delivered the opinion of the court:
This patent case, the subject of two earlier opinions of the court, comes before us once again, this time for a final determination of damages. Our first opinion, reported at 195 Ct. Cl. 53, 445 F.2d 323 (1971), held that defendant’s purchase of oscilloscopes (scopes) from third-party defendants (TPDs) infringed valid claims in eight patents held by plaintiff. Our second opinion, reported at 213 Ct. Cl. 257, 552 F.2d 343 (1977), modified by order of June 24, 1977 (reported at 213 Ct. Cl. 307, 557 F.2d 265), directed the trial judge in his computation of the damages to which plaintiff *147was entitled. As modified, the latter opinion held that a 10-percent royalty rate should be applied to a compensation base which was to include scopes intended for commercial use, scopes intended for military use, and "plug-ins” (unpatented components used with the scopes). Packaging, accessories such as probes and carts, and other miscellaneous costs were not to be included in the compensation base. The trial judge was instructed to determine delay compensation by applying the interest rates prescribed in Pitcairn v. United States, 212 Ct. Cl. 168, 547 F.2d 1106 (1976), cert. denied, 434 U.S. 1051 (1978), unless, for years after 1975, either party affirmatively demonstrated that a different rate should be applied. Trial Judge Browne has now rendered an opinion which addresses these issues. We are in substantial agreement with the trial judge’s result, although we take a different view regarding the timing of delay damage computation. The three key issues discussed in the trial judge’s opinion are reviewed in turn below.
Under the terms of our June 24, 1977, order in this case cited above, the trial judge was instructed "to recompute, on the basis of the present record, the royalty bases for both 'commercial’ and 'militarized’ scopes to include only the scopes themselves and all the plug-ins.” In making such a determination, the trial judge was permitted to call for such submissions from the parties as he deemed helpful to expeditious computation of damages.
The trial judge did receive submissions from the parties and considered these in concluding that the compensation base was $11,442,014 for commercial scopes and $10,990,187 for militarized scopes, for a total of $22,432,201. Defendant objects to this conclusion, arguing that the trial judge should have accepted, without modification, figures presented by defendant’s expert witness, Mr. Lawrence Glassman. Those exhibits, however, do not reflect the prices actually paid by defendant for the scopes and plug-ins alone but instead offer Glassman’s attempt to determine those costs from contracts whose prices include elements which we have ordered to be excluded from the compensation base. Defendant’s contracts with TPDs quite naturally did not include separate prices for all components of the contract price; breaking down the contract *148prices into components such as packaging, accessories, probes, and scopes became necessary only because of the terms of our remand order of June 24, 1977. To determine what part of each contract price should be included in the compensation base, Glassman started with the contract price paid by defendant and then made deductions for items which the court had directed to be excluded from the compensation base. In so doing, he was compelled by the state of the evidence to rely to some extent on pricing by analogy and guesswork.
Glassman testified that he tried to give plaintiff the benefit of doubts in such computations, but it is clear that plaintiff was not always favored by the witness’s approach. For example, when deciding how much to deduct from the contract price to account for excludable costs, he sometimes used plaintiffs list price for individual items sold separately. But this approach ignored considerations which undermine the analysis.
For one thing, plaintiffs prices were not necessarily identical to those of Hickok, Jetronix, and Lavoie. Plaintiffs prices for scopes were substantially higher than those of the TPDs, and there is reason to conclude some differences would be found with respect to accessories. Plaintiff had to develop the scopes through an expensive process of research and development, while the parties from whom the Government bought merely copied them. Thus, plaintiff had higher indirect costs to recover, and its accessories might have been priced higher than, for example, Hickok’s. Defendant attempts to meet this argument by saying there would have been price competition on the unpatented items sold by plaintiff and TPDs, so higher indirect costs, if any, could not have increased the price for accessories. But we doubt that normal price competition would have obtained if plaintiffs patent rights had been respected instead of abused; the market power created by the patents would have afforded plaintiff some protection from price competition on the accessories. Packaging costs, of course, would likewise have enjoyed some immunity from price competition.
Another flaw in Mr. Glassman’s approach was that he did not always select the appropriate comparison between *149probes purchased from TPDs and those listed in plaintiffs catalogue. With respect to most contracts, Glassman made a $22 deduction for each probe, in reliance on plaintiffs list price of $22 for a probe having the necessary capability. But plaintiff has shown that the requirements of some of the contracts could have been met by using probes which plaintiff listed at $8.50, thus establishing that the deductions to be made from the compensation base were overstated by Glassman’s approach. In fact, the $22 price was used even where the contract and delivery antedated the development of that higher priced probe. These are but examples plaintiff has cited to show the imperfection of Glassman’s approach.
Also worthy of mention is another key defect in the approach taken by defendant’s expert. His use of plaintiffs list price to determine the amount to be deducted from contract prices for one of several items simultaneously purchased ignores the obvious problem that list prices for items sold separately undoubtedly are higher than the cost of including such items in a package deal. A simple analogy demonstrates the point; no one would establish the price for a 1978 automobile by adding up the list prices for replacement parts necessary to build the car. Hence, defendant’s use of plaintiffs catalogue list prices for accessories results in an inflation of the amount to be deducted from the contract prices paid by defendant, and it improperly reduces the compensation base on which plaintiffs damages must be premised. The trial judge thus had good reason to doubt the accuracy of the computations proferred by defendant, but he was limited by our remand order from reopening the record for new evidence, which might not have been helpful in any case. Accordingly, he made the "empirical” judgment that defendant’s exhibits should be used but that the deductions proposed therein for accessories and packaging should be reduced by 50 percent and 33 percent, respectively.
We uphold the trial judge’s determination of the compensation base. Defendant’s accounting exhibits were a valid starting point for the computation, but our remand order did not bind the trial judge to accept defendant’s figures as an accurate determination of the compensation *150base. Had those figures been accepted by the court as they stood, this issue would not have required remand to the Trial Division. The adjustments made by the trial judge have not, defendant points out, been justified with respect to each item and to the penny. Nonetheless, we accept the trial judge’s determination as being in the nature of a "jury verdict,” which on numerous occasions we have said is the best measure of damages that the record permitted. See, e.g., Calhoun v. United States, 197 Ct. Cl. 41, 49-50, 53-54, 453 F.2d 1385, 1389-90, 1392-93 (1972) (approving "reasonable approximation”). Cf. Tektronix, supra, 213 Ct. Cl. at 271, 552 F.2d at 351. Our desire to expedite the concluding phase of this ancient litigation, in which substantial interest costs are accruing daily, justified the decision to foreclose further additions to the record. Yet the trial judge was not precluded from adjusting the figures in defendant’s exhibits where reason for so doing was adequately established. Plaintiff has shown the reasonableness of the adjustments, and the compensation base computation of the trial judge is affirmed.
We have previously established the reasonable royalty rate to be applied in this case as 10 percent. Taking 10 percent of the figure computed by the trial judge, $22,432,201, we find the basic compensation (exclusive of delay compensation) to which plaintiff is entitled to be $2,243,220.
The second matter in dispute is the interest rate to be applied in the computation of delay compensation. Our 1977 opinion in this case instructed the trial judge to apply the rates established in Pitcairn, supra, "unless and until it is affirmatively demonstrated that under the theory of the Pitcairn opinion the rate for years after 1975 should differ from the 7%% rate set for 1971-1975.” 213 Ct. Cl. at 274, 552 F.2d at 352-53. The trial judge has found that plaintiff has made such an affirmative showing. Evidence of Moody’s Corporate Bond Yield Averages and of the current Treasury rate paid under 26 U.S.C. §6621 (Supp. V, 1975) for interest on overpayment of taxes were cited by the trial judge as warranting the application of an 8-percent rate beginning January 1, 1976. Plaintiff has offered other judicially noticeable evidence which strengthens the con-*151elusion advanced by the trial judge. Defendant’s principal objection to the 8-percent rate is that the Moody’s averages on which the trial judge primarily relied were in evidence in the Pitcairn case, where a 714-percent rate was established. But it should be noted that the Pitcairn opinion focused solely on the period before 1976. Although the 7%-percent rate was ultimately applied to a period extending beyond 1975, the Pitcairn opinion, and our 1977 opinion in this case, left the door open for proof that a different rate should be applied in years after 1975. Plaintiff has supplied such proof in this case, and thus the Pitcairn rates will not be applied for the period beginning January 1, 1976.
This result is not to be taken as a ruling that the 8-percent rate should be applied in other cases for the post 1975 period. The Pitcairn rates will continue to be applied unless a party affirmatively demonstrates that a different rate should be applied for the period after 1975. For example, at the end of the 1976-80 quinquennium, it would be possible to establish a uniform rate for the period, and all litigation then pending could be subject to that rate.
We cannot agree with the trial judge’s disposition of the third issue, the timing of delay compensation. Having established the compensation base, computed basic compensation, and determined the interest rates to be applied, the trial judge next determined the time from which interest should begin to accrue. We see two faults with his approach to this problem.
First, he regarded the dates of contract execution (between defendant and TPDs), rather than the dates the infringing items were delivered, as the critical dates for delay damage computation. But the statute by which the Government has expressed its consent to be sued for patent infringement provides that its liability should depend on the "use or manufacture” by or for the United States of a patented item. 28 U.S.C. §1498(a)(1970). There is no evidence to establish the dates of manufacture of the infringing items, but plaintiff asks us to accept the dates of contract execution as a substitute. This requires acceptance of the unlikely assumption that the items manufactured by TPDs were already manufactured before defendant contracted to buy them. The record does not establish the *152truth of that proposition. By contrast, there is solid evidence, undisputed by the parties, of the times when infringing items were delivered to defendant. In this case it is simple and logical to follow the general rule that "Liability ordinarily attaches when the device is delivered to the purchaser and hence is available for use.” Coakwell v. United States, 178 Ct. Cl. 654, 658-59, 372 F.2d 508, 511 (1967).
Plaintiff, however, attempts to derive support for the trial judge’s approach from our statement in Pitcairn, supra, 212 Ct. Cl. at 181, 547 F.2d at 1115, that a taking occurs when the Government procures or uses infringing items. Plaintiff understands the word "procure” to refer to the time when the contract is executed. But we understand the word procure to have its usual meaning. Webster’s Third New International Dictionary (1965) defines "procure” to mean: "la(l): to get possession of: obtain, acquire.” Such a reading of "procure” is necessary because this meaning alone can comport with our decisions concerning the statute of limitations in patent cases. See Coakwell, supra; Irving Air Chute Co. v. United States, 117 Ct. Cl. 799, 93 F.Supp 633 (1950). Should the cause of action for patent infringement arise at the time of contract execution, there would be uncertainty about the scope of the interest taken by the Government. See Irving Air Chute Co., supra. Change orders might either avoid or exacerbate the infringement. Too, there would be a danger in some cases that the Government’s infringement would go long undetected and might permit the running of the statute of limitations, since the Government’s execution of a contract for infringing items might not attract the attention that delivery to the Government of those items would. A careful reading of the numerous cases in which the time of procurement is a point of focus leaves us convinced that the court is using the word in accordance with common usage and the dictionary definition quoted above. Put simply, these considerations compel the conclusion that the key dates here are the delivery dates, not the contract dates. Plaintiff admitted at oral argument that it could find no case holding that interest should accrue from the date of *153execution of a contract which contemplated the delivery of infringing items.
Our second problem with the trial judge’s analysis was his failure to award interest as part of delay compensation for the period between the acts giving rise to liability for infringement and the end of the year in which each such act occurred. Plaintiff argued that, if the time of contract execution were not used as the starting point for interest accrual, then interest should commence on the "weighted average delivery date” for each contract. The weighted average is based on the number of items from each contract delivered on specific dates, from which a single, averaged delivery date is computed for each contract. The parties agree that the weighted average is an acceptable method of simplifying the accounting problem caused by scattered deliveries over long periods. Only plaintiff, however, argues that interest on each infringement should begin immediately after the weighted average delivery date, and, of course, this argument is regarded by plaintiff as a poor alternative to computation of delay compensation beginning with contract execution.
The trial judge rejected the theory of plaintiffs argument, saying that ease of computation justified adopting the Pitcairn approach. Interest was awarded not from the date of the taking but from the first day of the year following execution of the contract for each infringing item.1 Defendant seems to endorse this approach, since delaying the accrual of interest until the first year of infringement ends obviously reduces the interest which is paid. But we again must caution that the approach adopted in Pitcairn was premised on the evidence in that case. This was clear from our statement that the delay damage computation was based on "all the circumstances involved in this litigation.” 212 Ct. Cl. at 192, 547 F.2d at 1121. That statement immediately followed reference to one of plaintiffs important commercial royalty agreements, which *154expressly provided for some delay between the acts triggering the royalty obligation and the time when that royalty would be paid; plaintiffs commercial practice thus warranted delay damage computation based on use of plaintiffs patent within specified periods, and interest quite naturally was awarded only from the close of the period.2 In this case, by contrast, there is no evidence that plaintiffs normal licensing practice was to collect royalty payments only at the end of specified periods. Nor is there any difficulty in computing damages accumulating from the averaged delivery dates themselves. Thus, there is no justification or need for withholding from plaintiff the interest which accumulated during the remainder of each year for patents infringed during such years. Our statutory duty is to provide "reasonable and entire compensation” to plaintiff. 28 U.S.C. §1498 (1970). Under the facts of this case, that mandate requires that plaintiff be awarded interest beginning with the weighted average delivery date for each of the 71 contracts under which defendant obtained items which infringed plaintiffs patents.
It being settled, then, that plaintiff is entitled to delay damages computed from the weighted average delivery dates, computation of the damages requires a separate damage computation for each of the 71 contracts. Fortunately, such computations have already been made and were submitted to the trial judge in plaintiffs brief of July 11, 1977, as Alternative B.3 Defendant objected to those figures because it preferred the Pitcairn approach of looking to the first day of the year following infringement; no objection, however, was made to the mathematical accuracy of the figures. Plaintiffs brief to the court cited the total damages due under this method; again, defendant did not object.
We have computed the interest for several of the contracts and find plaintiffs factor approach trustworthy. Evidently, we applied slightly different accounting prem*155ises, for our results differed minutely from plaintiffs.4 But the differences were so small that the time spent recomputing the factors would permit far more interest to accumulate than is at stake in the minor readjustments we would make. It should not be forgotten that interest is accumulating on this case at the rate of hundreds of dollars each day. Nor should it be forgotten that the record here required some empirical judgments which introduced some estimation into the damage computation process; the compensation base was fixed as a "reasonable approximation” only. These circumstances require that we accept the figures supplied by plaintiff. Accordingly, plaintiff is entitled to $1,995,087 for interest to June 30, 1977, plus $491.66 per day thereafter.
Judgment is entered for plaintiff in the amount of $4,238,307, plus $491.66 per day from July 1, 1977, until payment.
The trial judge did not, however, follow his own theory, since his opinion appears to be based on a submission of plaintiffs, in which factors were supplied for computation of delay damages. Those factors, which the trial judge incorporated in his opinion, were based both on contract execution dates and weighted average delivery dates, not the first day of the years following.
The court appears to have regarded immediate payment at the end of an annual period as the substantial equivalent of the royalty agreement’s provision for payments 45 days after the close of semi-annual periods.
Alternative A reflected plaintiffs preference for the date of contract execution rather than a delivery-focused approach.
For example, our computation of the interest due on the first contract listed in Schedule II reaches the result of $51,141.10, while use of plaintiffs factor produces a result of $51,146.65, a difference of $5.55 (just over one-hundredth of one percent).