon:
asbestos fibers are combined with cotton and rayon and entered into a carding machine which rolls and combs the fibers. The fibers form a web which is then rubbed by a machine to form asbestos roving. The roving is twisted on a spinning machine to form asbestos yarn. The spinning process locks in the fibers so that either no fibers or minimum quantities of fibers are released during use or handling of the product. The yarns are then woven into the end products. This weaving process further locks in the fibers.
Gideon, 761 F.2d at 1144. This process was described to the Dartez jury in the deposition testimony of Milton Scowcroft, a marketing consultant for Raymark.
Scowcroft also described a series of tests performed in 1975 and 1976 by North Brothers, an organization hired by Ray-mark. These studies analyzed the number of fibers emitted during various uses of Raymark’s asbestos cloth, including cutting the cloth and wrapping it around pipe joints. The highest fiber count in any sample was .35 fibers per cubic centimeter. At that time, the Occupational Safety and Health Administration (OSHA) set the exposure limit at 2 fibers per cubic centimeter. In 1983 this standard was lowered to .5 fibers per cubic centimeter. Gideon, 761 F.2d at 1144. The fiber count for Ray-mark’s product in the North Brothers’ study is below any threshold limit defined by OSHA. OSHA’s regulations regarding the need for warning labels on asbestos products exempt those products which do not emit fibers in excess of OSHA’s exposure limits.2
*471“Compliance with such government safety standards constitutes strong and substantial evidence that a product is not defective.” Gideon, 761 F.2d at 1144. Dartez did not introduce any testimony to counter this evidence, nor did he seek to demonstrate that exposure to levels of fibers below OSH A standards is dangerous. Even when we construe the evidence in the manner most favorable to' Dartez, as we are obliged to do, the record does not contain sufficient evidence to support a finding that Ray-mark’s products were unreasonably dangerous or that Raymark failed to warn of any scientifically discoverable risk associated with its cloth. Moreover the record does not contain any evidence that Dartez’s limited exposure — eighteen weeks out of a ten-year career — to the relatively small number of fibers emitted by this cloth under the most unfavorable conditions could be a producing cause of his injuries. See Gideon, 761 F.2d at 1145. The judgment against Raymark cannot stand.
C.
The record contains sufficient evidence to support the verdict against Celotex, Fi-breboard, Owens-Illinois, and Pittsburgh Corning, and these corporations must remain as defendants in the new trial. The judgment against Raymark is reversed and Raymark is dismissed from the action.
V. COMMENTS BY THE TRIAL JUDGE
A United States District Judge is far from a figure-head or a passive arbitrator. He is obliged to act when necessary to ensure that the trial is properly conducted and not subject to delay. Fed.R.Evid. 611(a). Cranberg v. Consumers Union, 756 F.2d 382, 391 (5th Cir.1985). However, he must always be careful to maintain both his objectivity and the appearance of neutrality when he does intervene in a proceeding. Dixon v. International Harvester Co., 754 F.2d 573, 585 (5th Cir.1985). Because of his position the trial judge necessarily exercises a great deal of influence over the jury. They may consider his lightest word or intimation to be controlling. Newman v. A.E. Staley Manufacturing Co., 648 F.2d 330, 334-35 (5th Cir.1981).
Defendants contend that the remarks of the trial judge here were so biased against them as to affect the basic fairness of the trial by leading the jury to conclude that the judge favored plaintiff’s position. Id. at 336. To support this assertion, they stress the following excerpts from the trial record:
1) During the reading of Dr. Smith’s deposition, counsel for Fibreboard had the following exchanges with the Court: With regard to the subject of Dr. Smith’s testimony:
MR. WELLER: Your Honor, may I have the objection that the doctor is describing employees in mining operations and not insulators?
THE COURT: Well, he is trying to prove that asbestos is dangerous. And if the defendants will admit that, well we can do away with this type of testimony. MR. WELLER: I object to the court’s remark before the jury and ask that the court instruct the jury not to consider the remarks.
THE COURT:' The court has tried many of these cases, and counsel knows and counsel has tried many of them as the court knows and this is the issue of the background of the state of the art in regard to asbestosis. [If] counsel is willing to concede that it’s dangerous, well then we can do away with this testimony.
MR. WELLER: Your Honor, I would ask the court to withdraw the jury in order that I might make a motion at this time.
THE COURT: I understand what you are saying and the court will instruct *472the jury. But I just want to try to silence you, if I possibly can, from making foolish objections. And you can object to that too if you wish.
:k * $ * * *
THE COURT: Why do you continue to stand?
MR. WELLER: Your Honor, I was listening to the court and I still—
THE COURT: You cannot listen sitting down? Just be seated.
While Dr. Smith’s deposition was being read, the following exchange occurred after the reader mistakenly read “this” for “his”:
MR. WELLER: [X-ray films were] taken “in this town” and not “his town”. I would ask the witness be required to read the script.
THE COURT: Your objection [is] to a mistake that he made in the reading[?] MR. WELLER: He said “his town” and it was “this town”.
MR. WELLER: I would like for him to stay in the script.
THE COURT: How absurd can you be, Mr. Weller? This has reached a real high for the absurdity which you seem to take pleasure in. Objection will be overruled.
2) Counsel for Owens-Illinois was involved in this exchange with regard to the subject of Dr. Smith’s testimony:
MR. STAMM: Excuse me, counsel. I have a new objection that I would pose to this. And that is that it is immaterial and irrelevant to any issue in this case what JM [Johns-Manville] knew or should have known. The issues in this case is what these defendants knew or should have known. Unless there is some showing in that connection I object as to that it is irrelevant.
THE COURT: Make your objection and quit arguing.
MR. STAMM: I object as to immateriality.
THE COURT: Objection is overruled. That sounds very elementary. It sounds like an objection that a first year student out of law school would make. For a man of your appearance, I am surprised that you would make such an objection.
MR. STAMM: Your Honor, may I have, when the jury is out, the same objection that Mr. Weller has that I would like to make?
THE COURT: You can make all of the objections you want to.
¿if * * * ¡fc
THE COURT: From your appearance, you look to me like a man about fifty years old and that you would make such an objection on the basis of materiality when the court has already stated the purpose for this offer of the deposition and the many objections that have already been made. All right. Well if you get any pleasure out of it, make another one.
3) During the cross-examination of the plaintiff by counsel for Fibreboard the following occurred:
MR. WELLER: Did you also tell him you were drinking six beers a day?
PLAINTIFF: Yes, I probably told him that too.
THE COURT: What has beer have to do with asbestosis?
MR. WELLER: Your Honor, we are trying a health problem here as far as this witness is concerned.
THE COURT: Is beer bad for your health[?] If it is, do you drink it?
MR. WELLER: I drink it on very rare occasions, your Honor.
THE COURT: For your health purposes.
MR. WELLER: No. My doctor prohibits me from drinking.
THE COURT: All right. Any reference to beer will be stricken from the record.
MR. WELLER: If the court please I would like to develop my testimony from the witness.
THE COURT: The court will not permit you to develop that sort of absurd testimony. We are talking about asbesto*473sis. Maybe cigarettes might have something to do with it but not beer. You couldn’t have asked that question except to be unfair to the witness.
******
THE COURT: In some way to embarrass him. But you know, that is not— MR. WELLER: I again object to the remarks of the court.
THE COURT: I don’t care how much you object to it. You are going to be quiet until the court finishes talking.
MR. WELLER: All right.
THE COURT: The court feels that the only reason you ask questions of that type is in some way to — because you are representing a client that this man has sued, to some way embarrass him. Because it doesn’t have any bearing whatever on the subject matter.
All right. Now you may do all of the objecting you want to.
4) While cross examining the witness Hartwell to elicit whether proper identification could be made of the manufacturer of the specific product involved, Ray-mark’s counsel asked the following questions:
MS. THOMPSON: How can you tell “A” cloth, who manufactured a particular piece of “A” cloth?
******
Are there more than two manufacturers of “A” cloth?
******
Was it pretty important to see the tag attached to the cloth that you would use in the field?
******
So there would be occasions when [the insulator] would not see the tag attached to the cloth, is that correct?
In the course of your lengthy career, have you used a number of Johns-Man-ville products?
After which the court interjected:
THE COURT: So far you haven’t asked any pertinent or significant questions.
Plaintiff contends that any error in the remarks quoted above was cured by the following portion of the court’s charge to the jury:
It is incumbent upon the court to do the best that he can to expedite the presentation of a case in order to conserve the time of the court and the jury ... if the court feels that objections are unnecessary in that they have already been made, the court might make a comment concerning the matters and might reprimand counsel. The court wishes the jury to know however that there is nothing personal in any comment that the court has made and we do not wish the jury to infer from any comment that the court has made that he feels that one side or the other should prevail in the law-suit____ The court would say this, une-quivocably that this case has been well tried by excellent attorneys, ... who are very capable and who have discharged their responsibility in representing their respective clients and they have a right to make objections and although the court may have reprimanded them, they still may have that right____ But the court has made comments only in order to try to expedite the trial of the case.
Because a retrial is required by our evi-dentiary rulings, it is not necessary to resolve this issue. However, if the case is to be retried by the same judge, sound judicial husbandry counsels that we indicate these remarks went far beyond the bounds of exemplary trial conduct by the district judge. Despite the admonishment in the instruction at the close of the case, these comments could be interpreted by the jury as attributing an unwarranted degree of incompetence and obfuscation of issues to defense counsel that could weigh against accepting the defenses offered on behalf of their clients. Perhaps the recitation of these highlights will suffice to prevent any recurrence.
*474VI. DIVISION OF DAMAGES
In determining the monetary liability of each defendant, the district court subtracted the $26,375 Dartez had received from the three settling defendants from the $200,000 jury award and entered judgment against the five appellants for the remaining $173,625. Defendants contend that the court should have divided the award between the eight joint tortfeasors on a pro rata basis so that each tortfeasor, including the settling defendants, would be liable for one-eighth of the $200,000 award or $25,-000 apiece. According to their argument, the settlement agreement discharged three-eighths of the plaintiff’s claim, which would limit his proper recovery to five-eighths of any jury award. Although this issue is technically mooted by our vacation of the judgment, we address it here because of the likelihood it will recur in the new trial.
In a decision rendered after the district court had entered judgment in Dartez, the Texas Supreme Court rejected both the pro rata credit advocated by defendants and the dollar credit applied by the district judge. Duncan v. Cessna Aircraft Co., 665 S.W.2d 414, 430 (Tex.1984). In Duncan, the Texas court adopted the theory of comparative causation for strict liability in tort. Id. at 427. The jury must apportion the liability between “all whose action or products combined to cause the entirety of the plaintiff’s injury.” Id. at 428.
A settlement with one tortfeasor reduces the liability of the nonsettling defendants by the percentage of causation allotted to the settling defendant. Id. at 429. “If the settling tortfeasor's liability is not established, the defendant will not be entitled to any benefit from the settlement agreement.” Id. at 434. The court specified that comparative causation would govern all future product liability cases, “including all cases remanded for retrial for whatever reason.” Id. Accordingly, if Dartez is successful at his second trial, the principles enunciated in Duncan will govern the apportionment of liability and the treatment of the settlement payments.
Although we have no record basis for deciding precisely how Duncan should be applied to Dartez’s case, we do note that its application could be complicated by the fact that two initial defendants, Johns-Manville and Unarco, whose products assertedly contributed to plaintiff’s injury, were severed from the action when they commenced bankruptcy proceedings. The stay imposed by the Bankruptcy Code, 11 U.S.C. § 362, has halted all legal action against these defendants.
Under Duncan the issue of an absent settling tortfeasor’s share of causation may be submitted to and determined by the jury. Acord v. General Motors Corp., 669 S.W.2d 111, 117 (Tex.1984). However, § 362(a)(6) of the Bankruptcy Code, which prohibits “any act to assess a claim against the debtor ... ”, seems to prevent any determination in the current action of the percentage of liability attributable to Johns-Manville and Unarco.
This reality creates somewhat of a conflict between two of the controlling principles of Duncan: (1) the jury must assess the liability of each tortfeasor based on the percentage of causation attributable to that entity, 665 S.W.2d at 428, and (2) each defendant will be jointly and severally liable for the total damages to ensure that the plaintiff receives full compensation in the event that one defendant should prove insolvent. Id. at 429. Thus, Duncan gives plaintiff the right to full recovery, while giving defendants the right to limit their percentage of the liability if they can prove that their share of causation was also limited. See Shipp v. General Motors Corp., 750 F.2d 418, 425 (5th Cir.1985). The Texas court has decided that plaintiff’s right to recovery should take precedence over defendants’ right to limit liability if one defendant is insolvent. However, a substantial question may exist as to whether Johns-Manville or Unarco are insolvent. The circumstances under which they commenced Chapter 11 proceedings are unusual, to say the least. It may be that at some point in their reorganization these *475companies will be able to pay their share of the liability.
Because Duncan requires a jury determination of what that share would be, a question appears to exist as to whether a final judgment may be entered against the nonsevered defendants in this case without assessing the portion of liability attributable to Johns-Manville and Unareo. In Wedgeworth v. Fibreboard, 706 F.2d 541, 544-546 (5th Cir.1983), this court refused to grant a discretionary stay to the other asbestos companies until the two bankruptcy proceedings are concluded. Duncan may have altered the balance of equities discussed in Wedgeworth. The parties have not raised or briefed Duncan’s impact in this regard. Whether its advent indicates that the application of Wedgeworth to this case should result in a different evaluation of the issue of a discretionary stay must be left to further development on remand. We intimate no view whatever on this matter.
VII. CONCLUSION
The judgment against Raymark is reversed and that corporation is dismissed from this action. Because of the evidentia-ry errors described herein the judgment is vacated and this case is remanded for a new trial of all issues as against Celotex, Owens-Illinois, Fibreboard, and Pittsburgh Corning.
REVERSED as to RAYMARK and VACATED and REMANDED FOR A NEW TRIAL as to APPELLANTS, CELOTEX, OWENS-ILLINOIS, FIBREBOARD and PITTSBURGH CORNING.
. We have analyzed the etiology of these diseases and their relationship to asbestos exposure in many previous cases and will not repeat that analysis here. See Gideon v. Johns-Manville Sales Corp., 761 F.2d 1129, 1135 n. 2 (5th Cir.1985).
. Caution labels shall be affixed to all raw materials, mixtures, scrap, waste, debris, and other products containing asbestos fibers, or to their containers, except that no label is required where asbestos fibers have been modified by a bonding agent, coating, binder, or other material so that during any reasonably foreseeable use, handling, storage, disposal, *471processing, or transportation, no airborne concentrations of asbestos fibers in excess of the exposure limits prescribed in paragraph (b) of this section will be released. 29 C.F.R. § 1910.1001(g)(2)(i) 1983.