786 Opinion of MARSHALL, J. join Part V of the opinion of the Court. I also agree that the introduction of the shotgun into evidence at petitioner's trial did not violate the Fourteenth Amendment.1 But, I believe that in failing to disclose to petitioner certain evidence that might well have been of substantial assistance to the defense, the State denied him a fair trial. The opinion of the Court relates at some length the facts relating to the crime with which petitioner was charged, the circumstances of his arrest, the course of the trial, and the developments at the post-conviction hearing. As these facts are complicated and quite confus 1I find the constitutional question presented by the introduction of this evidence to be much harder than the majority seems to. It was uncontradicted at trial that the weapon introduced against petitioner had no bearing on the crime with which he was charged. It was, in fact, clear that the shotgun admitted into evidence was a 16-gauge gun, whereas the murder weapon was a 12-gauge gun. Despite the fact that the prosecution conceded this in a pretrial bill of particulars, it did everything possible to obfuscate the fact that the weapon admitted into evidence was not the murder weapon. This was highly improper. The record also indicates that the trial judge was confused as to why he thought the weapon should be admitted. At one point he said, "There was testimony here that this was a shotgun killing. And I can see nothing wrong if they say that this defendant, who will be identified by other people, was apprehended with this gun." Abstract of Record (Abs.), 65. If the trial judge meant to imply that because the crime was committed with a shotgun, it was sufficient to prove that the petitioner possessed any shotgun, whether or not it was the murder weapon, he surely erred. But it is impossible to tell from the record in this case precisely what was intended, or whether the judge confused the jury when he admitted the weapon. Although this highly prejudicial and irrelevant evidence was introduced, and although the prosecution did its best to lead the jury to believe that there was a relationship between the murder weapon and the shotgun in evidence, the fact that petitioner's counsel explained to the jury that the two weapons were not identical is, on the very closest balance, enough to warrant our finding that the jury was not improperly misled as to the nature of the evidence before it. Opinion of MARSHALL, J. 408 U.S. ing, I have not reiterated them here. Rather, I have emphasized those that seem to me to be particularly important and I have added several details that are omitted from the Court's opinion. Two interrelated defenses were raised against the charge of murder-alibi and misidentification. Petitioner's theory of the case was that he was not at the scene when the murder was committed and that those witnesses who testified that they saw him there were confusing him with someone else. Only two witnesses affirmatively asserted at trial that they saw the murder and that they could identify petitioner as the assailant. They were Patricia Hill, a waitress in the victim's bar, and Henley Powell, a customer. Aside from their testimony, the only other evidence introduced against petitioner related to statements that he allegedly made two days after the murder. There is a problem with the eyewitness testimony of Powell that did not become apparent until the postconviction hearing in the trial court. At trial he testified as follows: "The defendant (indicating) came into the tavern while I was at the table. I first saw him when he walked in the door with a shotgun. I was sitting at the table along the wall. I was facing where the bartender was standing and I also had a view of the man that walked in the door. I was looking to the west." Abs. 32. But at the post-conviction hearing it was discovered that police officers who had investigated the murder possessed a statement by one Charles Mayer, who had been sitting with Powell at a table in the bar, which contained a diagram indicating that Powell was seated in a direction opposite that indicated in his trial testi 786 Opinion of MARSHALL, J. mony. This diagram was never made available to defense counsel.2 Donald O'Brien, who had also been seated at Powell and Mayer's table, testified at trial and contradicted the testimony of both Powell and Patricia Hill. Although O'Brien admitted that he did not actually see the shooting because his back was to the bar, he was certain that petitioner was not the man who had been ejected from the victim's bar only an hour before the killing. O'Brien's testimony greatly undercut the apparent retaliatory motive that the prosecution attributed to petitioner.3 2 It is true, as the Court states, that following the shooting Powell followed the assailant into the street, but it is also true that he never got closer than 50 to 60 feet of the murderer. Abs. 32. The strength of his testimony lay in the alleged opportunity he had for close observation of the murderer while the crime was committed. Footnote 6 of the Court's opinion implies that during the trial the prosecution turned over Mayer's diagram to defense counsel. But there is absolutely no support for this implication in the record. While it is true that the diagram was drawn on the back of the original statement given by Mayer to the police, there is nothing to indicate that it was ever recopied and made a part of any reproductions of Mayer's statement. All indications are that it was not reproduced. At the post-conviction hearing the following testimony was adduced: the police officer who aided the prosecution at trial indicated that he had the original diagram in his file, Abs. 244-249; the two lawyers who had represented petitioner at trial both swore that they were given only Mayer's statement, not his diagram, Abs. 307, 328; and the prosecutor testified that he did not know for sure whether he gave the diagram to defense counsel, but that it was certain that he did not supply the diagram if it was not in his file. Abs. 324. Since the diagram was in the police officer's file, not the prosecutor's, it is clear that it was never made available to defense counsel, even though the prosecutor was aware of its contents. See infra, at 809. 3 The Court asserts that O'Brien may have been drunk. His testimony at trial made it clear beyond doubt that when the victim ejected the man alleged to be the petitioner from the bar, this wit Opinion of MARSHALL, J. 408 U.S. Because of the contradictory testimony of those persons who were present at the scene of the murder, the statements allegedly made by the petitioner after the crime were crucial to the prosecution's case. The key prosecution witness in this regard was Virgle Sanders. He testified that two days after the murder he was in the Ponderosa Tavern, that petitioner (whom he knew as "Slick") was there also, and that petitioner said "[s]omething about it's season or open season on bartenders or something like that." Abs. 44. The bartender also testified that he recognized petitioner as being present at the same time as Sanders. And the owner of the tavern stated that he gave petitioner and petitioner's friend a short ride in his automobile, at the end of which the friend mentioned something about "trouble with the bartender." Abs. 52. After his trial and conviction petitioner learned that five days after the murder, Sanders gave a statement to the police in which he said that he had met "Slick" for the first time about six months before he spoke to him in the Ponderosa Tavern. As the Court notes, it would have been impossible for Sanders to have met the petitioner at the time specified, because petitioner was in federal prison at that time. At the post-conviction hearing, Sanders said that he was not positive when he first met the man known as "Slick," but that he definitely knew it was before Christmas 1961. Petitioner was not released from federal custody until March 1962. When confronted with this fact, Sanders indicated that it was impossible that petitioner was the man with whom he had spoken in the Ponderosa Tavern. Abs. 296. Sanders' trial identification was further impeached at the post-trial hearing by testimony that on ness was perfectly sober. Later, especially after the killing, the witness drank heavily and became intoxicated. No one contradicted this at trial. 786 Opinion of MARSHALL, J. the day of trial he told police officers that petitioner was approximately 30 or 40 pounds lighter than he remembered "Slick" being. Abs. 294. Sanders' testimony that petitioner and "Slick" were not one and the same was corroborated at the hearing. The reason that Sanders could remember the first time that he had met "Slick" was that "Slick" had been involved in a scuffle with one William Thompson. Thompson testified at the hearing that he remembered the altercation, that he knew "Slick," that prior to the trial he had told police officers that petitioner was not "Slick," and that he remained certain that petitioner and "Slick" were different people. Finally, Sanders' testimony was corroborated by Delbert Jones, the owner of the tavern where "Slick" and Thompson scuffled. Jones testified that he was certain that petitioner was not the man known as "Slick." The fact is that Thompson and Jones were both familiar with one James E. Watts, whom they knew as "Slick," and who looked very much like the petitioner. The record makes clear that the police suspected Watts as the murderer and assigned a lieutenant to search for him. A raid of Jones' bar was even made in the hope of finding this suspect. Sanders' testimony at the post-conviction hearing indicates that it was Watts who bragged about the murder, not petitioner. It is true that the bartender and the owner of the Ponderosa Tavern testified at trial that it was petitioner who was in the bar with Sanders, but the bartender had never seen "Slick" before, and the owner was drinking the entire afternoon. Furthermore, the fact remains that petitioner and Watts look very much alike. Petitioner urges that when the State did not reveal to him Sanders' statement about meeting "Slick" at an earlier time and the corroborative statements of |