dissenting:
I respectfully dissent.
Erasmus is a tough act to follow.
But even tougher is the Supreme Court of the United States.1
In Bessemer City, of course, the Court emphasized that the findings of fact of a district court, especially when the judge has heard the witnesses ore tenus in open court, shall not be set aside unless clearly erroneous, with due regard being given to the opportunity of the trial court to judge the credibility of the witnesses. In that decision, the Court reminded us of several principles that should be repeated here: “a reviewing court [must not] reverse the *471finding of the trier of fact simply because it is convinced that it would have decided the case differently,” 470 U.S. at 573, 105 S.Ct. at 1511; “appellate courts must constantly have in mind that their function is not to decide factual issues de novo,” 470 U.S. at 573, 105 S.Ct. at 1511; “[w]here there are two permissible views of evidence, the factfinder’s choice between them cannot be clearly erroneous,” 470 U.S. at 574, 105 S.Ct. at 1511; and “[t]his is so even when the district court’s findings do not rest on credibility determinations, but are based instead on physical or documentary evidence or inferences from other facts.” 470 U.S. at 574, 105 S.Ct. at 1511.
In deciding this case, we must couple the deferential standard of review of factual findings set forth in Bessemer City with the law, unmentioned in the majority opinion, that a finding of abandonment is a factual finding subject to the clearly erroneous standard. Nunley v. M/V DAUNTLESS COLOCOTRONIS, 863 F.2d 1190, 1198 (5th Cir.1989); see also Bunge Corp. v. Agri-Trans Corp., 542 F.Supp. 961, 969 (N.D.Miss.1982), aff’d in part, vacated in part on other grounds, 721 F.2d 1005 (5th Cir.1983); Williamson v. Mennella, 248 A.D. 911, 290 N.Y.S. 645 (1936); The Port Hunter, 6 F.Supp. 1009, 1011 (D.Mass.1934); De Bardeleben Coal Co. v. Cox, 16 Ala.App. 172, 76 So. 409, 410 (1917), cert. denied, 200 Ala. 553, 76 So. 911 (1917); Steinbraker v. Crouse, 169 Md. 453, 182 A. 448 (1936); H.B. Chermside, Jr., Annotation, Rights In and Ownership of Wrecked or Derelict Vessels and Their Contents Not Cast Upon the Shore, 63 A.L.R.2d 1369, 1373 3[b]; 1 C.J.S. Abandonment § 11; City of New London v. Pequot Point Beach Co., 112 Conn. 340, 152 A. 136, 138 (1930) (real property); Bennett v. Bowers, 238 Iowa 702, 28 N.W.2d 618, 620 (1947) (real property); Ray Coal Mining Co. v. Ross, 169 Iowa 210, 151 N.W. 63, 65 (1915) (lease); Cadillac Oil & Gas Co. v. Harrison, 196 Ky. 290, 244 S.W. 669, 671 (1922) (lease); Anson v. Tietze, 354 Mo. 552, 190 S.W.2d 193, 196 (1945) (real property); Newman Signs, Inc. v. Hjelle, 317 N.W.2d 810, 817 (N.D.1982) (signs); Llewellyn v. Philadelphia & Reading Coal & Iron Co., 308 Pa. 497, 162 A. 429, 430 (1932) (personal property); McMillin v. Titus, 222 Pa. 500, 72 A. 240, 244 (1909) (lease); Jordan v. State, 107 Tex.Crim. 414, 296 S.W. 585, 587 (1927) (personal property); Garrett v. South Penn Oil Co., 66 W.Va. 587, 66 S.E. 741, 745 (1909) (lease). I am convinced that the majority decision is incorrect in its principal holding that the district court’s factual finding of abandonment was clearly erroneous; in applying the law of salvage to a long lost wreck; in making factual findings as an appellate court; and in erroneously stating the issue to be decided by the district court on remand. In addition, I would hold that the district court did not abuse its discretion in attaching conditions to the untimely motions for intervention.
Finding of Abandonment
In my view, throughout its opinion, the majority has disregarded the rules set forth by the Supreme Court in Bessemer City. The majority, in contravention of the mandate that “[wjhere there are two permissible views of evidence, the factfinder’s choice between them cannot be clearly erroneous,” reverses the district court’s factual finding as clearly erroneous simply because, in my opinion, “it would have decided the case differently.” Bessemer City, 470 U.S. at 573-74, 105 S.Ct. at 1511-12. ■
The district court found as a fact that the property taken from the CENTRAL AMERICA had been abandoned by its owners. I would hold that the district court was not clearly erroneous in making that determination. The circumstances of the insurance companies’ failure to retain any indicia of ownership, such as bills of lading or commercial invoices, or to make any attempt to recover the property since 1858, coupled with the passage of more than 100 years, is ample evidence (not to mention details) to support the district court’s finding that any claim the insurance companies may have had in the cargo of the CENTRAL AMERICA has been abandoned.
The majority holds that the district court committed clear error in finding abandon*472ment because it was “unable to find the requisite evidence that could lead a court to conclude that the underwriters affirmatively abandoned their interest in the gold,” and specifically, the majority could not “find any evidence that the underwriters intentionally or deliberately destroyed any of their documents about the CENTRAL AMERICA.” Majority op. at 467-468. In my opinion, both the district court’s finding of abandonment and its specific finding that the underwriters intentionally destroyed the documents are amply supported by the evidence.2 The ■ majority’s holding, by necessary implication, must be that abandonment can only be shown by an express statement by the owner indicating his intention to abandon. This holding, in my opinion, cannot be reconciled with established precedent. Abandonment may be inferred from all of the relevant facts and circumstances and may be determined on the basis of circumstantial evidence. See Halteras, Inc. v. The U.S.S. Hatteras, 1984 A.M.C. 1094, 1097 n. 5 (S.D.Tex.1981) (“A formal declaration is not necessary; abandonment may be inferred from acts and conduct of an owner clearly inconsistent with an intention to return to the property, and from the nature and situation of the property.”); City of New London v. Pequot Point Beach Co., 112 Conn. 340, 152 A. 136, 138 (1930) (real property) (“intent [to abandon] may be inferred as a fact from the surrounding circumstances”); Llewellyn v. Philadelphia & Reading Coal & Iron Co., 308 Pa. 497, 162 A. 429, 430 (1932) (personal property) (“intention [to abandon] may, and indeed often must be, inferred from acts”); Pierce v. Bemis (The Lusitania), [1986] 1 Q.B. 384, 389 (“So far as the owners of the contents are concerned, it is a necessary inference from the agreed facts and from the lapse of 67 years before any attempt was made to salve the contents that the owners of the contents abandoned their property”); 1 Am.Jur.2d Abandoned, Lost, and Unclaimed Property § 40 (“it is not necessary to prove intention to abandon by express declarations or other direct evidence”).
Testimony was presented, for example, that the practice of the marine insurance industry is to require that the insured present a commercial invoice and a bill of lading prior to a claim being paid, and that the practice in that industry is much the same today as it was in 1857. The insurance company then retains those documents as proof of ownership with potential salvors. None of the insurers in this action produced any invoices or bills of lading regarding shipments on the CENTRAL AMERICA. (Atlantic Mutual — J.A. 991, 999; London Assurance and the Alliance Assurance Co. — J.A. 1054,1069; Royal Exchange Assurance — J.A. 1105-1105A; Indemnity Marine Assurance Company — J.A. 1131; Salvage Association — J.A. 1153; Insurance Company of America — J.A. 1195). The marine insurance companies involved routinely destroy their records on a periodic basis. However, they do not destroy documents so long as they retain an interest in the cargo of which the documents reflect own*473ership.3 It is doubtlessly expensive to maintain all records which an insurance business generates, therefore the companies have made the business decision not to keep documents which are unnecessary. The obvious, most logical, and clearly permissible inference from the fact that the insurance companies did not retain any documents pertaining to the CENTRAL AMERICA is that, at some point during the past hundred and thirty-odd years, the companies, giving up any expectation of ever recovering the property, destroyed their documents and, in so doing, evidenced their abandonment of their claims to the CENTRAL AMERICA.4 See Wyman v. Hurlburt, 12 Ohio 81, 87 (1843) (vessel is abandoned when “all hope, expectation, and intention to recover the property were utterly and entirely relinquished” by the owner); U.S. v. Smiley, 27 F.Cas. 1132, 1134 (Cir.Ct.N.D.Cal.1864) (Justice Field) (cargo from shipwreck was abandoned when “all efforts to recover the property had been given up” by the owners.)
Indeed, I further submit that the specific finding that the insurance companies destroyed their documents is amply supported in the record by both direct and circumstantial evidence. The following testimony of one Richard Smith, Atlantic Mutual’s representative at trial, is sufficient, in and of itself, to support the district court’s finding:
Mr. Davey (Attorney for Columbus-America): You don’t have any of the policies today that may have covered goods on board the CENTRAL AMERICA, do you?
Mr. Smith: At present we do not.
Mr. Davey: To the best of your knowledge, they have all been routinely destroyed along with the other business records of the company?
Mr. Smith: That’s correct.
■jC * * * }}: *
Mr. Davey: So that the bills of lading which would have been taken up as part of the proof of loss have been destroyed as a matter of course?
Mr. Smith: If they had been taken up, they probably would have been destroyed in that and/or we moved several times. They could have been missing in that. Probably in the destruction- when we destroyed the claim files if they existed. They would have been destroyed.
Mr. Davey: But any record of sale would have been destroyed in the ordinary course of your business just like the insurance policies, the bills of lading and commercial invoices and everything else? Mr. Smith: Yes, sir, if one existed.
Evidently, the majority did not consider this testimony, as it states that the district court held that the documents were intentionally destroyed “because an insurance executive testified that the usual practice today is for insurance companies to destroy worthless documents after five *474years_” Majority op. at 466 (emphasis supplied by the majority). Even that contention of the majority is also incorrect, I submit, because testimony was presented, not just as to the current practice, but as to the document destruction policies of the companies for more than 40 years. The insurance companies’ representatives testified that, for as long as they had worked for the companies (which in some cases was over 40 years), the companies had document destruction policies in place.5
The majority’s holding that “it seems as, if not more, likely that the documents were lost or unintentionally destroyed, rather than being intentionally destroyed,” illustrates the majority's misconception of our role on appeal. It is not our role to decide a case on appeal based on which factual inference we would draw from the evidence. See Bessemer City, 470 U.S. at 573-74, 105 S.Ct. at 1511-12. Rather we should only determine if the factual inferences drawn by the district court are supported by the evidence. In addition to the clear testimony of Atlantic Mutual’s representative that the policies covering goods on the CENTRAL AMERICA were destroyed and that the bills of lading were probably destroyed, the evidence showed that not one of the numerous insurance companies involved in this litigation was able to produce any of the documents that would have to be produced prior to a claim being paid, and that the companies destroy documents on a routine basis. In my opinion, these facts fully support the inference drawn by the district court, but rejected by the majority, that the documents were intentionally destroyed.
In addition, the passage of over 130 years since the cargo was lost is strong circumstantial evidence of abandonment.6 And, as the First Circuit has stated in Martha’s Vinyard, 833 F.2d at 1065, when “articles have been so long lost that time can be presumed to have eroded any realistic claim of original title, and ... those articles are now in hand, having been actually recovered, the better-reasoned authorities agree that the law of finds may appropriately be utilized.” See Chance v. Certain Artifacts Found and Salvaged, 606 F.Supp. 801, 804 (S.D.Ga.1984), aff’d without opinion, 775 F.2d 302 (11th Cir.1985) (“Under the law of finds, the inference of *475abandonment may arise from lapse of time and nonuse of the property ... since THE NASHVILLE has rested at the bottom of the Ogeechee, unclaimed by its owner since 1863, this court concludes that the law of finds applies.”) As the amount of time since the loss of the ship increases, the strength of the presumption that the property has been abandoned, stated above, increases, I suggest. In this ease, more than 130 years is sufficient to establish a very strong presumption that the property has been abandoned.7
At the risk or repetition, I again state that I think the majority’s out-of-hand dismissal as clearly erroneous the district court’s finding of abandonment is quite contrary to law, for the inferences and factual conclusions made by the district court are supported by the evidence. Indeed, even as to lack of weight, the weakness of the evidence relied upon by the majority is exposed by only casual analysis. The majority writes that “almost all of the evidence in the record actually seems to indicate a specific predisposition on the underwriters’ part not to abandon the treasure.” Majority op. at 467. However, all the majority points to in support of this proposition is evidence of salvage negotiations in the 1860’s involving two of the insurers, the fact that Atlantic Mutual, in its official history, written in 1967, devoted a couple of pages to the tragedy, and the insurers’ actions from the late 1970’s to the present upon being contacted by potential salvors concerning their potential claims. Majority op. at 467-468.
Evidence of salvage negotiations from the 1860’s is, of course, of little consequence in determining if the insurers abandoned their claims subsequent to the 1860’s. And the very fact that the 1860’s efforts at salvage were either abandoned or not undertaken is, indeed, evidence of abandonment. In fact, it is highly likely that at the time the insurers explored the 1860’s salvage options, they possessed all the documentation concerning their claims. They knew those documents would be necessary to establish the amounts of their claims. It was probably not until after the salvage operations proved unsuccessful that the insurers abandoned their claims. They destroyed the only documentation that would be necessary to prove the amounts of their claims when, at a later time, they lost all hope of recovering the property. In so doing, the district court held there was evidence of abandonment, and I agree.
The fact that Atlantic Mutual mentioned the CENTRAL AMERICA tragedy in its official history is, in my view, no evidence of a “specific predisposition” not to abandon the property. It indicates no more for Atlantic Mutual than that Atlantic Mutual, which believed it was an insurer of some of the cargo, found the sinking of the CENTRAL AMERICA to have been an interesting historical matter in the company’s history. Indeed, the fact that that history shows that Atlantic Mutual has even now retained some papers referring to the CENTRAL AMERICA, but not the papers necessary to prove a claim or even the amount of the loss, is evidence that the matter was later abandoned.
Finally, evidence of the insurers’ actions concerning their potential claims from the 1970’s to the present is of little consequence in determining whether the insurers abandoned their claims prior to the 1970’s. Property, once abandoned, cannot be reclaimed absent possession. See 1 C.J.S. Abandonment § 12. It is apparent that *476the insurers abandoned the property because at the time they did so, they had no hope of ever recovering it. Clearly, once the technology became available following World War II to make recovery a possibility, the insurers would not agree that they had abandoned any claim. This explains their activity in the recent past and their establishment of files. However, since property, once abandoned, cannot be reclaimed absent possession, in my view their 1970’s and subsequent activity is of little consequence in determining whether they abandoned their claims at some point during the previous 120 years. A property interest abandoned as valueless cannot be revived simply because new technology, some 120 years later, has made the abandonment seem to have been error. In fact, an equally or more valid inference than the one drawn by the majority is that the insurance companies' creation of files in the 1970’s and '80’s, pertaining to the CENTRAL AMERICA, supports the conclusion that they had abandoned their claims.8 These files were created when potential salvors contacted the insurance companies in the 1970’s and '80’s.9 Had each insurer not abandoned its claim earlier, there would have been no need to create a file because one would have existed since 1857. By reasonable inference, each insurer believed the cargo it insured on the CENTRAL AMERICA was irretrievably lost and any property interest it had in the cargo was worthless. As a consequence, each insurer abandoned what to them was a worthless interest in the cargo.
Also, it is interesting that, even after the technology for recovering the CENTRAL AMERICA became available, the insurers did not enter into any salvage contracts concerning its recovery. They did nothing to recover the property, just as they had done for the past 130-plus years.
As I have shown, the evidence offered by the majority provides little support for its argument that the insurers did not abandon their claims. In my opinion, the majority’s substitution of its judgment as to the facts is even more improper when, as here, there is ample support in the record for the district court’s factual finding of abandonment.
Finds v. Salvage
I disagree with the majority’s decision to apply the law of salvage to the facts of this case. It appears to base its decision on the erroneous belief that courts favor applying salvage law rather than the law of finds. Majority op. at 460. In fact, in the context of long lost wrecks, such as the instant case, courts, almost without exception, apply the law of finds. Martha’s Vineyard Scuba Headquarters v. Unidentified, Wrecked and Abandoned Steam Vessel, 833 F.2d 1059, 1065 (1st Cir.1987) (ocean liner sunk in 1909); Klein v. Unidentified Wrecked and Abandoned Sailing Vessel, 758 F.2d 1511, 1514 (11th Cir.1985) (English ship sunk in eighteenth century); Treasure Salvors, Inc. v. Unidentified Wrecked and Abandoned Sailing Vessel, 569 F.2d 330, 337 (5th Cir.1978) (Spanish galleon sunk in 1622); Sub-Sal, Inc. v. The DeBraak, No. 84-296, 1992 WL 39050, at * 1-2, 1992 U.S. Dist. LEXIS 2461, at *6-7 (D.Del.1992) (British ship sunk in 1798); Zych v. Unidentified, Wrecked and Abandoned Vessel, 746 F.Supp. 1334, 1343-44 n. 12 (N.D.Ill.1990), vacated, 941 F.2d 525 (7th Cir.1991) (steamer sunk in 1860); Jupiter Wreck, Inc. v. Unidentified Sailing Vessel, 691 F.Supp. 1377, 1385 (S.D.Fla.1988) (Spanish galleon sunk in seventeenth century); Indian River Recovery Co. v. The China, 645 F.Supp. 141, 144 (D.Del.1986) (ship sunk in nineteenth century); Chance v. Certain Artifacts Found and Salvaged from the NASHVILLE, 606 F.Supp. 801, 804 (S.D.Ga.1984) (Confederate raider sunk in 1863); Wiggins v. 1100 Tons, More or Less, of Italian Marble, 186 F.Supp. 452, *477456 (E.D.Va.1960) (Norwegian ship sunk in 1894); Commonwealth v. Maritime Underwater Surveys, Inc., 403 Mass. 501, 531 N.E.2d 549, 552 (1988) (pirate ship sunk in 1717); see also Schoenbaum, Admiralty and Maritime Law § 15-7, p. 514 (1987) (“In virtually all of the treasure salvage cases involving wrecks of great antiquity, the law of finds, not salvage, is appropriate ... ”) Even Benedict on Admiralty, from which much of the majority’s opinion appears to have been derived, would apply finds law to “long lost wrecks.” See 3A Norris, Benedict on Admiralty, § 158, p. 11-17 (1992).
Modern courts have rejected application of the law of salvage in the context of long lost wrecks because, under the law of salvage, even an abandonment does not divest the owner of title. See 3A Norris, Benedict on Admiralty, § 150, p. 11-1 (1992) (“Should a vessel be abandoned without hope of recovery or return, the right of property still remains in her owner.”); Sub-Sal, Inc. v. The DeBraak, No. 84-296, 1992 WL 39050, at * 1-2, 1992 U.S. Dist. LEXIS 2461, at * 6-7 (D.Del.1992) (“Cognizant that the application of salvage law involves a presumption that the owner of an abandoned vessel has not been divested of title, this Court rejects the salvage law theory.”). The major premise of salvage law, that the property to be salved is owned by persons other than the salvor, is inapplicable in the case of long lost wrecks. See Schoenbaum, Admiralty and Maritime Law, § 15-7, p. 512 (1987).
The majority opinion breaks with the law of every other circuit that has decided the question and holds salvage law to be applicable to long lost wrecks. See Martha’s Vineyard Scuba Headquarters v. Unidentified, Wrecked and Abandoned Steam Vessel, 833 F.2d 1059, 1065 (1st Cir.1987) (ocean liner sunk in 1909); Klein v. Unidentified Wrecked and Abandoned Sailing Vessel, 758 F.2d 1511, 1514 (11th Cir.1985) (English ship sunk in eighteenth century); Treasure Salvors, Inc. v. Unidentified Wrecked and Abandoned Sailing Vessel, 569 F.2d 330, 337 (5th Cir.1978) (Spanish galleon sunk in 1622). I agrt.e with our sister circuits that “[disposition of a wrecked vessel whose very location has been lost for centuries as though its owner were still in existence stretches a fiction to absurd lengths.” Treasure Salvors, Inc., 569 F.2d at 337.
.Impermissible Factfinding
Also, the majority has disregarded well-established law that appellate courts should not engage in factfinding. The majority has decided, on appeal, that the insured shipment was worth $1,219,189 (1857 valuation) when the amount of the insured shipment was not an issue before the district court. Majority op. at 465 n. 6.
The trial before the district court was to be bifurcated. The first portion was to decide if the insurers or anyone else could establish a claim to portions of the cargo. The second part of the trial was to involve a determination of the amount to be awarded to any claimants who established they had a valid claim. However, the second part of the trial was never held since the district court determined that no one other than Columbus-America had established a right to any of the cargo. Therefore, the question of how much of the shipment was insured was not before the district court and should not be decided in the first instance by this court.
In addition, the majority finds as a fact that the insurers had established an ownership interest in the cargo. Majority op. at 465-466. The majority attempts to disguise its factfinding by attributing a holding to the district court that was, in fact, never made by it. The majority writes that “the district court did not err when it held that the underwriters who are now parties, or their predecessors in interests, paid off claims upon and became the owners of the commercial shipment of gold in 1857.” Majority op. at 465. The district court did not hold that the underwriters had paid claims and become owners of the commercial shipment. The district court did not decide the factual issue of whether the insurers had proven that they had an ownership interest in the cargo because the district court clearly was satisfied, as am I, that in any *478event the cargo had been abandoned.10 Since the district court did not decide the factual question of whether the insurers had established that the claims were ever paid, it is entirely improper for this court to make that factual ruling at this juncture. On remand, the district court should be given an opportunity to fully address the issue of whether the insurers met their burden of proof in establishing an ownership interest in the cargo and, in its proper role as a finder of fact, to decide that issue in the first instance.
Issue on Remand
In my view, the majority has incorrectly stated the issue to be determined on remand. It writes that the district court “must determine what percentage of the gold each underwriter insured.” Majority op. at 468. This ignores the fact that the insurers only became owners of cargo to the extent of their payment of claims. The amount they insured is irrelevant. The crucial fact is how much cargo they actually paid claims on. The burden is on the insurance companies to establish the amount of claims that they actually paid.
Intervention Motion
I also disagree with the majority’s holding that the district court abused its discretion by allowing Columbia University, Harry John, and Jack Grimm to intervene on the eve of trial only on the condition that the trial go ahead as scheduled. The trial was set for April 3, 1990. On March 29, 1990, John and Grimm filed their motion to intervene. Columbia filed its motion on April 2, 1990. It would have been well within the district court’s discretion to deny the motions to intervene as untimely. See Gould v. Alleco, Inc., 883 F.2d 281, 286 (4th Cir.1989), cert. denied, 493 U.S. 1068, 110 S.Ct. 870, 107 L.Ed.2d 953 (1990) (“This Circuit has stressed the importance of timeliness and the wide discretion afforded the district courts.”); NAACP v. New York, 413 U.S. 345, 365-66, 93 S.Ct. 2591, 2602-03, 37 L.Ed.2d 648 (1973) (district court’s ruling on timeliness will not be disturbed on appeal unless the court abused its discretion). In determining whether a motion to intervene is timely, courts look at how far the suit has progressed, the prejudice which delay might cause other parties, and the reason for the tardiness in moving to intervene. Gould, 883 F.2d at 286. In this case, the motions for intervention were filed on the eve of trial after all pre-trial discovery had been performed by the parties then in the suit. Any delay in the commencement of the trial would have prejudiced Columbus-America as its ability to attract funding was directly impacted by resolution of the issues concerning ownership of the cargo, and key members of the recovery team were required at trial and a delay in the trial would have interfered with Columbus-America’s recovery efforts. Finally, there was no excuse for the inter-venors’ tardiness in moving to intervene. They were put on constructive notice, at the latest, on September 7, 1989, when Columbus-America published its Rule C notice in the Virginian Pilot and Ledger *479Star stating the coordinates of the area where the CENTRAL AMERICA was found. Moreover, the record reveals that Columbia has been aware of, and indeed a part of, this litigation almost from the first moment that Columbus-America first set foot in court. Less than two months after filing its original in rem complaint on July 17, 1987, Columbus-America sought and obtained a preliminary injunction forbidding other would-be recoverers of the treasure from an area then thought to contain the wreck of the CENTRAL AMERICA. See Columbus-America Deep Search, Inc. v. The Unidentified Wrecked and Abandoned Sailing Vessel, Civil No. 87-363-N (E.D.Va. July 17, 1987); App. 53. Not only was Columbia a party in that proceeding and specifically addressed in the district court’s order, it appealed and unsuccessfully sought a stay of the injunction in this court. See order of August 3, 1987 entered in our case no. 87-3606, which appeal was later dismissed by stipulation on December 18, 1987. The fact that the CENTRAL AMERICA later was found within the area established by the injunction of August 1989, see Civil No. 87-363-N (E.D.Va. Aug. 18, 1989), does not later the inescapable fact that, at the latest in July of 1987, Columbia had actual knowledge of Columbus-America’s assertion of rights over the very wreck that was the subject of Dr. Ryan’s work and of the very lawsuit in which Columbia later sought to intervene on the eve of trial. See 742 F.Supp. 1327, 1334 (E.D.Va.1990); App. 438. Given these facts, it would not have been an abuse of discretion for the district court to have denied the motions for intervention as untimely.
However, rather than deny the motions, as it could have done, the district court conditionally permitted the intervention on the intervenors’ willingness to go forward with the trial at its previously scheduled time. Conditions can be imposed on an intervenor even when he intervenes as a matter of right under Rule 24(a). See Stringfellow v. Concerned Neighbors in Action, 480 U.S. 370, 383, 107 S.Ct. 1177, 1185, 94 L.Ed.2d 389 (1987) (Brennan, J., concurring) (“restrictions on participation may also be placed on an intervenor of right”); United States v. South Florida Water Management Dist., 922 F.2d 704, 710 & n. 9 (11th Cir.), cert. denied sub nom. Western Palm Beach County Farm Bureau, Inc. v. United States, — U.S. -, 112 S.Ct. 407, 116 L.Ed.2d 356 (1991); Harris v. Pernsley, 820 F.2d 592, 599 n. 11 (3d Cir.1987), cert. denied sub nom. Castille v. Harris, 484 U.S. 947, 108 S.Ct. 336, 98 L.Ed.2d 363 (1987); United States v. Hooker Chemicals & Plastic, 749 F.2d 968, 992-93 n. 22 (2d Cir.1984); Morgan v. McDonough, 726 F.2d 11, 14 (1st Cir.1984). The conditions imposed by the district court were not onerous considering the untimeliness of the motions for intervention. In fact, the court even allowed the intervenors to perform limited discovery during trial.
If the intervenors were unhappy with the conditions imposed, they had the option of declining the court’s offer. Had they done so, their motion to intervene would have been denied and that denial could have been appealed to this court. .The majority’s holding thus allows the intervenors two days in court instead of the timehonored one. Because they accepted the court’s conditions, they were able to participate in the trial. At trial, they lost. Now the majority’s holding will allow them to try their case a second time. I would hold that the district court did not err in conditioning John, Grimm, and Columbia’s untimely intervention on their willingness to participate in the trial at its previously scheduled date. The district court, by being overly accommodating to the intervenors, and not requiring them to take the case as they found it, if intervention was granted, as it had every right to do, is now wrongfully found guilty of error, I submit, for its perfectly proper rulings.
The effect of the majority’s decision is to allow a prospective intervenor to delay a motion to intervene until the eve of a trial it has known about for months, and then force either a delay in the trial or error. I submit that Rule 24(a) properly admits of no such result.
*480 Conclusion
In sum, I would affirm the district court’s judgment. I would hold that the district court was not clearly erroneous in finding that the insurance companies had abandoned all claims they may have had to the cargo in the CENTRAL AMERICA and that the district court did not abuse its discretion in placing conditions on granting of the untimely motions to intervene.
. Anderson v. Bessemer City, 470 U.S. 564, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985).
. The majority makes much of the fact that the documents may have been lost or stolen. As I have stated, I think the evidence establishes that the most likely inference is that the documents were destroyed, in the regular course of business, as worthless. In any event, I believe resolution of the question of whether the documents were destroyed or lost is unnecessary to a finding of abandonment. The insurance companies affirmatively abandoned any property interest they had in the cargo of the CENTRAL AMERICA by not maintaining the indicia of ownership they may have had at one time pertaining to the cargo on the CENTRAL AMERICA. We cannot know, more than 100 years after the fact, why the relevant documents are no longer in the possession of the insurance companies. However, we do know that any documentation that may have existed was in the control of the insurance companies and they no longer have these documents. Had they believed that a situation would arise where they would need to prove that they had paid claims on the cargo, they certainly would have taken steps to maintain and protect the documents pertaining to the CENTRAL AMERICA. The fact that the companies do not know when the documents were last in their possession or the circumstances of their removal is sufficient evidence that they viewed the documents as worthless. The obvious inference to be drawn from the fact that the insurance companies did not maintain their documents or any record of their documents is that they had abandoned their claims to the property-
. Testimony at trial concerning the companies’ document retention policies was typified by an insurance company representative who was asked, "And if you thought there was any hope of recovering goods, you would not let those documents be destroyed, would you?” He replied, “No, I wouldn't.”
London Assurance’s representative testified that a marine insurance company would keep the documents for so long as the company retained an interest in lost cargo. Representatives for Royal Exchange Assurance and Indemnity Marine Assurance each testified that their company’s policy was to retain documents for so long as it feels they are necessary and disposed of only when they serve no purpose whatsoever.
. The majority writes that because of the long passage of time, "it seems as, if not more, likely that the documents were lost or unintentionally destroyed, rather than being intentionally destroyed." Maj. op. at 466. Not only is this a factfinding by a court of appeals, opposite to the inference drawn by the district court and contrary to Bessemer City, the majority also puts the onus on the finders to prove that the insurance companies’ destruction of documents pertaining to the CENTRAL AMERICA was deliberate. In the absence of any showing to the contrary, I would place the burden on those who last had possession of the documents to establish that they were "lost or unintentionally destroyed.” Indeed, standard learning is that the failure of a party who possesses a document to produce it permits an inference that its tenor is unfavorable to the party’s cause. See Missouri, K. & T. Ry. Co. v. Elliott, 102 F. 96, 101-103 (8th Cir.1900); Wigmore on Evidence, Chadbourn (1979) 285, 291.
. Alan John Birch, who appeared as a representative of claimants London Assurance and the Alliance Assurance Company, began in the marine claims business 43 years ago for London Assurance. He testified that the longest period of document retention that he was aware of at the company was 10 years, and that so long as the company retained an interest in lost cargo, it would not destroy the documents with respect to the loss.
Richard Smith, Atlantic Mutual’s representative, has worked for the company for 30 years. He testified that during his entire tenure at Atlantic Mutual, the company has had a document destruction policy, under which it destroys claim files and other documents after from five to 10 years.
. I would hold that long-lost wrecks such as the one in question here are presumptively abandoned. See Martha’s Vineyard Scuba Headquarters, Inc. v. Unidentified, Wrecked and Abandoned Steam Vessel, 833 F.2d 1059, 1065 (1st Cir.1987) (law of finds applies when “articles have been presumptively abandoned, i.e., either affirmatively renounced, or so long lost that time can be presumed to have eroded any realistic claim of original title....") (emphasis added). "Disposition of a wrecked vessel whose very location has been lost for centuries as though its owner were still in existence stretches a fiction to absurd lengths.” Treasure Salvors, Inc. v. Unidentified Wrecked and Abandoned Sailing Vessel, 569 F.2d 330, 337 (5th Cir.1978). In Treasure Salvors, the Fifth Circuit held that disposition of the treasure retrieved from a Spanish galleon, the NUESTRA SEÑORA DE ATOCHA, sunk in 1622 and found near the Marquesas Keys in the Gulf of Mexico, should be governed by the law of finds rather than salvage law.
Under the law of finds, title is vested in "the first finder lawfully and fairly appropriating it and reducing it to possession, with the intention to become its owner.” 569 F.2d at 337, quoting Rickard v. Pringle, 293 F.Supp. 981, 984 (E.D.N.Y.1968). Salvage law provides that the original owner maintains ownership and the recoverer only receives a salvage award.
The majority attempts to distinguish Treasure Salvors from the instant case on the fact that the parties in Treasure Salvors had stipulated that the ATOCHA had been abandoned by its original owner, the Spanish government. Nothing in the court’s opinion indicates that the decision would have been any different had the Spanish government appeared in court to assert a claim. The parties merely stipulated to an obvious fact, that the long lost ship was abandoned.
. The difficulty in adjudicating rights concerning cargo on a ship lost for 135 years is underscored by the Herculean task facing the district court on remand. Contrary to the majority’s assertion that "[w]ithout doubt, most, if not all, of the claims were promptly paid off by the underwriters,” at 456-57, there is doubt as to how much of the cargo was actually insured and, just as importantly, which insurance claims were actually paid. The only evidence available is speculative • newspaper accounts from the 1850’s. On remand, the district court will be faced with the difficult, if not impossible, task of determining which claims were paid by the insurance companies and in what amounts. This problem is caused by the insurance companies’ failure to maintain the records necessary to establish their claims. Had the insurance companies not abandoned their claims they surely would have maintained the records necessary to establish their amount.
. For the most part, these files contained nothing more than newspaper clippings and letters from potential salvors.
. The fact that would-be salvors contacted some of the insurance companies is, of course, of no consequence in determining whether the property was abandoned. A reasonable person might seek to avoid litigation with potential claimants by obtaining assignments of any possible claims they may have.
. As evidenced by the following passage in the district court’s opinion, the court clearly was unimpressed with the evidence presented by the insurance companies as to ownership:
Insurance Companies rely solely upon the newspaper accounts for evidence that they were insurers, the name of the insured, the amount of the insurance, the value of the shipment, whether a claim was made, what sum was supposedly paid, and to whom. They well must rely on the newspaper articles for they have no copies of any insurance policies, invoices for shipments, proof of loss, amounts paid or other records. Without the newspaper accounts, the evidence is totally insufficient to establish any claim of the insurance companies.
Columbus-America Discovery Group, Inc. v. Unidentified, Wrecked and Abandoned Sailing Vessel, 742 F.Supp. 1327, 1341 (E.D.Va.1990).
The majority appears to have concluded that since the district court stated that the evidence was "totally insufficient” without the newspaper accounts, the court was implicitly stating that the evidence was sufficient with those accounts. I do not agree with this conclusion. I believe the district court was merely stating that without the newspaper accounts there was ho evidence to support the insurance companies’ claims and with the accounts there was some evidence. The cpurt did not find that the insurance companies had met their burden of proof in establishing that they had paid claims and therefore had acquired an ownership interest in the cargo.