BREYER, Chief Judge. On September 21, 1987, Timothy Walsh died as a result of a fire that destroyed his motor home (technically called a "recreational vehicle" or "RV"). On September 20, 1989, Walsh's personal representative (the plaintiff) brought this diversity wrongful death action against the RV's maker (Fleetwood), the manufacturer of the RV's refrigerator (Stolle), and the maker and distributor of a gas detector (CCI Controls and Manchester Tank). Plaintiff claimed that a broken valve in the RV's refrigerator leaked propane into the refrigerator, from which it leaked into the RV itself (without setting off the gas detector alarm) and exploded, causing the fatal fire. The defendants replied that any such chain of events was highly improbable; they blamed the fire on gasoline, which they said Walsh brought into the RV, perhaps to clean his carpet.
Before trial, the district court granted summary judgment for the refrigerator's manufacturer, Stolle, on statute of limitations grounds. After plaintiff presented his case, the court directed verdicts for the gas detector defendants, CCI and Manchester Tank, because there was insufficient evidence that they made or sold this particular gas detector. At the end of the trial, the jury found in favor of the only remaining defendant, the RV manufacturer, Fleetwood.
Plaintiff appeals all these adverse judgments. After reviewing the plaintiff's arguments and the record, we find plaintiff's claims unconvincing. We affirm the judgments.
The Exclusion of Expert Testimony
Plaintiff's first set of arguments focuses on the district court's decision not to permit plaintiff's expert, Wayne Buck, to testify. In effect, the court's decision to exclude that testimony sanctioned the plaintiff for having waited until the very eve of trial before telling the defendants what Buck would testify about. The plaintiff argues here that the court, in deciding that a sanction was appropriate and in choosing exclusion as a sanction, acted beyond its legal powers.
To understand why we reject these arguments, the reader must keep the following background in mind. First, the court believed that Buck's testimony could prove relevant to three different issues:
i) On February 22, 1991, eleven days after trial began, plaintiff made an explicit offer of proof in which he told the court that Buck would testify that "he did an analysis of the polystyrene composition . . found in the [RV] . . . and that the polystyrene component is a solid chemical incendiary that acts as an accelerant" (the "polystyrene issue").
ii) At the same time, on February 22, 1991, plaintiff told the court that Buck would also testify that he "did a variety of gas chromatographic tests on debris" that the plaintiff had obtained from the RV, which tests revealed that that debris "did not contain gasoline" (the "plaintiff's samples issue").
iii) The next day Buck indicated (in a court-ordered mid-trial deposition) that he also would testify that Fleetwood's test results on debris samples that Fleetwood had obtained from the RV showed so much gasoline that someone must have added gasoline to them after the fire (the "defendant's samples issue").
Second, the initial issue, the "polystyrene issue," was not relevant to the case. The district court excluded Buck's testimony about polystyrene because, by the time of trial, the issue had evaporated. The court believed that polystyrene had nothing significant to do with the dispute, and the plaintiff does not contest that Conclusion on this appeal.
Third, the parties agree that defendants learned nothing about the other two subjects of Buck's testimony until just before (or during) trial. The relevant pretrial events include the following:
i) in September 1987 the accident took place,
ii) in November 1987 plaintiff obtained his samples;
iii) in September 1989 plaintiff filed this lawsuit;
iv) in September and October 1989 plaintiff provided Rule 26(b) expert "subject matter" disclosure, which did not mention gasoline, which indicated Buck would testify about the (irrelevant) "polystyrene issue," and which added that Buck had "absolutely . . . no opinions" about how the accident happened;
v) in November 1990 plaintiff learned, in some detail, about defendant's tests of defendant's samples (finding gasoline to be present);
vi) in December 1990 Buck tested plaintiff's samples for gasoline; and
vii) in February 1991 defendants learned about Buck's testimony on the "plaintiff's samples" and "defendant's samples" issues; also, the trial began.
Fourth, the legal standards that govern our consideration of the exclusion of Buck's "gasoline issue" testimony are straightforward:
i) Fed. R. Civ. P. 26(e) requires a party "seasonably to supplement" information initially given to the opposing party about "the subject matter" on which an "expert witness . . . is expected to testify . . .,"
ii) The district court has broad legal power to impose sanctions, including exclusion, for violations of this duty, see Advisory Committee Note to Rule 26(e), and
iii) This appellate court will set aside such a sanction as outside the district court's broad legal powers only when it finds "manifest -- and damaging -- error." Northern Heel Corp. v. Compo Industries, Inc., 851 F.2d 456, 468 (1st Cir. 1988) (citation omitted).
Hence, the legal questions now before us are whether, in respect to the gasoline issues (the "plaintiff's samples" issue and the "defendant's samples" issue), plaintiff failed "seasonably to supplement" his earlier disclosure, and, if so, whether the court's decision to exclude that testimony as a sanction, amounted to "manifest -- and damaging -- error." We shall explain briefly, in respect to each of the two gasoline issues, why we conclude that the district court acted lawfully.
A. The plaintiff's samples. The facts adequately support the district court's Conclusion that plaintiff failed "seasonably" to supplement his disclosure in respect to Buck's failure to find gasoline in plaintiff's samples. Fed. R. Civ. P. 26(e)(1). Buck first tested those samples in December 1990. Plaintiff made his first effort to supplement on February 7, 1991, four days before trial. And that supplementation simply said that "Buck has performed tests . . . on plaintiff's exhibits 1 - 23 . . . , all contained in metal cans," and that "no gasoline was found." It gave no details that would help defendants assess the validity of the testing; and it misleadingly suggested Buck would rely on negative tests of all the samples (although, in fact, he intended to rely only on the samples taken from the RV's dinette area). On the day of trial plaintiff gave the defendants an additional three-page document, but it added nothing to the earlier disclosure. The defendants did not learn of the substance of Buck's testimony until they deposed him on February 23, in the midst of trial. In our view, the district court could fairly conclude that plaintiff's February 7 and 11 disclosures were too little and too late; and that the disclosure by means of the February 23 deposition, while adequate in scope and substance, occurred much later and certainly was not "seasonable."
The appropriateness of excluding Buck's testimony as a sanction presents a closer question, but one that we must resolve in the district court's favor. On the one hand, plaintiff's disclosure delay potentially prejudiced the defendants significantly. Had they known about Buck's testimony sooner, they might well have decided to counter it, through cross-examination or other expert testimony, suggesting, for example, that gasoline in samples can evaporate after three years, or that the tested samples came from a special place in the RV (under the floor boards in the dinette), and were consistent with gasoline being present elsewhere. Such refutation might well have required time to prepare and experts to obtain. A continuance on the eve of trial would have offered defendants only a Hobson's choice of inadequate preparation or costly delay; and a continuance midway through the trial could have meant losing the benefits of previously cross-examined witnesses as well. At the same time, plaintiff presented no plausible excuse for his delay. To the contrary, it is unclear why he waited several years to have Buck test his own samples.
On the other hand, the exclusion deprived plaintiff of potentially valuable evidence. Yet, the record suggests that prepared cross-examination and additional experts (who might elaborate, for example, on the fact that the tested samples all came from under floorboards in one area) could have reduced that value significantly.
On balance, we believe that the district court could reasonably have concluded that plaintiff improperly delayed disclosure, leading to surprise and prejudice, without any practical cure. Under these circumstances, we can find no "manifest" error in the district court's choice of sanction. See Alimenta (U.S.A.), Inc. v. Anheuser-Busch Cos., 803 F.2d 1160 (11th Cir. 1986) (affirming exclusion of expert testimony for failure to supplement disclosure where defendant allowed to depose expert only the day before trial, leaving inadequate time to prepare to meet his testimony); Smith v. Ford Motor Co., 626 F.2d 784 (10th Cir. 1980) (affirming exclusion of expert testimony where pre-trial disclosure did not reveal substance of testimony, resulting in surprise and prejudice to opposing party which continuance could not correct and which disrupted trial), cert. denied, 450 U.S. 918, 67 L. Ed. 2d 344 , 101 S. Ct. 1363 (1981); see also Jenkins v. Whittaker Corp., 785 F.2d 720 (9th Cir.), cert. denied, 479 U.S. 918, 93 L. Ed. 2d 296 , 107 S. Ct. 324 (1986); Holiday Inns, Inc. v. Robertshaw Controls Co., 560 F.2d 856 (7th Cir. 1977).
B. The Defendant's Samples. The facts still more strongly support the district court's Conclusion that plaintiff failed "seasonably" to tell the defendants about Buck's proposed testimony that their samples had been doctored. Plaintiff had known of the defendant's test results since the previous November. Plaintiff heard the district court state on February 11, the first day of trial, that "if an expert witness's testimony" had "not been clearly disclosed in a timely manner," it would not "permit counsel to disadvantage the opposition by bringing up something new . . . ." Yet, plaintiff said nothing about Buck's theory that gasoline had been added to the samples. Buck revealed that theory for the first time during his court-ordered mid-trial deposition on February 23. The plaintiff offered no satisfactory explanation for the delay. The district court could have concluded that plaintiff had not supplemented his disclosure "seasonably," indeed, that he had not made the required disclosure at all.
The sanction, too, seems appropriate. The potential prejudice to defendants seemed serious. They had planned, and tried, their case on the assumption that their expert's testimony, about gasoline in the samples, would strongly support their theory of the fire. They did not think they would have to meet contrary expert views about those samples. Had they known, they might well have conducted their cross-examination of plaintiff's witnesses differently. A continuance might have permitted them to prepare, but it still would have denied the defendants those benefits that a different course of cross-examination of earlier witnesses (or the weakness of plaintiff's case) might have produced.
On the other hand, exclusion as a sanction deprived the plaintiff of potentially useful evidence (although refutation by other experts might have significantly reduced its value). Nonetheless, on balance, we believe the district court could have found an unjustified failure by plaintiff to meet his discovery obligations, along with surprise and prejudice to defendants, and an inability to devise a fair, practical cure. The resulting sanction is well within the district court's discretion. See, e.g., United States v. Ladd, 885 F.2d 954, 959 (1st Cir. 1989); Sexton v. Gulf Oil Corp., 809 F.2d 167, 170 (1st Cir. 1987); and other cases cited at pp. 8 - 9, supra.
Plaintiff next claims that the district court should have permitted him to introduce three documents that he hoped would help rebut the testimony of a Fleetwood gas system expert, Lester McLaughlin, who had examined the RV after the fire. To understand the merits of his claim, one must first understand the trial context. McLaughlin had testified ("with a reasonable degree of professional certainty") basically as follows:
i) the gas tanks supplying the RV with propane were located outside the RV;
ii) the tanks fed propane into small, flexible pipelines that ran through the back of the RV into the RV's refrigerator;
iii) he had examined the pipelines and found no leaks;
iv) if the valve inside the refrigerator had leaked, the gas would simply have moved up a vent (running from the back of the refrigerator through the RV roof) connecting the inside of the refrigerator with the open air; and,
v) consequently, "the refrigerator did not" play "any role" in the explosion or fire.
On cross-examination, plaintiff established that McLaughlin's Conclusion rested on his belief (based on having seen the same kind of refrigerator in another RV) that the back of the refrigerator (containing machinery and the gas-connecting valve) was sealed and connected to the exhaust vent. Plaintiff's counsel then asked:
Q. You have come to the Conclusion there cannot be any leakage; is that your Conclusion?
A. It is very improbable. I mean, if the thing were hit in an accident and hit that seal; sure.
Q. Wait a minute. You admit there is some possibility first of all?
A. You could create a hypothetical situation, I suppose there could be some leakage.
At this point plaintiff asked the court for permission to use three documents (apparently for purposes of cross-examination and rebuttal) to show that propane, released by a faulty valve, might escape into the RV, rather than traveling up the vent into the open air. They included:
i) Several pages from plaintiff's deposition of John Moody, a Fleetwood industrial designer;
ii) correspondence and diagrams about a 1979 "voluntary recall" of 188 Fleetwood "travel trailers" with a "gas shut off valve" safety problem; and,
iii) correspondence and diagrams about a 1977 "voluntary recall" of 570 Fleetwood "Travel Trailers which may have Damaged Refrigerator Gas Valves."
The court, after taking a recess and reviewing the documents, held all three items inadmissible.
Plaintiff now asks us to decide whether the court acted within its legal powers. We believe that it did so. See, e.g., Fed. R. Evid. 403; cf. United States v. Malik, 928 F.2d 17, 19-20 (1st Cir. 1991) (citations omitted) (trial Judge has wide latitude to impose reasonable limits on cross-examination to avoid jury confusion).
a. The Moody Deposition. The plaintiff wished to use certain cited pages of the Moody deposition. He claims that those pages showed a Fleetwood "admission" that the gas could escape into the trailer. We have examined the pages, and they do not show any such admission. On those pages Mr. Moody says that he is a Fleetwood designer, that the "major structural difference" between Walsh's RV model (called "B-24") and "earlier models" was the "sidewall construction," and that he knew nothing about earlier recalls. On other pages, on other lines, which plaintiff tells us (but which he did not tell the district court) are relevant, Moody says,
Q. That [recall] information [that Moody had just said he knew nothing about] had nothing to do then with your design of the B-24?
What this has to do with the issue escapes us. The district court told counsel that he was "wasting the Court's time . . ., and I'm tired of it." In respect to the Moody deposition, that seems a fair comment.
b. The 1977 and 1979 Recall Documents. The relevance question, in respect to the recall documents, is a closer one. The 1979 documents say, in part, that a refrigerator "gas line" might "become loose from the thermostat, allowing gas to escape into the inside of the travel trailer," thereby causing an explosion or fire. The 1977 documents say, in part, that some refrigerators' "gas shut-off valves that have been damaged during installation in a "travel trailer" could cause "gas leakage into the refrigerator compartment." These statements tend to refute McLaughlin's opinion that gas leakage into Walsh's RV was unlikely, provided, of course, that they refer to the same kind of RV refrigerator. If they do not, if the systems to which they refer are not "substantially similar" to those in Walsh's RV, then introduction of the evidence might simply confuse the jury, or prejudice the defendants, without offsetting probative value. In that event, the district court could exclude the proposed evidence. McKinnon v. Skil Corp., 638 F.2d 270, 277 (1st Cir. 1981) (citations omitted).
The district court found that the systems in question lacked the requisite "similarity." And, that Conclusion has adequate support in the record. The recalled refrigerators were Magic Chef and Dometic brands; Walsh's refrigerator was a Norcold. The recalls involved towed trailers; Walsh's motor home was a self-propelled RV. The recalls involved vehicles made in 1977 and 1979; Walsh's RV was made in 1985. And, all the vehicles were manufactured in different states. Different brands of refrigerator, different kinds of vehicle, different years and places of manufacture -- taken together, these show dissimilarity sufficient to undermine the probative value of the offered documents.
Plaintiff tries to rehabilitate his offer of proof here by arguing that the Moody deposition (taken together with the recall documents) shows that "there was a continuous and substantial similarity of product" between the 1977 and 1979 travel trailer refrigerators and Walsh's. But, the cited pages say only that Moody had worked on the design of the earlier vehicles, that Walsh's RV reflected an "evolution" from the earlier vehicles, and that Moody had "incorporated" his earlier experience into his later design. These statements strike us as irrelevant, as does plaintiff's further suggestion that we compare diagrams of all the refrigerators. We have tried to do so, but we have found only Magic Chef and Dometic diagrams on the cited pages.
In sum, the decision to keep these documents from the jury falls squarely within the scope of the district court's trial-management powers.
At the close of plaintiff's case the district court found that plaintiff had not presented sufficient evidence that defendants CCI Controls and Manchester Tank had made, or distributed, the particular gas detector found in Walsh's RV. Consequently, it directed a verdict for those two defendants. Plaintiff argues that the district court failed to consider one important piece of evidence, namely CCI's original (but subsequently superseded) answers to plaintiff's interrogatories. Plaintiff says that CCI, in those answers, admitted that it made the gas detector in question.
In our view, the court properly decided not to treat the original answers as an "admission," for we cannot find any such admission in the interrogatory answers.
Plaintiff argues the contrary by pointing to interrogatory answers in which CCI said "yes," to questions asking, for example, whether it "designed the product." (Emphasis ours.) He notes that, in the introduction to its 94 interrogatories, he defined "the product" as
CCI Model # 7719-Pretell 2, which caused injury or damage to Timothy Walsh as alleged in plaintiff's pleading.
And, he concludes that CCI thereby admitted that its product "caused injury or damage to Timothy Walsh."
The flaw in this argument, however, is that it does not read the interrogatories fairly. The questions to which plaintiff points do not directly ask CCI to admit that the RV's gas detector was its product. Nor does the introductory definition, preceding the specific questions, suggest that such is the intent of every later question that uses the word "product." In fact, most of the specific questions, such as whether CCI had made
the following product: CCI gas detector, Model #7719 Pretell 2
seem to be about the products that CCI makes, not about the particular gas detector found in Walsh's RV. It is not reasonable to assume that, in answering such questions affirmatively, CCI was admitting anything about the particular detector. Indeed, when plaintiff asked CCI directly to admit that it had made Walsh's gas detector, it refused to do so.
Since the answers to which plaintiff points are not admissions, they are not relevant, and the district court properly refused to consider them.
Statute of Limitations
The district court dismissed plaintiff's complaint against Stolle Corporation, the refrigerator manufacturer, for a technical reason. Plaintiff filed his complaint on September 20, 1989, one day before Maine's two-year statute of limitations expired. Me. Rev. Stat. Ann. tit. 18-A, § 2-804(B). But, plaintiff named the wrong refrigerator defendant. Instead of Stolle, the plaintiff sued Norcold, a division of Stolle. Norcold is an entity that has no legal existence. It is nothing but a name.
By the following March plaintiff recognized his mistake. He amended his complaint, this time naming Stolle. By then, the statute of limitations had long since expired. But, plaintiff argues the amendment should "relate back" to the previous September. The courts should treat the complaint as if it had named the right defendant at the right time.
The Federal Rules of Civil Procedure specifically permit "relation back" of an amendment in order to cure this "wrong name" problem. Fed. R. Civ. P. 15(c). But, unfortunately for plaintiff, the relevant rule imposes several conditions, including one that plaintiff cannot meet. The Rule permits "relation back" of an amendment only when the named party (here Stolle) receives notice of the lawsuit within the limitations period. In this case, plaintiff did not provide notice to anyone (neither Norcold nor Stolle) until December 15, 1989, nearly three months after expiration of the limitations period.
Plaintiff argues that we should find some kind of exception, for the rule works a hardship. We cannot do so. The Supreme Court has held that this rule means what it says. In Schiavone v. Fortune, 477 U.S. 21, 91 L. Ed. 2d 18 , 106 S. Ct. 2379 (1986), the Supreme Court considered an almost identical fact situation -- a plaintiff sued a wrongly-named defendant, filing the suit on time, but serving the defendant after expiration of the limitations period -- and it held that a "renaming" amendment cannot "relate back." We cannot find any meaningful difference between Schiavone and this case.
In Schiavone, the plaintiff brought a libel action, within the statute of limitations period, against a defendant that it named in the caption of its complaint as "Fortune," which company the complaint described as a "corporation having its principal offices at Time and Life Building, Sixth Avenue and Fifth Street, New York City." In this case, plaintiff brought a tort action, within the statute of limitations period, against a defendant that it named as "Norcold," which company the complaint described as a "corporation . . . with a principal place of business in Sydney, Ohio."
In Schiavone, it turned out that Fortune had no independent corporate existence; it was "an internal division of Time, Incorporated." The magazine itself said, on page 2, "FORTUNE (ISSN 0015-8259), May 31, 1982, Vol. 105, No. 11. Issued biweekly by Time Inc., . . . Principal offices: Time & Life Building . . . . FORTUNE is a registered trademark of Time, Incorporated." In this case, it turned out that Norcold had no independent corporate existence; it is an internal division of Stolle Corporation. The first sentence of the Limited Warranty for the RV's refrigerator says, "This Limited Warranty is given by NORCOLD, Div. of The Stolle Corporation ("Company") . . . ."
In Schiavone, Time officials received the plaintiff's process (addressed to FORTUNE). The Court assumed this service gave Time actual notice of the lawsuit. But Time received process only after the statute of limitations period had run. In this case, we shall assume that Stolle officials received actual notice of the suit through receipt of the process addressed to Norcold. But Stolle received any such notice only after the statute of limitations period had run.
In Schiavone, the Court refused to permit a "relation back" of the change-of-party amendment because Rule 15(c) explicitly states that an
amendment changing the party against whom a claim is asserted relates back [to the date of the original pleadings, only if several conditions are satisfied, including the condition that] within the period provided by law for commencing the action against the party to be brought in by amendment that party . . . has received . . . notice . . . .
The Schiavone Court could find no notice to any party, Fortune or Time, "within the period provided by law for commencing the action." And, neither is it possible for us to find any such timely notice here.
Seeking to distinguish Schiavone, plaintiff tries to find a waiver. He claims that Stolle waived its objection to the relation back of his amendment when Norcold filed a paper on March 5, 1990, stating that plaintiff was "filing an agreed to amendment . . to correct the name of the defendant sued as Norcold to the Stolle Corporation." However, that very same piece of paper adds that "[o]nce that amendment has been granted by the Court, the Defendant Stolle Corporation will file its answer with affirmative defenses." And, when Stolle filed that answer, it asserted, as an affirmative defense, expiration of the limitations period. We therefore cannot find a waiver. Pessotti v. Eagle Manufacturing Company, 946 F.2d 974, 979 (1st Cir. 1991) (defendant did not forfeit statute of limitations defense, later raised in answer to amended complaint, by failing to contest court order allowing complaint to be amended). Plaintiff also makes several attempts to argue under Maine law that relation back is permitted or that the statute of limitations was tolled, but federal law controls here. Hanna v. Plumer, 380 U.S. 460, 14 L. Ed. 2d 8 , 85 S. Ct. 1136 (1965).
On our own motion, we have considered the effect of an amendment to Rule 15(c), that took effect on December 1, 1991, and which was expressly designed to change the result in Schiavone. The amendment provides that
amendment of a pleading relates back to the date of the original pleading when . . . the amendment changes the party or the naming of the party against whom a claim is asserted if . . . within the period provided by Rule 4(m) for service of the summons and complaint [i.e. 120 days after the filing of the complaint], the party to be brought in by amendment (A) has received such notice of the institution of the action that the party will not be prejudiced in maintaining a defense on the merits, and (B) knew or should have known that, but for a mistake concerning the identity of the proper party, the action would have been brought against the party.
(Emphasis ours). Since Norcold (and hence Stolle) received the summons and complaint on December 15, 1989, within 120 days of filing, this amendment could make a critical difference. We therefore, with the help of new briefs that we requested, considered whether the amendment should apply to this case.
Often an appellate court will apply the law in effect at the time of the appeal, Bradley v. Richmond School Board, 416 U.S. 696, 711-12, 40 L. Ed. 2d 476 , 94 S. Ct. 2006 (1974), quoting United States v. Schooner Peggy, 1 Cranch 102 (1801) (Marshall, C.J.), at least where Congress so intends. Kaiser Aluminum & Chemical Corp. v. Bonjorno, 494 U.S. 827, 110 S. Ct. 1570, 1577 , 108 L. Ed. 2d 842 (1990); Bowen v. Georgetown University Hospital, 488 U.S. 204, 109 S. Ct. 468, 471 , 102 L. Ed. 2d 493 (1988). But the court will not do so if that application would work a "manifest inJustice." Bradley, 416 U.S. at 716-21, citing Thorpe v. Housing Authority of Durham, 393 U.S. 268, 21 L. Ed. 2d 474 , 89 S. Ct. 518 (1969). We need not decide whether or not a procedural rule change taking effect after oral argument in a court of appeals normally should or should not apply retroactively to require a new trial, compare Atlantis Development Corp. v. United States, 379 F.2d 818 (5th Cir. 1967) with Hashemi v. Campaigner Publications, Inc., 784 F.2d 1581 (11th Cir. 1986), for we conclude that application of the amended rule in this case would work a "manifest inJustice."
For one thing, after reading the entire record, we find that plaintiff had a full and fair opportunity to present to the jury the very case he wishes us to allow him to present again, against Stolle. Yet the jury found against him. There is no reason to think that a second trial would produce a different result, and no reason to force Stolle, who "played by the rules" in effect at the time, to endure the expense and inconvenience of a trial likely to produce that outcome.
For another thing, we would likely have concluded appellate review of this case and affirmed the district court before the effective date of the new rule if the plaintiff had presented his arguments, all of them weak, in a more forthright manner, without numerous questionable citations to portions of the record that offer him no significant support. By requiring us to untangle his arguments and, in effect, construct his case for him, plaintiff has delayed the decision in this case, and that fact should not entitle him to an otherwise unattainable windfall due to a change in a procedural rule. For these reasons, we find a significant "public interest" in not permitting what would be, in effect, a retrial of the case. See New England Power Co. v. States, 693 F.2d 239, 245 (1st Cir. 1982) (considering public and private interests in determination of "manifest inJustice").
For the reasons given above, the judgments of the district court are
* Of the Fifth Circuit, sitting by designation.