DUNIWAY, Circuit Judge:

This case involves a collision at sea between the SS PRESIDENT GRANT and the fishing vessel MARTIN HIGGINS, which sank. After a trial of the claim of the HIGGINS' insurers against the GRANT and her owner, American President Lines, the district court found the GRANT to be 85% at fault in the collision and the HIGGINS 15% at fault, and assessed damages accordingly. The GRANT appeals, and we affirm.

On cross-appeal, the HIGGINS contends that the district court erred under the "Major-Minor Fault Rule" in allocating 15% of the fault to her. Again, we affirm. The HIGGINS also contends that the district court abused its discretion in awarding pre-judgment interest at the rate of 8%. We agree, and remand to the district court to redetermine the pre-judgment interest award.

I. FACTS.

The fog was thick in the approach to San Francisco's Golden Gate on October 22, 1980. Visibility was 100 to 200 yards. The HIGGINS, an 80-foot fishing boat, was outbound from Fisherman's Wharf. When she entered the fog, she reduced her speed from 11 or 12 knots to 4 or 5 knots. She had radar, which she was using. Meanwhile, the GRANT, an 820-foot container ship, was headed inbound at sea speed, about 18 knots. At this rate, she travelled 100 to 200 yards in 10 to 20 seconds. From her bridge, the look-out could not see through the fog beyond her bow. The GRANT had a highly qualified crew and extensive radar capacity, most of which she did not use. At 1:48 p.m. (1348), about four miles outside the Gate, the GRANT struck and sank the HIGGINS. A lifeboat from the GRANT rescued the HIGGINS' two crew members.

The owners, underwriters, and crew of the HIGGINS sued the GRANT and her owner, American President Lines, in federal district court. The HIGGINS' owner, Western Pacific Fisheries, settled its claims for uninsured losses. The crew members, Block and Matthews, settled their claims for personal injuries. After a ten-day trial, the district court awarded 85% of the HIGGINS' $485,000 insured value, plus $1,200 paid to a diver to see if the wreck obstructed navigation, plus 8% pre-judgment interest, to the underwriters. The district court denied the GRANT's motion to reopen the trial to hear evidence of marijuana trafficking by the HIGGINS.

Four issues are presented:

1. Whether certain findings as to the fault of the GRANT are clearly erroneous and require reversal.

2. Whether it was error to refuse to let the GRANT show that the HIGGINS was transporting marijuana.

3. Whether it was error to refuse to apply the "Major-Minor Fault Rule" and to find the HIGGINS 15% at fault.

4. Whether it was error to use an 8% rate in awarding pre-judgment interest.

We set out the facts in more detail in considering the first issue.

II. FINDINGS OF FACT AS TO THE COLLISION.

A traffic separation scheme establishes lanes, marked by buoys and on charts, for incoming and outgoing ships in the San Francisco Main Ship Channel. The GRANT claims that she was properly in the incoming (south) lane before and at the time of the collision. The HIGGINS claimed, and the district court found, that the GRANT was in the wrong, outgoing (north) lane and that the collision occurred in the outgoing lane. On appeal, the GRANT's principal argument is that the district court clearly erred in making this finding of fact. See Fed.R.Civ.P. 52(a). We quote the most pertinent portions of Judge Patel's careful and detailed findings of fact:

The exact location of the collision, which occurred near the entrance of the San Francisco Bay, has been the subject of much conflicting testimony and is still in doubt. Plaintiff's expert, Captain Robert Slack, after several depositions, calculations, and recalculations, came up with at least five different tracks, or courses, for the Grant. Another expert, David Sears, using a radar simulator arrived at another track and location of the collision. Dr. William Webster, defendant's expert, reached a different conclusion and, in the process, his work yielded at least two different tracks. (ER 86)

. . . . .

Despite sophisticated equipment and well-trained, experienced personnel, there is insufficient data from which to conclude with certainty the Grant's course. Chief Officer McCloud summed it up best when he testified that "there weren't but a couple of people even on the ship, I think, that had any knowledge where the ship was." (Deposition 62:20-22) Surprisingly, in light of these variances, the tracks suggested by each of the experts are relatively close. It is not necessary for this Court to completely resolve the mystery of the courses taken by the Grant and the Higgins and the exact location of the collision in order to determine whether plaintiff has established liability by a preponderance of the evidence. (ER 87)

THE CONDUCT OF THE HIGGINS.

. . . . .

Blocker [one of the owners of the HIGGINS] was at the radar and manning the wheel nearly all of the time, including all times after the Grant was picked up on the radar. When he started out from the fuel dock the radar was on the 12-mile range. He changed to the 6-mile range and finally to the 3-mile range as the Grant approached. He took two radar fixes prior to the collision. He could see the channel buoys on the radar but could not see all of them all of the time.

The Higgins was proceeding at approximately 11 to 12 knots through the water until it encountered fog. It then reduced its speed to 4 or 5 knots, which, with the 2 to 3 knot current, gave it a speed of 6 to 8 knots over the ground. (ER 90)

. . . . .

The Higgins was outside the fog when it first sighted the Grant just outside the buoy channel. Blocker observed the Grant coming through the buoy channel. At first it appeared to be in the center of the channel. Then it appeared to be going north into the outbound lane. The Higgins at all times prior to the collision remained in the outbound lane north of the separation line.

Blocker first determined a possible risk of collision at 4 to 5 miles from the Grant. He never marked the Grant's position or plotted to determine course and speed. At 3 miles off the Grant the Higgins made a 15 ° course change to the right; at one quarter mile distance it made a 10 ° right course change. It still appeared as if the vessels were bow to bow. Two or three seconds before the collision the Higgins visually sighted the Grant. The Higgins never executed a hard right or any port turns. (ER 91)

. . . . .

THE CONDUCT OF THE GRANT.

. . . . .

No one person was assigned radar watch. Third Mate Miles and Captain Colivas took turns observing it; neither of them was at the radar continuously.

The 10-cm radar had true motion capability, that is, the radar displays vessels, including own ship, moving at their true courses and speeds. Objects having no motion appear stationary. The 10-cm radar was also equipped with an automatic plotting device, which could be operated only when the stabilized relative motion mode was in use. On the day in question the 10-cm radar was in the unstabilized, head-up relative mode. In the relative mode, whether head-up, or North-up, own ship appears at the center of the scope, and motions of other vessels relative to own ship can be observed. With the stabilized North-up display courses and bearings and closest point of approach can be plotted. No plotting was done. The automatic plotting device was not used. (ER 92-93)

. . . . .

The GRANT proceeded at full speed, about 18 knots.

. . . . .

The Grant first observed the Higgins after the Grant had passed through the buoys. This was approximately three minutes before the collision. It was noted on the 10-cm radar by Captain Fowler who believed it was between Mile Rock and Buoy 8, south of the inbound lane. It was also noted by Captain Colivas, Chief Mate Arechavala, and Third Mate Miles. It was noted by them as a "pip" on the 3-cm radar. Captain Colivas estimated the Grant to be near Buoy 8 at that time and the other vessel, the Higgins, to be three to four miles away and approximately 10 ° off the starboard bow. Other ships' personnel estimated it at 30 ° to 35 °. They also believed it to be dead in the water or moving very slowly.

No plotting was done to determine the CPA, or true course and speed. The cursor and variable range marker were put on the "pip". The target or "pip" began to fall of to the right. The cursor was not moved. No ranges or bearings were taken or charted.

Approximately two minutes after the citing, [sic] the fishing boat came out of the fog and could be seen from the starboard wing of the bridge. It was 150 to 200 feet away and just about 30 to 35 degrees off the starboard bow of the Grant. (ER 93-94)

. . . . .

The Grant made two course changes to the left after the "pip" appeared on the radar. The first was at 1345.5 and the second at 1347. Captain Fowler estimated by eye only that the closest point of approach was 500 yards. The heading of the Grant at the time of the first course change was 070 °. After the first course change it was 060 °, and then 050 ° after the second change.

Immediately preceding the impact, a hard left order was given. The heading of the Grant at the time of impact was 040 °.

At no time from the first sighting of the target that turned out to be the Higgins until the collision did the Grant reduce its speed from 18 knots. (ER 94)

. . . . .

The Grant did not have a lookout on the bow of the ship. The lookout was on the starboard wing of the bridge. Other ship's personnel occasionally went out on the starboard wing to observe. The distance from the bridge to the bow of the Grant is 113 feet. In the visibility conditions obtaining at the time, the lookout on the wing of the bridge could not see beyond the bow of the vessel. The weather and sea conditions were not so hazardous or inclement that a lookout could not be stationed on the bow. Visibility was better on the deck level than on the bridge level. Any targets observed on the radar could have been communicated from the bridge to the bow lookout by telephone. (ER 95)

. . . . .

No fixes had been taken aboard the Grant after 1124, when the ship passed Pigeon Point. The next fix was taken by Third Mate Miles at 1352. That fix was the actual location of the collision as determined by the personnel aboard the Grant. Although there is conflicting testimony on the significance of this fix, the persuasive evidence is that the fix was not the location of the Grant at 1352 but rather the site of the collision at 1348, which was later determined at 1352. That location was fixed at 37 ° 4.5' N. latitude, 122 ° 33.8' W. longitude. This places the collision in the westbound traffic lane just south of the northern boundary of the lane. (ER 95-96) (We refer to this as the "1352 finding.")

THE FAULTS. Based on the foregoing, the Court finds that a preponderance of the evidence shows that the collision was caused primarily by the failure of those in charge of the Grant to navigate the vessel in a reasonably prudent manner. The Grant was negligent in proceeding at a speed unsafe for the conditions obtaining on October 22. Speed should have been reduced because of low visibility. The stopping distance of the Grant at the speed it was pursuing was 3365 feet, well in excess of one half the distance of visibility. There is no evidence that the Higgins' speed was imprudent for the conditions.

In addition, when another vessel appeared in close range without certainty of its course and speed, the Grant should have reduced speed or stopped until it could accurately determine the course and speed and the closest point of approach. Instead, the Grant continued to proceed at 18 knots.

The Court further concludes that had the Grant plotted the course and speed of the target or otherwise used its radar capabilities it would have been able to determine whether the Higgins was in fact on its starboard or port bow and what course changes if any should be made. By so doing, it is more likely than not that the Grant could have avoided the collision. (ER 99-100)

Counsel for the GRANT point their big guns at the 1352 finding and, having demolished it to their satisfaction, argue that the judgment must be reversed because Judge Patel's findings that the HIGGINS was in the outbound lane and the GRANT entered that lane and struck the HIGGINS, rest entirely on the 1352 finding. We cannot agree. Judge Patel did not say that she relied entirely, or at all, on that finding. She did not even mention it in establishing fault. She said that it was not necessary to establish the exact location in order to determine liability.

There was considerable other evidence as to the location of the HIGGINS when she was struck. There was testimony by the crew of the HIGGINS as to its course. There was latitude and longitude indicated by the Master of the GRANT, Captain Colivas, on Coast Guard Form 2692. There was expert extrapolation from the 1410 (C.P.O. Rapp) and 1440 (Miles) anchorage fixes, which were 900 to 1000 yards north of the 1348 fix position (Seals and Slack testimony). There was expert extrapolation from the location of the wreck of the HIGGINS, about 1000 yards northwest of buoy 7 (which marks the north side of the outgoing lane well west of the collision site). The only evidence expressly rejected by Judge Patel was the radar observations by Coast Guard Vessel Traffic Service personnel, which purported to show that the collision occurred in the inbound lane.

A judge is not required, in making findings, to mention every item of evidence and either adopt it or reject it. We presume that the judge considers all of the evidence, and relies on so much of it as supports the finding and rejects what does not support the finding, unless the judge states otherwise. See generally 5A J. Moore, Moore's Federal Practice ¶ 52.06[1] (2d ed. 1982), esp. at 2715-16.

Judge Patel could have said that the GRANT was at fault for being in the wrong lane. She did not. Rather, she concluded that the GRANT had violated the following rules of the International Regulations Preventing Collisions at Sea (COLREGs), adopted as United States law by 33 U.S.C. § 1602, and the Navigation Safety Regulations, 33 C.F.R. § 164.01 et seq., some of which the HIGGINS (which the judge found to be 15% at fault) also violated:

— Rule 5: proper look-out;

— Rule 6: safe speed under prevailing circumstances and conditions;

— Rule 7: use of radar equipment and other available means to determine risk of collision;

— Rule 8: action to avoid collision;

— Rule 9: narrow channels;

— Rule 10: traffic separation schemes;

— Rule 19: conduct of vessels in restricted visibility;

— N.S.R.: radar fix and plotting, use of equipment, reliance on buoys, evaluation of radar information, setting of speed.

The location of the vessels or the collision is potentially relevant to only three of these violations. Rule 9(a) requires that vessels keep to the outer limit of narrow channels to the extent safe and practicable. Violation of Rule 9 by the GRANT is not predicated on the collision being at the 1352 finding position. The GRANT violated the rule by not keeping to the right side of, or to the south of, the inbound channel and by her course changes to port, even if she was on the course described by the GRANT's expert witness, Dr. Webster. That course had the GRANT heading into the outbound lane at the time of the collision. The GRANT's mistaken belief that this course was necessary to avoid the collision is no excuse because the mistake arose from the GRANT's failure to track the HIGGINS, as required by Rule 7(b).

Judge Patel's finding that the HIGGINS did not violate Rule 9 because she stayed to the right of the channel does not depend on the location of the collision. There is other evidence that supports the finding. Thus, this conclusion is not tainted by the 1352 finding.

Considering Rule 10, which governs traffic separation schemes, Judge Patel found that the GRANT violated good seamanship and prudence when she "proceeded too close to the center line and then crossed over into the outbound lane." As with the Rule 9 violation, the judge is supported by the track proposed by Dr. Webster, even if we accept the 1352 finding as establishing the 1352 position rather than the collision position. Indeed, the GRANT argues that the court's conclusion on this point is inconsistent with the 1352 finding, which undercuts the proposition that this violation depends on the 1352 finding.

The HIGGINS was not subject to the San Francisco traffic separation scheme because she weighed less than 300 gross tons, i.e., 121 gross tons. Judge Patel concluded that the HIGGINS acted prudently in light of the traffic separation scheme because she was in the outbound lane, a finding which does not depend on the 1352 finding.

Judge Patel concluded that the GRANT violated Rule 19 in five respects, one of which was improper turns to port. The GRANT argues that the turns were proper to avoid collision, given her course in the inbound lane. Rule 19(d) requires that a vessel alter course to avoid collision when her radar detects a risk of collision in restricted visibility conditions, and provides a general stricture against turns to port. The GRANT's turn to port did not avoid the collision. Rather, in conjunction with the failure to track the HIGGINS on radar as required by Rule 7, it was the actual cause of the collision. Again, the conclusion does not depend on the 1352 finding.

Judge Patel was well aware of the conflicting evidence and uncertainty regarding the location of the collision. To a large degree, she took this uncertainty into account when, applying her discretion as fact-finder, she allocated fault between the two colliding vessels. The GRANT's speed and failure to take proper precautions provided ample basis for the finding that she was primarily at fault, even if we ignore any violations possibly connected with the 1352 finding. We need not decide whether the 1352 finding is clearly erroneous. If it were, there is still ample evidence to support the findings and conclusions as to fault and degree of fault.

III. EXCLUSION OF EVIDENCE OF DRUG TRAFFICKING AND FORFEITABILITY OF THE HIGGINS.

The GRANT argues that the judge's refusal to admit evidence and denial of motions to reopen the trial to take additional testimony regarding alleged drug violations by the HIGGINS was erroneous as a matter of law, and was also an abuse of discretion. The GRANT's theory is that if the HIGGINS was carrying marijuana in violation of the Controlled Substances Act, 21 U.S.C. § 801 et seq., she was forfeit to the United States under 21 U.S.C. § 881(a)(4). Such forfeiture, the GRANT contends, was unconditional, automatic, and took place at the moment of violation. Therefore, says the GRANT, Western Pacific Fisheries and the underwriters lost their rights in the HIGGINS and had no standing to sue for damages.

This argument is without merit, for exactly the reasons stated by Judge Patel in her order denying the motion for a new trial. (ER 80, 83) There is clear authority that when the government initiates forfeiture proceedings, the forfeiture relates back to the time of the violation. However, the forfeiture must be "consummated" by a judgment or decree of a court. Ivers v. United States, 9 Cir., 1978, 581 F.2d 1362, 1367; United States v. Stowell, 1890, 133 U.S. 1, 16-17, 10 S.Ct. 244, 247, 33 L.Ed. 555; Confiscation Cases, 1868, 74 U.S. (7 Wall.) 454, 460-61, 19 L.Ed. 196. See, e.g., United States v. One 56-Foot Yacht Named the TAHUNA, 9 Cir., 1983, 702 F.2d 1276, 1279: "The government's legal title is not established until after it has complied with the forfeiture procedures mandated by statute."

Here, the government has not initiated forfeiture proceedings and there is no suggestion it will do so. The statute of limitations will run in December, 1985. 28 U.S.C. § 2462. Because the HIGGINS is in Davy Jones' locker, seizure is unlikely, to say the least. The GRANT cites, and we find, no authority for the crucial proposition that a private party may, in the absence of a government-initiated forfeiture action, invoke the drug forfeiture statute to defeat an adverse party's standing. We find no equitable reason for the GRANT to avoid liability for sinking the HIGGINS because a third party, the United States government, may have an unasserted interest in the HIGGINS. Western Pacific Fisheries and the underwriters retain good title as against the GRANT.

IV. THE MAJOR-MINOR FAULT RULE.

The HIGGINS, in her cross-appeal, argues that Judge Patel's finding that the HIGGINS violated Rule 8 and was therefore 15% at fault violates the Major-Minor Fault Rule. The HIGGINS' fault is said to be so minor that, under this rule, it was presumptively not a cause of the collision. She argues that the Major-Minor Fault Rule survives the Supreme Court's landmark decision in United States v. Reliable Transfer, 1975, 421 U.S. 397, 95 S.Ct. 1708, 44 L.Ed.2d 251, citing our decision in State of California v. Italian Motorship ILICE, 9 Cir., 1976, 534 F.2d 836, 840-41. We disagree and hold that the Major-Minor Fault Rule has no place in the system of comparative fault established by Reliable Transfer.

The "Mutual Fault-Equal Contribution Rule" established that, when there is mutual fault, the damages must be equally allocated regardless of the degree of comparative fault. ILICE, 534 F.2d at 840; The Schooner CATHARINE v. Dickinson, 1855, 58 U.S. (17 How.) 170, 177, 15 L.Ed. 233. In providing for allocation of liability for property damage proportionate to the comparative degrees of fault, Reliable Transfer expressly abolished the Mutual Fault-Equal Contribution Rule. 421 U.S. at 410-11, 95 S.Ct. at 1715; ILICE, 534 F.2d at 840; Alkmeon Naviera, S.A. v. M/V MARINA L, 9 Cir., 1980, 633 F.2d 789, 798 n. 12. This is not questioned by the HIGGINS.

The "Major-Minor Fault Rule" established a rebuttable presumption in favor of a vessel at minor fault that the other vessel at major fault (or grossly negligent) was the exclusive cause of the collision. ILICE, 534 F.2d at 840; The CITY OF NEW YORK, 1893, 147 U.S. 72, 85, 13 S.Ct. 211, 216, 37 L.Ed. 84; The GREAT REPUBLIC, 1874, 90 U.S. (23 Wall.) 20, 34-35, 23 L.Ed. 55. Reliable Transfer did not expressly reject the Major-Minor Fault Rule, but did treat the rule as an adjunct to the Mutual Fault Rule. It referred to the Major-Minor Fault Rule as being similarly unfair and "inherently unreliable." 421 U.S. at 406, 95 S.Ct. at 1713. The rationale of the Court's rejection of the Mutual Fault-Equal Contribution Rule clearly implies a rejection of the Major-Minor Fault Rule as well. The rules are, respectively, the "all" and the "nothing" of the all-or-nothing approach that Reliable Transfer replaced with a comparative fault approach. Accord Getty Oil Co. v. SS PONCE DE LEON, 2 Cir., 1977, 555 F.2d 328, 332-33.

The HIGGINS cites, and we discover, no authority holding that the Major-Minor Fault Rule survives Reliable Transfer. ILICE recites the rule but does not apply it because the district court found equal, not major or exclusive, fault. ILICE neither says nor shows that the Major-Minor Fault Rule remains viable. One can plausibly read ILICE to imply in dicta that, because the Major-Minor Fault Rule is a rule of evidence, while the Mutual Fault Rule is "substantive," Reliable Transfer can be held to preserve, or at least not to reject, the Major-Minor Fault Rule. We expressly reject this possibility. Accord PONCE DE LEON, 555 F.2d at 333 n. 4.

Another case cited by the HIGGINS, Board of Comm'rs of the Port of New Orleans v. M/V FARMSUM, 5 Cir., 1978, 574 F.2d 289, 297, is not in point. FARMSUM relied on the concept of proximate cause, not the Major-Minor Fault Rule, to rebut the Pennsylvania presumption. It thus does not demonstrate the viability of the Major-Minor Fault Rule after Reliable Transfer.

V. PRE-JUDGMENT INTEREST RATE.

The district court awarded damages plus pre-judgment interest from the date of payment by the underwriters at the rate of 8% per annum. (Findings Conclusions at 22; ER 107.) The judgment included this interest, calculated to a dollar sum, and also assessed post-judgment interest at the rate of 9.29%, in accordance with 28 U.S.C. § 1961. (Judgment at 2; ER 79.) The HIGGINS, in her cross appeal, argues that the award of pre-judgment interest at the rate of 8% was inadequate to compensate the underwriters.

It is well-established that compensatory damages in maritime cases normally include pre-judgment interest. M/V MARINA L, 633 F.2d at 797; The PRESIDENT MADISON, 9 Cir., 1937, 91 F.2d 835, 845; see generally Turner v. Japan Lines, Ltd., 9 Cir., 1983, 702 F.2d 752, 756 n. 4. Although the award of pre-judgment interest is within the discretion of the trial judge, this discretion must be exercised with a view to the fact that pre-judgment interest is an element of compensation, not a penalty. M/V MARINA L, 633 F.2d at 797; Rosa v. Insurance Co. of Pennsylvania, 9 Cir., 1970, 421 F.2d 390, 393. An award of pre-judgment interest at below market rates does not fully compensate the prevailing party and, in addition, tends to discourage prompt litigation because delay gives an economic benefit to the debtor. Sea-Land Service, Inc. v. Eagle Terminal Tankers, Inc., W.D.Wash., 1977, 443 F. Supp. 532, 534.

There was uncontroverted evidence that when the underwriters paid Western Pacific Fisheries for the loss of the HIGGINS, in December 1980, the prevailing prime rate of the major New York banks was 20.000%, that the rates on 90-day certificates of deposit ranged from a high of 17.625% (in December 1980) to a low of 11.620% (in December 1981), and that the actual cost of money to the underwriters was (in aggregate) 19.53% calculated on a "funds borrowed" basis or 17.85% on a "funds invested" basis. See Kingsley letter of February 25, 1982 (Exh. A-49, SER 43). Calculated on these respective bases, the pre-judgment interest award would have been $116,887.36 or $106,677.76. ( Id. at 8, ER 50.) This is significantly higher than the $63,639.81 awarded based on an 8% rate. (Judgment at 2; ER 79.)

The GRANT argues that the rate of pre-judgment interest was within the discretion of the trial court and should not be rejected without a showing of abuse. The very case that the GRANT cites, however, emphasizes that the district court should exercise its discretion to award "the income which the monetary damages would have earned, and that should be measured by interest on short-term, risk-free obligations." Independent Bulk Transport v. Vessel MORANIA ABACO, 2 Cir., 1982, 676 F.2d 23, 27. The many authorities cited by both parties are consistent with this approach to pre-judgment interest. See, e.g., Edinburgh Assur. Co. v. R.L. Burns Corp., 9 Cir., 1982, 669 F.2d 1259, 1262-63; Sauers v. Alaska Barge Transport, Inc., 9 Cir., 1979, 600 F.2d 238, 248. We are especially unpersuaded by the GRANT's argument that since we awarded pre-judgment interest at the rate of 8% in Sauers for the period of 1973 to 1979, an award at 8% is appropriate to the period of 1980 to 1982, when market interest rates were significantly higher.

In 1982, Congress amended 28 U.S.C. § 1961 to provide that district courts must set the interest accrued after civil judgment by reference to 52-week U.S. Treasury bill rates. Pub.L. 97-164, § 302. Congress intended to remove the economic incentives to delay that exist when judicially-awarded interest rates are less than the contemporary cost of money. S.Rep. No. 97-275, 97th Cong., 2d Sess. 30, reprinted at 1982 U.S.Code Cong. Ad.News 40. Although § 1961 does not provide for pre-judgment interest, it is entirely compatible with awards of such interest. Id.; Brick-layers' Pension Trust Fund v. Taiariol, 6 Cir., 1982, 671 F.2d 988, 989; Jarvis v. Johnson, 3 Cir., 1982, 668 F.2d 740, 741 n. 1.

We conclude that the measure of interest rates prescribed for post-judgment interest in 28 U.S.C. § 1961(a) is also appropriate for fixing the rate for pre-judgment interest in cases such as this, where pre-judgment interest may be awarded, unless the trial judge finds, on substantial evidence, that the equities of the particular case require a different rate. Absent such a finding, pre-judgment interest in this case should be calculated from the date of payment by the underwriters "at a rate equal to the coupon issue yield equivalent (as determined by the Secretary of the Treasury) of the average accepted auction price for the last auction of fifty-two week United States Treasury bills settled immediately prior to the date" of payment by the underwriters. 28 U.S.C. § 1961(a).

VI. CONCLUSION.

The judgment for pre-judgment interest is reversed, and the case is remanded for further proceedings relating to pre-judgment interest. In all other respects, the judgment is affirmed.

In appeal No. 82-4726, costs are awarded to plaintiffs-appellees. In appeal No. 83-1523, each party shall bear its own costs.