TABLE OF CONTENTS
I. INTRODUCTION 561
II. THE INTERVENORS’ CLAIM 562
A A Tripartite Alliance 562
B. Ryan’s search 568
C. The Columbus-Ameriea Effort 564
D. From the Control Room to the Courtroom 565
1. The Intervenors’ Proprietary Interest in the CONRAD Data 566
2. Columbus-Ameriea’s Alleged Misuse of the Data 567
III. THE SALVAGE AWARD 568
A The Doctrine of Unclean Hands 569
1. The CONRAD Data Revisited 569
2. Columbus-America’s Disclosures to the District Court 569
B. The Moiety Rule 570
C. The BLACKWALL Analysis 571
1. The Labor Expended by the Salvor 571
2. The Salvor’s Promptitude, Skill, and Energy 571
3. The Value of the Property Employed by the Salvors 572
*5614. The Risk Incurred by the Salvors Oí .-q tO
5. The Value of the Property Saved C7i «q tO
6. The Degree of Danger to the Salvaged Property Or -q tO
7. The Salvor’s Preservation of the Historical and Archaeological Value of the Wreck and Cargo CH —q CO
8. Other Considerations ch -q 0^
IV. THE MARKETING RULING Cn «o ^
V. EVIDENCE OF THE UNDERWRITERS’ OWNERSHIP Oí *-q CR
VI.CONCLUSION üi ~q Oi
Before RUSSELL, WIDENER, and HALL, Circuit Judges.Affirmed and remanded by published opinion. Judge HALL wrote the opinion, in which Judge RUSSELL and Judge WIDENER joined.
OPINION
K.K. HALL, Circuit Judge:Quid, non mortalia pectora cogis,
Ami sacra fames!1
—Virgil
I.
INTRODUCTION
In September 1988, a long and difficult search ended for Columbus-America Discovery Group soon after the crew aboard the recovery vessel ARCTIC DISCOVERER had lowered NEMO, Columbus-America’s aptly named undersea robot, almost one and one-half miles into the depths of the Atlantic Ocean. As NEMO’s camera scanned wood, coal, and other debris along the ocean floor, it suddenly revealed the distinctive side-wheels of the S.S. CENTRAL AMERICA, a nineteenth-century passenger ship. A mighty hurricane had sunk the CENTRAL AMERICA about 160 miles east of Charleston, South Carolina, on September 12, 1857, nearly four years before the first shots were fired on Fort Sumter. The doomed ship had taken with her more than 400 lives and over $1 million worth of California gold — an astounding sum in those days.
Columbus-America commenced an in rem proceeding in the district court to establish its rights regarding the CENTRAL AMERICA, and, in 1989, moved the court to declare that it owned the gold and other artifacts aboard the ship. British and American insurers and their successors-in-interest opposed the motion, each asserting that it had underwritten the risk of loss of a portion of the gold and had paid claims filed in the aftermath of the disaster.2 The district court scheduled a trial of the matter to begin on April 3, 1990.
On the very eve of trial, the district court permitted Columbia University, Jack F. Grimm, and Harry G. John to intervene as of right.3 In 1983-84, these three had cooperated in an effort to locate the CENTRAL AMERICA. During the search, sonar aboard the research vessel CONRAD imaged a target very near to where Columbus-America ultimately found the CENTRAL AMERICA. The intervenors asserted a proprietary interest in the information generated as a result of the CONRAD expedition, and they alleged that Columbus-America improperly obtained and relied on that information to make its discovery.
At the trial’s conclusion, the district court ruled that the underwriters had abandoned any interest in the gold, and Columbus-America was therefore entitled to keep it all. The court also found that there was no evidence to support the intervenors’ claim. The underwriters and the intervenors appealed, and a divided panel of this court held that the district court erred in applying the law of finds, rather than the law of salvage, to determine the parties’ respective rights in the gold. Columbus-America Discovery *562Group v. Atlantic Mut. Ins. Co., 974 F.2d 450, 468 (4th Cir.1992), cert. denied, — U.S. -, 113 S.Ct. 1625, 123 L.Ed.2d 183 (1993) (CADG I).4 We held farther that the court had abused its discretion by refusing the intervenors an opportunity for discovery pri- or to trial. Id. at 470. We remanded the ease to the district court to apply the law of salvage and determine an appropriate award for Columbus-America. We also instructed the court to conduct a new trial on the inter-venors’ misappropriation claim, after allowing them a reasonable time for discovery.
On remand, the district court determined that Columbus-America was entitled to a salvage award of ninety percent. By separate order, the district court denied the underwriters’ request to divide the gold in specie so that each party could market its own share; the court instead assigned to Columbus-America the task of selling all of the gold, with the proceeds to be thereafter divided among it and the underwriters. The district court, apparently uncertain of the scope of our remand, declined to make specific findings as to whether any or all of the underwriters had sufficiently proved what, if any, portion of the gold aboard the CENTRAL AMERICA each had paid claims on and thereby owned as subrogees. As for the intervenors’5 claim, the district court again held the proof insufficient to establish that Columbus-America had availed itself of any of the data gathered during the CONRAD expedition.
The underwriters and the intervenors again appeal (Nos. 94-1105 and 94-1200). In light of the district court’s decision to defer, pending this appeal, an examination of the validity of each underwriter’s ownership claim, Columbus-America cross-appeals (No. 94-1106), urging us to direct the court to put the underwriters to their proof. Because various matters respecting the continuing rights and responsibilities of the parties remain to be decided in the district court, jurisdiction over these appeals lies in this court under 28 U.S.C. § 1292(a)(3).6
We now affirm the district court’s judgment in nearly all respects. We hold that sufficient evidence supports the court’s finding that Columbus-America did not misappropriate the CONRAD data. We also agree with the district court that, in return for recovering the gold aboard the CENTRAL AMERICA, Columbus-America deserves a ninety percent salvage award. We hold further that the court correctly decided that the gold should be marketed as a whole, with Columbus-America designated as the marketing authority.
However, that Columbus-America is entitled to a ninety percent salvage award does not automatically mean that the underwriters may, merely by virtue of bringing suit, walk out the courthouse door with the remaining ten percent. We therefore instruct the district court, on remand, to consider the evidence and decide whether each underwriter actually owns the portion of the gold that it claims.
II.
THE INTERVENORS’ CLAIM
A. A Tripartite Alliance
On September 16, 1983, Thomas G. Thompson of Columbus, Ohio, flew to Mil*563waukee at the behest of Harry G. John to discuss locating and salvaging the CENTRAL AMERICA. Thompson, a research scientist and engineer for the Battelle Memorial Institute’s laboratories in Columbus, became interested in the CENTRAL AMERICA in the 1970s. John knew of Thompson’s interest and that Thompson had researched the shipwreck’s potential location. Thompson, for his part, wanted to organize a search for the CENTRAL AMERICA and knew that John had money to invest.
As a result of their meeting, John, on behalf of his company, Santa Fe Communications, Inc., commissioned Thompson to conduct further research on the CENTRAL AMERICA and submit the names of persons who might be equipped to look for it. In late 1983, John and Thompson flew to Seattle to meet with Michael Williamson, an expert in sonar surveys whose skills Thompson hoped to employ in the search. Together, the three traveled to International Submarine Technology in nearby Redmond to inspect its manufacturing facilities and discuss the technical aspects of a new generation of side-looking sonars, known as the SeaMARCs,7 that 1ST had developed.
Thompson had provided John with alternative contacts, including Columbia University’s Lamont-Doherty Geological Observatory, home to one of the three largest institutions for oceanographies in the United States. In 1980, Lamont-Doherty, in collaboration with 1ST, built the first SeaMARC. The SeaMARC I allowed Columbia’s oceanographers to obtain high resolution sonar imaging of the ocean floor and objects on it in a swath some ten times wider than the preexisting technology had permitted. Lamont-Doherty frequently accepted research grants to map areas of the ocean floor, but the demand for SeaMARC I’s services meant that prospective clients were ordinarily forced to endure at least a two-year wait.
The wait was two years, that is, unless the client was Texas oilman Jack F. Grimm. In 1980, Grimm, who was then searching for the TITANIC, had given Lamon1>-Doherty approximately one-third of a million dollars to purchase the SeaMARC I and other equipment; in return, Grimm received the right to use the sonar for five years. John learned of this arrangement and contacted Grimm, who consented to help John negotiate with Columbia’s trustees for a better place in line. These negotiations were a success; John, on behalf of Santa Fe, agreed to fund a $300,000 grant to have Lamont-Doherty survey 400 square miles of ocean floor with the SeaM-ARC I aboard the research vessel ROBERT D. CONRAD.
B. Ryan’s search
The CONRAD expedition, under the direction of Dr. William Ryan, a Senior Research Scientist at Lamonb-Doherty, set out from Norfolk, Virginia, on February 26,1984, and docked in Bermuda two weeks later. During the trip, the SeaMARC I imaged seven targets on the ocean floor within the survey area.
Ryan interpreted the data gathered by the SeaMARC, and, by letter dated April 27, 1984, informed John of the results. Ryan concluded that, of the seven targets, “[o]nly one, target #4, is a candidate feature that correlates with the expected distribution of wreckage of the S.S. CENTRAL AMERICA. ... Target # 4 at 2180m (7200 feet) water depth is 260 to 300 feet long and of very high reflectivity.” Ryan noted that the size estimate was consistent with the CENTRAL AMERICA’S 280-foot length. He wrote:
[TJhis target is almost certainly the scattered debris of a shipwreck.... I would speculate that the cluster of debris upwind from target # 4 could be the materials cut loose from the decks, spars, etc. of the S.S. CENTRAL AMERICA in the attempt to try to fuel the boilers.... In my opinion, you and Mr. Jack Grimm have a likely candidate for further exploration. As you know, we were unable to undertake any photography ... after completing the sonar operations due to gale force winds and seas. Target # 4 is sufficiently well located (to plus or minus one nautical mile) to make it sensible and cost-effective to next *564use ... submersibles for the subsequent stage of target development.
The “subsequent stage” was never reached. A scant eight months later, Santa Fe deeded “all its right, title, and interest to any and all claims arising out of its interests in undersea salvage operations, including ... its partnership interest in Central America Limited Partnership and ... research performed by Columbia University incident thereto” to Catholic monks of the Capuchin Order at St. Benedict Friary in Milwaukee.8
C. The Columbus-America Effort
Meanwhile, Thompson’s interest in the CENTRAL AMERICA had not diminished, though he and John had parted ways soon after their 1983 encounter. Thompson continued to research historical accounts of the disaster, but also began to compile information from more contemporary sources. He gave particular attention to how hurricanes affected navigational accuracy at sea, as well as the effect of such storms on drift variables, such as prevailing currents and wind direction, that might determine the path of a swamped vessel whose crew was far less concerned about steering than just staying afloat.
Thompson was still determined to mount his own search for the shipwreck, and, when his proposal to John for financing fell through, he solicited venture capital from private investors who lived, for the most part, in and around Columbus. In May 1985, Thompson and some of his investors formed Recovery Limited Partnership, with Thompson as the general partner. RLP’s assets derived solely from its ongoing sale of interests in the partnership; these funds were used to finance Thompson’s company, Columbus-Ameriea Discovery Group, Inc., in its continuing efforts to locate and salvage the CENTRAL AMERICA.
Thompson was aware that John had contracted with Lamont-Doherty to conduct a survey in early 1984. He also knew, just one day after the CONRAD’s arrival in Bermuda, that the survey had failed to positively identify the CENTRAL AMERICA.9 Thompson contacted Ryan at Lamont-Doherty soon thereafter to discuss sonar in general and the SeaMARC in particular, and how the data from the SeaMARC was analyzed to enable the user to recognize and define anomalies, or targets, on the ocean floor. Between 1984 and 1986, Thompson and Ryan had numerous conversations concerning various sonar searches that Lamont-Doherty had conducted using the SeaMARC I, culminating in Thompson’s February 12, 1986, request for copies of the sonar record and contour map data from the 1984 CONRAD survey.
Ryan passed on Thompson’s request to his Division Administrator, who, on March 19, 1986, wrote a letter to Thompson at Battelle granting the request on the condition that the data would be for Thompson’s sole use and not reproduced for others, because “the data ... is not in the public domain.” The letter also stated that Columbia would not relinquish its exclusive right to publish the data until its graduate students who were working with the data had published their dissertations. Thompson agreed to Columbia’s terms and received the data, which, in addition to the sonar photographs and the contour chart, included the CONRAD’s ship *565track lines, indicating the vessel’s path within the survey boundaries.10
Columbus-America hired Williamson’s company to perform a survey, using the SeaMARC IA, during June and July 1986. Thompson and the other project directors worked with Dr. Lawrence D. Stone, a search theorist, to create a “probability map” from which an efficient search plan could be devised. The map divided almost 1500 square miles of ocean into two-mile by two-mile cells and assigned each cell a probability. The numbers in the cells reflected a mosaic of sorts — an amalgam derived from the three major eyewitness accounts of the disaster, with each scenario accorded a greater or lesser weight depending on a subjective determination of its likely accuracy. The calculations were then refined by introducing the drift variables that Thompson had earlier researched.
The search plan was designed to survey all of the high- and medium-probability cells, and as many low-probability cells as possible. During the survey, however, Williamson had difficulty adhering exactly to the course dictated by the plan. Like a gardener who has overestimated the cutting width of his lawnmower, Williamson was forced to make additional sonar runs between the planned track lines to ensure that there were no “holidays,” or gaps, in the survey. It was near the terminus of one of these extra runs, very close to the southwest extreme of the probability map and in a cell with an assigned probability of less than one in one thousand, that the sonar recorded an image of what turned out to be the CENTRAL AMERICA
Of course, because a sonar image has appreciably less resolution than a conventional photograph, positive identification of any of the targets discovered during the survey had to wait until cameras could be lowered into the depths. The following summer, Thompson’s group went to sea aboard the ARCTIC DISCOVERER. The group, however, did not immediately proceed to the “correct” target; they instead traveled to a similar target in a more promising, high-probability cell over thirty miles to the northeast. After recovering several lumps of coal, along with some pottery and other artifacts from the site, Columbus-America received an injunction from the district court granting them the exclusive right to operate in the immediate area.
Columbus-America explored the site throughout the summer of 1987, but was unable to obtain any proof of the wreck’s identity. That winter, the group reexamined the survey data and noticed similarities between the target they had been exploring and the target thirty miles to the southwest. When, during the summer of 1988, Columbus-America lowered NEMO to the ocean floor near this new target, the robot’s almost immediate revelation of the ship’s sidewheels provoked an exuberant uproar in the ARCTIC DISCOVERER’S control room. Not long thereafter, the group found the proof that it had been seeking — the bell of the CENTRAL AMERICA.
D. From the Control Room to the Courtroom
Unfortunately for Columbus-America, the CENTRAL AMERICA turned out to be located very near to where the CONRAD had, in 1984, imaged Target # 4 along the eastern edge of the Lamont-Doherty survey. The intervenors assert that Columbus-America improperly used the materials that Ryan sent Thompson in 1986 to locate the shipwreck. Upon remanding this case for a new trial, we advised that “[sjhould it be found that Columbus-America used the interve-nors’ information and such use contributed to the discovery of the CENTRAL AMERICA, a salvage award [for the intervenors] could be proper.” CADG I at 470.
On retrial, the district court concluded that Columbus-America did not rely on the CONRAD data to find the CENTRAL AMERICA. In weighing the evidence, the court was acting pursuant to its role as factfinder, *566and we may set aside the court’s finding only if it is clearly erroneous. Fed.R.Civ.P. 52(a); see McAllister v. United States, 348 U.S. 19, 20, 75 S.Ct. 6, 7-8, 99 L.Ed. 20 (1954) (“No greater scope of review is exercised by the appellate tribunals in admiralty cases than they exercise under Rule 52(a)_”). A finding is clearly erroneous when, “although there is evidence to support it, the reviewing court on the entire evidence is left with a definite and firm conviction that a mistake has been committed.” Id. (citations omitted); see also Concrete Pipe & Prod. of Calif., Inc. v. Constr. Laborers Pension Trust, — U.S. -, -, 113 S.Ct. 2264, 2279, 124 L.Ed.2d 539 (1993).
To prove that Columbus-America misappropriated the CONRAD data, logic dictates that the intervenors had to demonstrate (1) that Columbus-America obtained the data, (2) that the intervenors had a proprietary interest in the data, and (3) that Columbus-America, without authority, actually used the data in a manner inconsistent with the intervenors’ proprietary interest. Further, we are of the opinion that, as a prerequisite to receiving a salvage award, the intervenors were required to show also that the data materially assisted Columbus-America in -finding the CENTRAL AMERICA.
Columbus-America readily admits that it received, through Thompson, the sonar photographs, contour charts, and track lines from the CONRAD expedition; thus, the first part of the test is satisfied. The district court determined, however, that the CONRAD information, contrary to the assertions made by Lamont-Doherty’s Division Administrator in her March 1986 letter, see Section II-C, supra, was in the public domain, and, therefore, the intervenors could claim no proprietary interest in it. The court also found that Columbus-America made no use of the data; that the data were of no material assistance is, of course, a corollary to that finding.11
The district court’s ultimate finding that Columbus-America did not misappropriate the CONRAD data can be clear error only if all three of the finding’s factual underpinnings are likewise clearly erroneous. To the contrary, we believe that the court’s findings are sound.
1. The Intervenors’ Proprietary Interest in the CONRAD Data
Lamont-Doherty had not published the CONRAD data and, indeed, had clearly reserved its exclusive right to initial publication. Though Ryan testified that, pursuant to a preexisting agreement between Columbia and the government, he probably submitted a report of the survey to the Office of Naval Research, the evidence indicated that the report would not, for national security reasons, have been susceptible to disclosure under the Freedom of Information Act. Because there is no evidence that, as of March 1986, the CONRAD data were otherwise freely available, it is unlikely that the information was actually in the public domain.12
It does not necessarily follow, however, that the intervenors had a proprietary interest in the data; in fact, we shall see that they did not. As the district court pointed out, Article VII of the survey agreement between Columbia’s trustees and John and Grimm provided that “[i]t is agreed that Columbia will be free to publish the results of this *567research without restriction, but not less than one year after termination of contract.” The one-year restriction was added to the agreement by John and Grimm, and both initialed the entire provision. The agreement terminated, according to its terms, on September 30, 1984, and, as previously detailed, see Section II-C, supra, Thompson did not request the CONRAD data until February 1986, well after the one-year restriction had expired.
Had Columbia decided to publish the CONRAD data at that time, John and Grimm, according to the agreement’s unambiguous terms, would have had no say in the matter, and thus could not have been heard to complain had Thompson gone to his local library and availed himself of the information. That Columbia instead “published” the data directly to Thompson does not alter the inescapable conclusion that, if any party had a proprietary interest in the use of the CONRAD data, it was Columbia — and Columbia has already settled its claim. See note 5, supra. The intervenors’ resultant lack of a proprietary interest in the data provides an independent ground for our affirmance of the district court’s judgment.
2. Columbus-America’s Alleged Misuse of the Data
The district court considered the testimony of Thompson and nine of his associates involved in searching for the CENTRAL AMERICA. Although Thompson, Williamson, and Evans each testified that he had seen the CONRAD data before or during the 1986 survey, see note 10, supra, all denied making any use of the data; the other seven, each an important member of the search team, all denied having seen the data prior to 1990. As all, or nearly all, of the Columbus-America witnesses testified in person before the court at one time or another, it Vitas afforded an ample opportunity to evaluate their comportment and demeanor, and, despite some inconsistencies in Thompson’s testimony,13 the district court expressly found Columbus-America’s witnesses to be credible in the main.
Absent extraordinary circumstances, we will not disturb a factfinder’s credibility determinations. Fed.R.Civ.P. 52(a) (“due- regard shall be given to the opportunity of the trial court to judge ... the credibility of the witnesses.”); see United States v. Saunders, 886 F.2d 56, 60 (4th Cir.1989); see also Concrete Pipe & Products, supra:
[I]n the usual ease, the factfinder is in a better position to make judgments about the reliability of some forms of evidence than a reviewing body acting solely on the basis of a written record of that evidence. Evaluation of the credibility of a live witness is the most obvious example.
— U.S. at -, 113 S.Ct. at 2280.
Against the uniform denials of the Columbus-America witnesses, the intervenors offered a theory of the case that hinged on what could charitably be called circumstantial evidence. The intervenors asked the district court to believe, in essence, that Thompson contacted Ryan in early 1984 with an eye toward gaining the latter’s confidence so that, a full two years later, Thompson could casually request data that would definitively reveal the location of the CENTRAL AMERICA — though the intervenors themselves had not acted on this “definitive” information,14 and had, in fact, felt so little *568stirred by it that they gave all their rights in it to a midwestern monastery.
The intervenors further suggested that, in a master stroke of deception, Columbus-America spent millions of dollars to devise an enormously intricate — but nevertheless pre-textual — search plan and conduct a bogus sonar survey for several weeks during the summer of 1986; of course, it was relatively easy to manufacture an excuse for not precisely following the recorded plan. According to the intervenors, Thompson and his cohorts then intentionally searched thirty miles away from the CENTRAL AMERICA’S actual location for several months in 1987. Finally, the piece de resistance: Columbus-America, despite having fought off other would-be salvors who attempted to violate their court-protected search area, waited over a year to return so that the shipwreck could appear to have been “accidentally” found somewhere outside the search area.15
While we would not foreclose all possibility that persons with the demonstrated intelligence, initiative, resourcefulness, and determination of those associated with Columbus-America could be capable of such Machiavellian deceit, we are far from being left with “a definite and firm conviction” that the district court made a mistake by finding to the contrary. We therefore affirm the court’s judgment in favor of Columbus-America on the intervenors’ claims.16
III.
THE SALVAGE AWARD
We remanded this ease so that the district court could, through the exercise of its equitable powers, “determine the proper salvage award for Columbus-America.” CADG I at 468. We suggested that the award might be expressed as a percentage of the total recovery, and we instructed the court that it could adjust the award upward or downward to account for the merit of the salvor’s services in relation to the property salvaged. See id. at 468-69. We neither interposed a floor nor erected a ceiling to limit the district court’s discretion to set an award, except to opine that “we are hazarding but little to say that Columbus-America should, and will, receive by far the largest share of the treasure,” and, later, “[f|rom a distance, it appears that ... considerations militate towards awarding Columbus-America a significant portion of the recovered gold.” Id. at 468. We nevertheless made it quite clear that it was the dis*569trict court’s job to take the evidence and make an appropriate decision.
To inform the court’s analysis, we described the six factors, announced by the Supreme Court in THE BLACKWALL, 77 U.S. (10 Wall.) 1, 13-14, 19 L.Ed. 870 (1869), that have been traditionally applied by admiralty courts to determine salvage awards.17 In light of the shipwreck’s age, we directed the district court to consider a seventh factor: the degree to which the salvors have worked to protect the historical and archaeological value of the wreck and items salvaged.
On remand, the district court considered the BLACKWALL factors and concluded that Columbus-America’s reward for salvaging the underwriters’ property should be ninety percent of the value of that property. Our review of that decision is limited, as salvage awards are left to the district court’s wide discretion. Waterman S.S. Corp. v. Dean, 171 F.2d 408, 411 (4th Cir.1948), cert. denied, 337 U.S. 924, 69 S.Ct. 1168, 1171, 93 L.Ed. 1732 (1949). The amount “is primarily a matter of judgment to be exercised by the trial eourt[,] and[,] beyond a careful examination of the facts[,] little remains for the appellate court except to determine whether the judgment has been exercised in accordance with the general principles respecting salvage....” Id.
In the face of this extremely deferential standard of review, the underwriters nevertheless appeal what they believe to be an excessive award. We will examine the district court’s findings, but we must first briefly address two arguments posed by the underwriters that, if correct, would have circumscribed the court’s discretion from the outset.
A. The Doctrine of Unclean Hands
“The salvor who seeks a reward for his services in rescuing property on navigable waters[ ] must act in entire good faith and with honesty of purpose. He must come into court with clean hands.” 3A Martin J. Norris, Benedict on Admiralty § 99 (7th ed.1993) [.hereinafter Benedict]. The underwriters maintain that the salvage award — regardless of the amount finally approved — should be reduced, because, they allege, Columbus-Ameriea (1) deceitfully obtained and used the CONRAD data, see Section II, supra, and (2) misrepresented to the district court that it had found the CENTRAL AMERICA within the original injunction box in 1987 so that it could retain its exclusive right to search the area, thereby purchasing the additional year it needed to finish building NEMO.
1. The CONRAD Data Revisited
To decide whether Columbus-America acted in good faith regarding the CONRAD data, we employ an analytical framework much like the one we used to decide the intervenors’ misappropriation claim, see Section II-D, supra. The difference is that we focus more closely on the question of whether Columbus-Ameriea culpably intended to use, and did use, the data in a manner inconsistent with any colorable proprietary interest that it perceived the intervenors to have, and far less closely on the intervenors’ actual rights in the data and whether the information materially assisted Columbus-America’s efforts.
Although the analytical framework may be somewhat different, we nevertheless reach the same conclusion. The district court’s finding that Columbus-Ameriea made no use of the CONRAD data is based on what it reasonably perceived to be credible evidence and, hence, is not clearly erroneous.
2. Columbus-America’s Disclosures to the District Court
The underwriters’ second assertion that Columbus-Ameriea lacked “honesty of purpose” concerns the latter’s alleged manipulation of the district court to keep other potential salvors from reaching the ship*570wreck first. Like the intervenors, the underwriters contend that Williamson’s 1986 survey was a sham, intended only to obfuscate Columbus-America’s true purpose — to use the CONRAD data to find Target # 4, then obtain a high-resolution sonar image to verify that it was the CENTRAL AMERICA.
During the winter of 1986-87, according to the underwriters, Columbus-America learned that other potential salvors were planning to survey the area. This potential competition was a problem, because, though Columbus-America had the ability to lower cameras into the deep water, it was not yet able to recover any objects from the ocean floor.
The solution, according to the underwriters, was for Columbus-America to conduct additional sham operations the following summer and, by informing the court that it had almost certainly found the CENTRAL AMERICA at a particular site — when the wreck was actually far distant — obtain exclusive rights to an injunction area. Any injunction area would suffice, because the appearance that the wreck had been found, lent credibility by the court’s decree, would serve to dissuade potential competitors. Only in the summer of 1988, when NEMO had finally been constructed and tested, could Columbus-America risk traveling to the CENTRAL AMERICA’S actual location to obtain the evidence of provenance necessary to secure a new injunction.
The district court rejected the underwriters’ theory. It instead credited the testimony of the Columbus-America witnesses that several targets imaged during the 1986 survey were good candidates for further exploration. The court accepted that Thompson and the rest had sincerely believed that they had located the shipwreck in 1987, and that, until 1988, when the group reprocessed its sonar data using new software with improved image processing capabilities, what turned out to be the CENTRAL AMERICA had previously been thought to be a sonar image of a geological formation.
The district court’s conclusion was based upon its belief that the witnesses for Columbus-America were credible and truthful. The record, viewed in its entirety, does not demonstrate that the court’s belief was an unreasonable one. Its account of the evidence was plausible, and is thus beyond our power to reproach. See Davis v. Food Lion, 792 F.2d 1274, 1277 (4th Cir.1986) (“[W]e can find no clear error if there are two permissible views of the evidence, and the district court as factfinder chooses one over the other.”) (citing Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 573-74, 105 S.Ct. 1504, 1511-12, 84 L.Ed.2d 518 (1985)). We therefore affirm the district court’s rejection of the underwriters’ “unclean hands” theory.
B. The Moiety Rule
In a typical salvage ease, a seaworthy vessel renders assistance to one that has become disabled, but has not yet sunk. Depending on the severity of the trouble, the salvor may tow the disabled vessel back to port, it may take possession of perishables or other cargo, or it may even rescue the crew and passengers from imminent peril.
Many years ago, when the cash value of a salvaged ship’s cargo was difficult to ascertain, distributions in kind were more common, and courts consequently eschewed attempts to evaluate, on a case-by-case basis, the value of the services rendered by an individual salvor. See Benedict, § 240. The moiety rule emerged. It awarded a successful salvor, as a matter of course, one-half of the imperiled goods salvaged from a derelict vessel.
Most courts found this rule unsatisfactory, because it failed to account for the particular salvor’s skill, the labor expended, and the relative difficulty of the operation. Id. One authority has stated that the rule’s abandonment
is usually explained by the fact that, as steam replaced sail, the value of ships became so great that awards of a moiety would have been absurdly inflated.... [T]he award will never be for more than half of the value of the property, so that the moiety has become a ceiling instead of a floor. In fact, except where the property saved is of trifling value, the award of anywhere near 50% would be exceptional.
*571Grant Gilmore & Charles L. Black, Jr., The Law of Admiralty, § 8-10 (2d ed.1975) [hereinafter Gilmore & Black].
The underwriters urge that the moiety rule curbs the district court’s discretion in this case to impose a salvage award of greater than fifty percent. We disagree. What the underwriters have failed to appreciate is that, even long after “steam replaced sail,” the ability of salvors to rescue sunken cargo was severely limited by their inability to delve beyond only the most shallow reaches of the ocean. The vast majority of salvage operations were then conducted at sea level, and, even today, full-blown salvage of a shipwreck in even moderately deep water (let alone one and one-half miles) is a rare occurrence. See Benedict, App. D (documenting American salvage cases from the eighteenth century to the present); Gilmore & Black, § 8-2 (“The prototypical act is rescuing a ship in peril at sea and towing her to a place of safety.”).
While it may offend our sensibilities to allow a salvor who has provided but minimal towing services to receive one-half of a distressed ship’s cargo, especially where that cargo is particularly valuable,18 we are in no sense offended by the possibility that a sal-vor who has “boldly gone where no one has gone before” — indeed, where no one could have gone when the moiety rule unfolded— might be entitled to the lion’s share of a long-lost treasure. We recognized almost fifty years ago that the moiety rule, in lieu of a reasoned analysis of the circumstances surrounding a particular case, cannot serve as a floor on a salvage award. See note 18, supra. Today we recognize that the rule is equally inapt as a ceiling.
C. The BLACKWALL Analysis
We now begin our review of the district court’s exercise of its equitable judgment in accordance with the general principles respecting salvage. See note 17, supra, and accompanying text.
1. The Labor Expended by the Salvor
Addressing the first BLACKWALL factor, the district court found the following:
In addition to surveying more than 1,400 [square] miles of ocean bottom, an area the size of Rhode Island, hundreds of runs were made to recheck objects and to photograph them. The salvors spent 487 days at sea working shifts of 12 hours each. Personnel of [Columbus-America] logged some 411,295 hours of labor from 1985 through 1992, at a cost for labor of $8,421,-733.66. This figure does not record the fact that Thompson had been doing research since the early 1970[ ]s and Evans for some period prior to 1985.
Our research has revealed only two other sunken property cases in which it was reported that the salvor’s operations lasted longer than a month.19 The vast majority of such eases involve operations lasting only a few hours or days. See Benedict, App. D § 7. Even if a salvor had expended but a fraction of the labor detailed in the above account, that fraction alone would represent a huge investment. In the actual case, it is apparent that Columbus-America’s effort was monumental.
2. The Salvor’s Promptitude, Skill, and Energy
Regarding the second BLACKWALL factor, the district court had this to say:
A considerable amount of equipment was specially designed, manufactured and assembled for the project. Selection of the personnel to assist was from those most highly qualified and experienced in the contemplated activities- From the inception of the project, [Columbus-America] had assembled experts in the various fields of science, archeology, history, maritime, marketing and publicity, to advise *572and assist in the use, care and preservation of the shipwreck site.... It is sufficient to say they are the leaders in their field[s] of science and calling.
When we write an opinion, our tendency to emphasize those discrete facts necessary to our legal conclusions sometimes serves to distort the broader perspective. Were we to allow that to occur in this case, it would be a disservice to Columbus-America and the significance of its accomplishments. Hence, a brief review is in order.
Thompson, a man of modest means, became fascinated with the CENTRAL AMERICA and its history and resolved to find it. When he could not obtain financing from traditional sources, he made his pitch locally, stringing together a syndicate of small investors. Faced with searching in waters well over a mile deep, he hired a company with access to some of the best sonar equipment available. To find an exceedingly small needle in a dauntingly large haystack, he painstakingly researched the available information and consulted a leading expert in mathematical search theory. Upon discovering that no existing machine would allow an efficient recovery of items from the ocean bottom while simultaneously maintaining the integrity of even the smallest of those items, Thompson’s response was, in effect, “No problem — we’ll build one.” Thompson and Columbus-America indeed “built one,” and it worked.
In short, we cannot imagine anyone demonstrating more diligence, skill, and energy than Columbus-America has shown here. Its efforts provide a standard against which all others should be judged.
3.The Value of the Property Employed by the Salvors
The record evidence is that NEMO, the ARCTIC DISCOVERER, and the rest of Columbus-America’s recovery equipment are worth in excess of $6 million. Columbus-Ameriea rented additional equipment and incurred related expenses of millions more. We once more refer to salvage cases involving sunken property, of which — we must again point out — there are relatively few, and note that no ease reports a salvor to have employed property of a comparable value. See Benedict, App. D. § 7.
4. The Risk Incurred by the Salvors
This factor is often important where a distressed vessel’s immediate peril likewise imperiled its attempted savior. This scenario was not present here. Nevertheless, the crew of the ARCTIC DISCOVERER constantly dealt with heavy equipment — not the least of which was the six-ton NEMO — and the vessel’s 160-mile distance from shore meant that treatment for the most severe injuries was hours away.
While many salvage operations have occurred at a considerable distance from shore, none appear to have been conducted for such an extended period. As the district court noted, “[ajnytime a ship goes to sea, there is danger to the ship and the persons aboard.” Although the risk to the crew of the ARCTIC DISCOVERER at any particular time may not have been unusually great, the 487 days that it spent at sea multiplied the intrinsic risks accordingly.
5. The Value of the Property Saved
The testimony of the trial experts varied substantially as to the precise value of the CENTRAL AMERICA’S cargo; however, it now appears that the treasure’s value will not approach one billion dollars, as earlier speculated. See, e.g., CADG I at 454. It is nevertheless clear that, when the salvage is completed and all the treasure is eventually sold, the haul will be one of the largest in history. One commentator has suggested that “[t]he combination of high values and highly meritorious service should result in a high award.” Benedict, § 258. We entirely agree.
6.The Degree of Danger to the Salvaged Property
As we have pointed out, this is not the typical case where a rescuer chances upon a vessel in immediate peril. The gold and other cargo aboard the CENTRAL AMERICA came to rest about one-and-one-half miles beneath the ocean, and, because of the prevailing conditions, was found to be *573relatively free of accumulated sediment, notwithstanding that 131 years had passed.
Whether property that is greatly endangered because it is about to sink becomes any less imperiled once it actually slips below the ocean’s surface — with all possibility of an immediate recovery disappearing along with it — is an interesting question. The argument for an affirmative answer appears to be that property progresses from being endangered, as it is sinking, to some state of oblivion once it has sunk. The question of the degree of danger posed to the property would then be a matter of evaluating the conditions within the oblivion once the property is rediscovered.
We reject this argument. While it is true that the ocean itself presents no danger to the essential nature of gold and similar substances, it is also true that any value that our society attributes to gold depends entirely on the ability of someone to assert a property interest in it. Because property is far less certain of being recovered once it has sunk, especially when it has sunk in deep water, we perceive that its sinking sharply increases the degree of danger to its continued existence and utility as property. We have little doubt that, if the BLACKWALL Court were transported over 125 years into the future to decide this case, it would consider Columbus-America’s salvage of the CENTRAL AMERICA to be the ultimate rescue from the ultimate peril.
7. The Salvor’s Preservation of the Historical and Archaeological Value of the Wreck and Cargo
The district court was convinced that Columbus-America had taken extraordinary care in preserving the CENTRAL AMERICA, benefiting a range of sciences and disciplines:
[Columbus-America’s expert] witnesses described the care and means used to preserve the artifacts and objects recovered, how [Columbus-America] maintained monitoring and schedules, and how [it] followed their advice and suggestions for handling and preserving the articles. They described the scientific value of such articles, the discovery of new species, and of the unparalleled opportunities while aboard the research vessel to see first-hand the operation and collection of objects ... [,] species never before seen and artifacts.
The court continued:
Video tapes of the operation of things recovered and the means used to care for such articles were shown and then introduced in evidence as exhibits. Likewise, samples of the articles and artifacts recovered were presented for inspection. They demonstrated the particular care exercised in recovering and handling delicate items such as jewelry, china, cloth, papers and so on. One of the items was a cigar, which appeared in perfect condition.20
The district court noted further that Columbus-America had published a book and promoted a television account of its endeavors, and had provided educational materials to schools interested in teaching their students about the CENTRAL AMERICA and its history. The court found that “the efforts to preserve the site and the artifacts have not been equaled in any other case,” and no evidence to the contrary is before us.21
It will not be often, if ever, that all of the principles governing salvage awards will so overwhelmingly militate in the salvor’s favor. It will also be rare that a ship and cargo will have been imperiled for so long, and will be so difficult to objectively value, that a flat percentage award will be preferred to a sum certain. Nevertheless, this is that exceptional case, and though the district court’s ninety percent award is a generous one, we cannot say that it is excessive.22
*5748. Other Considerations
“It is ... the policy of the admiralty law to make salvage awards sufficiently liberal so as to encourage salvors to take the risks involved. ... It has been declared that the underlying theory of salvage is based on the encouragement of voluntary exertions for the saving of imperiled ships and cargoes.” Benedict, §§ 233-34 (citations and internal quotation marks omitted). It is undoubtedly sound public policy to set salvage awards high enough to provide an inducement for potential salvors to seek and recover property lost at sea.
To ensure that its award provided sufficient incentive for future endeavors of like kind, the district court, after obtaining a base figure by calculating the expenses incurred by Columbus-Ameriea, considered the opportunity cost of the venture capital provided by the RLP investors and the high-risk nature of the investment. We believe that the court’s actions were entirely appropriate.
Neither a salvor’s expenses nor its expected return on investment are factors to be taken into account in determining whether the salvage services rendered are meritorious. However, once the BLACKWALL and other relevant factors establish that the sal-vor merits an award, expenses plus a minimally sufficient inducement bonus may act to serve as a floor for that award. To say that a salvor’s services may be so outstanding as to merit an amount above that deemed minimally sufficient is but to say that admiralty courts should provide an incentive for salvors to conduct their operations in a more than minimally sufficient manner.23
IV.
THE MARKETING RULING
On January 13, 1994, the district court entered an order that, inter alia, designated Columbus-Ameriea as the central marketing authority for all of the gold salvaged from the CENTRAL AMERICA. The underwriters contend that the district court was without jurisdiction to deprive them of the right, as owners, to possess their share, and they urge us to direct the court to divide the gold in specie, so that they may dispose of their property as they see fit.
Ordinarily, we might agree with the underwriters’ position. The difficulty in this case, however, is that there is just too much gold involved to allow more than one party to market it. The loss of the gold aboard the CENTRAL AMERICA significantly affected the precious metal and financial markets in 1857, and there is evidence that the gold’s recovery could, upon its re-entry into the market, have a significant effect today.
Our review of the expert testimony introduced at trial convinces us that a unified plan *575is necessary to ensure the maximum return from the sale of the gold. Where there can be but one plan, we think it eminently sensible to allow the party with by far the greatest interest in the plan’s success to coordinate it; that party, of course, is Columbus-America.24
We do not share the underwriters’ alleged concern, expressed at oral argument, that “their share” will be sold at bargain basement prices to straw parties in collusion with Columbus-America. To the contrary, to the extent that the marketing proposal accords the underwriters a “share” of anything, it is a share of the proceeds from the sale of the gold, and not the gold itself. If the gold must eventually be sold at a discount, it is Columbus-America that will suffer no less than ninety percent of the damage. We note also that, because of the ongoing nature of these admiralty proceedings, the district court will retain jurisdiction over the marketing of the gold. Any allegations of impropriety on the part of Columbus-America can be investigated and adjudicated there.25
V.
EVIDENCE OF THE UNDERWRITERS’ OWNERSHIP
The final matter concerns Columbus-America’s cross-appeal. From the time that the underwriters filed their claim on September 29, 1989, the district court has never made findings concerning the amount of gold aboard the CENTRAL AMERICA that each underwriter — or its predecessor-in-interest — had insured and obtained by subrogation upon paying a claim of loss. The court was not dilatory; until CADG I was decided in 1992, the court believed that there was no need to determine the ownership interests of the individual underwriters because it had held that, under the law of finds, all of the underwriters had abandoned any interest they may have had.
On remand, Columbus-America asked the court to address the ownership issue. The court noted that it was reluctant to do so, because it believed that our opinion in CADG I could be construed to have decided that each underwriter had established the validity of its claim. For example, we stated in the earlier opinion that
The lower court ... found “prima facially” that the underwriters did insure the treasure and that they received ownership interests in the gold once the claims were paid.... [W]e find that the district court did not err when it held that the underwriters who are now parties, or their predecessors in interest, paid off claims upon and become the owners of the commercial shipment of gold in 1857.
CADG I at 465. Earlier, we had said that “[Sjhould more than $1,219,189 (1857 valuation) of gold be rescued, Columbus-America should be found the owner of any surplus.” Id. at 465 n. 6.
The underwriters maintain that the above-quoted language demonstrates that we “found,” in resolving the previous appeal, that they collectively owned the $1.2 million of insured gold that had been reported lost. They interpret our directive that “[o]n remand, the district court ... must determine what percentage of the gold each underwriter insured,” CADG I at 468, to mean that the court was required merely to divide the ownership interest in the gold among the individual underwriters to ensure that all $1.2 million was distributed and that each underwriter got something.26
It is a basic tenet of our legal system that, although appellate courts often review facts *576found by a judge or jury to ensure that they are not clearly erroneous, they do not make such findings in the first instance. Maine v. Taylor, 477 U.S. 131, 144-45, 106 S.Ct. 2440, 2450-51, 91 L.Ed.2d 110 (1986) (“Factfinding is the basic responsibility of district courts, rather than appellate courts”), quoting Pullman-Standard v. Swint, 456 U.S. 273, 291, 102 S.Ct. 1781, 1791-92, 72 L.Ed.2d 66 (1982) (internal citation omitted); Minnesota Mining and Mfg. Co. v. Blume, 684 F.2d 1166, 1170 (6th Cir.1982), (“Fact-finding is entrusted to the district court .... ”), cert. denied, 460 U.S. 1047, 103 S.Ct. 1449, 75 L.Ed.2d 803 (1983) and 461 U.S. 939, 103 S.Ct. 2110, 77 L.Ed.2d 314 (1983).
Of course, it was necessary for the district court, when faced with claims from over thirty underwriters, to make an initial finding that the claimants had standing to pursue their claims. As the above-quoted language from our prior opinion indicates, the district court was satisfied that the underwriters who were parties to the appeal had made at least a prima facie showing of ownership, and we were satisfied that the district court had not clearly erred in thus finding. Indeed, the claims of several underwriters were dismissed prior to the first trial, presumably because they could not make a prima facie showing.
Our prior opinion should not be read, however, to hold that a prima facie showing is all that is required for the individual underwriters to prove their claims. A pair of excerpts from the opinion serve to clarify the matter:
[T]he district court had bifurcated the trial, so that the first part would concern only whether Columbus-America was entitled to finder or salvor status.... If the latter scenario were found to be true, a second phase of the trial would be necessary, wherein the Court would have to determine what each underwriter had insured. ...
and
The Court had earlier bifurcated the trial so that the first phase only concerned whether the underwriters had abandoned the gold; should no abandonment be found, a second phase, examining what the underwriters actually insured ... would also be necessary.
CADG I at 459 and 470. We suppose that we might have more clearly expressed ourselves, but it ought to be crystal clear that a claimant cannot prove his or her claim merely by filing a lawsuit.
The district court must, therefore, consider each underwriter’s admissible evidence of the validity of its claim to a portion of the gold salvaged from the CENTRAL AMERICA. Columbus-America is entitled to introduce admissible evidence, if any, to rebut these claims. The district court must then weigh all of the evidence and decide whether each underwriter has proved all or any portion of its claim. If the total amount of gold proved owned by the underwriters constitutes less than one hundred percent of the gold salvaged — either because the proof of ownership fails with regard to one or more underwriters, or because a significant portion of uninsured gold is recovered, or for any other reason — Columbus-America may retain the excess.27
VI.
CONCLUSION
What Thompson and Columbus-America have accomplished is, by any measure, extraordinary. We can say without hesitation that their story is a paradigm of American initiative, ingenuity, and determination. Almost as extraordinary, perhaps, have been the efforts of the district judge, who has intrepidly waded through the morass of records and filings, and who has consistently evidenced good humor, notwithstanding the occasional contentiousness among the parties. He is to be commended.
The judgment of the district court is affirmed, and the case is remanded for further proceedings consistent with this opinion.
AFFIRMED AND REMANDED.
. "To what cannot you compel the hearts of men, 0 cursed lust for gold!” Aeneid, III. 56.
. We will refer to the claimant insurers and their successors collectively as the "underwriters.”
. See Fed.R.Civ.P. 24(a).
. The circumstances surrounding the last voyage of the CENTRAL AMERICA, along with extensive background information concerning the parties to this litigation, are detailed in CADG I, 974 F.2d at 454-59, and are not, except where necessary, repeated here.
. Only Grimm and Joanne Lampe Charlton, representing John's estate, remain as intervenors. Columbia settled with Columbus-America on June 29, 1993 — the day retrial began.
. Section 1292(a)(3) confers jurisdiction on the courts of appeals over "[ijnterlocutory decrees of ... district courts or the judges thereof determining the rights and liabilities of the parties to admiralty cases in which appeals from final decrees are allowed.” Decrees that “determinfe] the rights and liabilities of the parties” are those that determine substantive rights and liabilities, and not those that merely resolve procedural, tactical, or adjectival matters. E.g., Martha's Vineyard Scuba HQ, Inc. v. Unidentified, Wrecked and Abandoned Steam Vessel, 833 F.2d 1059, 1063 (1st Cir.1987); Miskiewicz v. Goodman, 341 F.2d 828, 830-31 (4th Cir.1965); see also 9 James W. Moore et al., Moore’s Federal Practice W 110.19[3] (2d ed.1995). The orders on appeal before us amply qualify.
. SeaMARC is an acronym for "Sea Mapping and Remote Characterization.”
. The Capuchins never realized any benefit from the gift until they gave it back to John in exchange for ten dollars on March 22, 1990, less than two weeks prior to the first trial in the district court. As we noted in CADG I, "John did nothing to enlighten [the Capuchins] on the discovery of the ship or the upcoming trial.... [T]he monks must have protested, for on April 10 both John and Grimm signed an agreement with the Capuchins giving the Order one-third of any judgment they ... would recover." 974 F.2d at 459.
. Employees of Lamont-Doherty and 1ST had worked together closely to develop the SeaMARC I. Dale Chayes, who had been aboard the CONRAD, told John Shaw, IST's General Manager, about the survey. Shaw, in turn, told Williamson, who passed the information on to Thompson.
Thompson then telephoned Chayes. Thompson’s recollection of the conversation was hazy, but he testified that he called Chayes "probably ... to see if the [CONRAD] data was going to become available.” Chayes referred Thompson to Ryan. Chayes testified that he told Ryan that he "had gotten a call from this guy who was at Battelle, ... he was asking about shipwrecks."
. Thompson violated, at least technically, Columbia’s conditions for disclosing the CONRAD data by copying the materials for inclusion in a Columbus-America file. That file was in Thompson's possession aboard Williamson’s vessel during the 1986 survey, detailed infra By sharing the data with Williamson and Robert D. Evans, one of Columbus-America’s project directors, Thompson may have committed further violations.
. The court nevertheless specifically stated that “there is a lack of evidence, factual or circumstantial, to show ... that the Columbia data assisted [Columbus-America] in locating the CENTRAL AMERICA.... [Such a conclusion] would be pure speculation or conjecture....”
. The intervenors contend that "[t]he District Court did not decide whether the [CONRAD data were] in the public domain....” Intervenors' Opening and Answering Brief at 30. The court's statement in its opinion that it “need not decide this question on whether the information was in the public domain” appears, at first blush, to support the intervenors’ contention. However, we believe that the unequivocal pronouncements of the court, e.g., "[Lamont-Doherty's] statement that the information was not in the public domain was not true,” and that the Observatory's assertion to the contrary "was a misstatement,” as well as its comprehensive analysis of the issue, make clear that “this question" to which the district court referred was the larger question of Columbus-America’s alleged misappropriation, and that, had the court instead resolved the public domain issue in the intervenors' favor, such a resolution would not have been determinative of their claim.
. Thompson testified during the first trial that he did not discuss the CONRAD data with Williamson or his associates, and that he was unsure that the data were even aboard Williamson’s vessel, the PINE RIVER, during the 1986 survey. On retrial, Thompson testified during cross-examination that he had shown the sonar records (but not the contour chart or the ship track lines) to Williamson aboard the PINE RIVER, and that the two had briefly discussed some general characteristics of the records and the imagery contained on some of the photographs, including particular targets. The matter was fully before the district court, which also had to consider certain inconsistencies in the testimony of the intervenors' witnesses. For instance, Ryan admitted during the first trial that he had made an inaccurate statement in an earlier affidavit, and Grimm himself offered testimony of dubious credibility concerning his purported efforts in 1987 to follow up on the 1984 CONRAD survey.
. Actually, the district court found that CONRAD Target # 4, despite its apparent similarity in size and admitted proximity to the proven location of the CENTRAL AMERICA, was not in fact the shipwreck. Because we affirm the court's denial of the intervenors' claim on sufficient alternative grounds, we need not review this particular finding.
. One possible motive for this alleged ruse, detailed in Section III-A, infra, is that Columbus-America needed time to develop NEMO, its undersea robot.
. The intervenors contend that they were forced to rely on circumstantial evidence because the district court erred by failing to direct Columbus-America to produce certain navigational records, diving logs, and various other materials, and by allowing Columbus-America to redact some of the discovery that it did produce. The intervenors also complain that they were denied inquiry into the last-minute settlement agreement between Columbia and Columbus-America, see note 5, supra, which, they assert, was designed in part to keep Ryan from testifying at retrial and thereby obstruct the truth-finding process.
The scope and conduct of discovery are within the sound discretion of the district court. Erdmann v. Preferred Research, Inc., 852 F.2d 788, 792 (4th Cir.1988). As to the navigational and related records, the court was well aware that Columbus-America had been forced to actively defend its injunction area, graphically demonstrating the need to keep the precise location of its ongoing salvage efforts a secret. The court might have nevertheless compelled additional discoveiy if it expected that relevant evidence might be produced; however, in light of the rather fantastic theory relied on by the interve-nors, the court did not abuse its discretion by striking the balance in Columbus-America's favor. Likewise, denying inquiry into the settlement agreement was hardly an abuse of discretion, inasmuch as Ryan's deposition and trial testimony from the previous proceeding were already in the record.
A final matter concerning the intervenors' appeal remains to be addressed. One of the trial exhibits that Columbus-America designated for inclusion in the Joint Appendix was a color copy of a slide taken from a computer monitor, showing the CENTRAL AMERICA as imaged by Williamson’s radar. When the Joint Appendix was compiled, the exhibit was apparently recopied using ordinary, non-color xerographic methods, resulting in an unsatisfactory reproduction. The intervenors objected and filed a motion asking that Columbus-America be directed to submit the original exhibit. While the motion was pending, Columbus-America represented to the Court that it could not find the original and instead submitted a duplicate. The intervenors have not objected to this submission; hence, we consider the matter closed.
. These factors are (1) the labor expended by the salvors in rendering the salvage service; (2) the promptitude, skill, and energy displayed in rendering the service and saving the property; (3) the value of the property employed by the salvors in rendering the service, and the danger to which such property was exposed; (4) the risk incurred by the salvors in securing the property from the impending peril; (5) the value of the property salvaged; and (6) the degree of danger from which the property was rescued.
. See Waterman, supra, at 412 (moiety of $450,-000 urged by salvors who had floated a vessel that had run aground would be "absurdly excessive”).
. See Treasure Salvors, Inc. v. Unidentified Wrecked and Abandoned Sailing Vessel, 569 F.2d 330 (5th Cir.1978) (several months spent salvaging the galleon NUESTRA SEÑORA DE LA ATOCHA); Rickard v. Pringle, 293 F.Supp. 981, 984 (E.D.N.Y.1968) (recovery of ACARA's propeller over ten-month period).
. We were given the opportunity at oral argument to examine the very cigar of which the district court spoke.
. The National Maritime Historical Society, the National Association of Academies of Science, and the Explorers Club have submitted briefs as amicus curiae in support of Columbus-America.
.As have the intervenors, see note 16, supra, the underwriters contend that the district court abused its discretion by allowing the curtailment or outright denial of discovery into various matters. The underwriters primarily object to the court permitting Columbus-America to produce only edited videotapes and selected photographs of the wreck site, and to the court's refusal to *574allow discovery into RLP’s profits or the identity of its investors. See Section II-C, supra. Because it appears from the credible evidence that the requested videotapes and photographs would not have mandated a different conclusion as to the preservation issue, and because it appears unlikely that inquiry into RLP’s profits or investors would have produced relevant evidence, the district court’s discovery rulings were not an abuse of discretion.
. An answer to the question, “How large should an award be to serve as an inducement?” may be more easily determined if the owner has previously attempted to negotiate a contract for the salvage of his or her property. We noted in our previous opinion that one underwriter had attempted to negotiate a salvage of the CENTRAL AMERICA whereby the salvor would receive 98% of the treasure. CADG I at 467.
Two other matters require our attention at this juncture. The first concerns the underwriters' claim to some of the silver — in addition to the gold — aboard the CENTRAL AMERICA. The underwriters first brought this claim to the district court’s attention late in the proceedings, more than three months after the retrial had concluded. The court ruled that the underwriters had essentially waived the claim, and that, in any event, the claim had no support in the evidence. We see no reason to disturb the court’s ruling on either point.
The second matter concerns the underwriters’ motion on appeal to strike Columbus-America's reply brief, which, at forty-seven pages, is twenty-two pages too long on its face. See Fed. R.App. P. 28(g) (“Except by permission of the court ... reply briefs must not exceed 25 pages....”). Columbus-Ameriea has conceded that it did not obtain the court's prior permission to exceed the maximum page limit. Accordingly, we strike and refuse to consider Pages 26-47 of Columbus-America’s reply brief, thereby granting the underwriters' motion as modified.
. To the extent that Columbus-America's designation as the central marketing authority for all of the gold appears to unduly interfere with the underwriters' ownership rights, it must be remembered that separate marketing of the gold would be less efficient and inevitably result in a lower total return. Thus, Columbus-America would be left with nine-tenths of a smaller pie, to the detriment of their eventual ownership rights.
. With regard to the marketing and valuation issues, the underwriters have moved to supplement the record with documents obtained from the public record of unrelated judicial proceedings, involving a witness who testified before the district court in the present matter. We grant the motion.
.The underwriters have agreed among themselves on how to distribute the CENTRAL AMERICA proceeds, and have submitted their agreement to the district court.
. See CADG I at 465 ("[A]n abandonment may be found, and Columbus-America may be declared the finder and sole owner, as to any recovered parts of the ship, all passenger possessions, and any cargo besides the insured shipments.").