The owner of the Ore Chief, a bulk ore carrier that went aground in the Delaware River, sued the federal government for negligence, alleging that a channel depth statement the Corps of Engineers had published was inaccurate. The district judge found for the defendant, because he was “not convinced by a preponderance of the evidence presented that the Ore Chief went aground because of the government’s conduct in issuing a channel depth statement.” 388 F. Supp. 276, 288 (E.D.Pa.1974). The plaintiff has appealed. We affirm.
We have considered the six contentions raised by appellant and have concluded that the facts found in this non-jury trial were not clearly erroneous, Krasnov v. Dinan, 465 F.2d 1298 (3d Cir. 1972), and that the district court did not err in its legal conclusions drawn therefrom, as more particularly set forth in the opinion of Judge Daniel H. Huyett, III, supra.
The maritime complaint filed on December 29, 1966, and finally adjudicated December 19, 1974, alleged 11 specific acts of negligence, as well as negligence “in other respects which will be shown at trial.” Plaintiff tried the case on the theory that the Ore Chief struck large indigenous rocks in the Delaware River bed; that the channel depth above the rocks was inaccurately reflected in the Corps of Engineers’ published channel depth statement; and that the government was, therefore, liable in negligence for the damage from the accident. The government, for its part, made no serious attempt to offer an alternative explanation of the accident; instead it simply sought to discredit plaintiff’s theory of the accident, and relied on the plaintiff’s burden to prove negligence by a preponderance of the evidence.
The plaintiff was very specific in the theory under which it put the case. It argued that the Ore Chief struck particular rocks, at a particular point in the channel. It sought to establish this to show both that the government’s depth statement was inaccurate, and that the inaccuracy caused the accident. Plaintiff’s trial memorandum filed November 5, 1973, stated that certain “witnesses will establish the location of the rocks which caused the damage. Certain tests were performed by Ambric Testing Laboratory as to fresh paint on the rocks and a representative from the Laboratory will testify that paint on the rocks was the same as that on the bottom of the ORE CHIEF.” (Trial Mem. 3.) The memorandum also stated: “The constant surveying of this area by the United States with the use of electronic sounding machines quite obviously failed to detect the rocks. Such a repeated surveying by the United States of this area constitutes constructive notice of the presence of the undetected indigenous rocks as part of the bottom, since by proper methods the United States should have known of their presence and their heights and the depth of the water over the rocks so that the depths published would be accurately reported to mariners.” (Trial Mem. 13.)
In its Proposed Findings of Fact Before Trial (P.F.B.T.), the plaintiff asked the court to find the presence in the channel of three specific rocks: the “principal” or “log” rock, the “plateau” or “table” rock, and a third, unnamed rock. Plaintiff also asked the court to find that the “top surface of the log rock was newly chipped or ruptured in two separate areas,” and that “[sjamples obtained from the log rock and the third *75rock contained fresh paint scrapings.” P.F.B.T. 13. Plaintiff sought further to support its theory that the Ore Chief collided with these rocks by proposed findings 15 and 16, which stated:
15. The ORE CHIEF thus came into contact at least with the log rock and the third rock. Both were uncharted obstacles within the 40' x 400' channel. Before it could be determined if the vessel had also came [sic] into contact with the plateau rock, the rock was destroyed .
16. The weight and force of the ORE CHIEF knocked loose the upper portion of these rocks which caused the grounding.
The plaintiff did not change its theory after the close of testimony. In its Proposed Findings of Fact After Trial (P.F. A.T.) it asked the court to find that there were two indigenous rocks — log and table — plus a third rock; and that “[a]ll three of these rocks showed fresh abrasions, indicating recent contact with the bottom of the ORE CHIEF”. P.F. A.T. 27. Plaintiff also asked for a finding that “[i]t was uncontradicted that [a log] rock sample had red ship’s paint on its surface. . . . Ambric Testing Laboratories, Inc. analyzed the sample and concluded that the red paint on the rock was the same paint as applied to the bottom of the ORE CHIEF on the occasion of her last drydocking in July, 1964 prior to the grounding.” P.F.A.T. 28. Additional suggested findings related to the size and location of all three rocks, P.F.A.T. 27; the obtaining of the log rock sample, P.F.A.T. 29; and the indigenous nature of the log and the table rocks, P.F.A.T. 27 & 31. After thus requesting numerous specific findings describing these three rocks in detail, plaintiff requested a finding that:
The uncontradicted preponderance of credible evidence established that the weight and force of the ORE CHIEF collided with and fragmented the upper portion of these indigenous rocks which caused the grounding.
P.F.A.T. 35.
In its reply brief on appeal plaintiff advances an argument in the nature of res ipsa loquitur quite at variance with its trial strategy. Plaintiff asserts that the “most persuasive argument” for reversal “is based on certain of the Court’s findings which were fully supported by the record: (1) the vessel grounded in the channel and (2) the vessel sustained the alleged damage on that occasion.” Reply Brief at 1. We do not decide whether there were such “findings”,1 or whether they were “fully supported by the record,” because, whatever merit such a theory might have had at trial, it has none now.
Federal procedure relies on notice pleading rather than fact pleading, but at a minimum, as explained by the Advisory Committee on Civil Rules in October, 1955, the pleader is required “to disclose adequate information as the basis of his claim for relief”. 2A J. Moore, Federal Practice If 8.01[3] (2d ed. 1974). Conley v. Gibson, 355 U.S. 41, 47, 78 S.Ct. 99, 103, 2 L.Ed.2d 80 (1957), teaches that “the defendant [be given] fair notice of what the plaintiff’s claim is and *76the grounds upon which it rests,” (emphasis supplied), and we have recently reiterated that the defendant is entitled to “fair notice of the claim asserted.” Joiner Systems, Inc. v. AVM Corp., 517 F.2d 45, 47 (3d Cir. 1975). Rule 15 of the Federal Rules of Civil Procedure fits the pattern, allowing liberal amendment of pleadings to conform to issues actually tried by express or implied consent of the parties. However, a different theory of recovery may not be urged on appeal where prejudice would result to the other party.2 The test for prejudice in this context is whether the other party “had a fair opportunity to defend and whether it could offer any additional evidence on the different theory.” Jurinko v. Edwin L. Wiegand Co., 477 F.2d 1038, 1045 (3d Cir.), vacated on other grounds, 414 U.S. 970, 94 S.Ct. 293, 38 L.Ed.2d 214 (1973); 3 J. Moore, Federal Practice 1115.13[2], at 993 (2d ed. 1974); see United States v. 47 Bottles, More or Less, 320 F.2d 564, 573 (3d Cir.), cert. denied, 375 U.S. 953, 84 S.Ct. 444, 11 L.Ed.2d 313 (1963).3
We feel that consideration of plaintiff’s argument in the nature of res ipsa loquitur would be prejudicial to the government. At the very least; if the government had been aware of this theory at trial it might have been more contentious about the facts of the accident. The government stipulated to several facts, including the Ore Chief’s draft, and the condition of the tide, and, as indicated, relied primarily on the plaintiff’s burden of proving the government negligent rather than offering its own version of the accident. Accordingly, we decline the invitation to reverse the district court on the theory in the nature of res ipsa loquitur, which was not raised in the district court. See TMA Fund, Inc. v. Biever, 520 F.2d 639, 641 (3d Cir. 1975); United States v. Greenlee, 517 F.2d 899, 905 (3d Cir.), cert. denied,U.S. -, 96 S.Ct. 391, 46 L.Ed.2d 301 (1975); Rhoads v. Ford Motor Co., 514 F.2d 931, 933-34 (3d Cir. 1975).4
In its role as a fact finder the court summarized the duty before it:
It is clear from the record in this case that the government has assumed the duty of surveying portions of the Delaware River. The question before us is whether or not the government breached that duty.
Our resolution of that question is almost exclusively a factual one. Except for the claim that a negative inference should be drawn against the government because of the removal of the rocks, this case requires applying the preponderance of the evidence standard to the facts before us.
388 F.Supp. at 285 — 86 (footnotes omitted). The court performed its duty, ar*77ticulating clearly both its fact finding and its legal conclusion:
We have not found convincing the plaintiff’s propositions concerning the existence of the rocks the Ore Chief is said to have hit and the nature of the damage to the rock samples produced by plaintiff. We credit the testimony of the government’s geologist that the paint on one rock sample could not have been placed there by the kind of impact which would have resulted if the Ore Chief had hit it. Further, plaintiff could not establish convincingly that the rocks were indigenous to the river bed.
Ibid, at 288.
While we consider the evidence supporting plaintiff’s theory of the case to be substantial, we have concluded, after a careful consideration of the nature and quality of plaintiff’s evidence, that plaintiff has failed to prove it more probable than not that the Ore Chief went aground in a collision with rocks the water above which had a depth less than that reported by the government in its notices to mariners. We will, therefore, enter judgment in favor of the government.
Ibid, at 277-78.
Upon an independent examination of the record, we do not consider the court’s fact finding clearly erroneous. These facts being vital to the theory under which plaintiff tried its case, there was no error in the trial court’s conclusion.
The judgment of the district court will be affirmed.
. Plaintiff asserts that the trial court “found definitively that the ORE CHIEF had grounded within the 40 foot project channel,” citing Finding 43. Reply Brief at 2. This misreads that finding. Finding 43 merely sets forth testimony of plaintiff’s witnesses about the accident; it does not adopt the testimony as the court’s finding.
In its reliance on Findings 43 and 46, the dissent has fetched up on the same shoals. Findings 43 through 49 all begin as follows: “The testimony of plaintiff’s witnesses was that . . Finding 42 states that “[t]he nature of the damage to the Ore Chief was irreconcilable with the testimony of plaintiffs witnesses as to how the accident occurred.” When a trial court sitting as a fact finder has refused to find a fact, no appellate court, nor judge thereof, can elevate to “fact” that which the fact finder refused to find. The device of setting plaintiffs contentions in quotation marks and italics, while evidencing enthusiasm, (see dissent pages 78, 79), lacks jurisprudential potency to convert a non-fact into a fact-found.
. Recently, the Supreme Court had occasion to comment on the propriety of entertaining a point not raised in a petition for certiorari or a cross-petition. “Because it would alter the judgment of the Court of Appeals, which like that of the District Court had effectively put an end to the litigation, rather than providing an alternative ground for affirming it, we will not consider the argument when raised in this manner.” Rondeau v. Mosinee Paper Corp., 422 U.S. 49, 61-62 n. 11, 95 S.Ct. 2069, 2077, 45 L.Ed.2d 12 (1975).
. We recognize that, in Jurinko, we considered the different theory on appeal. There, however, we found no prejudice to the defendant. Moreover, the theory was advanced to affirm the judgment of the district court, rather than to reverse it. Jurinko, supra, 477 F.2d at 1045. Compare n. 2 supra.
. Apparently disagreeing with the Supreme Court’s decision in Conley v. Gibson, supra, the dissent does not choose to meet the issue jurisprudentially or prudentially; instead it makes accusations of “gamesmanship” and, from a jural quarterdeck, observes that res ipsa loquitur “was in the case all along”. Although this theory was not invoked by the plaintiff’s attorneys at trial and escaped detection by the district judge, the dissent would discover it in the factual allegations and consider it on appeal. This was not pro se litigation. Plaintiff’s lawyers had this case in the district court for eight years — from December 29, 1966, to December 19, 1974. The dissent faults us for expecting too much — for expecting that some time during those eight years plaintiff’s counsel would have shared with the trial court and the defendant a theory they disclosed for the first time in a reply brief on appeal.