In May 1978 Virginia and Leonard Benedict filed a complaint against the United States under the Federal Tort Claims Act, 28 U.S.C. § 2671 et seq., and the National Swine Flu Immunization Program of 1976, 42 U.S.C. § 247b(k)(2)(A), alleging that a swine flu vaccination administered pursuant to a federal program caused Mrs. Benedict to contract Guillain-Barre Syndrome (GBS). After a bench trial in early 1986 the district court entered judgment for the United States Benedict v. United States, 634 F.Supp. 123 (N.D.Ohio 1986). The sole issue on appeal is whether the district court abused its discretion in not permitting the Benedicts to offer certain rebuttal testimony at trial. For reasons stated below, we conclude that the district court did abuse its discretion in not permitting portions of the proffered rebuttal testimony.
The parties stipulated that Mrs. Benedict was vaccinated for swine flu and then contracted GBS some nine and one-half weeks later; at trial the issues were limited to causation and damages. Inasmuch as it is medically impossible to determine the etiology of a given case of GBS, the evidence of causation necessarily rests on expert testimony regarding the events that typically precede GBS and that link the swine flu vaccine with GBS. The Benedicts’ sole expert witness in their case-in-chief was Dr. Foley, a board certified neurologist familiar with GBS. Dr. Foley testified that after examining Mrs. Benedict, her medical history, and records, he concluded that the swine flu vaccination was “the contributing and probably determining factor” in her GBS. Dr. Foley’s opinion was based on the temporal sequence of three events: her vaccination on November 6, 1976; her subsequent development of a non-specific, flu-like illness, which he attributed to an immunological reaction; and the onset of symptoms which culminated in a typical case of GBS on or around January 13, 1977. Although Dr. Foley testified that he could make such a diagnosis without reference to epidemiological data, he did acknowledge the value of epidemiological data as general proof of the relationship between GBS and the swine flu vaccine. In particular, he relied on Dr. Lawrence Schonberger’s study which found an increased risk of GBS for up to ten weeks after vaccination.
The Government then called three expert witnesses: Dr. Victor, a board certified neurologist who participated in a special panel commissioned by the government to conduct an epidemiological study of the relationship between GBS and the swine flu vaccine; Dr. Cole, a board certified neurologist; and Dr. Nathanson, an epidemiologist and co-author with Dr. Victor of the special panel report regarding GBS and the swine flu vaccine. The government’s three experts testified that, given the unavailability of medical tests to determine the origin of GBS in given cases, epidemiological studies are the sole means of demonstrating an association between GBS and certain antecedent events, e.g., the swine flu vaccine. They all testified that key epidemiological studies published since the Schonberger study refine Schonberger’s data and show that the period of increased risk for GBS lasts for a maximum of eight weeks after vaccination. Dr. Nathanson in particular testified to the methodology and data which underlay the special panel report, and how that methodology differed from Dr. Schonberger’s. The government also presented Dr. Victor’s expert testimony that the most common antecedent events causally related to GBS are upper respiratory infections, which precipitate ap*1428proximately one-half of GBS cases. Dr. Foley also acknowledged this during his own testimony, as well as the fact that Mrs. Benedict appeared to have an upper respiratory infection a few weeks before the onset of her GBS.
At the close of the government’s case, the Benedicts sought to introduce Dr. Goldfield, an epidemiologist, as an expert rebuttal witness. The government objected that Dr. Goldfield should have been called during the Benedicts’ case-in-chief. The Benedicts’ counsel argued that he needed to call Dr. Goldfield as a rebuttal witness because at trial Dr. Cole changed his opinion regarding the extent of the risk period. In 1980 Dr. Cole had agreed with the Schonberger study which established a ten week risk period. However, he had since changed his opinion to conform to the special panel report authored by Drs. Nathan-son and Victor which reduced the risk period to eight weeks. After the government pointed out that the Benedicts had been or should have been on notice that Dr. Cole changed his opinion by early 1983, the Ben-edicts’ counsel shifted his argument to state that he believed he had a right to rebut the government’s epidemiological evidence “as it relate[d] to certain tests or documents.” The district court ruled that Dr. Goldfield’s rebuttal testimony would not be allowed, because it logically belonged in the case-in-chief and went to the case’s central issue of causation.1 The Benedicts’ attorney then proffered for the record that had Dr. Goldfield been permitted to testify he would have stated that Dr. Nathanson’s analysis of the special panel report was erroneous, that his analysis of Dr. Nathanson’s data and of his own data indicated a ten week risk period, and he believed, based upon the epidemiological data, that the vaccination proximately caused Mrs. Benedict’s GBS.
II.
A trial judge’s determinations regarding the order of proof and scope of rebuttal testimony will not be disturbed absent an abuse of discretion. Geders v. United States, 425 U.S. 80, 86, 96 S.Ct. 1330, 1334, 47 L.Ed.2d 592 (1976). As succinctly stated by this court in Martin v. Weaver, 666 F.2d 1013 (6th Cir.1981), cert. denied, 456 U.S. 962, 102 S.Ct. 2038, 72 L.Ed.2d 485 (1982):
In the exercise of sound discretion, the district court may limit the scope of rebuttal testimony, Geders v. United States, supra; United States v. Algie, 503 F.Supp. 783, 793 (E.D.Ky.1980), to that which is directed to rebut new evidence or new theories proffered in the defendant’s case-in-chief. See, e.g., Bowman v. General Motors Corp., 427 F.Supp. 234, 240 (E.D.Pa.1977). However, ‘[wjhere ... [the] evidence is real rebuttal evidence, the fact that it might have been offered in chief does not preclude its admission in rebuttal.’ National Surety Corp. v. Heinbokel, 154 F.2d 266, 268 (3d Cir.1946). Furthermore, with respect to ‘real rebuttal evidence,’ the plaintiff has no duty to anticipate or to negate a defense theory in plaintiff’s case-in-chief. Weiss v. Chrysler Motors Corp., 515 F.2d 449, 458-59 (2d Cir.1975).
Id. at 1020.
The Benedicts argue that Dr. Goldfield’s proffered testimony was “real rebuttal evidence,” as contemplated in Martin. Id. The Benedicts assert that they established in their case-in-chief a causal relationship between Mrs. Benedict’s GBS and the vaccination through Dr. Foley’s clinical analysis that relied principally on the temporal sequence of the vaccination and onset of her GBS symptoms. As a clinician, Dr. Foley viewed epidemiology only as additional proof of a general relationship between the swine flu vaccine and GBS. When the government then defended by introducing expert testimony that a causal relationship can be established only through epidemiological data, not clinical analysis, and that the best epidemiological data reflects a maximum risk period of *1429eight weeks, the Benedicts assert that the government introduced new evidence or a new theory which they had a right to rebut through Dr. Goldfield. We agree to a point.
Despite the Benedicts’ shifting reasons for calling Dr. Goldfield, their actual proffer plainly indicated that their purpose was twofold: to dispute the accuracy of Dr. Nathanson’s methodology and data, and to reinforce their initial premise that the risk period lasted ten weeks. To the extent that Dr. Goldfield would have testified that his own data showed that the risk period was ten weeks and that he believed the vaccination proximately caused Mrs. Benedict’s GBS, the district court acted within its discretion to deny his testimony in rebuttal. Such testimony would have been cumulative of Dr. Foley’s testimony, and would merely have served to reinforce the Benedicts’ case-in-chief, not rebut new evidence. However, insofar as Dr. Goldfield’s testimony would have been directed to disproving the accuracy of Dr. Nathanson’s methodology and data, it was proper rebuttal and should have been allowed.
The government argues that Dr. Nathan-son’s epidemiological evidence was not new because both parties had long been aware of the central role and plethora of epidemiological data in swine flu vaccine cases. As such, the government contends that Dr. Goldfield’s testimony should have been presented in the Benedicts’ case-in-chief, and was in fact necessary to establishing their prima facie case. However, we consider that argument to be flawed.
First, the Benedicts were under no compulsion to present one particular theory of causation. Their burden was to prove by a preponderance of the evidence that the swine flu vaccine administered pursuant to a federal program proximately caused Mrs. Benedict’s GBS, Saxe v. United States, 577 F.Supp. 135, 144 (N.D.Ohio 1983), aff'd, 751 F.2d 386 (6th Cir.1984); their choice of theories by which to prove causation, persuasive or not, was their own to make. They chose Dr. Foley’s clinical analysis. When the government defended by introducing an epidemiological theory that cast doubt on the Benedicts’ clinical approach, new evidence was injected into the case. Although the epidemiological evidence was not “new” since the parties knew of its existence prior to trial, it was new for rebuttal purposes. “[Ejvidence is new if, under all the facts and circumstances, ... the evidence was not fairly and adequately presented to the trier of fact before the defendant’s case-in-chief.” Rodriguez v. Olin Corp., 780 F.2d 491, 496 (5th Cir.1986). Goldfield’s testimony regarding the accuracy of the methodology would have “served the permissible rebuttal function of counteracting the testimony of the opposing expert witness.” United States v. Posey, 647 F.2d 1048, 1052 (10th Cir.1981). In Posey, the government’s expert witness used a particular chemical test to confirm that the substance in issue was L-cocaine. The defense in response presented expert testimony that the government’s test was unreliable and less accurate than another method, the polarimeter test. The government on rebuttal was allowed to reintroduce its expert who used the polarimeter test suggested by the defense, and reported results consistent with her previous findings. Id. at 1050. See also Sanchez v. Safeway Stores, Inc., 451 F.2d 998, 999-1000 (10th Cir.1971) (district court properly allowed the plaintiff’s expert to testify in rebuttal to defendant’s introduction of photographic evidence, but not to testify as to results of his own independent investigation). Similarly, in this case the government sought to disprove the Benedicts’ theory of causation by introducing contradictory data and methodology; the Benedicts had a right to counteract the opposing expert’s testimony by putting on their own expert to controvert the accuracy and reliability of his methodology and data.
The government’s second argument that the data had been available and known during the pretrial period is irrelevant. Under the law in this Circuit the Benedicts had no duty to anticipate the government’s defense or to negate in its own case-in-chief a theory that would later be raised by the *1430government. Martin, 666 F.2d at 555. See also Rodriguez, 780 F.2d at 496.
The government’s remaining argument that Dr. Goldfield’s proffered testimony was necessary to establish the Benedicts’ prima facie case and was therefore properly excluded from rebuttal, see 6 J. Wigmore, Evidence § 1873 at 678 (Chadbourne rev. 1976), is also unpersuasive. Although the government correctly observes that a “[m]ere temporal relationship between a Swine Flu vaccination and a subsequent illness is not sufficient, in itself, to establish causation,” Saxe, 577 F.Supp. at 144, the Benedicts presented more than this in their case-in-chief. While Dr. Foley did emphasize the importance of the temporal relation between Mrs. Benedict’s vaccination and the onset of her GBS in forming his opinion of causation, he also expressly testified to and relied in part on Dr. Schonberger’s epidemiological study. In our view the Benedicts thus established a prima facie case. When the government, which did not move for a directed verdict, then introduced its own epidemiological study that refuted the Schonberger study and Dr. Foley’s conclusions, the Benedicts had a right to challenge the accuracy of the government’s data through rebuttal testimony.
In sum, Dr. Goldfield’s testimony which would have attacked the reliability of Dr. Nathanson’s data was proper rebuttal. The district court partially based its decision on its finding of fact that the government’s studies were reliable. It concluded that the Benedicts had not supported their attack on those studies with “hard data.” These circumstances force us to conclude that the court’s failure to permit the Benedicts the opportunity to rebut the government’s data regarding this critical issue was an abuse of discretion.
Reversed and remanded for a new trial and proceedings consistent with this opinion.
. Although Dr. Goldfield was not listed as a witness as required in Judge Batchelder’s pretrial order, Judge Batchelder did not exclude him for that reason and we will not consider it on appeal.