LEVIN H. CAMPBELL, Circuit Judge.

Marion Coffran has permanent injuries caused by hepatitis, which was first diagnosed on her admission to the Mary Hitchcock Memorial Hospital on June 18, 1970. Mrs. Coffran had two hernia operations at the Hospital in 1970, on April 15 and June 5. These were performed by doctors who were members of the Hitchcock Clinic, Inc. In 1975 Mrs. Coffran, with her husband as co-plaintiff, commenced this diversity action in the United States District Court for the District of New Hampshire against the Clinic and the Hospital, contending that her hepatitis was caused by the anesthetic “halothane,” which was used for both the April and June operations. She asserted that these administrations of halothane were inconsistent with reasonable medical care. An 11-day jury trial was held, which resulted in verdicts for both defendants. The district court, however, granted plaintiffs’ motion for a new trial in respect to the defendant Clinic, appellant herein. The court denied plaintiffs’ motion for judgment n. o. v., and therefore its decision to grant the new trial was interlocutory and unappealable at the time. A second jury trial was held, before another judge, resulting in verdicts and sizeable judgments for the plaintiffs. On appeal from those judgments, defendant focuses exclusively on the issue of whether the judge who presided over the first trial had abused his discretion in ordering the new trial.

The district court granted the new trial on the ground that the verdict in favor of the Clinic was against the weight of the evidence. This court has described the proper standards in that situation, both for the trial judge’s decision and for our own review, in Borras v. Sea-Land Service, Inc., 586 F.2d 881, 886-87 (1st Cir. 1978). We said there that the judge has a duty to set aside the verdict and grant a new trial if he is of the opinion that the verdict is against the clear weight of the evidence, or is based upon evidence which is false, or will result in a clear miscarriage of justice, quoting Aetna Casualty & Surety Co. v. Yeatts, 122 F.2d 350, 352 (4th Cir. 1941). Only if the trial judge commits an abuse of discretion in granting or denying a- motion of this character will an appellate court reverse his decision. We went on to observe in Borras, however, that a “trial judge should not act merely as a ‘13th juror’ and set a verdict aside simply because he would have reached a different result had he been the trier of fact.” 586 F.2d at 887. Rather the judge’s duty is to exercise a more limited discretion. He should not interfere with the verdict “ ‘unless it is quite clear that the jury has reached a seriously erroneous result.’ ” Id., at 887, quoting 6A Moore’s Federal Practice ¶ 59.08[5], at 59-160. We observed that the modern trend when reviewing the granting of new trials on the ground that the verdict is contrary to the great weight of the evidence is away from earlier decisions that the trial judge’s discretion is virtually unlimited. Thus, the trial judge’s discretion, although great, must be exercised with due regard to the rights of both parties to have questions which are fairly open resolved finally by the jury at a single trial. See Borras, 586 F.2d at 887, quoting Lind v. Schenley Industries, Inc., 278 F.2d 79, 90 (3d Cir.), cert. denied, 364 U.S. 835, 81 S.Ct. 58, 5 L.Ed.2d 60 (1960). 1 Applying these stan *7 dards after a careful review of the record, we are constrained to conclude that the district court erred in granting plaintiffs’ motion for a new trial, and that the jury’s verdict in favor of the defendant Clinic must be reinstated. It is true that the court below expressly found that “the jury reached a seriously erroneous result — a result which was contrary to the clear weight of the evidence and which, if allowed to stand, would constitute a patent and grave miscarriage of justice.” But in our opinion the record does not support that conclusion. Rather, while it was established that Mrs. Coffran’s care was deficient in certain respects, the medical evidence was conflicting as to whether the hepatitis she contracted was causally related to those deficiencies. On the evidence presented, we do not see how a defendant’s verdict could be considered as contrary to the clear weight of the evidence. Rather it reflected a jury’s resolution of what was factually a very close and difficult case.

A brief review of the facts will be necessary before analyzing the district court’s decision and reasoning in more detail. Mrs. Coffran was admitted to the Hospital for her hernia problem on April 14, 1970. She was operated on the next day, at which time she received halothane, an anesthetic which defendant’s expert testified is used in one-third of all operations, and is considered in most instances particularly safe and effective. This was the second time in her life that halothane had been administered to her, the first being in 1964. The April hernia repair was not completely satisfactory, and Mrs. Coffran was re-admitted on June 4, 1970 for further surgery. During the operation on June 5, she again received halothane. After returning home, she became jaundiced about June 15. She returned to the Hospital on June 18, and was diagnosed as suffering from hepatitis. Her condition worsened, resulting in some brain damage due to hepatic encephalopathy. Her treating physician, Dr. McCleery, diagnosed Mrs. Coffran as suffering from hepatitis caused by the June exposure to halo-thane.

Although it was known in 1970 that a tiny number of patients who received halo-thane later contracted hepatitis, there was some question among experts within the medical profession whether the hepatitis was actually caused by the halothane. Those who believed that halothane caused hepatitis — and the evidence suggests that the believers outnumbered the disbelievers — posited a “sensitization” theory, as follows. A patient who is susceptible to halo-thane hepatitis is sensitized to halothane by exposure; there is no hepatitis at this time. After a subsequent exposure to halothane, however, the sensitized patient does contract hepatitis. This sensitization process is to be distinguished from direct toxic disease reactions, whereby liver damage is caused by the first exposure to a drug, and by a very different mechanism. Those doctors who did not accept the validity of the halo-thane sensitization thesis believed that halo-thane had been mistakenly incriminated in cases of viral hepatitis which became clinically evident after operations using halo-thane. Regardless of whether the halo-thane actually caused the hepatitis, or was merely innocently associated with it, the incidence of hepatitis following use of halo-thane was very rare: one study of fatal hepatic occurrences reported only nine otherwise unexplained fatal hepatic occurrences out of 250,000 administrations of halo-thane. Two points are vital. First, there was not at the time of trial, or in 1970, a preoperative test which could definitively identify those patients who would be susceptible to halothane sensitization, or who already were sensitized. Second, once hepatitis was contracted, there was no test which could distinguish between “halothane *8 hepatitis” (if indeed it existed) and other forms, such as viral hepatitis.

Plaintiffs’ theory was that Mrs. Coffran’s hepatitis was caused by her exposure to halothane in June, following sensitization by the April 1970 or 1964 administrations of halothane. They argued that while there was no single test to predict halothane sensitivity, Mrs. Coffran exhibited certain “markers” of sensitivity that should in the exercise of due care have been recognized, resulting in the choice of a different anesthetic (or in at least the administration of additional tests that might have further suggested sensitivity). Plaintiffs also asserted that the Clinic negligently omitted certain pre-operative tests in June which, had they been done, would have shown halothane sensitivity. The Clinic contended that Mrs. Coffran’s hepatitis was not caused by halothane; that even if it was, its doctors were not negligent in failing to recognize her alleged markers of sensitivity; and that even if it was negligent in failing to perform the pre-operative tests, this failure was not a cause of Mrs. Coffran’s hepatitis, since the tests would not have shown anything to contraindicate the use of halo-thane.

Plaintiffs presented four expert witnesses at trial: Dr. Bricker, a general practitioner and expert in “legal medicine”; Dr. McCleery, a gastroenterologist 2 and member of the Clinic, who was Mrs. Coffran’s primary treating physician since the onset of her hepatitis; Dr. Sy, 3 the anesthesiologist who prescribed halothane for Mrs. Coffran in April and June; and Dr. Barnett Greene, an anesthesiologist who was widely published, but had written nothing specifically about halothane. Defendant presented one expert, Dr. Nicholas Greene, the chairman of the Yale Medical School’s Department of Anesthesiology, whose publications included work on halothane.

Plaintiffs’ evidence suggested a number of different negligent acts or omissions by the Clinic with respect to the April operation. These, however, could not have directly caused Mrs. Coffran’s injury since it was not plaintiffs’ theory that Mrs. Coffran contracted halothane hepatitis at that time. 4 Plaintiffs also presented evidence relating to the Clinic’s alleged failure to notice purported markers of Mrs. Coffran’s sensitivity. However, there was a great deal of evidence — including testimony by Dr. McCleery, plaintiffs’ own witness— from which a jury might reasonably have determined that the purported “markers” (such as “unexplained post-operative fever”) were not in fact such and that there was no negligence in prescribing halothane in June in light of what the Clinic physicians knew of Mrs. Coffran’s condition and medical history. The district court, therefore, quite properly did not rely on any activities related to the April operation, or the evidence regarding markers of sensitivity, in stating its reasons for granting a new trial.

Rather, the district court focused on two things: the Clinic’s failure to perform an “eosinophil count” before the June operation, and the possibility that Mrs. Coffran had contracted viral hepatitis before the June operation which was aggravated by readministration of halothane. We discuss the “pre-existing viral hepatitis” theory first.

The district court’s reference to this theory is puzzling for several reasons. First, the theory was never presented to the jury. The jury was instructed without objection by plaintiffs that the plaintiffs needed to prove that halothane was “the probable cause of Mrs. Coffran’s hepatitis.” At a conference in chambers, plaintiffs’ counsel explicitly agreed that if the jury did not find halothane caused her hepatitis, the verdict would be for the defendant. The jury was never instructed that it could consider whether the halothane aggravated a preex *9 isting hepatic condition. Further, the court believed that an “Australian antigen test” had ruled out viral hepatitis. Given this belief, it is.difficult to understand how the court could have also thought it possible that Mrs. Coffran did in fact suffer from viral hepatitis. Finally, the court’s interpretation of the Australian antigen test was mistaken. The evidence was uncontradicted that Mrs. Coffran’s negative result did not rule out viral hepatitis. Only a positive result would have been conclusive; nothing could be inferred either way from a negative result. In any case, the belief that Mrs. Coffran might have contracted hepatitis before June could not have warranted a new trial given the theory on which the case was presented and tried.

We are not certain of the extent to which the court’s erroneous reliance on the above theory contributed to its conclusion to grant a new trial. In any case, the court’s primary reliance was clearly on the Clinic’s failure to perform an eosinophil test in June. This is a test which determines the percentage of eosinophils — a type of white cell — in the blood. Plaintiffs argued that if it had been performed, it would have warned of Mrs. Coffran’s sensitivity to halothane, and thus would have led to more tests and the choice of a different anesthetic in June. Three distinct questions are presented in connection with this theory, two related to causation and one to the standard of care. First, does Mrs. Coffran suffer from halothane hepatitis? If not, then the test is irrelevant since the administration of halothane would not be a cause of her hepatitis. Second, did the omission of this test constitute a failure to provide reasonable medical care? Third, if the test had been done, would it have shown halothane sensitivity? If not, again its omission would not be a cause of Mrs. Coffran’s hepatitis, since even if it had been performed, her sensitivity would have gone undetected and she would have been given the halothane.

The evidence was conflicting as to whether Mrs. Coffran’s hepatitis was caused by halothane. All of plaintiffs’ experts testified that it was. This included both Dr. McCleery and Dr. Sy, both of whom were members of the Clinic during the time they treated Mrs. Coffran. Mrs. Coffran’s medical records showed that all fourteen of the doctors (many just interns and residents) who saw her over the years following her June 18 admission wrote down that her problem was halothane hepatitis. While the district court thought this was “significant,” the testimony showed that most of these doctors simply recopied the original diagnosis without any thorough independent evaluation of their own.

Dr. Nicholas Greene testified that in his opinion Mrs. Coffran was suffering not from halothane hepatitis, but from viral hepatitis which became evident after the operation. 5 He gave two reasons for this belief. First, the course of Mrs. Coffran’s illness was not typical of those cases usually ascribed to halothane hepatitis. Second, Dr. Greene was not convinced that halo-thane ever actually caused hepatitis. Rather, he thought the sensitization theory the product of a statistical “mirage” caused by the coincidence of increasing use of halo-thane and increasing, but unrelated, instances of viral hepatitis. Dr. Greene was not alone in his belief. Plaintiffs’ own experts testified to an ongoing debate within the profession over whether halothane does cause hepatitis.

Given Dr. Nicholas Greene’s exceptional credentials and expertise on the subject of halothane, and his reasoned explanation of his conclusions, it is perhaps doubtful whether a jury could properly be held to have seriously erred in crediting his testimony over that of the other witnesses. As Chairman of the Yale Medical School’s Anesthesiology Department, and the witness who had done the most research on the precise issue of the causal connection between halothane and hepatitis, Dr. Greene’s *10 views cannot so easily be written off as the views of the more questionable expert in Borras, 586 F.2d at 887. On the other hand, the district judge was entitled to note that the Clinic’s own doctors, Dr. McCleery and Dr. Sy, diagnosed the case as one of halo-thane-induced hepatitis. Dr. Sy was the very anesthesiologist here charged with negligence. Additionally, as the court noted, the manufacturer of halothane gave written warnings of the possibility of hepatitis after use of halothane at least as early as 1968. Given these, and a number of other factors which we need not take the time to relate, we shall for present purposes accept, as not exceeding the trial court’s discretion, its conclusion that it would be against the clear weight of the evidence to find that Mrs. Coffran’s hepatitis was not caused by halothane.

Turning next to whether the failure to give the eosinophil test constituted negligence, the evidence was very strong that it did. Even Dr. Nicholas Greene admitted that it was negligent not to give Mrs. Coffran a blood test in June; the only question is whether that test should have included an eosinophil count. Dr.' Greene testified that only certain hemoglobin tests were needed, but plaintiffs’ experts, including Dr. McCleery and Dr. Sy, testified that a “complete blood count,” including an eosinophil test, was required. While the content of a complete blood count evidently varies from institution to institution, the testimony of Drs. McCleery and Sy indicates that at least at Hitchcock, it included the eosinophil test. Therefore, when they said that it was negligent not to give a complete blood count, they were including the eosinophil test. We thus believe that the district court was on firm ground in concluding that the failure to give Mrs. Coffran this test was so plainly negligent that a contrary jury finding would be against the clear weight of the evidence.

The most difficult problem, and the one which causes us to disagree with the district court, is presented by the last issue: whether the clear weight of the evidence was that had the test been done, it would have shown an abnormal reading which, in turn, would have triggered an examination leading to the use of another anesthetic. Plaintiffs’ theory was (1) that an abnormally high eosinophil reading may reflect sensitivity to a variety of different drugs, as well as other conditions; (2) that Mrs. Coffran was sensitized to halothane; (3) that therefore if the test had been done, it would have shown an abnormal reading; and (4) that this abnormal reading, while not in itself specifically pointing to halothane as the cause, would have been reason to postpone the operation, make further tests, and ultimately decide on a different anesthetic. The first proposition was undisputed, and, while the fourth is more speculative, we shall accept it for present purposes. As to the second, Dr. Greene testified to an opinion that Mrs. Coffran did not have halo-thane hepatitis, and that, indeed, halothane would not produce sensitization at all. However, as indicated above, we shall assume, if with some doubts, that the judge was entitled to discount this view as contrary to the great weight of the evidence. Thus we shall proceed on the supposition that the weight of the evidence demonstrated that her hepatitis was caused by the halothane, and that she was sensitized to it prior to June.

Our difficulty, then, focuses on the third step in the argument: assuming she was sensitized, did the clear weight of the evidence demonstrate that, in her case, the test would have been abnormal? If not, the causal link is broken; it is fundamental that negligence without causation is not actionable. See, e.g., Restatement (Second) of Torts § 432. The burden of proof is on the plaintiff to establish causation by a preponderance of the evidence. Id., § 433B & comment a; Jutras v. J. Scanlon Co., 106 N.H. 130, 133, 206 A.2d 615 (1965). The district judge, while dealing with the other points, made no mention of this one, suggesting that after concluding, correctly, that it was poor medical practice to omit the test, he failed to consider the degree to which the evidence demonstrated that, had the test been given, it would have been abnormal and thus served as a “red flag.” *11 We do not think that plaintiffs carried their burden so convincingly on this point as to justify the grant of a new trial.

Plaintiffs’ expert anesthesiologist, Dr. Barnett Greene, testified that 50 percent of those who are sensitized to halothane exhibit “eosinophilia,” or an abnormally high eosinophil count. He also testified that a high eosinophil count Mrs. Coffran had in July 1970 was evidence that she was among the 50 percent who do show eosinophilia as a marker of halothane sensitization. Defendant’s expert, Dr. Nicholas Greene, disputed this evidence, however. Assuming sensitization ever occurs, he testified that only 33 percent of those who are sensitized exhibit eosinophilia. Furthermore, he testified repeatedly that the failure to do blood tests in June was not causally related to Mrs. Coffran’s hepatitis because their results would not have affected the choice of anesthetic.

Given Dr. Nicholas Greene’s testimony, we do not think a new trial could properly have been granted on the ground that the jury was seriously mistaken if it disbelieved Dr. Barnett Greene’s testimony implying that Mrs. Coffran would have shown eosinophilia had the test been done in June. In the absence of special circumstances not present here, such as a vast disparity in qualifications or the presence of strong impeaching factors, we do not think it can be said that a verdict merely crediting one witness over one other witness is against the clear weight of the evidence. To be sure, the evidence did show that there was a 33-50 percent chance her eosinophil count would have been abnormal. In cases such as this, when the defendant omitted a safety precaution which would have had at least a substantial likelihood of saving the plaintiff from harm, courts have been generous in permitting the evidence to go to the jury over a defendant’s motion for a directed verdict, even “though it is often a pretty speculative matter whether the precaution would in fact have saved the victim.” 2 F. Harper & F. James, The Law of Torts § 20.2, at 1113 (1956); see, e.g., Zinnel v. United States Shipping Board Emergency Fleet Corp., 10 F.2d 47, 49 (2d Cir. 1925) (L. Hand, J.); Kallenberg v. Beth Israel Hospital, 45 A.D.2d 177, 357 N.Y.S.2d 508 (1974), aff’d, 37 N.Y.2d 719, 374 N.Y.S.2d 615, 337 N.E.2d 128 (1975); Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978). This standard for overcoming a motion for a directed verdict does not mean, however, that the plaintiff is relieved of his burden of proving causation: the jury must still find that defendant’s negligence caused plaintiff’s injury. See Kimball v. Scors, 59 A.D.2d 984, 399 N.Y.S.2d 350 (1977), explaining Kallenberg, supra; Hamil, supra, 392 A.2d at 1288 n.9. New Hampshire has also emphasized that causation is a question for the jury. See, e.g., Tullgren v. Amoskeag Manufacturing Co., 82 N.H. 268, 274, 133 A. 4 (1926). Under this standard, we must conclude that the jury could not be characterized as seriously mistaken if it chose to believe that the failure to perform the eosinophil test in June was not a cause of Mrs. Coffran’s injuries.

Given the credible expert testimony that the Clinic’s failure to perform an eosinophil test in June was not causally related to Mrs. Coffran’s hepatitis, we must conclude that the district court abused its discretion in granting a new trial on the ground that the verdict in favor of the Clinic was against the weight of the evidence. We find no basis in the record on which it could logically be concluded that the clear weight of the evidence supported an inference that had an eosinophil test been given, Mrs. Coffran’s alleged susceptibility to halothane hepatitis would have been revealed.

We are left with no alternative but to vacate the order granting a new trial and the judgments entered in the second trial. The case is remanded with directions that judgment be entered on the verdict in favor of the defendant Hitchcock Clinic, Inc.

So ordered.

. The dissenters in Lind v. Schenley Industries, Inc., 278 F.2d 79, 91 (3d Cir.), cert. denied, 364 *7 U.S. 835, 81 S.Ct. 58, 5 L.Ed.2d 60 (1960), argued that the district court’s ordering of a new trial on the ground that the verdict was against the weight of the evidence should be upheld if the appellate tribunal finds “any basis in reason for the trial judge’s conclusion as to the weight of the evidence and the injustice of the verdict.” This court sided with the Lind majority in Borras, although we emphasize that the district court’s discretion to grant a new trial remains large, and will not be disturbed if reasonably based.

. Gastroenterology is a sub-specialty which includes the study of liver illnesses.

. Dr. Sy’s testimony was presented via transcribed and video-taped depositions.

. Even if Mrs. Coffran was sensitized in April, that alone caused her no injury absent a subsequent exposure to halothane.

. Dr. Greene did not testify that the halothane aggravated this hepatitis in any way. He did, however, testify that the hernia operation and administration of any anesthetic would have been contraindicated by the known presence of hepatitis.