This is an appeal from a judgment by the trial court which granted relief to the plaintiffs in an independent action to set aside, on the ground of fraud, a judgment against them entered by the same court in 1956. Relief was also sought under Rule 60(b). The original action was a Tort Claims Act suit for the loss of sheep asserted to have been caused by radiation from atomic tests at the Las Vegas Test Site in 1953.
In the original case the findings and conclusions were that the sheep had not died as a consequence of the atomic tests but had died from other causes.
The original Tort Claims Act suit is referred to by the parties as Bulloch I. A very small portion of the testimony was transcribed and the notes have been lost. Only a few of the exhibits are available. There are some depositions and briefs which have survived. The trial judge in his oral and written findings and conclusions summarized the testimony. The opinion appears at 145 F.Supp. 824.
*1117The plaintiffs in the appeal before us assert that there was information withheld from them in the original suit, that witnesses for the Government were pressured, and there was fraud on the court. The plaintiffs also filed a motion to set aside the original judgment. The trial court entered judgment in the separate action and ordered that the judgment of 1956 be set aside by reason of fraud on the court. The court also awarded attorney fees to the plaintiffs. The Government has taken this appeal.
The record in Bulloch I which has survived is very limited. As mentioned, a very small part of the testimony as transcribed is available. From the depositions, briefs and memoranda and with the trial court’s summary, the trial as to the specific points relied on by the trial court in Bulloch II can be reconstructed with some confidence. Most of the witnesses from Bulloch I testified again.
The trial court in its conclusions and findings in Bulloch II makes several general statements as to fraud and withholding of information in Bulloch I. The trial court makes reference to four specific events at trial of Bulloch I or items of proof. Presumably the general observations or statements as to fraud are drawn from these particulars.
The four particular items mentioned by the trial court are:
(1) The Bustad Report — a comparison of radiation tests on sheep at Hanford with the data on the Utah sheep. The Hanford experiments had been made several years before the Utah suit. The trial court concluded that the comparison omitted essential facts.
(2) That the Government had given misleading information as to measures of radiation at certain places — radiation dosages.
(3) That witnesses, especially the veterinarians, had been “pressured” to testify in a certain way.
(4) That the Government had given misleading answers to interrogatories.
Estimates of radiation dosages were, of course, a significant part of the Bulloch I trial. The amount of radiation to which plaintiffs’ sheep had been exposed was a factual issue at the first trial. Plaintiffs’ Trial Memorandum in Bulloch I details the radiation dosages which were the subject of their claim, and an exhibit was presented by plaintiffs to demonstrate the point. The plaintiffs had available and used the same basic data used by the Government to make its estimates. The plaintiffs also presented at the original trial additional radiation maps apparently derived from their own sources. The record thus shows that all the information the Government had as to dosages was available to plaintiffs, and that they also had other data.
The data and maps presented the material in a detailed manner. It is apparent that during the investigation by the Government preceding the trial the dosages were revised as it progressed. These revisions were always upward and were, according to the testimony, to give maximum levels and thus to be above “probable average values.” The plaintiffs were aware of these revisions and the reasons long before the trial. Since all data was available the parties could trace its development during the investigations. There was also available the reports, with the opinions of the experts, which will be mentioned later but which show a change of opinion as the investigations progressed and as experiments were conducted. All of this was before Bulloch I, and revealed in detail to plaintiffs. There was no evidence presented at Bulloch II that the maps or data were incorrect.
The plaintiffs assert that the Government knew at the time of the first trial that there were unexplained small areas of higher radiation within generally lower areas. These were referred to in the second trial as “hot spots.” The plaintiffs assert that this phenomenon was not revealed at the first trial, and it was not. The plaintiffs state that this condition was discovered with the detonation of the Boltzman *1118shot, and the record so reflects, but this event was not until 1957, and hence after the first trial, so nothing was withheld as to this condition at Bulloch I as no one knew of its existence.
All the witnesses who testified in Bulloch I who were supposed to have participated in the suppression of evidence testified in Bulloch II. They also there testified specifically that there had been no suppression whatever of evidence at Bulloch I. They likewise testified that there had been no attempt to pressure them as to their opinions or as to their testimony.
It may be repetitious, but the most significant aspect of the records in Bulloch I and II is that all the information, data and witnesses were available to plaintiffs at Bulloch I. The plaintiffs chose to use some of the data and not other parts; they likewise selected among the witnesses who were made available for depositions and for trial; also they limited the extent of their consideration of the reports and studies which were provided to them. Plaintiffs in Bulloch I thus tried the case the way they wanted. They made decisions on what to put on as trial strategy and limited their review of data reported. It appears that they would not try the case the same way should they have another chance although the material is the same.
As mentioned, the record shows that at Bulloch I the Government provided all data plaintiffs asked for. They were given the AEC file data on the investigation of the sheep deaths. The Government made an offer to make available the individuals who had participated in the study, and any others. Other reports of experiments were provided and the experts were available for depositions. The Government did have all the technical data in 1953 and 1954, and conducted the experiments and investigations, but again this was made available to plaintiffs. This is all the data that is now available.
The plaintiffs had knowledge of the series of reports made by the Government veterinarians during the course of the investigations. They were aware that the conclusions of each were not the same, and were aware that positions changed as the pretrial investigation progressed. We find nothing whatever that could have misled plaintiffs as they could trace the progress of the investigation from the reports which were provided to them. Again, they had the data the Government had. If they did not choose to use it that was a decision they made. The reports of the investigation were quoted at length in plaintiffs’ Bulloch I trial memorandum, and the depositions show they had the documents and the information. Dr. Holmes, one of the persons participating in the investigation, was called as a witness by plaintiffs, and they had several pre-Bulloch I trial conversations with him. Dr. Holmes had died before Bulloch II. The plaintiffs assert that the Government presented an untrue representation as to the views of Dr. Holmes at Bulloch I, but again, he testified for plaintiffs and they had a full opportunity to interview him. The plaintiffs assert that the correct version of his views was revealed by certain documents they now refer to; however, these documents were available to them at the first trial.
The record shows that the trial court in Bulloch I considered the views of the veterinarians Veenstra, Holmes and Thompsett, and their development during the investigation. There appears to be nothing unusual about this.
The AEC facilities at Hanford, Washington had done some experiments from 1949 to 1953 (before Bulloch I) to study radiation damage to sheep and lambs. The sheep were fed radioactive material in this experiment. The Government asked the officials at Hanford to compare the data they had obtained with the data on the Utah sheep. The comparison was made, and the report was known at Bulloch I as the Bustad Comparative Study. This report concluded that the “Utah sheep showed no evidence of the radiation damage observed in experimentally treated sheep.”
The experts testified that if the Hanford ewes ingested radioactive material they *1119would develop a hypothyroid condition, and the lambs would be weak and die not long after birth. The Utah ewes however had not developed any abnormalities of the thyroid, and thus the experts found no significant similarity between the two groups and concluded the Utah sheep and lambs had not died from radioactive exposure. The trial court did not follow the testimony of the experts in the second trial and relied only on the fact that the lambs were weak and died young in both groups to establish a similarity. The trial court thus concluded that the Bustad report did not reveal all it should have by not comparing the early death of lambs as the important factor and thus something was withheld. The experts relied on the thyroid comparison as the significant factor.
All the Hanford data was published and was available to the public in 1953. This comparison listed the previous reports and it referred the readers to the listed reports for the “details of the experiments.” The Report recited that it only purported to be a summary of the experiments and conclusions. The basic data was cited. The plaintiffs thus had available all the pertinent data on the experiments with clear references-citations to the details. The damages caused by radiation at the Han-ford experiments were there clearly identifiable, as mentioned, but there was no such damage in the Utah sheep.
We find nothing to demonstrate that misleading answers were made by the Government in Bulloch I. The plaintiffs were familiar with all the background data as hereinabove described. With this familiarity they chose not to seek additional answers or clarifications.
The Comparative Study referred to above became a central issue in Bulloch II in the trial court’s finding of failure to disclose by the Government, but there is no basis in the record to suggest that anything was withheld.
The several specific items referred to by the trial court in its findings in the case before us have been considered in the foregoing description. As mentioned at the outset, these are the only particular matters referred to and the balance of the trial court’s findings and conclusions are in general terms, and could only be based on these factors.
The trial court thus (1) mentioned the Bustad report and asserted that information had been withheld; (2) it mentioned that the Government had given misleading information as to radiation dosages; (3) that the veterinarians had been pressured by the Government; and (4) that misleading answers had been given to interrogatories by the Government. Each of these specifics have been considered in relation to the record in the proceedings appealed from and to what original record and briefs are extant. We must conclude that none of these several findings or conclusions are supported by the record. The information, data, reports, maps, experiments and witnesses were all available to plaintiffs at the first trial. The data is the same now. They tried their case on the basis of their then best judgment in the selection of witnesses, interviews, the evaluation of data and reports, and the consideration of the position of the experts.
The trial court in its Memorandum shows that it would draw different conclusions and inferences from the testimony, and from the developing opinions of the veterinarians.
We have not mentioned the Los Alamos Scientific Laboratory comparative study of radiation burns (in experiments) and the lesions on the Utah sheep. The analysis of tissues, organs, bones and blood samples made by Doctors Comar, Rust and Trum showed no evidence of radiation injury to the sheep in the flocks of plaintiffs where the losses took place. A tissue analysis was made on the sheep by the Utah University College of Medicine and the conclusion was reached that the sheep had been exposed to radiation but the examination concluded that there had been no acute effects from the exposure. The Hanford comparative study, as mentioned, concluded that the sheep had not died from exposure to radiation.
*1120In Bulloch I the conclusion of the experts was that the sheep and lambs had died as a result of unprecedented cold weather during the lambing and shearing, together with malnutrition, also to generally poor range conditions and to common diseases which caused the skin lesions and deaths. These conditions were established by the evidence. The losses began after an initial detonation but before the detonation of the “Harry” test. The expert witnesses who were the leading authorities on the subject thus concluded that the losses were not due to radiation. There was no evidence presented that humans in the areas with the sheep nor animals such as the sheep dogs or horses, which were with the sheep, were affected.
It is apparent that in Bulloch II the assertions of fraud against the Government, the witnesses, the Government employees, the experts and the attorneys for the Government were all fully tried and considered. As stated above, there was no evidence whatever of fraud developed in the court hearings which covered all the details. The plaintiffs in Bulloch II again put on the evidence on this issue and were unable to make a case against anyone concerned. The non-party individuals were there considered by name. The hearings were conducted on the new complaint of plaintiffs and the trial court directed its decision to the complaint rather than to the separate motion to reopen the old case.
It appears that a Congressional hearing relating to the test firings was held several years before the plaintiffs filed this second suit. Some of the plaintiffs appeared at the hearings and made statements and their attorney filed a “report”. Only the complainants’ side of the dispute was so presented. The trial court makes reference to these 1979 hearings and apparently gave them some weight in its findings. The plaintiffs put great emphasis on the Congressional hearings and all or much of the report of the hearings was made a part of the record. We must instead rely entirely on the record made during the proceedings in Bulloch I and ÍI within the Rules of Federal Procedure.
The public discussions of atomic energy and weapons have heated up to a great degree since Bulloch I 25 years ago. There is now a much greater public awareness of issues relating to atomic bombs and radiation. This is certainly as it should be.
This recent increase in the heat, the light, and awareness of atomic energy is all worthwhile, but we are unable to find in it any reason, as the plaintiffs apparently would have us do, to overturn the considered judgment of the court reached 25 years ago.
The plaintiffs also have a heavy burden to demonstrate why they waited these many years to make the assertions they have made in this action. The only event of any consequence which has taken place and which seems to be relied on by plaintiffs were the 1979 Congressional hearings referred to above. The trial court also seems to have given this event some weight, as mentioned. This cannot be enough alone nor with the other assertions. The hearings purported to consider only one side of the incidents and only the plaintiffs and their attorney sought to use it to present their claims and personal views. If this was the source of the data upon which this proceeding is sought to be based, it must fail.
The judgment in Bulloch I was entered in the fall of 1956, as mentioned, and by the same court from which this appeal is taken and by the same judge. The parties are the same as well. This action to set aside the judgment was filed in February of 1981. As mentioned, it is basically a separate or independent action to set aside the judgment on equitable grounds, although it does mention Rule 60(b). Before judgment was entered in 1956 the plaintiffs were asserting that there had been some sort of “cover-up”. However, they acknowledged that they did nothing about it for 20 or 25 years. Plaintiffs have produced nothing to explain why they waited for 25 years to file this action and of more importance they have come forward with nothing of any consequence which was not available to *1121them at Bulloch I. This has been referred to above in the consideration of the specific points mentioned by the court. But to repeat, everything was then available to plaintiffs and they made their choice as to what to do and what to use. The trial court in its memorandum opinion in Bulloch II describes much of the material, reexamines it and draws conclusions.
During the 25-year period witnesses Thomsett and Holmes have died, and there is no physical evidence available. The principal witness for plaintiffs in Bulloch II, Mr. Knapp, acknowledged that nothing would be accomplished by any additional study of the circumstances. Thus it must be concluded that the plaintiffs would have the court reexamine the same material that was available in 1956. As mentioned, the public attitude toward atomic testing and atomic energy has changed very much since 1956 as has public awareness. The Congressional hearings in 1979 demonstrated this quite clearly, and the plaintiffs would use them as a reason to retry their case.
This case demonstrates the very good reasons why judgments should be final and should not be disturbed. The plaintiffs to prevail must have shown by clear and convincing evidence that there was fraud on the court, and all doubts must be resolved in favor of the finality of the judgment.
It is beyond question that a federal court may investigate a question as to whether there was fraud in the procurement of a judgment. Universal Oil Products Co. v. Root Refining Co., 328 U.S. 575, 66 S.Ct. 1176, 90 L.Ed. 1447. This is to be done in adversary proceedings as in the case before us. See Hazel-Atlas Glass Co. v. Hartford-Empire Co., 322 U.S. 238, 64 S.Ct. 997, 88 L.Ed. 1250; Sprague v. Ticonic National Bank, 307 U.S. 161, 59 S.Ct. 777, 83 L.Ed. 1184; and United States v. Throckmorton, 98 U.S. (8 Otto) 61, 25 L.Ed. 93.
Fraud on the court (other than fraud as to jurisdiction) is fraud which is directed to the judicial machinery itself and is not fraud between the parties or fraudulent documents, false statements or perjury. It has been held that allegations of nondisclosure in pretrial discovery will not support an action for fraud on the court. H.K. Porter Co., Inc. v. Goodyear Tire & Rubber Co., 536 F.2d 1115 (6th Cir.). It is thus fraud where the court or a member is corrupted or influenced or influence is attempted or where the judge has not performed his judicial function — thus where the impartial functions of the court have been directly corrupted.
The basic decisions of the Supreme Court are Throckmorton, Hazel-Atlas, and Universal Oil Products, cited above. These cases considered the basic issues raised in cases to set aside judgments and demonstrate with Marshall v. Holmes, 141 U.S. 589, 12 S.Ct. 62, 35 L.Ed. 870, the nature of the fraud and the proof required for relief as set out in the preceding paragraph.
As to actions for relief from fraud on the court it is generally held that the doctrine of laches as such does not apply, but unexplained delays bear on the basic concept of the finality of judgments and the proof. Rule 60(b) does not impose a time limit on motions asserting fraud on the court. The rule also, of course, recites that it does not limit the power of the courts to consider independent actions to relieve a party from a judgment.
In Wilkin v. Sunbeam Corp., 466 F.2d 714 (10th Cir.), we considered a claim under Rule 60(b). It was there said that: “Relief under the rule may be granted when the application is clearly substantiated by adequate proof.” There also it appeared that the documents complained of were in whole or in part available at trial but no use was made of them. In Chisholm v. House, 160 F.2d 632 (10th Cir.), we considered equitable relief from a judgment for fraud, and considered the intrinsic and extrinsic distinction, citing United States v. Throckmorton, 98 U.S. (8 Otto) 61.
The plaintiffs in their independent action asserted fraud on the court in the *1122withholding of data, records and letters. They have set forth extravagant statements in their briefs which have not been borne out by the record. The showing made by plaintiffs in the Bulloch II hearings falls far short of proof of fraud on the court or any other kind of fraud.
As mentioned, the trial court concluded in Bulloch II that there had been a fraud on the court (the same judge) in Bulloch I. In so doing, the court seems to have placed much, if not controlling, weight on the hereinabove described development (prior to Bulloch I) of the opinions of the veterinarians during the course of the investigations. We have considered this carefully (with the other factors raised by the trial court) and must conclude that nothing was demonstrated which would constitute fraud on the court.
We must thus conclude that the trial court was so in error and its conclusion and judgment constituted an abuse of discretion. The award of attorney fees must also be set aside.
The judgment of the trial court is set aside. The case is REVERSED and REMANDED.
McKAY, Circuit Judge,with whom
SEYMOUR, Circuit Judge, joins, dissenting.For the first time in the history of the Tenth Circuit, we have overturned a district court’s grant of a new trial under Rule 60(b) of the Federal Rules of Civil Procedure. What makes this action even more remarkable is that the decision of the trial, judge was based on his finding that a fraud had been committed on his court in a case tried to the bench.
Not only has this circuit never reversed a trial court’s grant of a new trial under 60(b) but, in reviewing cases decided by the other circuits in the past five years, I could find only one in which such a decision was reversed.* Further, my review of the past forty years of Supreme Court cases revealed no case in which the Court reversed a circuit court’s affirmance of the grant of a 60(b) new trial. See generally 11 Wright and Miller, Federal Practice and Procedure § 2872 (1973). Admittedly, the dearth of appellate decisions overturning the grant of 60(b) new trials does not establish a principle that it should not be done, but it does starkly highlight the severity of the constraints against so doing.
The historical reticence of appellate courts to overturn the grant of a new trial is not surprising in view of the applicable appellate standard. A 60(b) motion is addressed to the sound discretion of the trial court and will not be disturbed unless the trial court has manifestly abused its discretion. Caribou Four Corners, Inc. v. Truck Insurance Exchange, 443 F.2d 796, 799 (10th Cir.1971). Implicit in the notion of trial court discretion is the recognition that “[a]s an appellate court it is not for us to determine whether the trial court reached the correct decision, but whether it reached a permissible one in light of the evidence.” Higgins v. Oklahoma, 642 F.2d 1199, 1202 (10th Cir.1981).
Further contributing to the dearth of decisions overturning the grant of a new trial under 60(b) is the well-established principle of appellate review that:
When a case is tried to the district court, the resolution of conflicting evidence and the determination of credibility are matters particularly within the province of the trial judge who heard and observed the demeanor of the witnesses.
Dowell v. United States, 553 F.2d 1233, 1235 (10th Cir.1977). This principle has its roots not in any complicated jurisprudential theory but in the simple, commonsense notion that a judge who observed the trial is in a better position to assess credibility than a judge who did not. The Supreme *1123Court recently vigorously reminded the circuit courts of the limited scope of their review of trial court factual findings in Anderson v. City of Bessemer City, North Carolina, — U.S.-, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985). The Court noted that even district court findings based on physical or documentary evidence or inferences from other facts could not be overturned unless clearly erroneous. The Court further held that:
When findings are based on determinations regarding the credibility of witnesses, Rule 52 demands even greater deference to the trial court’s findings; for only the trial judge can be aware of the variations in demeanor and tone of voice that bear so heavily on the listener’s understanding of and belief in what is said____ [W]hen a trial judge’s finding is based on his decision to credit the testimony of one of two or more witnesses, each of whom has told a coherent and facially plausible story that is not contradicted by extrinsic evidence, that finding, if not internally inconsistent, can virtually never be clear error.
— U.S. at -, 105 S.Ct. at 1512-13.
This principle of deference, coupled with the abuse of discretion standard of review, makes the majority’s opinion doubly perplexing. In leaping an unprecedented hurdle — the overturning of the grant of a new trial under Rule 60(b) — the majority employs an impermissible means — the substitution of its own assessment of credibility for that of the trial judge. I will not attempt to chronicle every instance of such substitution. A few examples should suffice to illustrate the error.
Plaintiffs initially sued the government in 1955 when an unusually high number of their sheep died after having grazed in an area adjacent to the Nevada test site following a series of atomic tests. The district court found for the government. Bulloch I, 145 F.Supp. 824 (D.Utah 1956). In the years following this decision, the government’s virtual monopoly on information regarding the effects of radiation ended. In 1979 Congress conducted public hearings regarding the health effects of the Nevada nuclear tests of the 1950’s and 1960’s. Through these hearings, plaintiffs learned information regarding their claim of which they had previously been unaware. On the basis of this information plaintiffs made further investigations that led to the commencement of Bulloch II, 95 F.R.D. 123 (D.Utah 1982), in which plaintiffs sought to have the prior judgments vacated for fraud on the court. The trial judge found that fraud had been committed on his court in the prior action, and ordered a new trial.
Plaintiffs alleged in Bulloch II that the government fraudulently withheld significant information in the so-called “Bustad Comparative Study” offered into evidence in Bulloch I. The study, which was conducted by Dr. Bustad for the Atomic Energy Commission, examined the effect of radioactive iodine on sheep. It concluded that “the Utah sheep showed no evidence of the radiation damage observed in experimentally treated sheep.” Bulloch II, 95 F.R.D. at 131-32. Plaintiffs’ attack on the study in Bulloch II was two-fold. They alleged that the study was misleading because its conclusion was untrue and because information so indicating — regarding the substantial similarities between Utah lambs exposed during gestation, so-called “fetal lambs,” and experimental fetal lambs — was suppressed.
The government argues that the conclusion of the Comparative Study was correct because, even though there were similarities between the Utah fetal lambs and the experimental fetal lambs, those similarities were appropriately viewed as coincidental. The government contends this to be the case because of what it views as a determinative cfe similarity between the thyroids of plaintiffs’ sheep and the experimental animals. The government does not deny that the information regarding fetal lambs would have been important to the plaintiffs’ investigation and to the district court’s consideration of the issue of causation. Dr. Bustad so indicated in his testimony in Bulloch II, Appellees’ Appendix, *1124vol. 1, at 275, 280, as did Dr. Trum. Appellees’ Appendix, vol. 1, at 165. Those two scientists explained the omission of this information from the Comparative Study on the grounds that the Study was made merely for the information of the Atomic Energy Commission, which already had the fetal lamb report, the fetal lamb report was cited in footnotes to the Comparative Study, and the report would have confused the district court.
The trial court held that the fetal lamb reports had been wrongfully withheld at trial, were of substantial significance and weight in favor of plaintiffs’ claims, and were essential to a fair and proper consideration of the merits of those claims by the court. The court further found that regardless of whether or not Dr. Bustad buried the information in a footnote reference as a ploy, those who managed the government’s case knew of the information and wrongfully withheld it at trial. In overturning the district court, the majority finds that “the plaintiffs ... had available all the pertinent data on the experiments with clear réferences-citations to the details.” Maj.Op. at 1119. This analysis wholly misconceives the nature of our appellate inquiry. As the Supreme Court noted in Hazel-Atlas Co. v. Hartford Co., 322 U.S. 238, 246, 64 S.Ct. 997, 1001, 88 L.Ed. 1250, reh’g denied, 322 U.S. 772, 64 S.Ct. 1281, 88 L.Ed. 1596 (1944), a case that also involved an allegedly fraudulent document:
even if Hazel did not exercise the highest degree of diligence, Hartford’s fraud cannot be condoned for that reason alone____ [Tampering with the administration of justice in the manner indisputably shown here involves far more than an injury to a single litigant. It is a wrong against the institutions set up to protect and safeguard the public, institutions in which fraud cannot complacently be tolerated consistently with the good order of society. Surely it cannot be that preservation of the integrity of the judicial process must always wait upon the diligence of litigants. The public welfare demands that the agencies of public justice be not so impotent that they must always be mute and helpless victims of deception and fraud.
Thus, the issue on review is not whether plaintiffs could have found the fetal lamb information if they had been more diligent but, rather, whether the government sought to diminish the likelihood of their finding the information. If the intention of those managing the government's case in Bulloch I was indeed to suppress the fetal lamb report, it is not unthinkable that they would have sought the middle ground between complete suppression and complete disclosure that an artfully disguised footnote can provide. Regardless, it is not our function to determine whether or not the government intended to suppress the fetal lamb report. Bulloch I was tried to the bench. The district court not only observed the witnesses in that first trial, but is singularly aware of the effect the fetal lamb information would have had on his decision. That same district judge heard contradictory testimony in Bulloch II regarding the reasons for the conclusion of the Comparative Study and for exclusion of the fetal lamb report from that Study. Having observed the witnesses and examined the evidence, the district court chose to believe the plaintiffs’ witnesses and the plaintiffs’ evidence. The majority, armed only with the cold record, chooses to believe the government’s witnesses and the government’s evidence; this, despite the constraints against so doing inherent in the abuse of discretion standard of review and the well-established principle that the assessment of credibility is particularly within the province of the trial judge.
The majority errs in the same way with regard to the issue of whether the government gave misleading answers to interrogatories. Among the interrogatories and answers were these:
Interrogatory # 28: Did anyone involved in the investigation disagree with the report? [AEC report denying causal connection between radiation and sheep injury.]
Answer: We are not aware of anyone who is involved in the Commission’s in*1125vestigation of the alleged sheep losses who now disagree with the report issued by the AEC.
Interrogatory # 29: Did anyone conclude that radioactive fallout was a possible cause of the injury to the sheep? Answer: We do not know of anyone connected with the AEC’s investigation of the alleged sheep losses who has now concluded that radioactive fallout was a possible or probable cause of the injury to the sheep.
The district court found that the Government’s answers were both intentionally evasive and untrue:
The answers were highly deceptive in that they led plaintiffs to believe that those on whom reliance had been placed for making out a prima facie case had now changed their minds in unanimous support of the views of the defendant and discouraged the exploration in depth by plaintiffs of their original opinions by inducing apprehension on the part of plaintiffs’ counsel that this would only emphasize the adverse effect of the changed opinions.
Bulloch II, 95 F.R.D. at 139.
The majority gives short shrift to plaintiffs’ contentions with regard to the interrogatories:
We find nothing to demonstrate that misleading answers were made by the Government in Bulloch I. The plaintiffs were familiar with all the background data as hereinabove described. With this familiarity they chose not to seek additional answers or clarifications.
Maj.Op. at 1119.
The majority’s finding that there is nothing to demonstrate that the government gave misleading answers is unfounded. Dr. Veenstra testified in Bulloch II that he never changed his original opinion that radiation may have been a contributing cause of the injuries to the sheep. Appellees’ Appendix, vol. 2, at 431, 464, 466-67. There are strong indications in the record that Dr. Thompsett did not actually change his original position that radiation was a contributing cause of the damage to the sheep, despite his deposition testimony and letter to the contrary. Appellees’ Appendix, vols. 1 & 2, at 229-30, 233-34, 644-45, 672-74. Dr. Trum, who assisted in the preparation of the interrogatory answers, has indicated, with some equivocation, that he was aware that neither Dr. Veenstra nor Dr. Thompsett had changed his original position. Appellees’ Appendix, vol. 1, at 209-11, 227-30. Further, the record indicates that government attorneys were aware of these dissenting opinions. Appellees’ Appendix, vol. 2, at 685-87.
Moreover, the majority’s focus on plaintiffs’ lack of diligence is again misplaced. The issue is not whether plaintiffs could have discovered the government’s answers to have been misleading but, rather, whether the government intentionally responded incompletely and untruthfully to the plaintiffs’ queries. The trial court’s finding that the government engaged in intentional concealment is bolstered by the fact that government attorneys refused to reveal relevant facts even in the face of plaintiffs’ explicit claim in Bulloch I that they were concealing such facts. Bulloch II, 95 F.R.D. at 127-28. This concealment is made even more egregious by the fact that government lawyers in civil actions have the “responsibility to seek justice and to develop a full and fair record____” Model Code of Professional Responsibility, EC 7-14 (1980).
Finally, the majority suggests that even if the district court judge was correct in findings that the government had intended to suppress the fetal lamb reports and to respond untruthfully to the interrogatories, such findings do not support his holding that a fraud had been committed on his court. The majority argues that fraud on the court is limited to instances “where the impartial functions of the court have been directly corrupted.” The majority further asserts that fraud on the court “is not fraud between the parties or fraudulent documents, false statements or perjury.”
In support of this curiously narrow construction of what constitutes fraud on the court, the majority offers two cases that *1126not only do not support it, but directly contradict it. The majority points to Hazel-Atlas Co. v. Hartford Co., 322 U.S. 238, 64 S.Ct. 997, 88 L.Ed. 1250 (1944) as authority for the proposition that fraudulent documents cannot constitute fraud upon the court. This ignores the fact that the Hazel-Atlas Court found precisely the opposite — that fraudulent documents, in that case an article whose authorship had been misattributed, can constitute such fraud. Id. at 247, 250, 64 S.Ct. at 1001-1003. The case of H.K. Porter Co., Inc. v. Goodyear Tire & Rubber Co., 536 F.2d 1115 (6th Cir.1976) does, as the majority argues, stand for the proposition that nondisclosure by a party or witness is not grounds for an action for fraud upon the court. The Porter court goes on to make clear, however, that if Porter’s attorneys had known of and sponsored such misdeeds as nondisclosure, misrepresentation or perjury, it would have been quite another matter: “Since attorneys are officers of the court, their conduct, if dishonest, would constitute fraud on the court.” Porter, 536 F.2d at 1119. In the case at hand, the district court stated, with ample support in the record, that:
it appears clearly and convincingly from all of the circumstances of the litigation as they now appear, and I find, that one or more of defendant’s attorneys knowingly participated in a program for the concealment from the Court of facts which he or they knew or in good conscience should have known the Court was entitled to have placed before it in order properly to rule upon the determinative issues of the case.
Bulloch II, 95 F.R.D. at 142. If, as the district court found, the government attorneys in this case intentionally withheld information, offered into evidence an intentionally misleading document, and pressured witnesses to testify in a certain way, it can hardly be argued that such actions do not constitute fraud on the court under Rule 60(b).
While I have reviewed only two of the district court’s findings in detail, it is important to make mention of the full scope of his findings in support of his holding that fraud had been committed, and to place that holding in its historical context. The district court found that the government intentionally deceived the court; improperly pressured witnesses not to testify as to their real opinions or to discount their qualifications and opinions; intentionally withheld a vital report; presented information in another report in such a manner as to be deceitful, misleading, or only half true; deceptively answered interrogatories; and deliberately concealed significant facts regarding the possible effects of radiation upon the plaintiffs’ sheep. The court found that by these actions the government manipulated the processes of the court to its improper and unacceptable advantage. Bulloch II, 95 F.R.D. at 144. Further, the trial court’s opinion is informed by a sense of historical perspective wholly lacking in the majority’s opinion. The trial court recognized that the government committed this fraud at a time when “the AEC and those associated with it enjoyed a virtual monopoly of knowledge in comparison to that independently available to the plaintiff sheepowners’, their attorneys and, indeed the Court____” Bulloch II, 95 F.R.D. at 144. Thus, the government’s concealment and deception was even more egregious at the time of Bulloch I than it would be today.
In conclusion, I again wish to point to the Supreme Court’s recent opinion in Anderson. The Court noted that:
“a finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.” ____ This standard plainly does not entitle a reviewing court to reverse the finding of the trier of fact simply because it is convinced that it would have decided the case differently. The reviewing court oversteps the bounds of its duty under Rule 52 if it undertakes to duplicate the role of the lower court. “In applying the clearly erroneous standard to the findings of a district court sitting without a jury, appellate courts *1127must constantly have in mind that their function is not to decide factual issues de novo.”____ If the district court’s account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently. Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.
— U.S. at-, 105 S.Ct. at 1511-12 (cita-
tions omitted). Thus, our task as an appellate court is not to determine whether the radiation caused the damage to plaintiffs’ sheep. Nor is it to assess the diligence of plaintiffs’ counsel. Nor is it even to determine whether the government committed a fraud on the district court. Our task is to determine whether the district court’s finding of fraud was so clearly erroneous as to constitute a manifest abuse of discretion.
The testimony regarding whether the government intentionally tried to mislead the plaintiffs and the district court was contradictory. In the end, the issue turns on the assessment of credibility. Plaintiffs have presented substantial evidence that, if believed, supports a finding of fraud on the district court. In view of the trial court's better perspective for making the assessment of which evidence should be believed, and the constraints inherent in the abuse-of-discretion appellate standard, I cannot agree that this district court judge abused his discretion in overturning his previous decision because it had resulted from the commission of a fraud on his court. I would affirm.
In Corex Corp. v. United States, 638 F.2d 119 (9th Cir.1981), the Ninth Circuit reversed a trial court’s grant of a new trial under Rule 60(b). Unlike the case at hand, however, the trial court’s decision was not based upon the ground of fraud, but upon the ground that new evidence had been discovered. Because all of the new evidence concerned events that had occurred after the end of the trial, the Ninth Circuit held that it was "not newly discovered evidence within the meaning of the rules.”