¶1concurring in the judgment:
¶2I concur in the judgment of the majority that the jury verdict for the plaintiffs should be overturned. I write separately, however, because I believe that such a result can be justified only if this court reviews the verdict under the special independent review standard set out in Bose Corp. v. Consumers Union of United States, Inc., 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984).
¶3Both the majority and dissenting opinions leave the distinct impression that neither opinion adheres to the standard of review it claims to be employing. The majority purports to apply the traditional standard for overturning a jury verdict, i.e., the reviewing court must accept all “permissible” inferences the jury might draw from the objective facts it finds. See maj. op. at 777-78. Nonetheless, throughout its opinion it scrutinizes the inferences to be drawn from evidence of actual malice far more rigorously than that ordinary judgment n.o.v. standard would permit. The dissent, on the other hand, accepts the independent review principle of Bose as applicable, but then proceeds to interpret it as requiring that an appellate court give complete deference not only to the jury’s presumed factual findings but also to any reasonable inferences from those findings that might support the verdict. The court’s independent review is to be restricted to the “ultimate constitutional issue.” See diss. op. at 816-17, 830.
¶4Because I believe that this highly publicized and controversial case should be reviewed according to an articulated standard that comports with what the court actually does, I will set out the standard I think should apply here. I find this necessary not only because I do not think an ordinary judgment n.o.v. standard, or even the preRose independent review standard the majority claims to apply
¶5*82At a minimum, I believe Bose requires an appellate court independently to apply the legal standard of actual malice to the facts established by the evidence; this independent review is “a constitutional responsibility that cannot be delegated to the trier of fact.” Bose, 446 U.S. at 501, 104 S.Ct. at 1959.
¶6Were we to follow the dissenting opinion’s interpretation of Bose, I fear it would prove impossible in virtually all libel cases for a reviewing court to defer to all permissible inferences the jury may be presumed to have drawn, and then to conduct a truly independent review of the jury’s ultimate determination that the defendant acted with actual malice. The human mind is just not that agile, and indeed that is why judgments n.o.v. are so unusual; something must be very wrong before a reviewing court can give the jury the benefit of every reasonable inference and still find evidence lacking for their verdict, even under the preponderance standard. For example, in this case, once the dissent accepts the jury's “reasonable” inferences that the Peterson memorandum “raise[d] in [the defendants’] minds subjective doubts as to the accuracy of the story,” that “the defendants knew that Comnas was not trustworthy,” that the Post’s “extreme pressure for sensationalistic stories [led] to distortion of the truth,” and so on through each of the allegations made by the plaintiffs, see diss. op. at 829-39, there is nothing left for an appellate court to review: the conclusion that the defendants acted with actual malice follows inevitably. Independent review of the jury’s inferences concerning the existence of actual malice is necessary if we are to fulfill our responsibility to decide “independently ... whether the evidence in the record is sufficient to cross the constitutional threshold that bars the entry of any judgment that is not supported by clear and convincing proof of *83‘actual malice.’ ” Bose, 466 U.S. at 511, 104 S.Ct. at 1965.
¶7The majority, on the other hand, while stating that it need not apply any special post-Rose “independent review” standard to arrive at the determination that the defendants did not act with actual malice, in fact does reexamine and reject “permissible” inferences which the jury might have drawn to support their verdict. For example, the majority acknowledges that the jury could find that copy editor Christine Peterson had conveyed to the defendants that she found “the story’s theme impossible to believe,” and it recognizes that “the Peterson memorandum is evidence of ... actual malice.” Maj. op at 794. Nevertheless, the majority concludes that this evidence will not support an inference of actual malice with respect to the allegation that Tavoulareas “personally urged” Comnas to accept Peter as a partner, the one statement in the article which the majority holds to be both false and defamatory. See id.at 794 n. 44. Similarly, the majority concedes that there was evidence casting doubt on the reliability of Comnas as a source for the article, including “evidence of personal animus between Comnas and Tavoulareas,” but rejects the inference that reliance on Comnas could be viewed as indicative of actual malice. See maj. op. at 790-93. I could go on, but suffice it to say my own review of the record convinces me that the majority does in fact subject the evidence of actual malice to independent review of inferences to be drawn from it, as indeed it should under Bose, in order to assure that malice has been proven by “clear and convincing” evidence, and as I believe it must in order to arrive at its ultimate judgment, with which I agree. What I find disturbing, though, is its reluctance to acknowledge what it is doing or to grapple with the articulation of what is a proper application of Rose to the record. For future cases we badly need to clarify the appropriate role of a reviewing court in this volatile area of the law, and in this case, regrettably, we have lost a valuable opportunity to do so.
¶8. The majority states that it "conduct[s] the requisite independent review in this case, without relying on Bose — ” Maj. op. at 789 n. 37 (emphasis in original). Under this pre-Bose form of independent review, the majority explains, reviewing courts in defamation cases have traditionally "carefully and independently reviewed findings of fact and rejected jury verdicts when all favorable inferences did not constitute clear and convincing evidence of actual malice." Maj. op. at 778.
¶9.Although the issue in Bose was the standard of review to be applied to a trial court's determination that a statement was made with actual malice, the Court indicated that its holding applied to review of a jury verdict as well: "[Tjhe rule of independent review assigns to judges a constitutional responsibility that cannot be delegated to the trier of fact, whether the factfinding function be performed in the particular case by a jury or by a trial judge." 466 U.S. at 501, 104 S.Ct. at 1959; see also id.at 508 n. 27, 104 S.Ct. at 1964 n. 27 (Court announced rule of independent review in New York Times v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), in which it was reviewing a state court judgment entered on a jury verdict).
¶10In his dissenting opinion in Bose, Justice Rehnquist argued that review of a jury verdict "presented the strongest case for independent fact-finding by this Court:"
The fact-finding process engaged in by a jury rendering a general verdict is much less evident to the naked eye and thus more suspect than the fact-finding process engaged in by a trial judge who makes written ñndings as here. Justifying independent review of facts found by a jury is easier because of the absence of a distinct “yes” or "no" in a general jury verdict as to a particular factual inquiry—
¶11466 U.S. at 518 n. 2, 104 S.Ct. at 1953 n. 2 (Rehnquist, J., dissenting).
¶12. See Bose, 466 U.S. at 514 n. 31, 104 S.Ct. at 1967 n. 31 ("only those portions of the record which relate to the actual malice determination must be independently assessed”).
¶13. According to a leading commentator, the Supreme Court held in Bose that
[a]n appellate court cannot content itself with accepting the results of “reasonable" application of admittedly correct legal norms to the historical facts. The court’s responsibility is to scrutinize the record and marshal the evidence to see if it yields the characterization put on it by the court below.
¶14Monaghan, Constitutional Fact Review, 85 Colum.L.Rev. 229, 242 (1985) (footnote omitted).
¶15. This reading of Bose is consistent with the Supreme Court’s discussion of the evidence of actual malice in that case. The Court accepted the district court’s finding that a statement in the review challenged as libelous was an inaccurate description of what the engineer who prepared the report on which the review was based had actually perceived, but it rejected the district court’s inference from this evidence that the engineer "must have realized the statement was inaccurate at the time he wrote it.” 466 U.S. at 512, 104 S.Ct. at 1966.
¶16concurring:
¶17I concur in the court’s opinion holding that the evidence in this case was insufficient to warrant submission to a jury’s judgment. That decision, however, is more securely made at a distance from the trial itself. Immediately at the close of a hard-fought trial, the presiding judge, who must rule at once upon a motion for a directed verdict, is in a less advantageous position. He or she may prefer to abide the jury’s report and, if necessary, thereafter consider, with some space for reflection, whether to grant a requested judgment n.o.v. Despite our decision today, therefore, hotly contested cases of this genre are likely to reach juries again. With that prospect in view, I write separately to suggest what a trial judge might do to enhance the jurors’ capacity to comprehend and impartially apply the governing law.
¶18Keeping the jury on track poses a formidable challenge for the judge in a libel case governed by the “actual malice” standard of New York Times v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964). As Justices Black and Goldberg anticipated, the risk is considerable that jurors will not comprehend the difference between reckless disregard and mere neglect or carelessness, or will confuse or blend the separate issues of falsity and actual malice. See id. at 293, 84 S.Ct. at *84733 (Black, J., concurring) (malice “is an elusive, abstract concept, hard to prove and hard to disprove”); id.at 298 n. 2, 84 S.Ct. at 736 n. 2 (Goldberg, J., concurring) (“The requirement of proving actual malice or reckless disregard may, in the mind of the jury, add little to the requirement of proving falsity, a requirement which the court recognizes not to be an adequate safeguard.”). The challenge for the trial judge demands attention throughout the proceedings. It is not met by allowing the jurors to listen, without education, as the evidence unfolds and then submitting the case for their general verdict after “dous[ing] [them] with a kettleful of law during the charge that would make a third-year law-student blanch.” See Skidmore v. Baltimore & O.R.R., 167 F.2d 54, 64 (2d Cir.) (Frank, J.) (quoting Bok, I, Too, Nicodemus (1946)), cert. denied, 335 U.S. 816, 69 S.Ct. 34, 93 L.Ed. 371 (1948). See generally Schwarzer, Communicating with Juries: Problems and Remedies, 69 Calif.L.Rev. 731 (1981).
¶19Instead, the judge should act to reduce the risk that “the protective benefits of the Sullivanrule [will become] mythical.” Lewis, New York Times v. Sullivan Reconsidered: Time to Return to “The Central Meaning of the First Amendment,” 83 Colum.L.Rev. 603, 617 (1983). The trial judge can explain, instruct, or clarify continuously, from the commencement of the trial through to the jury’s deliberations. The means at hand include: instructing the
¶20jury, in plain English, at the opening of the case and at appropriate times during trial,
¶21I.
¶22Experience supports the intuition that jurors have considerable trouble “distinguishing actual malice from mere negligence.” Note, Punitive Damages and Libel Law, 85 Harv.L.Rev. 847, 860 (1985); see Lewis, supra, at 613-14. As Judge Bork has observed, “[t]he evidence is mounting that juries do not give adequate attention to limits imposed by the first amendment and are much more likely than judges to find for the plaintiff in a defamation case.” Ollman v. Evans, 750 F.2d 970, 1006 (D.C.Cir.1984) (concurring opinion); see Kaufman, Libel 1980-85: Promises and Realities, 90 Dickinson L.Rev. 542, 552 (1986) (citing studies by Libel Defense Resource Center documenting jury verdicts against media defendants at rates of between 60 and 90 percent); Franklin, Suing Media for Libel: A Litigation Study, 1981 Am.B.Found.Res.J. 797, 804, Table 7, nn. 19, 20 (contrasting decisions for libel plaintiffs in only two of five cases tried without a jury, *85with jury verdicts for plaintiffs in 20 of 24 cases).
¶23In the libel area particularly, it is not a large exaggeration to suggest that jurors “can easily misunderstand more law in a minute than the judge can explain in an hour.” Sunderland, Verdicts, General and Special, 29 Yale L.J. 253, 259 (1920); see Charrow & Charrow, Making Legal Language Understandable: A Psycholinguistic Study of Jury Instructions, 79 Colum.L.Rev. 1306, 1316 (1979) (jurors were able to paraphrase correctly only 54% of their instructions); see also Severance & Loftus, Improving the Ability of Jurors to Comprehend and Apply Criminal Jury Instructions, 17 L. & Soc.Rev. 153, 179-80, Table 2 (1982) (overall rate of incorrect responses for jurors with standard instructions was 29.6%); Note, Jury Instructions v. Jury Charges, 82 W.Va.L.Rev. 555, 558 (1980) (citing O’Mara & von Eckartsberg, Proposed Standard Jury Instructions— Evaluation of Usage and Understanding, 48 Pa.B.A.Q. 542, 550 (1977)) (in 1976 Pennsylvania study, only 68% of jurors could demonstrate actual comprehension of some standard instructions).
¶24II.
¶25To arm the jury with the information needed for the intelligent performance of its task, the judge might first endeavor to speak the language of the jurors, and avoid the jargon of the legal profession. Terms familiar to a lay audience, and shorter, less complex, more active sentences, increase the chance that the jurors will understand and recall the judge’s instructions. In addition, in lieu of a laundry list of law without discernible structure, a logically-ordered set of directions for consideration and decision of the separate issues should aid the jury to start out and remain on track. See Charrow & Charrow, supra, at 1326-27, 1338-39; Note, Jury Instructions, supra, at 558-59 & n. 17.
¶26Second, instructions should not be reserved exclusively for the trial’s finale. Enlightenment from the judge at the outset fosters “intentional” learning; by informing the jurors initially of the inquiries they are ultimately to undertake, early instructions may stimulate or facilitate the mental process involved in connecting up issues and key items of evidence. See Note, Jury Instructions, supra, at 562 & n. 37. Mid-stream instructions may be useful as well. They can serve as definers of difficult-to-grasp law at the moment that law bears on evidence produced. See id. at 565-67. The very repetition of instructions on elusive concepts should lessen the jury’s sense of insecurity or discomfort in dealing with those concepts.
¶27In addition, “[tjhere is a clear suggestion in the available research that access to written instructions in the jury room improves the quality of deliberations. This may be due in part to the optimal effects of presenting material in both visual and auditory modes.” Severance & Loftus, supra, at 155 n. 4 (citing Forston, Sense and NonSense: Jury Trial Communication, 1975 B.Y.U.L. Rev. 601 (1975)); see Cunningham, Should Instructions Go Into the Jury Room?, 33 Calif.St.B.J. 278 (1958); Katz, Reinstructing the Jury by Tape Recording, 41 J.Am.Jud.Soc’y 148 (1958). Detailed instructions delivered from the bench on the actual malice issue, for example, might accompany the jurors in written form as they proceed to deliberations.
¶28Finally, the special verdict or its cousin, the general verdict accompanied by answers to interrogatories, see Fed.R.Civ.P. 49, may be a particularly useful check against jury misconstruction or misapplication of a standard as uncommon as actual malice.
¶29The modern rule places the form of verdict in the hands of the judge and permits broad discretion in designing special verdicts, or general verdicts with interrogatories. In a case governed by the Sullivanstandard, separate questions on the issues of defamatory meaning, falsity, and actual malice would impel the jurors to advert to the framework within which the judge has instructed them to consider the case and could assist them to hold the distinct legal questions in clear and separate view.
¶30New York Times v. Sullivan presents a standard that may slip from the grasp of lay triers unfamiliar with legal concepts and perhaps unsympathetic to publishers who print statements shown to be false. Careful efforts by judges to make the legal rules genuinely accessible to jurors may reduce some of the turbulence in this unsettling area of the law.
¶31. I rely throughout on a concise paper prepared for the Federal Judicial Center describing why and how separate, seriatim verdicts were obtained from the jury in Sharon v. Time, Inc., No. 83-4660 (S.D.N.Y.) (Jan. 16, 1985: defamatory nature of publication), (Jan. 18, 1985: falsity), (Jan. 24, 1985: actual malice). See Sofaer, Jury Management in Sharon v. Time, Inc. (Sept. 15, 1985).
¶32. Some "instruction” might come from counsel rather than from the judge. In Westmoreland v. CBS, No. 82-7913 (S.D.N.Y., settled and dismissed Feb. 18, 1985), Judge Pierre N. Leval allowed counsel an opportunity to address the jury from time to time during the 62-day course of the trial. These "int-sums” were limited to a total of 120 minutes for each side. The lawyers had each used about 100 minutes in int-sums, averaging about two and a half minutes each, by the time the case was settled. Counsel used the int-sums to "explain the significance, strength, or weakness of proof; to point out confirmation or contradiction of other evidence; to introduce new themes; to respond to opposing arguments; and to challenge the adversary’s ability to prove his contentions." Le-val, From the Bench: Westmoreland v. CBS, 12 Litigation 7, 66 (1985). Although Judge Leval acknowledged that both the integrity of counsel and the vigilant control of the judge were necessary to prevent abuse of this system, see id. at 66, he found it a very useful way to keep the jury on track during the course of the long and complex Westmoreland trial. See id.
¶33An additional technique adopted by Judge Le-val in that case may also be useful in reducing jury confusion. The judge put a limit on the total number of hours each side could use to present its case, including both direct examination or introduction of evidence and cross-examination. This approach left the lawyers with complete control over how the time was allocated. See id. at 7-8. The lawyers in the Westmoreland case believed that they had "tried their case better as a result of the time limit, and that [trial time] was shorter by a half than it would have been.” Id. at 8. And the jury was spared the burden of attending to and remembering an even greater quantity of evidence put in over a longer period of time.
¶34. See Parker, Free Expression and the Function of the Jury, 65 B.U.L. Rev. 483, 550-56 (1985) (maintaining that "[clases involving free expression are precisely the type of cases that demand the use of general verdicts with interrogatories of rule 49(b)”).
¶35Objections to special jury findings center on their use in run-of-the-mine cases in which strict application of the rules of law is widely viewed as inconsonant with popular conceptions of "substantial justice." See, e.g., 9 Wright * Miller, Federal Practice and Procedure § 2503 (1971); James, Sufficiency of the Evidence and Jury-Control Devices Available Before Verdict, 47 Va.L.Rev. 218, 246-48 (1961); cf. 374 U.S. 865, 867-68 (1963) (Justices Black and Douglas op*86pose Rule 49 because it allows judges "to weaken the constitutional power of juries"). Compare Morris v. Pennsylvania R.R., 187 F.2d 837, 840-41 (2d Cir.1951) (declaring that courts should use special verdicts only with "discrimination and foresight”) with id. at 843 (Frank, J., concurring) (advocating frequent use of special verdicts contrary to the wishes of "those who appear to esteem the [jury] system, just because it gives rein to the passional element of our nature”) (quoting Skidmore v. Baltimore & O.R.R., 167 F.2d 54 (2d Cir.1948) (Hand, J., concurring)).
¶36. The trial judge in Sharon v. Time, Inc., Abraham D. Sofaer, reported the cooperation of counsel in allowing the jury to have with them during their deliberations detailed instructions on the actual malice issue as well as a jury verdict form which “took the jury step by step through each of the theories on which plaintiff relied." Sofaer, supra note 1, at 2-3. In addition, the parties prepared a chart, at the judge's request, “list[ing] every exhibit in evidence, describing] the exhibit in sufficient detail to enable the jur[ors] to identify the items they wished to see, and indicating] the issue or issues on which each exhibit was admissible.” Id. at 1-2.
¶37. Judge Sofaer observed that the use of separate, seriatim verdicts in the Sharon case made it possible to avoid definitive rulings on unsettled questions of law, e.g., whether “public figure plaintiffs were required to prove falsity by clear and convincing evidence.” Sofaer, supra note 1, at 3. He also noted his decision to bifurcate the trial so that evidence on reputation and damages would not influence the decision on liability. See id. at 4. While the structure he employed worked effectively in Sharon, Judge Sofaer suggested that other cases might call for a different ordering, for example, “the judge could have the juty pass on actual malice first, when that issue appears potentially dispositive and susceptible of decision without confusion.” Id. at 5.
¶38(dissenting).
¶39Rehearing en banc
¶40*87Effectively conceding this issue, Post counsel at oral argument retreated from its position that an appellate court’s independent review of the evidence of actual malice is “not dependent on anything below.” Transcript of Oral Argument at 25 (October 3,1985). Post counsel admitted that an appellate court’s independent review of the evidence of actual malice must “pay some obeisance” to the jury’s assessment of witness credibility. Id. at 26.
¶41The en banc majority opinion, presumably because of Post counsel's admissions at oral argument, refuses to address the important First Amendment issue ordered for briefing in this case — that is, the extent to which judicial deference to jury determinations of contested facts and credibility is called for under New York Times v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964) and Bose Corp. v. Consumers Union of United States, 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984). See Maj.Op. at 776-77. Reducing the conflict between the parties to a primarily factual dispute, the en banc majority fails to heed the principle that “appellate review serves a dual purpose: the correction of legal error and the establishment of legal rules for future guidance. Only the latter is ordinarily worth the attention of the full court.” Church of Scientology of California v. Internal Revenue Service, 792 F.2d 153, 155 n. 1 (D.C.Cir.1986) (en banc). The purpose for holding an en banc hearing— to establish legal rules for future guidance — has been forsaken by the majority in favor of a mere reversal of the panel opinion, in derogation of the rule that “[t]he function of en banc hearings is not to review alleged errors for the benefit of losing litigants.” United States v. Rosciano, 499 F.2d 173, 174 (7th Cir.1974) (per curiam).
¶42Rather than addressing the important First Amendment question presented by this case, the majority instead rejects entirely plausible (if not inevitable) interpretations of the article and then, by considering only a small fraction of the defamation charged, holds the jury’s verdict was invalid under what are alleged to be traditional judgment n.o.v. standards. In determining whether to grant judgment n.o.v., a reviewing court is “required to draw all inferences favorable to the plaintiff.” Carter v. Duncan-Huggins, Ltd., 727 F.2d 1225, 1234 (D.C.Cir.1984) (emphasis in original). “The weighing of conflicting evidence and the evaluation of witness credibility is exclusively within the jury’s province.” Id. See U.S. Industries v. Blake Construction Co., 671 F.2d 539, 550 (D.C.Cir.1982); Vander Zee v. Karabatsos, 589 F.2d 723, 726 (D.C.Cir.1978), cert. denied, 441 U.S. 962, 99 S.Ct. 2407, 60 L.Ed.2d 1066 (1979). After asserting that its review of the evidence in this case is made under “traditional j.n. o.v. standards,” the majority discards those traditional standards and proceeds to make its own assessment of witness credibility and to disregard controlling reasonable inferences in favor of the verdict.
¶43I respectfully dissent because I believe that the majority, in proceeding as above set forth, has denied plaintiff his constitutional right to a jury trial and has improperly failed to conform to the First Amend*88ment independent fact review as established by Bose Corp. v. Consumers Union of the United States, 466 U.S. 485, 511, 104 S.Ct. 1949, 1965, 80 L.Ed.2d 502 (1984).
¶44I. Counterstatement of Facts
¶45This is an unusual libel case and a unique majority opinion. When The Washington Post article at issue is read as a whole with all its innuendo, as it must be, it charges that William Tavoulareas (hereafter also “William”), as president of Mobil Oil Corporation, violated his fiduciary duty to the shareholders of the corporation by causing Mobil’s assets and personnel to be used with intent to benefit his son, Peter, to the detriment of Mobil. The case thus largely turns on whether Tavoulareas established the falsity of that charge. A jury has already determined that William has sustained the burden of proving the falsity of the Post article; and an appellate court under j.n.o.v. standards may only overturn that determination if it finds that no reasonable person could have believed that William proved the falsity of the article. The facts which follow must be read with this principle in mind.
¶46The business relationships that gave rise to this case began in early 1974, when Mobil, after an earlier decision to do so in 1973, bought a 30 percent interest in the Saudi Arabian Maritime Company (“Samar-co”), a joint venture that had been formed a short time before by members of a powerful Saudi Arabian family, the Alirezas, and Fairfield-Maxwell Ltd., a Japanese company that managed ships (RE 2470). The Saudis were the majority shareholders in Samarco. After first refusing an offer to join Samarco, Mobil decided to buy a minority interest after strong indications arose that the Saudi government was going to give oil shipping- preferences to Saudi-owned ships. It anticipated transferring operation of its fleet of oil tankers to Samarco in order to qualify for the expected shipping preference.
¶47The Samarco partners
¶48In the operating arrangement with the Saudis, Mobil agreed to “bareboat charter” its ships to Samarco (ie., charter the ships empty and unstaffed), and Samarco would in turn “time charter” them (ie., charter the vessels complete with crews and provisions) back to Mobil.
¶49Comnas, in setting up Atlas, asked two young coworkers at Lemos, Peter Tavoulareas and Ares Emmanuel, to join him as partners.
¶50Mobil, with the Samarco partners concurring, decided a year later, in 1975, after first giving Comnas an opportunity to improve his performance (Tr. 2266-67, 2325-26), to remove him from Atlas because of self-dealing, fraudulent practices, and dissatisfaction with his competence and performance (RE 1990-94). Emmanuel testified, based on firsthand knowledge, to facts that completely justified the termination of Comnas.
¶51The Mobil-Samarco-Atlas relationship, and the Tavoulareas-Atlas connection particularly, were explained to Mobil shareholders in a 1976 letter. At about the same time, a party or parties sent anonymous letters to a number of newspapers, alleging that the arrangement violated the United States securities laws. One of those letters went to reporter Robert Woodward of The Washington Post. Woodward and reporters for several other newspapers investigated the anonymous charges; none of them apparently found any wrongdoing and no stories were published. The Securities and Exchange Commission, which obtained a copy of one of the letters, conducted an informal investigation and called upon William Tavoulareas to testify in nonpublic proceedings. The SEC also found no impropriety.
¶52Evidence presented by the plaintiffs at trial showed that Mobil did not exert control over Samarco, in which Mobil was a 30% minority shareholder, and that all of the decisions relating to Atlas were approved by the controlling Saudi partners. The arrangement was fully disclosed to Mobil’s Conflict of Interest Committee, Mobil’s Board of Directors, and the company’s shareholders.
¶53The Post article had its genesis five years after the creation of Atlas and three years after disclosure to the shareholders. The article was investigated and written by Patrick Tyler, who had previously written an article on the nation’s oil crisis for which he was criticized by Mobil for misstating information given him by Mobil. Shortly after Mobil’s criticism of the Post for its earlier story by Tyler, Woodward gave Tyler his old file on Tavoulareas.
¶54In October 1979, Samuel “Sandy” Golden, a reporter for a suburban newspaper, The Montgomery [Maryland] Journal, met Philip Piro, a physician who was married to William Tavoulareas’ daughter. Piro was in the midst of a less-than-amicable divorce proceeding. Piro told Golden about the Atlas transactions. Golden testified that Piro told him that Mobil’s President William Tavoulareas had set up his son Peter in the business, making Peter an “overnight millionaire” (Tr. 166). Piro mentioned that Woodward had previously investigated the story.
¶55*91Golden, who had been angling for a job at the Post for some time, sensed a story. He unsuccessfully called Woodward several times and finally left a message that he “had a story about the president of Mobil setting up his son to become an overnight millionaire” (Tr. 169). Woodward, by then an editor at the Post, assigned reporter Patrick Tyler to return Golden’s call. Golden told Tyler that he had a hot source for the Tavoulareas story. Tyler, who already had Woodward’s file on the Tavoulareases, agreed that if Golden’s source led to a big story, Golden would share a Post byline.
¶56At this point, the two reporters began following separate trails. Golden concentrated on getting more information from Piro and conducted several telephone interviews with him. Although Golden repeatedly assured Piro that he was not recording their telephone conversations, he taped several of them. Golden also began seeking “possible enempes]” of Tavoulareas (Tr. 202). In his search, he met Peter Stockton, an investigator for Rep. John Dingell, Chairman of the House of Representatives Subcommittee on Energy and Power. Golden believed that Stockton, who had something of a reputation as a crusader, “hate[d] people with money” (Tr. 223). Golden played the Piro tapes for Stockton, who at some point relayed Golden’s information to Congressman Dingell. Stockton later contacted Tyler and passed other information to him. Dingell wrote to the SEC seeking an investigation of Tavoulareas.
¶57Tyler, meanwhile, tracked down Comnas and met with him in New York. Comnas, who initially requested anonymity, allegedly told Tyler about the Atlas arrangements. The Post contends that Comnas was the principal source for its stories. (Despite being listed as a witness by the Post, Comnas did not testify at trial, nor was his deposition introduced. Because Tyler does not use tape recorders, only Tyler’s account of these important conversations with Comnas is available — a fact that severely weakens the Post’s defense.) Tyler also talked to John Kousi, an executive of Fairfield-Maxwell, one of the Samar-co partners. Parts of Kousi’s deposition were admitted into evidence. Tyler requested personal interviews with William Tavoulareas and other Mobil executives, but generally was rebuffed because in his earlier story on the oil shortage he had misrepresented statements made to him by Mobil executives. Tavoulareas did, however, give written answers to questions that Tyler submitted in writing (RE 2343-46).
¶58Tyler wrote the story, and the Post submitted it to its editing process. During editing, parts that were favorable to Mobil and Tavoulareas were deleted — chiefly opinions of various persons in and out of Mobil that (1) the Atlas arrangement had benefitted the company, and (2) William Tavoulareas had not been directly involved in Atlas after its formation. The Post copy editor assigned to the story, Christine Peterson, read the Tyler story (as it was ultimately published), considered its implications, and wrote a memorandum to other editors stating in effect that she found it “impossible to believe” that William Tavoulareas had arranged the entire Samarco-Atlas venture in order to benefit his son.
¶59The first article appeared on November 30, 1979.
Mobil Oil Corp. president William P. Tavoulareas set up his son five years ago as a partner in a London-based shipping management firm that has since done millions of dollars in business operating Mobil-owned ships under exclusive, no-bid contracts.
¶60The article detailed the transactions that led to the creation of Samarco and Atlas and when read as a whole, as it must be, implied that William Tavoulareas misused his corporate position as president and breached his fiduciary duties to Mobil to benefit Peter.
¶61William and Peter Tavoulareas brought two actions: one for libel against The Washington Post Company
1. In the November 30 article the Post defendants defamed William Tavoulareas. The premise of the article was false and was published with knowledge or reckless disregard of its falsity. (The jury awarded Tavoulareas $250,000 in compensatory damages and $1,800,000 in punitive damages against the Post defendants.)
2. The Post defendants did not defame Peter Tavoulareas.
3. Piro defamed both William and Peter Tavoulareas. (The jury awarded compensatory damages of $5,000 to William and $1,000 to Peter.)
¶62After receipt of the verdicts, the district court set aside the jury verdict for William Tavoulareas and granted judgments n.o.v. for the Post defendants and for Piro. The trial judge ruled that Tavoulareas was a limited-purpose public figure and that there was no proof the defendants had acted in reckless disregard of the truth. While recognizing that “[t]he article in question falls far short of being a model of fair, unbiased, investigative journalism” the trial court concluded that “[tjhere is no evidence in the record ... to show that it contained knowing lies or statements made in reckless disregard of the truth.” 567 F.Supp. at 654. The district court affirmed the jury verdict in favor of Peter Tavoulareas against Piro. On appeal the court reinstated the verdict, 759 F.2d 90 (D.C.Cir. 1985), one judge dissenting.
¶63II. The Standard of Review
¶64The constitutional standard of liability *93for public figures
¶65Under Rule 50(b) of the Federal Rules of Civil Procedure, a motion for judgment n.o.v. is essentially a motion for a directed verdict made after the jury has returned its verdict. Consequently, “the standard for awarding a judgment n.o.v. is the same as that applied when ruling on a motion for a directed verdict.” Vander Zee v. Karabatsos, 589 F.2d 723, 726 (D.C.Cir.1978), cert. denied, 441 U.S. 962, 99 S.Ct. 2407, 60 L.Ed.2d 1066 (1979). Thus,
[ujnless the evidence, along with all inferences reasonably to be drawn therefrom, when viewed in the light most favorable to the plaintiff is such that reasonable jurors in fair and impartial exercise of their judgment could not reasonably disagree in finding for the defendant, the motion must be denied.
¶66Alden v. Providence Hospital, 382 F.2d 163, 165 (D.C.Cir.1967) (emphasis added) (footnote omitted). In considering a motion for judgment n.o.v., the court ordinarily must abstain from weighing the evidence and assessing the credibility of witnesses. Morelock v. NCR Corp., 586 F.2d 1096, 1104 (6th Cir.1978), cert. denied, 441 U.S. 906, 99 S.Ct. 1995, 60 L.Ed.2d 375 (1979); 5A J. Moore, Moore’s Federal Practice It 50.07[2] (2d ed. 1986).
¶67This deferential standard for reviewing jury determinations does not fully apply, however, in First Amendment defamation cases. In addition to the requirement that actual malice be proven with clear and convincing evidence, the Supreme Court held in New York Timesthat courts reviewing public figure defamation verdicts “must ‘make an independent examination of the whole record,’ ... so as to assure [themselves] that the judgment does not constitute a forbidden intrusion on the field of free expression.” 376 U.S. at 285, 84 S.Ct. at 729 (quoting Edwards v. South Carolina, 372 U.S. 229, 235, 83 S.Ct. 680, 683, 9 L.Ed.2d 697 (1963)) (emphasis added); see also Bose Corp. v. Consumers Union of United States, 466 U.S. 485, 511, 104 S.Ct. 1949, 1965, 80 L.Ed.2d 502 (1984). As the original panel opinion noted, it is necessary to reconcile the Supreme Court’s instruction to review the evidence in a First Amendment defamation case independently with the well-settled standard, reiterated in Anderson v. Liberty Lobby, 477 U.S. 242, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986) (a limited-purpose public figure defamation case), that “Credibility determinations, the weighing of evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of a judge, whether he is ruling on a motion for summary judgment or for a directed verdict.” 477 U.S. at-, 106 S.Ct. at 2513. Cf. Crane v. Kentucky, 476 U.S. 683,---, 106 S.Ct. 2142, 2145-46, 90 L.Ed.2d 636 (1986). I remain of the view that this reconciliation is to be achieved by limiting appellate courts’ independent review to the question of ultimate constitutional fact — i.e., whether, assuming the most favorable facts and “legitimate inferences” that could reasonably have been found by the jury, the constitutional threshold of clear and convincing proof of actual malice was passed. Since, as has already been noted, the en banc majority does not address this legal question, I will not belabor it here, referring the reader to the discussion contained in the panel opinion, Tavoulareas v. Piro, 759 F.2d 90, 106-09 (D.C.Cir.1985).
¶68III. Defamation and Falsity
¶69A. Defamatory Meaning
¶70To constitute actionable defamation, a published statement must be shown to be both defamatory and false. As the district court instructed the jury, a communication is considered defamatory if it “tends to injure [the] plaintiff in his trade, profession, or community standing, or lower him in the estimation of the community or subject him to scorn, ridicule, shame, contempt or embarassment” (Tr. 4548). See Afro-American Publishing Co. v. Jaffe, 366 F.2d 649, 654 (D.C.Cir.1966) (en banc). In determining the meaning of an allegedly defamatory communication, the “ ‘publication claimed to be defamatory must be read and construed in the sense in which the readers to whom it is addressed would ordinarily understand it.’ ” Washington Post Co. v. Chaloner, 250 U.S. 290, 293, 39 S.Ct. 448, 448, 63 L.Ed. 987 (1919) (citation omitted).
¶71The respective roles of judge and jury are as follows: It is for the court to determine whether a particular communication is capable of bearing a defamatory meaning. “ ‘It is only when the court can say that the publication is not reasonably capable of any defamatory meaning, and cannot reasonably be understood in any defamatory sense, that the court can rule as a matter of law, that the publication is not libellous.’ ” Gariepy v. Pearson, 207 F.2d 15, 16 (D.C.Cir.) (citation omitted) (emphasis added), cert. denied, 346 U.S. 909, 74 S.Ct. 241, 98 L.Ed. 407 (1953); McBride v. Merrell Dow & Pharmaceuticals, 717 F.2d 1460, 1465 (D.C.Cir.1983). It is thus for the jury to decide whether a communication, found capable of a defamatory meaning by the court, was in fact so understood by its recipient. Restatement (Second) of Torts § 614 (1977).
¶72The majority goes about its task of determining whether the Post article is capable of any defamatory meaning by understating and separately considering each of only three allegedly libelous statements taken out of context and by completely ignoring the basic theme of the article. The majority considers the defamatory implication of the “set up” charge in the Post article in an unreasonably narrow fashion, asserting that “[t]he statement that ‘a father set up his son in business’ would ordinarily mean to a reasonable reader that the father provided the son with [a business opportunity].” Maj. Op. at 780. In so defining “set up,” the majority ignores the full meaning of those words in the article.
¶73*95Such analysis by the majority of the defamatory meaning of the Post article is flawed both factually and as a matter of law. As a matter of law, the majority’s analysis violates the cardinal principle of defamation law that “words must be read in the context of the entire communication as a whole.” R. Sack, Libel, Slander, and Related Problems 52 (1980) (emphasis added). See Pierce v. Capital Cities Communications, 576 F.2d 495, 502 (3d Cir.), cert. denied, 439 U.S. 861, 99 S.Ct. 181, 58 L.Ed.2d 170 (1978) (“a court should not scrutinize simply the literal references of the language in question, but also should weigh the words ‘together with their context,’ Restatement, Second, Torts § 563, Comment (d)”); Curtis Publishing Co. v. Vaughan, 278 F.2d 23, 26 (D.C.Cir.), cert. denied, 364 U.S. 822, 81 S.Ct. 57, 5 L.Ed.2d 51 (1960); Williams v. Anti-Defamation League of B’nai B’rith, 185 F.2d 1005, 1007 (D.C.Cir.1950). Gariepy, 207 F.2d at 16; Thomson v. Cash, 119 N.H. 371, 402 A.2d 651, 653 (1979); James v. Gannett Co., 40 N.Y.2d 415, 419, 386 N.Y.S.2d 871, 874, 353 N.E.2d 834, 838 (1976); H.O. Merren Co. v. A.H. Bel. Corp., 228 F.Supp. 515, 518 (N.D.Tex.1964), aff'd, 346 F.2d 568 (5th Cir.1965); Westby v. Madison Newspapers, Inc., 81 Wis.2d 1, 6, 259 N.W.2d 691, 693 (1977). Cf. Ollman v. Evans, 750 F.2d 970, 982 (D.C.Cir.1984) (en banc) (a proper reading of an allegedly defamatory article “depends upon the article or column, taken as a whole, of which the statement is a part”) (emphasis added). A court should not isolate particular words or sentences and determine whether, considered alone, they are defamatory. To be sure, the words used must be considered — but they must be considered in their context. Words that standing alone may be understood to be defamatory may be so explained by their context as to negate any defamatory meaning. Conversely, words that standing alone convey no defamatory meaning may, when read in context, be understood to be defamatory. The charges in this article were published in context and it is error for the court not to determine their meaning in that context. The court’s task is to determine whether the publication taken as a whole is defamatory. See Hoffman v. Washington Post Co., 433 F.Supp. 600, 602 (D.D.C.1977), aff'd, 578 F.2d 442 (D.C.Cir.1978); Joseph v. Xerox Corp., 594 F.Supp. 330, 332 (D.D.C.1984).
¶74In addressing whether the article is capable of bearing a defamatory meaning, the majority creates an easy task for itself. In analyzing only a few of the defamatory statements and those out of context, the majority escapes confronting the basic defamatory theme of the article — that William Tavoulareas violated his fiduciary duty as president of Mobil and misused Mobil assets, personnel, property, and financing to advance the career of his son Peter.
¶75The headline and lead sentence of the November 30, 1979 Post article charges that William Tavoulareas “set up” his son. The majority admits that the “set up” allegation is defamatory in accusing “Tavoulareas with nepotism — furthering his son’s business career.” Maj. Op. at 780. By reading narrowly the defamatory character of the “set up” charge and of the article as a whole, the majority finds that the defamatory “set up” allegation is true.
¶76To construe the defamatory accusation as being limited to simple “nepotism” is farcical. The “set up” charge cannot be read in isolation to simply accuse a father of furthering his son’s career. The defamation is far more extensive. Read in the context here where the father’s ability to “set up” his son stems solely from his position as president of a large corporation, the term “set up” further connotes that the father exercised his powers as president to use corporate assets with intent to benefit his son in breach of his fiduciary duties to the corporation’s shareholders.
¶77The majority effectively reads Mobil out of the article, notwithstanding the fact that the article mentions Mobil over one hun*96dred times and identifies William Tavoulareas acting as president of Mobil ten times. Without the Mobil involvement the article would not have been a page 1 story. There are numerous statements in the article that can only be interpreted as intending to convey to the readers that William clearly violated his fiduciary duty as president of Mobil, by making possible Peter’s partnership in Atlas and thereafter using Mobil assets and personnel to ensure the financial success of Atlas. According to the article, William Tavoulareas used Mobil assets and personnel to give Atlas “millions of dollars in business operating Mobil-owned ships under exclusive, no-bid contracts,” (Paragraph 1); to “provide[ ] Atlas with office space,” (Paragraph 4); to “create work for Atlas at a time when the shipping industry was severely depressed,” (Paragraph 5) and when “Mobil could have managed the Samarco fleet itself,” (Paragraph 63); and then to “intervene[ ]” (Paragraph 79) by personally “forcing the resignation” (Paragraph 24) of Comnas and by “dispatch[ing]” Hoffmann, an officer from Mobil, to help run Atlas for six months (Paragraph 82), thus securing for Peter control of the company (Paragraph 76). Moreover, the article specifically attributes Peter’s partnership in Atlas to “the help of Mobil” (Paragraph 2). The Post story therefore does not merely accuse a father of nepotism, but charges that the father, as president of Mobil, accomplished the nepotism through breach of his fiduciary duties to Mobil.
That [William Tavoulareas] breached his fiduciary duties to Mobil. That he wasted and misused assets of Mobil. That he wrongfully diverted such assets to Peter Tavoulareas for his benefit, and four, that he committed criminal acts.
¶78(Tr. 4550).
¶79By ignoring the article’s defamatory implication that Tavoulareas misused Mobil assets and his position as president to advance Peter in Atlas, the majority quickly disposes of the defamatory allegation that Tavoulareas “dispatched one of his senior shipping executives, Herman [sic] F. Hoffmann, to London to help run Atlas” (Paragraph 82). The majority admits that the “dispatch” allegation might be false, but attempts to obviate its defamatory effect by concluding that the statement adds no further defamatory implication to the charge that Tavoulareas set Peter up in Atlas, a charge the majority declares is true. Maj. Op. at 788. The allegation that Tavoulareas “personally dispatched” Hoffmann, however, is defamatory in its implication that Tavoulareas misused corporate personnel to insure the success of Peter’s business venture.
¶80*97It is instructive to compare the refusal of the en banc majority to acknowledge that the Post article can reasonably be given this interpretation with this court’s opinion in McBride v. Merrell Dow & Pharmaceuticals, 717 F.2d 1460 (D.C.Cir.1983). In that case Dr. McBride, who had been an expert witness in a Florida lawsuit, brought a suit claiming that he was libeled by an article containing the following language: “McBride ... was paid $5,000 a day to testify in Orlando. In contrast, Richardson-Merrell pays witnesses $250 to $500 a day, and the most it has ever paid is $1,000 a day.” Id. at 1462. Dr. McBride contended that this language was defamatory because it implied that he was “ ‘willing to prostitute his professional expertise and testify on behalf of the highest bidder.’ ” Id.at 1463 (quoting Dr. McBride’s Complaint). The district court agreed with the defendant that this language was not defamatory, concluding that standing alone it suggested, if anything, that Dr. McBride possessed a high degree of professional accomplishment, and that it was “improbable” that anyone would draw defamatory inference from this language. McBride v. Merrell Dow & Pharmaceuticals, 540 F.Supp. 1252, 1255 (D.D.C.1982). We reversed. Stressing the fact that, “[ujnder District of Columbia defamation law, a court’s power to hold as a matter of law that a statement is not defamatory is very limited,” 717 F.2d at 1465, we concluded that the language could reasonably have been understood to imply that Dr. McBride’s testimony was for sale.
¶81It seems to me impossible to reconcile the approach of the en banc majority in this case with McBride. Surely the distance between the simple statement that one expert witness was paid significantly more than other witnesses and the inference that that witness’ testimony was for sale is far greater than the distance between the statement that the president of a corporation set his son up in a business with which the corporation then did substantial business and the inference that the president thereby misused corporate assets and breached his fiduciary duties to the corporation’s shareholders.
¶82While ignoring the article’s defamatory implication that William Tavoulareas abused his position as president and violated his fiduciary duties to Mobil in prodding Peter’s fortunes in Atlas, the majority eliminates another aspect of the article’s central theme that Tavoulareas abused his corporate position, concluding “as a matter of law” that the article is incapable of bearing the defamatory “interpretation that Tavoulareas ‘set up’ the entire Mobil-Samarco-Atlas relationship to benefit Peter.” Maj. Op. at 780.
¶83The entire discussion of the legitimate business reasons in the Post article is pejorative, however, suggesting that although they were asserted by Mobil, these business reasons fail to explain the Mobil-Samarco-Atlas arrangement.
¶84When read in the context of the article as a whole, the “set up” accusation could reasonably be read to suggest that William engineered the entire Mobil-Samarco-Atlas arrangement to benefit Peter. For example, the article comments that Mobil was originally not interested in participating in Samarco, but only later became interested (Paragraphs 50-51). The article then notes that Comnas, one of the negotiators sent by Mobil “to push Mobil as a partner” in Samarco, was also “recruited personally by Tavoulareas to advise Mobil on how to set up an independent shipping concern.” (Paragraph 51). The article further states that Tavoulareas recommended to the Samarco partners that Comnas run the management arm of Samarco. Tavoulareas is then said to have “personally urged that his son become a partner in Atlas.” (Paragraph 52). These details in the article could reasonably be read to imply that Mobil’s entry into Samarco was motivated by Tavoulareas’ intention that Comnas, who would negotiate the Saudi deal for Mobil, would later run Samarco’s management arm and offer Peter a partnership in the company. The majority, failing to recognize that the decision to enter Samarco was contemporaneous with the decision to establish a separate company to manage Samarco’s ships, strips away an integral part of the defamatory meaning surrounding the “set up” allegation.
¶85The majority’s cavalier rejection of the article’s defamatory meaning that Tavoulareas engineered the entire Mobil-Samar-co-Atlas arrangement for Peter’s benefit as a “matter of law” is at odds with the McBride rule that a court’s power to hold as a matter of law that a statement is not defamatory is limited to those occasions where the “ ‘publication is not reasonably capable of any defamatory meaning and cannot be understood in any defamatory sense.’ ” 717 F.2d at 1465 (citations omitted). Given evidence that the article was in fact “understood in [the] defamatory sense,” McBride, 717 F.2d at 1465, that William Tavoulareas’ nepotistic intent prompted Mobil’s entry into the SamarcoAtlas arrangement, the majority’s conclusion as a matter of law that the article does not, and cannot, bear the defamatory meaning that Tavoulareas “set up” the entire Mobil-Samarco-Atlas relationship to benefit Peter is all the more unjustifiable.
¶86For example, Christine Peterson, a copy editor at the Post, reviewed the article pri- or to publication. She sent a memorandum to assignment editor Peter Milius, which stated:
I’ve read the Mobil story several times, and while I’m impressed with the amount of work the reporter obviously did, I’m still left with an overwhelming sense of So What? Is there any way to give this story of high-level nepotism a dollars- and-cents angle? Did Mobil’s shareholders lose anything? Mobil’s customers? Parts of Tyler’s case against Tavoulareas seem tenuous, and the whole — a $680,000-a-year plaything for an indulged son, at worst — just seems like a withered peanut in an 84" gilded shell. A far more interesting angle, it seems to me, is Mobil’s concern about Saudi preference shipping — a concern so profound that it led to the formation of an entire dummy corporation. It’s impossible to believe that Tavoulareas alone could put together such a scheme for the sake of his son’s business career, or that he would want to.
¶87(RE 2486) (emphasis added). This indicates that Peterson interpreted the article as a whole to charge that William engineered the entire Mobil-Samarco-Atlas transaction as “a scheme for the sake of [Peter’s] business career.”
¶88Moreover, Post trial counsel in his closing argument to the jury positively and extensively interpreted the article in a similar manner:
*100I suppose, ladies and gentlemen, when you get right down to it the thing that Mr. Tavoulareas most vigorously complains of is the term “set up” in the headline and the lead of the story.
Well, lend me your ears for two or three minutes and let’s talk about the headline and the lead.
By now, I dare say it is clear to everybody in the courtroom that what happened was that Mr. Tavoulareas, seeing a good business reason to go into the Samarco venture, and seeing a good business reason to have an independent ship management firm run those ships, took advantage of those good business reasons to see to it that his son, Peter, was set up for the rest of his life. That is what he did, and the mechanism, the device by which this was done was setting up Atlas, putting Comnas in charge, but with Peter with him as a partner, and then in three or four months seeing to it that Comnas is out of the picture, leaving Peter in there as the major partner with Ares Emmanuel as a minor partner. That is what he did.
He did, as I understood Mr. Tavoulareas’ testimony, while he wouldn’t like the words that I used as to each event that took place in sequence, there is no real dispute these are the things that happened.
Samarco was set up. Atlas was set up. There was an Atlas/Samarco contract. Comnas and Peter took on the Atlas operation, and in April of 1975 Comnas was induced to resign. Remember those consulting contracts and the rest of it?
Now, the question is: How do you describe all of this? How do you describe it so that someone coming to it for the first time understood it? What words do you use to make it clear and to make it interesting to the newspaper reader?
And the words that Pat Tyler settled upon were not unusual words, words that people use all the time. The words were “set up. ”
I submit to you, ladies and gentlemen, that [“set up”] is a perfectly accurate way of describing the facts of the relationships among William Tavoulareas, Peter Tavoulareas, Mobil, Samarco and Atlas.
¶89(Tr. 4491-94) (emphasis added). Here Post counsel repeatedly characterizes the article as alleging that William intended, when he “took advantage” of the situation and “set up” Samarco, Atlas, and the Samarco-Atlas contract, to have Peter “set up for the rest of his life.”
¶90The majority deems that this evidence of how the article was interpreted is irrelevant to its consideration of the article’s defamatory meaning. Maj. Op. at 781. In so concluding, however, the majority ignores the principle that “[t]he meaning of a communication is that which the recipient *101correctly, or mistakenly but reasonably, understands that it was intended to express.” Restatement (Second) of Torts § 563 (1977). That Christine Peterson and Post trial counsel understood the article to charge William Tavoulareas with setting up Samarco and Atlas for Peter’s benefit makes unsupportable the majority’s discarding of this interpretation “as a matter of law.”
¶91In avoiding the article’s defamatory implications, the majority charges the dissent with a “tortured attempt to discern some dark, hidden meaning in the ‘tone’ of the article.” Maj. Op. at 781. The article’s implication that William Tavoulareas’ nepotistic intent prompted Mobil’s entry into the Samarco-Atlas arrangement is not, as the majority asserts, “hidden.” The interpretation of the article offered by the dissent is that which Patrick Tyler himself ascribed to the story. In a telephone conversation prior to publication of the article, Tyler said to Peter Tavoulareas:
I will tell you what I think. I have read the files and am more familiar with this than anyone. I think your father set up Atlas and Samarco and that ... [Peter interrupts].
¶92(Tr. 4369). Tyler’s belief that William Tavoulareas “set up Atlas and Samarco” for Peter’s benefit surfaced in his article. See supra at 821-22. Moreover, Tyler interpreted the meaning of the news article in a memorandum to his Post superiors. Responding before publication to Christine Peterson’s critical comments, Tyler wrote:
Mobil undertook some incredibly fancy corporate footwork in the wake of the embargo to apparently accomplish two things: ingratiate itself with the Saudis ... and set up the son of Mobil’s president in a shipping business when business was bad and the business therefore stood little chance of prospering without Mobil’s help.
******
Our story does show that Mobil’s decisions in this case were not made for the traditional business reasons, or for the reasons stated by Mobil.
******
It should not seem impossible that Mobil — which originally turned down a chance to join Samarco — changed its mind [on joining Samarco] after it dawned on Mobil’s president that such a partnership would justify the creation of a small management firm at a time when Tavoulareas’ son was aspiring to such a career and was already at work with one of the Greek shippers. Samarco can be seen as nothing more complicated than a slight diversion from Mobil in the way it moved its crude oil from point A to point B.
The question of “why" Mobil’s president would want to orchestrate such a diversion [to Samarco ] when it would benefit his son’s business career is begging things a little.
¶93(RE 2488-90) (emphasis added). Thus, in addition to the interpretation of the article by the Post editor Peterson and the Post trial attorney, Tyler himself interpreted his article as attributing Mobil’s entry into Samarco to Tavoulareas’ nepotistic intent “to benefit his son’s business career ... after [that idea] dawned on Mobil’s president.” Id. The article, Tyler states, answers the basic subjective question “ ‘why’ Mobil’s president would want to orchestrate [the] diversion [of Mobil’s shipping to Samarco],” and the answer the article provides is that William Tavoulareas intended “[to] benefit his son’s business career — ” Id. The majority refers to the attention devoted in the article to Mobil’s legitimate business reasons for creating Samarco, but even Tyler in his memorandum recognizes that the discussion of valid business-reasons in the article was meant to disparage: “[0]ur story does show that Mobil’s decisions in this case were not made for the traditional business reasons, or for the reasons stated by Mobil” (RE 2489) (emphasis added).
¶94Tyler’s contemporaneous interpretation of the article he had written is powerful evidence that the article is best understood as implying that Mobil’s entry into the Samarco-Atlas arrangement was prompted by William Tavoulareas’ nepotistic intent — it is virtually dispositive evidence that the arti*102ele could reasonably be so understood. Nevertheless, the en banc majority asserts that Tyler’s memo is irrelevant because “[njothing in law or common sense supports saddling a libel defendant with civil liability for a defamatory implication nowhere to be found in the article itself.” Maj. Op. at 781.
¶95Such a myopic conclusion, however, ignores the principle that “defamatory imputation may be made by innuendo.” Restatement (Second) of Torts § 563(c). See McBride, 717 F.2d at 1465; Sellers v. Time, Inc., 423 F.2d 887 (3rd Cir.), cert. denied, 400 U.S. 830, 91 S.Ct. 61, 27 L.Ed.2d 61 (1970); Marcone v. Penthouse International, 533 F.Supp. 353 (E.D.Pa. 1982); Mihalik v. Duprey, 11 Mass.App. 602, 417 N.E.2d 1238, 1239 (1981); Thomson v. Cash, 119 N.H. 371, 402 A.2d 651, 654 (1979). Only by blinding itself to the defamatory implications Tyler himself ascribed to the article, and the clear innuendo of the article, can the majority conclude that the article did not charge Tavoulareas with setting up the Mobil-Samarco-Atlas scheme for Peter’s benefit.
¶96The defamatory meaning that the en banc majority does recognize is in a narrowly construed “set up” charge. The majority characterizes the “set up” charge as a defamatory accusation “that Tavoulareas made it possible for Peter to become a partner in Atlas and then helped to ensure that the business would prosper because Peter was his son.” Maj. Op. at 780.
¶97Consistent with its strategy to interpret the defamatory meaning of the article in an unreasonably narrow manner, the en banc majority insists that the article’s false defamatory statement that “Tavoulareas personally urged that his son be included as an equity partner in Atlas”
¶98B. Falsity
¶99To conclude that the “set up” charge is true, the majority claims reliance on “undisputed evidence of Tavoulareas’ personal involvement in the establishment and operation of Atlas to Peter's manifest benefit.” Maj. Op. at 786. The evidence upon which the majority relies, however, does not establish the truth of the “set up” charge as interpreted by the majority — i.e., that “Tavoulareas made it possible for Peter to become a partner in Atlas and then helped to ensure that the business would prosper because Peter was his son.” Maj. Op. at 780. For example, the majority claims that “the undisputed fact that Tavoulareas personally recruited Comnas to manage Samarco’s ships ... shortly after learning that Peter had an outstanding offer from Comnas goes far toward justifying the charge that Tavoulareas ‘set up’ his son in Atlas.” Maj. Op. at 784 (emphasis in original).
¶100In reaching this conclusion; however, the majority ignores the following crucial evidence:
• Rawleigh Warner, Chairman of Mobil, testified that when Comnas came to New York in January 1974, seeking to charter a ship to Mobil, Warner remarked that “George Comnas would be an absolute natural” to manage the partnership’s tankers in the event Mobil entered an agreement with the Saudis and Fairfield-Maxwell (Tr. 1515). Thus, by the time Tavoulareas and Walter MacDonald, then regional director of Mobil’s Middle East Department, asked Comnas in the spring of 1974 whether he was interested in managing Samarco’s ships (RE 2444), there was a consensus at Mobil with Warner’s earlier observation that Comnas was a good choice for the task. The majority’s assertion that Tavoulareas “personally recruited Comnas,” Maj. Op. at 784 (emphasis in original), simply does not prove anything beyond the fact that Tavoulareas acted on Mobil’s behalf when he later visited Comnas.35
• Before Rawleigh Warner suggested Comnas as “an absolute natural” to manage the ships, he was fully informed by Tavoulareas that Comnas was going to ask Ares Emmanuel and Peter to come to work for him. Peter, however, had not accepted a position in Comnas' company. Aware that Comnas might try to curry favor with Mobil, Warner and Tavoulareas both talked to the chairman of Mobil’s Conflict of Interest Committee about Peter’s possible affiliation with a company that might be doing business with Samarco. An arrangement was set up that would avoid a conflict of interest if Peter decided to accept Comnas’ offer (Tr. 1515-18).
*104• William Tavoulareas counselled Peter against leaving Lemos to join Comnas’ new company (Tr. 1297, RE 2426).
• Warner testified that “Tav did nothing personally to further the affairs of Peter in [Atlas].” (Tr. 1539).
¶101Given the abundant evidence of. Mobil’s interest in an independent shipping concern to manage Samarco’s ships, and in one headed by Comnas in particular, the fact that Tavoulareas asked Comnas if he was interested in managing Samarco’s ships does not establish the truth of the charge that “Tavoulareas made it possible for Peter to become a partner in Atlas.” Maj. Op. at 780. To the contrary, the evidence shows that Tavoulareas acted to carry out a decision made by Mobil for its benefit, and with great care to avoid possible conflicts of interest in the event that Peter, against his father’s advice, decided to join Comnas. In sum, the evidence is far from undisputed and does not support the accusation that William Tavoulareas “made it possible for Peter to become a partner in Atlas.”
¶102The majority further attempts to support the truth of the “set up” allegation by citing evidence for the charge that Tavoulareas “helped to ensure that [Atlas] would prosper because Peter was his son.” Maj. Op. at 780. Relying on evidence that Tavoulareas helped Atlas to “secure its management agreement with Samarco,” Maj. Op. at 785, and to “survive [and] prosper after Comnas left,” Maj. Op. at 786, the majority determines that the “set up” allegation is true. The majority, however, fails to acknowledge that due to Mobil’s and Fairfield-Maxwell’s possible conflicts of interest and the Saudi’s inexperience in ship management, Mobil was soundly committed to having an independent company manage Samarco’s ships (Tr. 1515, 1518). Moreover, it was in Mobil’s interests for Atlas and Samarco to reach agreement quickly (Tr. 1311). Thus, to the extent Tavoulareas “helped Atlas secure its management agreement with Samarco,” he was acting in Mobil’s interests to implement a Mobil decision.
¶103Furthermore, the majority mischaracterizes William Tavoulareas’ involvement with Atlas. After Peter decided in July 1974 to join Comnas, William formally removed himself from decisionmaking regarding Atlas (Tr. 1521, 1525). Although the majority cites meetings between Tavoulareas and the Alirezas as evidence of the truth of the “set up” allegation, Maj. Op. at 785, these meetings were fully consistent with the decision of the Conflict of Interest Committee and Rawleigh Warner that “Tav could involve himself in activities up to the decision level but he could not make that decision” (Tr. 1521). Of course, as president of Mobil and its longtime expert on Saudi affairs, Tavoulareas’ ability to carry out decisions made by the company was not required to be completely lost. He was free to negotiate on Mobil’s behalf (Tr. 4159). The jury could reasonably have concluded that Tavoulareas’ forthright communication to Mobil’s Conflict of Interest Committee immediately after he first acquired knowledge of a possible conflict of interest, his formal removal from decision-making months later when Peter joined Comnas, and the absence of any evidence showing that William made any decisions regarding Atlas after formally removing himself, negated the charge that he was acting on Peter’s behalf.
¶104Finally, the majority offers Tavoulareas’ involvement in the events surrounding Comnas’ departure from Atlas as establishing the truth of the “set up” charge. Maj. Op. at 785-86. Again the majority refuses to recognize that Tavoulareas complied with Mobil's conflict of interest guidelines and did not make any decisions concerning Atlas and Samarco after Peter joined Comnas. Paul Wolfe testified that it was he who recommended to Mobil Chairman Rawleigh Warner (in the presence of William Tavoulareas) that Comnas be replaced (Tr. 1069, 1070, 1098, 1186). Warner in turn made the decision that, for justifiable reasons, Comnas should be removed from Atlas (Tr. 1070,1326,1531-34). The other Samarco partners were also dissatisfied with Comnas. See supra at 812-13. Moreover, Warner suggested that Tavoulareas meet with Comnas in London *105to express Mobil’s disappointment in Comnas’ performance (Tr. 1070, 1333). In London, Harmon Hoffmann and Walter MacDonald negotiated in the morning with Comnas over the terms for his departure from Atlas. Tavoulareas met with Comnas, Hoffmann, and MacDonald at lunch, but the morning’s negotiations were not discussed (Tr. 3349). After reading the written agreement drafted by Hoffmann and MacDonald detailing the terms for Comnas’ departure, Tavoulareas deferred the decision on the proposed contract “to top management in New York” (Tr. 1339). These facts establish that Tavoulareas made no decisions concerning Comnas’ departure from Atlas.
¶105Thus, the majority offers no evidence establishing the truth of the charge that “Tavoulareas made it possible for Peter to become a partner in Atlas and then helped to ensure that the business would prosper because Peter was his son.” What the majority does submit is evidence that Mobil officials decided that it was in the best interests of Mobil that an independent company should manage Samarco’s ships and that the best candidate for the job was a company headed by George Comnas. That Peter Tavoulareas might incidentally benefit from this arrangement because of his later association with Comnas was understood by Mobil officials but considered acceptable in light of William Tavoulareas’ removal from decisionmaking regarding Atlas and Samarco. The majority torturously converts this evidence establishing that Mobil wanted Atlas to independently manage Samarco’s ships into evidence proving William Tavoulareas “set up” Peter.
¶106The majority has not explained why, in the face of overwhelming evidence that Mobil officials decided to have Atlas manage Samarco’s ships, Tavoulareas’ “personal involvement in the establishment and operation of Atlas,” Maj. Op. at 786, should be interpreted to be for “Peter’s manifest benefit.” Id. Given the evidence that William Tavoulareas complied with Mobil’s conflict of interest rules by removing himself from decisionmaking regarding Atlas and Samarco; that Mobil officials, independently of Tavoulareas, decided for good business reasons that Atlas should manage Samarco’s ships, and continued to do so after Comnas departed; and that the benefit redounding to Peter from this arrangement was incidental and not engineered by William, a reasonable jury could find that the “set up” charge was false.
¶107Because the evidence in the record demonstrates that a jury could reasonably have concluded that the Post article was both defamatory and false,
¶108IV. Reckless Disregard for the Truth — Actual Malice
¶109As noted in part II above, the duty of an appellate court in reviewing a defamation verdict is to decide, after viewing the evidence and all legitimate inferences in the light most favorable to the verdict, whether the ultimate constitutional conclusion was sufficiently proven at trial — i.e., in this case whether the evidence adduced at trial amounted to “clear and convincing” proof of reckless disregard for the truth. Although the Post argues that evidence may not be cumulated.to establish a defendant’s reckless disregard for truth or falsity, The Washington Post Petition for Rehearing and Suggestion for Rehearing En Banc at 7, the majority admits that a defendant’s reckless disregard for the truth can be proven both through direct evidence and through the cumulation of circumstantial evidence. Maj. Op. at 788-89. See Herbert v. Lando, 441 U.S. 153, 160, 99 S.Ct. 1635, 1640-41, 60 L.Ed.2d 115 (1979); Rosenbloom v. Metromedia, Inc., 403 U.S. 29, 55, 91 S.Ct. 1811, 1825, 29 L.Ed.2d 296 (1971); Bose Corp. v. Consumers Union of United States, 692 F.2d 189, 196 (1st Cir.1982), aff'd 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984); Goldwater v. Ginzburg, 414 F.2d 324, 342 (2d Cir.1969), cert. denied, 396 U.S. 1049, 90 S.Ct. 701, 24 L.Ed.2d 695 (1970); see also Tavoulareas v. Piro, 763 F.2d 1472, 1477-78 (D.C.Cir.1985) (panel opinion denying rehearing). What follows is a discussion of the evidence of reckless disregard for the truth that was developed at trial, along with the reasonable inferences that could be drawn from it. Based thereon, it is submitted that the evidence- adduced at trial constitutes clear and convincing proof that the Post article was published with reckless disregard of its truth or falsity, and that the jury’s verdict should be reinstated.
¶110A. The Christine Peterson Memorandum and Tyler’s Response
¶111Christine Peterson’s memorandum to her superiors at the Post stated that Tyler’s article was “impossible to believe.” See supra at 822-23. At trial, during which Peterson was still in the Post’s employ, Peterson testified that in using the phrase “impossible to believe,” she really did not mean what she said, and instead meant that “the focus of the story should have been the formation of Samarco in anticipation of preference shipping” (RE 1837).
¶112In considering all this evidence, our duty as an appellate court reviewing a jury’s verdict is to view the legitimate inferences from the evidence in the light most favorable to the verdict. We are “required to draw all inferences favorable to the [verdict],” Carter v. Duncan-Huggins, Ltd., 727 F.2d 1225, 1234 (D.C.Cir.1984) (emphasis in original). The inferences to be credited are not a matter for our judgment except to assure that they are legitimate; only the ultimate constitutional issue may be independently assessed. Thus, as to the Peterson memorandum we are required to assume: (1) The jury considered Peterson’s attempt to alter the obvious interpretation of her memorandum to be influenced by self-interest in her position at the Post and understood the memorandum to mean (what it said) that in her opinion the story was “impossible to believe,” and (2) the reasonable effect of Peterson’s communication of these doubts to those at the Post responsible for the publication was to raise in their minds subjective doubts as to the accuracy of the story. Indeed, the Peterson memorandum suggests that the central theme of the story is inherently improbable, and, as the Supreme Court stated in St. Amant v. Thompson, a jury may find actual malice where “the publisher’s allegations are so inherently improbable that only a reckless man would have put them in circulation.” 390 U.S. 727, 732, 88 S.Ct. 1323, 1326, 20 L.Ed.2d 262 (1968) (emphasis added). Ultimately, it is for the appellate court to independently decide whether the legitimate inferences from all the evidence constitute clear and convincing proof of recklessness.
¶113The en banc majority rejects the Peterson memorandum as evidence of actual malice, claiming that the plaintiffs did not connect Peterson’s “impossible to believe” language to a false defamatory statement in the article.
¶114Even if the majority is correct in rejecting the article’s broader defamatory meaning as a matter of law and in holding the “set up” allegation true, its dismissal of the Peterson memorandum remains faulty. Although admitting the memorandum can be read to express Peterson’s “disbelief that Tavoulareas ‘set up’ Peter,” the majority makes no mention of the effect of the Peterson memorandum on the false defamatory statement that William Tavoulareas *108“personally urged” Comnas to make Peter a partner in Atlas. Certainly, if Peterson doubted the truth of the story's charge that Tavoulareas “set up” Peter, her doubts would encompass the “personally urged” allegation as well, since, as the majority recognizes, “the personally urged allegation goes beyond the general charge that Tavoulareas ‘set up’ his son and suggests that Tavoulareas actively pressured Comnas to hire Peter rather than merely rewarded Comnas for doing Peter a favor.” Maj. Op. at 788.
¶115Thus, contrary to the majority's suggestions otherwise, there is no basis to disregard the “serious, doubts” raised by the Peterson memorandum. The memorandum and Patrick Tyler’s response admitting that “a good editor might say that part of our case against Tavolareas seems tenuous” therefore provide evidence that the Post acted with reckless disregard for truth or falsity in publishing the article.
¶116B. The Reliability of the Sources
¶117The Supreme Court has held that “recklessness may be found where there are obvious reasons to doubt the veracity of the informant or the accuracy of his reports.” St. Amant, 390 U.S. at 732, 88 S.Ct. at 1326. Considerable evidence was introduced at trial tending to show that Tyler had reason to doubt the veracity of the two principal sources he relied on for the article — Comnas and Piro.
¶1181. Comnas
¶119Comnas, according to the Post, was the principal source for the story, as the chief source for both the “set up” allegation and the “personally urged” charge and other details recounted in the story. Conflicting evidence was introduced at trial regarding Comnas’ reliability. On one hand, Piro testified he told Tyler that “George Comnas had been caught in some fraud involving Atlas, that he was forced to withdraw [from Atlas] and that if Comnas ever gave Peter and Atlas any trouble Comnas would be reported to the tax authorities” (RE 1857-58). See also supra at 812 n. 9. On the other hand, the Post contended that Comnas was reliable due to (1) his “Who’s Who” listing, indicating that he was an experienced shipping executive; (2) William’s SEC testimony, which “never hinted at any improprieties,” and stated “that Mobil ‘had every confidence in the world’ in Comnas;” and (3) some independent corroboration of much of what Comnas told the Post. Again, the evidence suggests that the Post knew of facts suggesting that Comnas was an unreliable source. The Post's failure to call him as a witness or to use his deposition, especially after listing him as a witness for trial, points to the glaring weakness of the Post’s defense.
¶120The majority opinion also places entirely too much emphasis on the fact that Tyler corroborated some of the information sup*109plied by Comnas. Maj. Op. at 790.
¶1212. Piro
¶122The evidence developed at trial regarding the Post’s reliance on Piro demonstrated that the Post was aware that Piro was engaged in a less-than-amicable divorce with William’s daughter and that Piro was admittedly ignorant of the companies and the complicated business relationships involved in the Mobil-Samarco-Atlas arrangement (Tr. 2899, 2901, 2908, 2910). The latter fact alone would have been sufficient grounds for a reasonable reporter to question Piro’s conclusions. Nevertheless, the Post, while denying it now, did rely on Piro for several parts of its story. Piro was solely responsible for two of the statements in the article: (1) that William “[gave] Peter a little nudge to get him along” in the shipping business,
¶123At trial, the defendants testified that they really did not rely on Piro in writing the article, even though he was cited and quoted several times. Under the circum*110stances, the jury was entitled to disbelieve this testimony, especially in light of the fact that Piro was the sole source for the two statements above. In viewing this evidence in the light most favorable to the verdict, one can only conclude that the jury found that the Post, in relying on Piro, relied on a source admittedly unknowledgeable in the details of the business transactions for which he was quoted.
¶124C. Defendants’ State of Mind For Knowing or Reckless Falsehood
¶125Because it provides a motive for knowing or reckless falsehood, evidence that a newspaper or a reporter followed a sensationalistic policy, or possessed ill-will toward the plaintiff, is evidence of actual malice. See, e.g., Curtis Publishing Co. v. Butts, 388 U.S. 130, 169, 87 S.Ct. 1975, 1998-99, 18 L.Ed.2d 1094 (1967) (Warren, C.J., concurring) (defendant had adopted a “program of ‘sophisticated muckraking,’ designed to ‘provoke people, make them mad' ”) (footnote omitted); Goldwater v. Ginzburg, 414 F.2d 324, 342 (2d Cir.1969) (“evidence of ... motive and intent” may help establish actual malice), cert. denied, 396 U.S. 1049, 90 S.Ct. 701, 24 L.Ed.2d 695 (1970); Bose Corp. v. Consumers Union of United States, 692 F.2d 189, 196 (1st Cir.1982) (same), aff'd, 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984); Cochran v. Indianapolis Newspapers, 175 Ind.App. 548, 560, 372 N.E.2d 1211, 1220 (1978) (desire to “get” plaintiff is “relevant and admissible as evidence in the determination of whether defendant possessed a state of mind highly conducive to reckless disregard of falsity”) (emphasis in original); R. Sack, supra, at 214 n. 168 (“Although common law actual malice — spite or ill-will — is not equivalent to or sufficient to prove constitutional ‘actual malice,’ evidence as to the former is admissible to prove the latter.”). As the Supreme Court of West Virginia explained:
[W]hen [Associated Press v. Walker, 388 U.S. 130, 87 S.Ct. 1975, 18 L.Ed.2d 1094 (1967)] and Curtis are read together, they still stand for the proposition that personal motives on the part of a newspaper, or participation by a newspaper in a plan or scheme to injure, is evidence of recklessness and willful disregard for truth which may be considered along with other evidence on the question of actual malice.
¶126Sprouse v. Clay Communication, 158 W.Va. 427, 211 S.E.2d 674, 688, cert. denied, 423 U.S. 882, 96 S.Ct. 145, 46 L.Ed.2d 107 (1975) (emphasis added). Cf. C. Mollenhoff, Investigative Reporting 358 (Macmillan Pub. Co. 1981).
¶127In this case, there is abundant evidence in the record demonstrating defendants’ state of mind for knowing or reckless falsehood.
¶1281. Tyler’s Desire to “Knock Off ’ Mobil and “Build a Case” against Tavoulareas
¶129Immediately after he first heard Piro’s story, Tyler seemingly accepted its truth and remarked to Golden: “It’s not every day you knock off one of the seven sisters.” (RE 789). Taking the evidence in the light most favorable to the plaintiff, this demonstrates an ambition at the very outset to “bring down” Mobil which could cause — and may properly be considered evidence tending to prove — a reckless disregard of the truth. The same antagonistic mind-set is reflected in Tyler’s reference, in his memorandum responding to Peterson’s criticism, to “our case against Tavoulareas” (RE 2489).
¶130The majority is playing with words when it says that these statements merely reveal that Tyler had adopted “an adversarial stance,” which is not indicative of reckless disregard of the truth. Taking an adversarial stance means treating someone like an opponent rather than impartially. While this may not conclusively demonstrate reckless disregard of the truth (any single piece of evidence rarely does), it unquestionably points in that direction. See Cochran v. Indianapolis Newspapers, 175 Ind.App. 548, 560, 372 N.E.2d 1211, 1220 (1978). An opponent is more likely to distort the facts than an impartial or objective reporter. I am surprised by the majority’s refusal to acknowledge the probative *111character of this evidence, since even the Post did not contest its relevance at trial.
¶1312. Tyler’s Willingness to Resort to Unlawful Means
¶132Tyler’s desire to “bring down” Mobil and “make a case” against Tavoulareas is particularly strong evidence of reckless disregard for truth or falsity when it is combined with evidence of Tyler’s willingness to use unethical, and indeed unlawful, means to achieve his ends. Taking the evidence in the light most favorable to plaintiffs, we must assume that Tyler inquired of Piro, with serious intent, whether Piro “knew of a family member who would rifle [Tavoulareas’] safe and [x]erox documents” (Tr. 179). Tyler’s willingness to resort to such felonious methods is consistent with a state of mind that would act recklessly in stretching the facts to “make a case.”
¶1333. The Post’s Pressure On Its Reporters
¶134One of the three issues the en banc court requested the parties to brief was “[w]hether managerial pressure upon reporters to produce sensationalistic stories is a factor that may be considered in the determination of actual malice?” The majority, apparently unconcerned by the fact that counsel for the Post conceded that evidence of such pressure was “relevant,” see Transcript of Oral Argument at 28-29 (October 3, 1985), concludes to the contrary. The majority does not dispute that Woodward instructed his reporters to produce “holy shit” stories — presumably stories so startling that they cause the reader to exclaim in this fashion. The majority attempts to discredit this evidence by replying that it does not indicate that the Post was unconcerned with the truth of its stories. Quite so, but totally irrelevant. The issue is not whether the Post subjectively desired false stories but whether extra-heavy pressure to produce sensationalistic stories could motivate reporters to stretch the truth. Of course it could, as any reasonable person, including the Post itself,
The Saturday Evening Post was anxious to change its image by instituting a policy of “sophisticated muckraking,” and the pressure to produce a successful expose might have induced a stretching of standards. In short, the evidence is ample to support a finding of highly unreasonable conduct constituting an extreme departure from the standards of investigation and reporting ordinarily adhered to by responsible publishers.
¶135388 U.S. at 158, 87 S.Ct. at 1993 (emphasis added). See also id. at 162, 169, 87 S.Ct. at 1995, 1998-99 (Warren, C.J. concurring).
¶136*112As to the effect of the Post’s pressure, that was for the jury to weigh. It is of course not inevitable that extreme pressure for sensationalistic stories will lead to distortion of the truth. And I cannot imagine that such pressure alone, without more, would support a finding of actual malice.
¶137D. Erroneous Recording of Interviewees’ Statements
¶138The record reveals that Tyler misstated in his notes and in the article responses by individuals whom he interviewed. For example, Tyler called Everett Checket, a Mobil vice-president, to inquire whether William Tavoulareas told Checket that he had sent Hoffmann to Atlas. Tyler wrote in his notes that Checket said the conversation sounded “somewhat familiar” (RE 2493). Checket testified, however, that he did not tell Tyler that the conversation sounded familiar. Indeed, Checket denied that the conversation ever occurred (Tr. 1844-48). The jury could reasonably have concluded from this evidence that Tyler either carelessly or intentionally misstated the contents of his interview with Checket.
¶139There is also evidence that Tyler failed to record accurately the statements of Peter Tavoulareas. Peter’s notes of his conversation with Tyler show that Peter stated “[a~\side from our management we have nothing to do with the Arabs or with Mobil” (RE 2550) (emphasis added). Tyler wrote in his notes “nothing to do with Samarco, nothing to do with Mobil” (RE 2518) and quoted Peter in the article as stating “Atlas has nothing to do with Mobil, it has nothing to do with Samarco and it has nothing to do with the Saudis” (Paragraph 33).
¶140Thus, the record supports a conclusion that, at best, Tyler was negligent in erroneously recording the contents of interviews. Although evidence of negligence does not alone support a finding of actual malice, such evidence can be probative of a reckless disregard for truth or falsity. Bose Corp. v. Consumers Union, 692 F.2d 189, 196 (1st Cir.1982), aff'd 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984); Goldwater v. Ginzburg, 414 F.2d 324, 343 (2d Cir.1969) ("[New York Times v. Sullivan ] does not hold that evidence of negligence is inadmissible”), cert. denied, 396 U.S. 1049, 90 S.Ct. 701, 24 L.Ed.2d 695 (1970).
¶141E. Resolution of Inferences Adverse to the Plaintiff
¶142It is well settled that in a situation where the facts are ambiguous, the mere selection *113of the most damaging inference by the reporter does not, alone, indicate reckless disregard for the truth. In Time, Inc. v. Pape, 401 U.S. 279, 91 S.Ct. 633, 28 L.Ed.2d 45 (1971), the Court held that Time magazine’s resolution of one such ambiguity in the light most damaging to the plaintiff did not amount to actual malice:
Time’s omission of the word “alleged” amounted to the adoption of one of a number of possible rational interpretations of a document that bristled with ambiguities. The deliberate choice of such an interpretation, though arguably reflecting a misconception, was not enough to create a jury issue of “malice” under New York Times.
¶143Id. at 290, 91 S.Ct. at 639. This holding makes obvious sense; since the focus of the malice inquiry is subjective, not objective, the mere selection of a single damaging inference is ordinarily not sufficient to prove actual malice.
¶144The defendants argue from Papethat the consistent selection of the inferences most damaging to the Tavoulareases would be irrelevant to the issue of actual malice. Appellee’s Brief at 62. This assertion, however, is entirely inconsistent with the subjective nature of the malice inquiry and goes far beyond the holding in Pape. A defendant’s single, or even occasional, selection of the most disparaging inference from among a number of possibilities is not evidence that it acted with malice. But a defendant’s consistent rejection of favorable inferences in favor of the most damaging inferences — inferences that in fact are false — does suggest a determination to incorrectly describe the plaintiff and place him in the most damaging light, and thus can demonstrate a reckless disregard for the truth.
¶145In this article the Post did not simply resolve a single ambiguity adversely to Tavoulareas. In building its “case against Tavoulareas,” it resolved practically all ambiguities against him, consistently acceptmg the most damaging statements from obviously biased and impeachable sources. It is absurd to suggest, as the majority does, Maj. Op. at 797, that this is not a basis for an inference of reckless disregard for the truth.
¶146F. Suppression of Information Favorable to the Plaintiff
¶147Evidence demonstrating the suppression of information favorable to the plaintiffs is closely related to evidence of the adoption of inferences adverse to the plaintiffs. Clearly, a writer who knowingly and consistently suppresses information favorable to the plaintiffs is more likely to have entertained subjective doubts as to the accuracy of his story. See, e.g., Time, Inc. v. Ragano, 427 F.2d 219, 221 (5th Cir.1970) (failure to include fact that plaintiff was attorney is evidence of actual malice when implication of article would lead people to believe he was organized crime figure); Wasserman v. Time, Inc., 424 F.2d 920, 922 (D.C.Cir.) (same), cert. denied, 398 U.S. 940, 90 S.Ct. 1844, 26 L.Ed.2d 273 (1970); Montandon v. Triangle Publications, Inc., 45 Cal.App.3d 938, 943-44, 120 Cal.Rptr. 186, 189 (omission of key fact in otherwise truthful statement held evidence of reckless disregard of truth), cert. denied, 423 U.S. 893, 96 S.Ct. 193, 46 L.Ed.2d 126 (1975); Indianapolis Newspapers v. Fields, 254 Ind. 219, 259 N.E.2d 651, 662 (time discrepancy known to reporter and witnesses’ repudiation of key fact mentioned only once during series of articles held evidence of actual malice), cert. denied, 400 U.S. 930, 91 S.Ct. 187, 27 L.Ed.2d 190 (1970).
¶148Evidence abounds in this record showing that the Post suppressed information favorable to the Tavoulareases. The district court examined several of these pieces of information which were known to Tyler but which were not included in the story:
1. Lewis Lapham, an outside director of Mobil, told Tyler that the Mobil board *114had “consistently reviewed the relationship between Mobil and Atlas and ... was completely satisfied with all aspects of it.” 567 F.Supp. at 658 (emphasis added).
2. Lapham told Tyler “that he did not believe that plaintiff [William Tavoulareas] played a personal role in Atlas and that at key board meetings plaintiff would leave the room to facilitate the opportunity for more open discussion of the subject.” Id.
3. There was considerable evidence that “Mobil profited significantly from [the Samarco-Atlas] relationship and that Atlas was not merely a fly-by-night organization set up solely for Peter’s benefit.” Id. at 659. Both John Kousi, a Samarco director, and Comnas, the defendants’ main source for the Post article, told Tyler that Atlas could run Mobil’s ships more cheaply than the oil company could.
¶149Early drafts of the story did contain some of this information, but it was eliminated from the final version during the editing process. One early draft, for example, stated:
During these briefings [on Atlas-Samarco dealings] by Warner, Lapham said, Tavoulareas would leave the room. Lapham said he still believes Warner’s account of events and rejects any suggestion that Tavoulareas took a personal role in Atlas.
“There’s nothing to any of that,” Lap-ham said. “It was explained to the satisfaction of the board.”
¶150(RE 2528). The defendants argue that while Lapham’s specific words were cut out, the point he made was not. But proof of the point was placed on weaker grounds. As a substitute for Lapham’s specific statement that Tavoulareas went so far as to leave the room during board discussions of Atlas, and his assurance that there was “nothing to [the suggestion that Tavoulareas took a personal role in Atlas],” the defendants point to the following passage from the article:
Mobil chairman Warner says he assured directors in board meetings that Tavoulareas “does not participate in any decisions” relating to Mobil’s business with Atlas.
¶151Thus, the article substituted an “assurance” by Warner (an inside director) for a more concrete and expansive statement by Lapham (an outside director) — a substitution that considerably weakens the point, particularly when an underlying theme of the Post article was that Tavoulareas' mere position as president constituted internal corporate influence that implicitly affected the Atlas operation.
¶152Next, in place of Lapham’s direct statement that the Mobil board “was completely satisfied” with the Samarco-Atlas relationship, the article stated:
The Mobil board of directors was told from the outset about the Atlas arrangement but was assured that company president Tavoulareas was not involved in his son’s venture in any way.
¶153Thus, instead of a direct assertion of an unimpeached outside director that the Board had investigated and was “completely satisfied,” the article states merely that the directors were simply “assured” by an unidentified “someone” that there was no impropriety. This grossly understates the force of the information provided by Lap-ham and operates to the disadvantage of Tavoulareas. Lapham’s statements that the Board had “consistently reviewed” the Atlas arrangements and that he was convinced that William Tavoulareas had played no personal role in Atlas were simply ignored and never printed.
¶154The Post article also disregarded other evidence in its possession that tended to describe the arrangement in a light more favorable to the plaintiffs and to Mobil. In evaluating the economic impact of the arrangement, for example, the Post ignored the information supplied by Kousi (whose favorable comments were deleted during the editing process) and even the comment furnished by its principal source, Comnas, that the Atlas arrangement saved Mobil money (RE 2501). The defendants alluded instead to “millions of dollars in business” and “exclusive, no-bid contracts,” which strongly implied that Mobil was being *115greatly exploited and gouged by an irregular arrangement in which Tavoulareas had “set up” his son. Similarly, the story failed to include or refer to two internal Mobil memoranda prepared at the time the Samarco-Atlas arrangement was created and which Tyler had in his possession. One memorandum instructed Mobil personnel that all business dealings with Atlas were to be conducted at “arm’s length; ” the other, written by William Tavoulareas, specifically directed Vice-President Wolfe to bypass Tavoulareas and report directly to Chairman Warner on any Samarco-Atlas matters.
¶155From this evidence, two reasonable conflicting factual inferences could be drawn — i.e., that Mobil was being gouged or that the Samarco-Atlas venture was financially beneficial to Mobil. But we are required to draw the legitimate inferences most favorable to the verdict. This evidence of the suppression of information favorable to the plaintiffs was directly probative of Tyler’s reckless disregard for the truth or falsity of the article.
¶156G. Conclusion
¶157The evidence in the record viewed in the light most favorable to the verdict establishes the following:
• Subjective doubts regarding the article’s truth were entertained by one person responsible for the publication (Peterson) and were brought by her to the attention of others responsible for the publication, who proceeded in disregard of her cautionary memorandum.
• The article relied on patently unreliable sources for its central allegations.
• Tyler wanted to “bring down” Mobil and “make a case” against Tavoulareas.
• Tyler was willing to resort to unethical, and indeed unlawful, means for that purpose.
• The reporting and editorial judgments made in connection with the article were affected by heavy managerial pressure to produce sensationalistic stories.
• Tyler either negligently or deliberately misstated remarks made by interviewees.
• The article regularly drew inferences adverse to the Tavoulareases where favorable inferences were equally available and justifiable.
• The article suppressed important information favorable to the Tavoulareases. In my view, this evidence establishes
¶158clearly and convincingly that the false statements and implications in the subject article were put forth with reckless disregard for their truth or falsity. The fact that Tyler faced no deadline pressure, and thus had plenty of time to reflect upon the story’s implausibility and conduct further investigation, buttresses this conclusion.
¶159The affirmative indications of good faith introduced by the defendants (which, under j.n.o.v. standards must be regarded in the least favorable light) are insubstantial and do not alter the conclusion. Primary among these is the fact that Tyler devoted a considerable amount of research to the story, working on it over a thirty-day period. Extensive efforts to get at the truth can be persuasive of good faith in some circumstances, but not when, as discussed above, the products of those efforts favorable to the plaintiff are systematically disregarded or excluded from the story, and adverse inferences systematically drawn. When such distortion of research appears, protracted investigation evidences only a dogged and thorough attempt to “get” the subject.
¶160The systematic suppression of favorable evidence likewise reduces to insignificance the potential probative effect of the defendants’ attempts to interview William Tavoulareas. The defendants were already aware that Tavoulareas had denied the substance of the charges, and there is no reason to believe that his exculpating statements with regard to any of the details would have been treated any more sympathetically than the exculpating statements Tyler received from more impartial sources such as Lapham — which, as we have seen above, were suppressed or altered. The Post did at least record Tavoulareas’ denial. The proof of malice would have been even more clear and convincing if the Post had not done so, but that minimal conces*116sion to fairness falls far short of redeeming a pattern of behavior that displays, clearly and convincingly, a reckless disregard for truth or falsity.
¶161Chief Judge Wald asserts that the foregoing analysis does not constitute an independent review of the evidence of actual malice. Concurring Op. at 805-06. To the contrary, the review employed here conforms with recent Supreme Court pronouncements on actual malice, which stress that “[t]he independent review function is not equivalent to a ‘de novo ’ review of the ultimate judgment itself,” Bose Corp. v. Consumers Union of the United States, 466 U.S. 485, 514 n. 31, 104 S.Ct. 1949, 1967 n. 31, 80 L.Ed.2d 502 (1984) and that “[credibility determinations, the weighing of the evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of a judge.” Anderson v. Liberty Lobby, 477 U.S. 242, -, 106 S.Ct. 2505, 2513, 91 L.Ed.2d 202 (1986). Judge Wald’s proposal that the court “independently review inferences that the jury may be presumed to have drawn,” Concurring Op. at 805, contradicts the Supreme Court’s recent reaffirmation in Andersonthat even when actual malice must be established (with clear and convincing proof), the drawing of legitimate inferences remains the function of the jury. Id.Independent review instead necessitates that the court evaluate the sufficiency of the facts in the record and the legitimacy of the inferences necessarily drawn by the jury from those facts in assessing whether the constitutional requirement of clear and convincing evidence of actual malice has been met. In such manner, the reviewing court respects both the right to jury trial under the Seventh Amendment
¶162The foregoing independent assessment of "those portions of the record which relate to the actual-malice determination,” Bose, 466 U.S. at 514 n. 31, 104 S.Ct. at 1967 n. 31, leads to the conclusion that the facts and the reasonable inferences drawn therefrom constitute clear and convincing proof that The Washington Post acted with actual malice when it published the November 30, 1979 article. The jury’s verdict should thus stand.
¶163V. Piro
¶164The plaintiffs sued Dr. Piro for slander and its foreseeable republication with respect to, among other things, his statement that William had “set up” Peter as a partner in Atlas. Complaint, Tavoulareas v. Piro, Civ. No. 80-2387 (D.D.C. Sept. 19, 1980) (RE 89, 90). The jury returned a verdict in favor of both plaintiffs, but the district court granted judgment n.o.v. in favor of Piro against William (although the court affirmed the jury verdict in favor of Peter). From the analysis set forth above, it is clear that a jury in William’s case could reasonably have found that this allegation by Piro was false and defamatory. The question then is whether it was uttered with reckless disregard for its truth or falsity.
¶165At trial and in his brief to this court, Piro admitted that he did not comprehend the details of the business arrangements involving Mobil, Samarco and Atlas. See Brief for Appellee Piro at 10; Piro’s testimony at Tr. 2899-2928. Indeed, he testified that he had “never heard of Samarco” (Tr. 2899). Since knowledge of those details was obviously crucial to the judgment *117whether William Tavoulareas breached his fiduciary obligation, Piro’s acknowledgment of ignorance was tantamount to an admission that in making that charge he acted recklessly. When that is combined with Piro’s hostility to the Tavoulareases resulting from his pending divorce, there was clear and convincing proof of Piro’s reckless disregard for truth or falsity.
¶166VI. Conclusion
¶167Great national issues are raised by the methods the majority employs to justify its reversal of the jury verdict in this case. By distorting the standards for applying j.n.o.v., the majority produces an opinion based on truncated facts and incomplete law. The outlandish refusal of the majority to consider the article as a whole surfaces throughout its discussion of defamatory meaning, falsity and actual malice. The majority goes to great lengths to avoid imposition of any liability on the Post — it whittles down the defamatory meaning of the article, stacks selective facts in an attempt to establish the truth of a stripped-down version of the “set up” charge and fails to acknowledge the clear and convincing evidence that the Post published the article with reckless disregard for its truth or falsity. In what it terms an “independent examination of the whole record,” the majority ignores very substantial amounts of testimony by credible witnesses, disregards facts contradicting its conclusion and in effect substitutes its own factual determinations for those obviously found by the jury. Such methods greatly exceed a court’s proper role in independently reviewing libel judgments under the First Amendment. If the majority’s analysis is to be upheld in the name of “independent review,” the Seventh Amendment right to a jury trial and reexamination of facts according to the rules of the common law will be obliterated. Is such evasive analysis to be upheld as the standard for the future?
¶168The proper course for a reviewing court is to defer to the jury’s factual findings and credibility determinations and reevaluate only the legitimacy of the inferences the jury necessarily drew in finding actual malice. The court then must determine whether the facts and reasonable inferences therefrom amount to clear and convincing evidence of reckless disregard for truth or falsity. This is the independent review required by Bose and the independent review conducted by this opinion. The jury’s verdict should be reinstated.
¶169 Judges Bork and Silberman were recused; Judges Buckley, Williams and D.H. Ginsburg were sworn in after the case was argued; and the participation of Judge Scalia was terminated by his appointment to the Supreme Court. See ante at 766.
¶170. Chief Judge Wald, concurring in the judgment of the en banc majority, agrees that the majority does not apply the judgment n.o.v. standard with integrity. Concurring Op. at 804-05.
¶171. The majority’s failure to disclose the principles guiding its review, moreover, is especially unfortunate in a case such as this, where
the predictability of decisions, which is of crucial importance in an area of law touching upon First Amendment values, is enhanced when the determination is made according to announced legal standards and when a body of public case law furnishes published examples of the manner in which these standards are to be applied.
¶172Ollman v. Evans, 750 F.2d 970, 978 (D.C.Cir.1984) (en banc), cert. denied, 471 U.S. 1127, 105 S.Ct. 2662, 86 L.Ed.2d 278 (1985).
¶173. “[S]ome 12 or 14 other [oil] companies [were] establishing similar organizations____” (Tr. 1090) .
¶174. The ownership of Samarco was ultimately divided as follows: the Alirezas, 45%; Prince Fahd, 15%; Mobil, 30%; Fairfield-Maxwell, Ltd., 10%.
¶175. One of the defendants’ principal witnesses, a former director of the shipping concern of Fair-field-Maxwell, Ltd., John Kousi, stated that such circuitous leasing, in contemplation of Saudi preference laws, "was a commonly understood business idea at the time," devised as a means to capitalize on an opportunity to make substantial profits by using surplus tankers laid up by the world-wide depression in oil (Tr. 3220-23, 3263, 3167), and that Atlas could operate ships at less cost than Mobil (Tr. 3246).
¶176.Mobil stated that it wanted an independent shipping operation to avoid a conflict of interest between the shipping interests of Mobil and those of Fairfield-Maxwell, Ltd. (Tr. 3126). This point was well taken.
¶177. The ownership of Atlas under Comnas was ultimately divided as follows: Comnas, 57%; Peter, 38%; Emmanuel, 5% (RE 2472-73).
¶178. In January, 1974, William Tavoulareas notified Mr. Warner, the Chairman of Mobil’s Board of Directors, immediately after being so advised personally by Comnas on January 9, 1974 (Tr. 1277), that "there was a chance of [his son Peter] participating in some kind of a joint venture with Comnas” (Tr. 4135). At the time, Comnas was seeking to charter a vessel to Mobil (Tr. 1277). Shortly thereafter, the possibility of Peter joining Comnas was disclosed to George Birrell, Chairman of Mobil’s Conflict of Interest Committee (RE 1458-61). The situation was subsequently discussed several times by the Conflict of Interest Committee and Chairman Warner. Procedures were approved under which William would be isolated from "decisionmaking," although, as was necessary given his responsibilities as president and his experience as Mobil’s expert on Saudi affairs, “he could involve himself up to the decision level,” and Warner so advised Comnas. The Post and the majority completely refuse to recognize this degree to which William was permitted to function and what was expected of him in this area (Tr. 1515-21). The organization of Atlas was completed in July, 1974 (Tr. 4427). “Peter joined Atlas [in] August 1974” (Tr. 4158). On August 2, 1974, William Tavoulareas wrote a letter stating, “I will not make decisions” (Tr. 4136). And on the same date Mr. Wolfe wrote a letter "notifying the people in the [Mobil] organization on a need to know basis of Peter's participation” (Tr. 4136).
¶179.According to Emmanuel’s testimony, Comnas was not at Atlas offices for any length of time (Tr. 2240, 2243). Comnas was off pursuing grandiose visionary projects (Tr. 2241, 2245); he could not be reached as he was traveling (Tr. 2247); he felt he could "do whatever he wanted without any regard to anybody" (Tr. 2244); he took $130,000 for salary and increased that by "expenses which could not be substantiated” (Tr. 2244); he would have deprived Samarco of their 5% discount on ship repairs, "which would have gone to some other source” (Tr. 2264) (emphasis added); and despite a conflict of interest as a stockholder in IOTRON he urged Atlas to buy ship equipment manufactured by IOTRON (Tr. 2265). Comnas also negotiated a deal with Japan Lines that would have been detrimental for Atlas if approved (Tr. 2250). Samarco and Mobil forced Comnas to reject it (Tr. 2256). In other words, the Samarco partners were completely justified in terminating Comnas’ employment.
¶180. Hoffmann’s name was incorrectly reported as Herman Hoffmann in the Post's November 30 article.
¶181. Although the majority asserts that "several of Mobil's outside directors raised objections to Peter’s involvement in Atlas,” Maj.Op. at 769, the record indicates that the one outside director who had originally expressed some reservations about the "appearance” of Peter’s involvement in Atlas, Ambassador McGhee, questioned neither the business justifications nor the ethical validity of Mobil’s entry into the Samar-co-Atlas arrangement. Following the SEC investigation of William Tavoulareas that resulted in no charges of wrongdoing, McGhee stated: "Congratulations, Tav, I knew you’d be cleared.” (Tr. 1813).
¶182.When the article appeared, Tyler was given sole byline credit for it, and Golden received credit as a "Special Correspondent” for the story.
¶183. "Seven Sisters” is a colloquial term for the seven largest oil companies, one of which is Mobil.
¶184. See infra at 778 for the text of Peterson’s memo.
¶185. A second article was published on December 1. but the jury found that it did not defame either of the plaintiffs.
¶186. The Washington Post Company is a large multimedia corporation which has substantial broadcast holdings and publishes magazines and newspapers, including the Post.
¶187. Also named in the complaint was Post President Katharine Graham, but during trial the district court dismissed the case against her, and the plaintiffs do not contest this. Accordingly, she is not a party to this appeal.
¶188. The panel denied rehearing and filed an opinion, 763 F.2d 1472 (D.C.Cir.1985), but with *93Judges Tamm and Scalia voting to deny, the court ordered rehearing en banc.
¶189. This opinion assumes, without deciding, that William Tavoulareas is a limited-purpose public figure.
¶190. It might be added, however, that the reconcil*94iation of independent review and deference to the fact-finder there set forth is in accordance with the reconciliation adopted by the Supreme Court in other First Amendment areas. For example, in Street v. New York, 394 U.S. 576, 89 S.Ct. 1354, 22 L.Ed.2d 572 (1969), the Court conducted an independent review of the ultimate constitutional fact whether the appellant’s words were "fighting words," so that the state could constitutionally punish their utterance. See id. at 592, 89 S.Ct. at 1365. In the process, however, the Court held that it must "consider[ ] [the evidence] in the light most favorable to the [verdict].” Id. at 590, 89 S.Ct. at 1365. See also Bose, 466 U.S. at 504-08, 104 S.Ct. at 1961-64.
¶191. See, for example, the interpretation given the article by Post counsel at trial, infra at 822-23.
¶192. As discussed infra at 782-85, the majority fails to credit all permissible inferences to Tavoulareas in holding the narrow "set up” allegation to be true — the jury could reasonably have found that Tavoulareas did not "set up" Peter in Atlas.
¶193. As discussed infra at 782-85, evidence at trial established that William Tavoulareas did not "set up" Peter in Atlas but rather acted purely in Mobil’s interests. The article’s implication that William breached his fiduciary duties to Mobil and wasted and misused assets of Mobil for Peter’s benefit could thus have been found false by a reasonable jury.
¶194. Both William Tavoulareas and Paul Wolfe, the Mobil executive with decisionmaking authority after William surrendered such authority with respect to Mobil-Samarco-Atlas, testified that Tavoulareas did not personally send Hoffmann to Atlas (Tr. 1098, 1440). The jury could therefore have reasonably concluded the defamatory "personally dispatched” charge was false.
¶195.That the article is capable of this defamatory meaning is established by the fact that the general public interpreted the article to charge William with abuse of his position as a corporate official. For example, the Service Union Reporter (a newspaper for unionized public employees) published a cartoon depicting William as an indulgent father giving some Mobil ships to Peter "to play with." In an article accompanying the cartoon, Peter's position in Atlas was directly attributed to “Daddy Tavoulareas — pull[ing] more than a few strings to put his son in the captain’s chair of a shipping empire." (RE 1560-^66, 2542). One former Commissioner of Internal Revenue called the Post article a story of "corporate incest ... [meaning] the *97relationship between the corporation and family members, or other corporations controlled by family members” (Tr. 1875). That the Post's readers interpreted the article to chronicle a corporate president’s violation of his fiduciary duties simply cannot be denied.
¶196. Seeking to distinguish McBride, the majority asserts that that decision "does not bear upon ... which of two defamatory meanings the [set up] statement in our case bears." Maj.Op. at 782 n. 23. The majority assumes that by finding that the “ ‘set up’ statement defames Tavoulareas by charging him with nepotism,” (a charge that the majority later finds is true), it can escape comparison to McBride. It is submitted, however, that McBride does illuminate the instant "dispute ... over the sweep of the [set up] statement.” Maj.Op. at 782 n. 23. In McBride we did not narrowly interpret the article’s fee comparison to merely accuse Dr. McBride of charging unreasonably high fees. Rather, we held that the article's statements on the expert witness fees were capable of a much broader defamatory meaning in that they could reasonably be read to imply that Dr. McBride’s testimony was for sale to the highest bidder. It is consistent with our interpretation of the defamatory statement in McBride to read the "set up" allegation here as capable, of bearing the defamatory meaning that Tavoulareas misused corporate assets and breached his fiduciary duties to the corporation’s shareholders by "setting up" Peter in Atlas.
¶197. Contrary to the majority’s protestations otherwise, Maj.Op. at 778 n. 16, there is an obvious distinction between the article’s accusation that Mobil joined Samarco once it occurred to William Tavoulareas that Mobil’s entry into Samar-co could ultimately benefit his son and the charge that, after Mobil decided to enter Samar-co, William wasted and misused Mobil assets, *98financing and personnel to prod his son's fortunes in Atlas. While both are aspects of the article's central defamatory theme that William violated his fiduciary duties as president of Mobil to benefit Peter, the charges are separately defamatory. The article’s charge that William’s actions with respect to Atlas violated his fiduciary duties remains even if the majority’s extreme contention were correct that the article cannot be read to attack Mobil’s motives for entering Samarco.
¶198The majority criticizes the foregoing analysis of the article’s defamatory meaning as one not advanced by Tavoulareas. Id. The majority’s attack completely ignores that the district court charged the jury to And whether, as “William Tavoulareas, in his complaint contendfedj ... the articles reasonably imply ... ’ [t]hat he breached his fiduciary duties to Mobil ... wasted and misused assets of Mobil ... [and] wrongfully diverted such assets to Peter for his benefit.” (Tr. 4549-50) (emphasis added). Moreover, the district court, in its memorandum accompanying the order of judgment n.o.v., considered Tavoulareas’ argument that the article implied that William "put together the Mobil-Atlas-Samarco arrangement solely for the sake of his son," 567 F.Supp. at 660, separately from Tavoulareas’ argument that the "article impliedly accused him of misusing Mobil’s assets." Id. Finally, in the decision reinstating the jury verdict, the panel wrote:
the article could reasonably be read to suggest that William Tavoulareas became involved with Atlas-Samarco primarily to benefit his son— At a minimum, the article implies that, irrespective of the motivation for creating the Mobil-Atlas-Samarco relationship, William misused Mobil assets and his position as president to advance his allegedly undeserving son — an act of corporate nepotism. Indeed, this interpretation of alleged nepotism was recognized by the district court as the basic
¶199theme of the article. 567 F.Supp. at 660. 759 F.2d at 111 (emphasis added). Quite clearly, the defamation analysis employed here is not new.
¶200Also erroneous is the majority’s contention that the defamatory implication that William Tavoulareas abused his fiduciary duties and misused corporate assets fails to add anything to the article's charge of nepotism. Even if one accepts, arguendo, the majority’s conclusion that William Tavoulareas "made it possible for Peter to become a partner in Atlas and then helped to ensure that the business would prosper because Peter was his son,” Maj.Op. at 780, the majority offers no evidence whatsoever that William Tavoulareas wasted Mobil assets, financing or personnel to the detriment of Mobil shareholders.
¶201Rather than conceding that Tavoulareas' actions were nepotistic, it is pointed out infra at 826-29 that the evidence is far from undisputed that William made it possible for Peter to become a partner in Atlas. Even less so is there evidence that William abused his corporate position and wasted Mobil resources in his dealings with Atlas. Not only does the majority fail to give this defamatory implication of the article the consideration it warrants, but the majority also dismisses evidence supporting this interpretation because the evidence does not sustain the separate defamatory implication that Tavoulareas engineered the entire Mobil-Samarco-Atlas venture to benefit Peter. See, e.g„ Maj.Op. at 782 n. 24. The majority’s odd tactic is, to say the least, unfair.
¶202. Indeed, Tyler himself so interpreted his article by stating in writing that ”[o]ur story does show that Mobil’s decisions in this case were not made for the traditional business reasons or for the reasons stated by Mobil." (RE 2489).
¶203. The majority seeks to avoid the conclusion that the article implies Samarco was “set up” for Peter by emphasizing the portion of the Post’s closing argument that acknowledges there were "good business reasons to go into the Samarco venture." Maj.Op. at 782 n. 25. While assuming that the article's mention of good business reasons for Mobil to enter Samarco precludes an inference from the article that Samarco was "set up" for Peter, the majority ironically concludes that Peter was “set up" in Atlas, despite evidence of good business reasons for Atlas to manage Samarco's ships. The majority cannot have it both ways — either the existence of "good business reasons" precludes a “set up" charge or it does not.
¶204Moreover, the majority's attempt to excuse the language used by Post counsel as "a single passage," Maj.Op. at 782-83 and uttered in the "heat of closing argument," id.at 782 n. 25, carries little force given the tremendous importance and forethought associated with summation to the jury.
¶205. In attempting to justify its refusal to consider the defamatory implications of the article, the majority makes the remarkable suggestion that defamation by implication cannot exist in an article that contains explicit statements unfavorable to the libel plaintiff. See Maj.Op. at 781 n. 21. Although the majority quotes selectively from the article to illustrate what it considers the "thrust of the piece" — i.e., Peter’s "meteoric rise to partnership in Atlas was an act of favoritism,” see id.— the majority astonishingly ignores huge portions of the article that concern only Mobil’s relationship with Samarco (Paragraphs 8-11, 39-52, 55-59). These portions of the article, read in the context of the whole story of "favoritism” (as the majority describes it), imply that William Tavoulareas reversed Mobil’s original decision and set up the entire Mobil-Samar-co-Atlas arrangement for Peter’s benefit. This defamatory implication is undeniably present, in addition to and notwithstanding the explicit statements in the article addressing Peter's "meteoric rise."
¶206. In light of the evidence at trial establishing that William Tavoulareas’ involvement in the Samarco-Atlas arrangement was in Mobil’s interest and with the concurrence of other top Mobil officials, see infra at 826-29, a reasonable jury could have found false the article's defamatory implication that Tavoulareas engineered the Mobil-Samarco-Atlas scheme for Peter’s benefit.
¶207. William Tavoulareas denied ever urging that his son be included in Atlas (Tr. 1293-94, 1296-97, 1433).
¶208The majority concedes, as it must in viewing the evidence most favorably to Tavoulareas under j.n.o.v. standards, that a jury reasonably could have found the "personally urged” charge to be false. Maj.Op. at 788.
¶209. Why the majority attempts to have the two charges considered separately becomes evident in its discussion on actual malice. After finding the "set up” allegation to be defamatory but true and the "personally urged” statement to be defamatory and possibly false, the majority assumes arguendo that Christine Peterson’s "memorandum is evidence of ‘serious doubt' or actual malice.” Maj.Op. at 794. The majority concludes, however, that Tavoulareas could not "connect Peterson’s disbelief to a false and defamatory meaning present in the article” because the "set up” allegation was found true. Of course, if the "personally urged” statement had been properly read as part of the “set up” charge that “Tavoulareas made it possible for Peter to become a partner in Atlas,” the majority would have been confronted with a false defamation that was published with actual malice. See Maj.Op. at 794 (“the [Peterson] memorandum *103could be interpreted ... to express disbelief that Tavoulareas “set up" Peter in Atlas”). By declaring the "personally urged” and "set up” allegations unrelated, the majority attempts to sidestep an actionable libel. The article, however, must be interpreted as a whole.
¶210. The district court recognized that the two statements are related. In concluding that the "set up" allegation was not a lie, the trial judge relied on Comnas’ statement that Tavoulareas had urged Comnas to bring Peter into Atlas. 567 F.Supp. at 659.
¶211. Twisting the evidence that William Tavoulareas had told Warner that Mobil was seriously considering joining Samarco and that Comnas was going to ask Peter Tavoulareas and Ares Emmanuel to join his new venture (Tr. 1514— 15), the majority torturously concludes that William stated “Comnas was likely to be tapped to manage Samarco’s ships.” Maj.Op. at 784 n. 29. The majority then attempts to discount Warner’s testimony that he thought of asking Comnas to manage Samarco's ships, claiming that ‘Warner was merely reacting to Tavoulareas’ statement” — a statement which Tavoulareas never uttered. Of course, the assessment of what Warner meant by his "absolute natural” statement is a determination to be made by the jury and not by a court claiming to review the verdict under judgment n.o.v. standards.
¶212.Tyler’s article stated that "[t]he elder Tavoulareas played a personal role in forcing the resignation ... of Comnas" (Paragraph 24), implying that William had caused Peter’s advancement in Atlas by forcing Comnas’ resignation. However, it is undisputed that Comnas’ removal was based on his unsatisfactory performance and was therefore justified. Thus, there is no ground for the article to suggest that William caused Peter’s rise in Atlas. Peter’s increased ownership resulted from Comnas’ justifiable removal and Hoffmann’s later decisions to refuse the offered equity interest in Atlas and to return to Mobil after he had determined that Peter and Emmanuel were qualified to manage Atlas.
¶213. The majority claims that the abundant evidence the dissent cites for the proposition that a reasonable jury could find the "set up” charge to be false is in reality only evidence that Tavoulareas complied with Mobil’s conflict of interest guidelines. Maj.Op. at 786 n. 33. To the contrary, the cited evidence also indicates that the benefit accruing to Peter from the Mobil-Samar-co-Atlas arrangement simply was not contrived by William. In sum, the evidence is far from "undisputed" that William "set up” Peter.
¶214. See also Maj.Op. at 778.
¶215. To be sure, portions of the non-defamatory statements in the Post’s article were true. “[T]he defamer may be the more successful when he baits the hook with truth.” Afro-American Publishing Co. v. Jaffee, 366 F.2d 649, 655 (D.C.Cir. 1966) (en banc). But a jury could readily have found false the defamatory implications of the article as a whole, as well as the specific allegations that "Tavoulareas set up his son,” “Tavoulareas personally urged that his son be included as an equity partner in Atlas” and "Tavoulareas dispatched ... Hoffmann to London to help run Atlas.”
¶216. This attempted qualification does not withstand analysis. Tyler wrote in response to Peterson’s memorandum that "a good editor might say that part of our case against Tavoulareas seems tenuous” (RE 2489) (emphasis added). Tyler thus recognized that Peterson’s memorandum interpreted the article to accuse William with organizing the Mobil-Samarco-Atlas “scheme for the sake of his son’s business career" and did not merely suggest a change in the "focus of the story.”
¶217. Without citation to any precedent, the majority asserts that "defamation plaintiffs cannot show actual malice in the abstract; they must demonstrate actual malice in conjunction with a false defamatory statement.” Maj.Op. at 794 (emphasis in original). The majority’s treatment of the evidence of actual malice demonstrates once again the failure of the majority to address the article as a whole.
¶218. The majority insists that because the Post included Tavoulareas' denial of the "personally urged” charge in the article, the statement could not have been published with actual malice. Maj.Op. at 779 n. 17 & 794 n. 44. It is noteworthy that in Mobil’s written response to Tyler’s questions, Tavoulareas did not only deny that he had, in April 1974, personally urged Comnas to take in Peter, but also affirmatively asserted that it was Comnas who had said he wanted Peter to join him in Atlas (RE 2344). Tyler’s failure to check with Comnas about the truth of Tavoulareas’ statement (which runs counter to Comnas' claim that William originated the idea of Peter joining Comnas’ new venture) indicates a reckless disregard for truth or falsity.
¶219The fact that Tyler published Tavoulareas’ denial of the "personally urged" allegation does not in any way vitiate the value of the evidence that Tyler published the charge with reckless disregard for truth or falsity. If the majority’s logic were followed, a newspaper that reported a man committed murder, knowing the report to be false or having serious doubts as to its truth, would be absolved from liability if the newspaper had also included the man's denial of the charge. Surely the First Amendment does not prevent a finding of actual malice whenever a defendant publishes false defamatory statements accompanied by denials.
¶220. In view of the Post's listing of Comnas as a witness (RE 2294) and then neither calling him nor introducing his deposition, the jury was left to evaluate the Post's defense without the testimony of the Post's principal source for the story. These circumstances must be taken into consideration in evaluating the credibility and inference determinations that the jury could legitimately draw from all the evidence.
¶221The fact that the Post neither introduced Comnas’ deposition nor called Comnas to testify is worthy of note even if reliance on a "missing *109witness” presumption would be inappropriate. That decision by the Post left its defense and the jury without any direct testimony from the principal source for the Post story. Comnas was the principal source for the story, and in place of Comnas’ direct testimony, or his deposition, the jury was left to consider primarily the repetition of his statements by Tyler whose credibility was weakened by obvious self interest, his questionable recording of conversations and other factors. See supra at 813-15, infra at 835.
¶222Moreover, it is far from clear that application of a missing witness instruction with respect to Comnas’ deposition testimony would be inappropriate in these circumstances. As we pointed out in United States v. Young, 463 F.2d 934, 942-43 (D.C.Cir.1972), the propriety of applying a missing witness presumption often turns on the missing witness’ "relationship to the parties." In this case Comnas must be considered as a witness favorable to the Post defendants. They relied upon him in the article. They listed him as one of their trial witnesses. At trial they relied upon his statements as repeated or allegedly recorded by others. Comnas’ statements were the very core of the Post’s defense; that defense was tremendously weakened by its unexplained reliance on second-hand evidence which resulted from the failure of the Post to call Comnas after listing him, or to use any part of his deposition. One must search far to find a weaker defense in a libel case. The true rule applicable to the situation presented here was set forth by Learned Hand, J., in United States v. Cotter, 60 F.2d 689, 692 (2d Cir.), cert. denied, 287 U.S. 666, 53 S.Ct. 291, 77 L.Ed. 575 (1932) and takes cognizance of the "relationship of the parties":
When both sides fail to call a witness who knows something of the facts, their conduct, like anything else they do, is a circumstance which a jury may use. If both can call him and he is impartial, ordinarily it will have little weight; if it appear that he would naturally side with one party, it is reasonable to expect that he does not use him for good reason; and that is fair argument for the other.
¶223(Emphasis added). The jury was perfectly free to come to a similar conclusion and no court should be required to rescue a defendant from such folly.
¶224. The majority stresses that Tyler received independent confirmation of Comnas’ charge from John Kousi. It is noteworthy that Tyler felt a greater need to obtain this "confirmation” after publication of the article than before. Tyler spoke with Kousi two or three times before the story appeared in print; ten to fifteen times afterwards (Tr. 3194). A more extensive interrogation of Kousi and others might have brought Tyler closer to the truth and given him a better understanding of the whole Mobil-Samarco-Atlas venture.
¶225. There is no basis in the record for concluding that this statement referred (as the article applied it) to William’s support for Peter’s joining Atlas and not to William’s support for Peter’s originally entering the shipping business with Lemos & Co.
¶226. The Post reached this precise conclusion in its published self-criticism following the so-called "Jimmy’s World" hoax, a Pulitzer Prize winning story published by the Post on September 28, 1980 concerning an alleged 8-year-old heroin addict. The story proved to be a complete fabrication. Following the public exposure of the story as false, the Post conducted an extensive internal investigation and, inter alia, published its finding regarding the effect of the pressure it placed upon its reporters. A part of that finding was as follows:
The troubling question is whether pressure on the staff distorts the news published in the paper____
Lewis Simons of Metro says: "Pressures are so great to produce, to go beyond excellence to the 'holy s — ' story. Everyone knows that’s what the editors want: The pressure is to get the incredible story, the extraordinary story.
“People want to succeed. They bust their ass to succeed here____ It can result in overselling a story____”
[S]ome reporters said they felt strongly that the "system” at the Post has editors making demands on reporters that cannot be met. That reporters are made to feel they are failures when they cannot meet those demands____
Editors are somewhat infected too.
¶227The Washington Post, Apr. 19, 1981, at A15.
¶228. The nature of the libel charged in the Buttsand Tavoulareas stories have many similarities. The media defendants in both cases misrepresented the intent with which those defamed acted. In Buttsall facts were construed to support a conclusion that the reason Butts, the Athletic Director at Georgia, and Bear Bryant, football coach at Alabama, were trading information was due to their "conspiring” (intent) "to fix” the upcoming Georgia-AIabama football game; in the Post article the alleged reason that Tavoulareas “set up" the Mobil-Samarco-Atlas venture (or just the "Atlas venture” as the majority would limit the charge) was because he intended to benefit Peter. In Butts, after denials *112by the plaintiff, it was thus for the jury to determine whether he had proved that he did not intend "to fix” the game as charged in the article. And here, after Tavoulareas’ denials, it was for the jury to determine whether he had "set up” and implemented the Mobil-SamarcoAtlas enterprise (or just the Atlas venture) with intent that violated his fiduciary duty to Mobil. In both cases, the basic facts — i.e., the existence of a phone conversation in Buttsand the entry by Mobil into the Samarco-Atlas venture — were true. The false defamation existed in the articles’ charges of corrupt intent.
¶229. Contrary to the majority’s characterization, I do not argue that a question of actual malice arises "whenever a libel plaintiff introduces evidence that the newspaper vigorously pursues high-impact stories of alleged wrongdoing.” Maj.Op. at 797. Rather, it is submitted that evidence of extreme pressure for sensationalistic stories may be probative of the issue whether a reporter was thereby motivated to deal recklessly with the facts.
¶230. The difference between Peter’s and Tyler’s versions is significant because the latter implies that Peter denied the management connection between Atlas and Samarco.
¶231. As Judge Gasch correctly charged the jury: In the course of your deliberations on the actual malice question, you [may consider] whether the Post defendants chose to resolve uncertainty in contradictions or ambiguities' in their information concerning the plaintiff in a way most harmful to the plaintiff.
¶232(Tr. 4555).
¶233. The Seventh Amendment provides:
In Suits at common law, where the value in controversy shall exceed twenty dollars, the right of trial by jury shall be preserved, and no fact tried by a jury shall be otherwise re-examined in any Court of the United States, than according to the rules of the common law.
¶234U.S. Const. Amend. VII.
¶235. Tavoulareas does not contest, and this opinion, along with the majority’s, see Maj.Op. at . 798 n. 53, assumes without deciding that Piro, although a non-media defendant, is nevertheless protected by the New York Times v. Sullivan requirement of proof of actual malice by clear and convincing evidence. This assumption also means that the court must exercise its independent judgment in determining whether the facts support a conclusion that the constitutional threshold of clear and convincing proof of actual malice has been met.
¶236. The jury awarded William Tavoulareas $250,-000 in compensatory damages and $1,800,000 in punitive damages against the Post. Post revenues in 1979 from circulation and advertising were over $240 million (Tr. 4017). Post-trial motions were made by the defendants in the district court, including motions for a new trial and motions to have the amount of compensatory damages reduced and the award of punitive damages set aside, or in the alternative, reduced. These motions were denied without prejudice when the judgments n.o.v. were granted. Those rulings are not before this court, and there is no occasion to pass on them at this time. However, I would reinstate the jury verdict against the Post defendants and Dr. Piro, and in that event the district court on remand would be required to reconsider these related motions. At this time no opinion is expressed on the issues raised by those motions.