This case is before us for the second time, and we assume familiarity with our first opinion, affirming Carl Leibowitz's conviction of a variety of crimes growing out of his efforts to procure the murder of a potential witness (his former partner in business and in crime). 857 F.2d 373 (7th Cir.1988). Leibowitz now is appealing from the district judge's refusal to grant him a new trial on the basis of affidavits in which the principal witness for the prosecution, Donald Wrobel, the hit man in the murder-for-hire scheme, recanted the testimony he had given at trial. The recantation occurred after our affirmance of Leibowitz's conviction. The motion for a new trial was brought under Rule 33 of the Federal Rules of Criminal Procedure, which allows two years from final judgment (following exhaustion of appellate remedies, United States v. Cook, 705 F.2d 350 (9th Cir.1983)) for the filing of the motion if the ground is newly discovered evidence, which, obviously, the recantation of a witness after trial is.

Judges view recantation dimly and this case shows why. In Leibowitz's trial Wrobel, a client of Leibowitz (Leibowitz is--or rather was--a lawyer; he has been disbarred), testified at length about the numerous efforts that he had made, upon Leibowitz's instructions, to rub out the witness, Gary Van Waeyenberghe. The most nearly successful occurred on September 21, 1986, at a roadside phone booth. Leibowitz had told Van Waeyenberghe to go to the booth at 10:00 p.m. to receive a call from him. Wrobel concealed himself at some distance from the booth, armed with both a pistol and .22 caliber rifle. Van Waeyenberghe arrived shortly before 10:00. At 10:00 the phone rang. It was Leibowitz, telling him, "I hear you've been talking to the feds." At that moment Wrobel fired the rifle. The bullet shattered the glass of the phone booth but missed Van Waeyenberghe.

Wrobel's testimony was corroborated by Van Waeyenberghe; by phone company records which revealed that calls from Leibowitz's phone had indeed been made to the phone booths to which, Wrobel testified, Leibowitz had directed Van Waeyeberghe in order to set him up for the assassination attempts, including a call to the roadside phone booth at 10:00 p.m. on September 21; and by tape-recorded conversations between Wrobel and Leibowitz that allude unmistakably to the existence of the murder-for-hire scheme. 857 F.2d at 381. The recantation occurred as follows. Wrobel had pleaded guilty to his part in the scheme and had been imprisoned, and by a regrettable mischance the Bureau of Prisons, inadvertently violating its own policies, had after our affirmance of Leibowitz's conviction transferred Leibowitz from the prison to which he had been originally assigned upon his conviction to the prison in which Wrobel was serving his sentence. The two met, and after several discussions Wrobel signed affidavits prepared by Leibowitz in which Wrobel recanted his trial testimony.

The affidavits are not worthy of belief. Apart from the questionable circumstances in which they were prepared, they contain fantastic assertions such as that the district judge conducted a mock trial of Wrobel in camera to persuade him to testify against Leibowitz, and they wholly fail to explain why Leibowitz lured Van Waeyenberghe to the phone booths. The "theory" propounded in the affidavits is that Leibowitz had hired Wrobel merely to trail Van Waeyenberghe. It fails to explain why, if that was the purpose, Leibowitz kept ordering Van Waeyenberghe to go to pay phones to receive calls from him with an armed Wrobel skulking in the bushes nearby.

It would create awful incentives to accept a recantation obtained by the defendant himself--a lawyer by training and former occupation--in face-to-face communication with another prisoner. Who knows what promises or threats Leibowitz made to Wrobel, whose psychological weaknesses were, in fact, a major defense theme at Leibowitz's trial?

The district judge conducted an evidentiary hearing before denying the motion for a new trial. Leibowitz's able counsel argues that the hearing was unduly truncated because the judge would not permit evidence related not to the truth or falsity of the recantation but to other questions bearing on Leibowitz's guilt, such as whether the glass in the phone booth would have shattered under the impact of a mere .22 caliber bullet, as Van Waeyenberghe testified it did, and whether a weather report showing that there were thunderstorms the night of September 21 was consistent with Wrobel's testimony. But all that was evidence that could have been discovered back in 1987, when Leibowitz was tried; the only newly discovered evidence fresh enough to warrant a hearing was the recantation. If the judge had believed Wrobel's affidavits he would then have had to decide whether there was enough other evidence of the murder-for-hire scheme to justify denying Leibowitz's motion, and in that setting other evidence bearing on guilt might have been admissible. And no doubt if the judge had been troubled by the question of the truth or falsity of Wrobel's recantation he might have decided to revisit the issue of guilt, since the strength of the other evidence against Leibowitz would certainly be relevant to the likelihood that the recantation was true. But he was not troubled; nor should he have been.

Only one point about the case troubles us and that is the insistence by the district judge and the United States Attorney that Wrobel's recantation, even if otherwise believable and believed, still could not be made a basis for a new trial unless Leibowitz was surprised by Wrobel's testimony at trial--the testimony he was trying to recant. In the first case in this circuit to announce a standard for when a witness's recantation should be accepted we said that the defendant must show three things: that the recantation is true; that the jury might have reached a different result if the witness in question had testified truthfully; and that the witness's false testimony took the defendant by surprise. Larrison v. United States, 24 F.2d 82, 87-88 (7th Cir.1928). Our subsequent cases repeat this formula--an example is United States v. Olson, 846 F.2d 1103, 1112 (7th Cir.1988)--as do a number of cases in other circuits. Annot., Recantation of Testimony of Witness as Grounds for New Trial--Federal Criminal Cases, 94 A.L.R.Fed. 60, 66-68 (1989). On the other hand, many federal and state cases do not include the third element of the Larrison test. Id. at 64-66; Annot., Standard for Granting or Denying New Trial in State Criminal Case on Basis of Recanted Testimony--Modern Cases, 77 A.L.R. 4th 1031, 1040, 1057 (1990).

Surprise is relevant, surely. If the defendant had every opportunity to meet the allegedly false testimony at trial, his failure to unmask its falsity at that time is some evidence that the testimony was true. But why the defendant should be required to demonstrate surprise in every case of recantation baffles us. In a case such as this in which the principal (though not the only) evidence of guilt is the testimony of an accomplice or eyewitness, the only resource of the defendant in unmasking the falsity, even with all the advance warning in the world, may be cross-examination, which--much mythology to the contrary notwithstanding--is not an infallible lie detector.

The tendency of judicial language to become ossified in rules is a familiar one, and it is illustrated by the promotion of surprise from a factor bearing on the acceptance of a recantation to an absolute requirement of such acceptance. We have found only one case in which absence of surprise alone barred acceptance of an otherwise truthful-seeming recantation--and that was a doozy. The defendant in United States v. Dworkin, 116 F.R.D. 29 (E.D.Va.1987), a doctor, had been prosecuted for parking in a space reserved for handicapped drivers, and had been acquitted on the basis of his testimony that he had parked in that space because of a medical emergency. Later he was investigated on suspicion of perjury, and while that investigation was in progress he filed a motion under Rule 33 for a new trial in the parking case on the ground that the testimony he had given in that case had been false. He hoped by expunging the testimony to beat a perjury rap. The district judge agreed that the testimony which the defendant was seeking to recant had been false, but pointed out that the defendant must have known this when he gave it and therefore could not plead that the false testimony--his own false testimony--had surprised him! Of course a witness who manufactures false evidence cannot use its falsity to get himself a new trial; that was Dworkin; but it is precipitate to infer that there can never be a case in which a defendant is entitled to a new trial because the false testimony that did him in, while indeed false, had not come as a surprise.

The rule of Larrison is dictum insofar as it teaches that surprise is a sine qua non of seeking a new trial on the basis of a witness's having recanted material testimony, and as such it is not binding. We add that the court in Larrison may have adopted the requirement of surprise in error. For just before formulating its three-part case the court cited with approval Martin v. United States, 17 F.2d 973 (5th Cir.1927)--which expressly rejects the requirement of surprise, remarking in the course of doing so that "notwithstanding the perjured testimony was contradicted at the trial, a new light is thrown on it by the admission that it was false." Id. at 976.

Our skepticism about the requirement of surprise can give no comfort to Leibowitz. The district judge made clear that, surprise or no surprise, Wrobel's recantation does not warrant a new trial.

AFFIRMED.