[EDITORS' NOTE: THIS PAGE CONTAINS HEADNOTES. HEADNOTES ARE NOT AN OFFICIAL PRODUCT OF THE COURT, THEREFORE THEY ARE NOT DISPLAYED.] *Page 667

On February 27, 1986, James Lewis Martin, Jr., was convicted in Montgomery County of murder during a robbery, in violation of §13A-5-40(a)(2), Ala. Code 1975. The prosecution presented evidence indicating that Martin and his codefendant, Edward Monteith, shot and killed Allen Powell and stole his wallet while Powell was test driving a vehicle with Kathy Ellison. The jury recommended that Martin receive the death penalty, and the circuit court sentenced Martin to death. This Court affirmed Martin's conviction and sentence, Martin v. State,548 So.2d 488 (Ala.Crim.App. 1988), affirmed, 548 So.2d 496 (Ala. 1989). The United States Supreme Court denied certiorari review. Martin v.Alabama, 493 U.S. 970 (1989).

Martin filed a petition pursuant to Rule 32, Ala.R.App.P., on September 28, 1990. Martin thereafter filed four amended petitions and conducted extensive discovery. On September 8-10, 1998, the Montgomery Circuit Court held an evidentiary hearing on the postconviction claims, except for the ineffective assistance-of-counsel claims, which Martin had withdrawn. On May 24, 1999, the circuit court entered an order denying Martin relief. The trial court rejected each of the claims Martin raised in his postconviction petitions. This appeal follows.

Martin contends that the trial court erred to reversal when it denied him relief on his Brady v. Maryland, 373 U.S. 83 (1963), claims. Martin argues to this Court, as he did in the court below, that, while conducting discovery during the postconviction proceedings, he found that the prosecution had withheld exculpatory evidence he was entitled to receive and that, therefore, he is entitled to postconviction relief and to a new trial. Martin argues that the prosecution withheld several key pieces of evidence. First, he argues that the prosecution withheld from him the fact that witness John Sims was hypnotized when he gave a statement describing the vehicle he saw and its occupants. *Page 668 Sims was the only eyewitness who identified Martin as one of the men he saw driving in the area at the time of the shooting. Second, Martin contends that the prosecution withheld information that Kathy Ellison had tentatively identified someone other than Martin in a lineup, and that she assisted the police in creating a composite drawing of the killer. Third, Martin argues that the prosecution failed to provide him with fingerprint evidence. He contends that he was not informed that a fingerprint on a receipt with information from the victim's account and generated by an automatic teller machine (ATM) recovered near the victim's wallet the day after the shooting did not match either Martin's or his codefendant's prints. Martin argues also that prints taken from the scope that could fit on the murder weapon, a rifle that was stolen approximately one month before the murder, did not match Martin's or his codefendant's. Fourth, Martin argues that he was not provided with a "highly suggestive photo array" shown to witness Stanley Wright. The array included photographs of Martin's vehicle and others, and Wright at trial identified the photograph of Martin's car as the one he had seen on the day of the murder. Fifth, Martin claims that the prosecution suppressed evidence showing that his former cell mate, Robert Elliot, who testified at trial that Martin had confessed to him that he had accidentally shot and killed Allen Powell, had been promised benefits in exchange for his testimony against Martin. Finally, Martin contends that the prosecution failed to timely disclose to him a composite created by Kathy Ellison, which, Martin says, bears little resemblance to him.

Martin presented these claims to the circuit court at the postconviction proceedings. In its order denying relief, the circuit court found that the various items of evidence had not been suppressed, or that if any of them were suppressed, the items were not material and, therefore, any failure to disclose them to Martin did not necessitate postconviction relief.

In Brady v. Maryland, 373 U.S. 83 (1963), the United States Supreme Court held that suppression of evidence favorable to the accused upon request violates due process when the undisclosed evidence is material either to guilt or to sentencing. Evidence is material if there is a reasonable probability that, if it had been disclosed to the defense, the result of the proceeding would have been different. United States v.Bagley, 473 U.S. 667, 682 (1985). The United States Supreme Court has explained that a "reasonable probability" of a different result arises when the suppression "undermines the confidence in the outcome of the trial." Kyles v. Whitley, 514 U.S. 419 (1995). A reviewing court must evaluate all of the withheld evidence together when determining whether it is material. Id. The prosecution is required to disclose impeachment evidence, in addition to evidence that is exculpatory. Bagley,473 U.S. at 676.

The State argues to this Court that the circuit court did not err in denying the Brady claim, because, it says, the production of the evidence would not have affected the outcome of either phase of Martin's trial.¹ Having reviewed the entire record *Page 669 in light of the relevant legal principles regarding the alleged suppression of Brady material, we must disagree with the State. Accordingly we reverse the circuit court's denial of the petition for postconviction relief.

A. Evidence at TrialDetermination of the significance of allegedly withheld evidence requires an understanding of the evidence that was presented at trial. The State's case was entirely circumstantial. Although Kathy Ellison was an eyewitness to the murder, she was unable to identify Martin as the perpetrator. John Sims, who lived near the site of the murder, testified that he saw two men drive past his house before he heard shots fired. Although he was unable in any of his pretrial statements to identify Martin as the driver of the car, he identified Martin at trial as the driver. He acknowledged that, before trial, he had seen photographs of Martin on television and had heard Martin identified as the suspect. Stanley Wright identified at trial a photograph of Martin's car as the one he had seen near the site of the murder. The prosecution elicited testimony that Martin had been seen in possession of a scope that had been stolen along with a rifle approximately one month before the murder. Ballistics evidence revealed that the bullet that killed the victim had been fired from the stolen rifle. Martin's fingerprints were not found on the rifle, and no one testified that Martin had ever been seen with the rifle itself. Robert Elliot, an inmate who shared a cell with Martin for approximately one month before trial, testified that Martin confessed to him his involvement in the crime. Elliot admitted that he hoped his testimony would result in favorable treatment from prison officials, but denied having received any promise of consideration in exchange for his testimony.

Each piece of allegedly withheld evidence is examined below, in light of the evidence actually presented at trial.

B. Witness John Sims's Statement Under HypnosisMartin contends that the State violated Brady because it failed to disclose to him a statement of witness John Sims that was given while Sims was hypnotized. Sims was the only witness at trial who identified Martin as one of the men he had seen near the site of the murder. He had been unable to identify Martin in any of his pretrial statements, but he did identify Martin at trial, much to Martin's surprise. Martin argues, therefore, that evidence indicating that Sims had given a statement while he was hypnotized would have rendered his in-court identification of Martin inadmissible or, at a minimum, would have impeached the credibility of Sims's in-court identification. The circuit court found the undisclosed evidence not material, because, it reasoned, the hypnosis did nothing to improve Sims's recall or his ability to identify Martin. Rather, the circuit court found that Sims was unable to identify Martin until shortly before trial, after he had seen Martin on television the second time, when Martin was depicted with short hair. Moreover, the trial court observed, the undisclosed statement did not differ significantly from Sims's other statements, which Martin did receive before and during trial. We disagree with the circuit court's analysis and its conclusions.

The State acknowledges that "the primary importance of Sims's testimony was that he identified Martin as the driver of the car he saw in his neighborhood shortly before the murder." (State's brief at p. 11.) The State concedes that at no time during trial were defense counsel informed that Sims had undergone hypnosis, nor were they provided with the *Page 670 transcript of the statement Sims made while under hypnosis. (State's brief at p. 11 n. 3.) Finally, the State acknowledges that law enforcement knew of the hypnosis² and that postconviction counsel discovered the statement in the prosecutor's files while conducting discovery in this proceeding. Id. The State maintains, however, that Martin did not prove that the prosecutors had knowledge of the statement before trial, and that Martin's assertion that the prosecutors had engaged in deliberate misconduct is "paranoid speculation." Id. We do not need to address this argument because the law clearly requires disclosure of exculpatory and impeachment evidence even if the evidence is "`known only to police investigators and not to the prosecutor.' [Kyles v. Whitley, 514 U.S. 419, 438 (1995)]. In order to comply withBrady, therefore, `the individual prosecutor has a duty to learn of any favorable evidence known to others acting on the government's behalf in this case, including the police.' Kyles, [514 U.S. at 437]." Stricklerv. Greene, 527 U.S. 263, 280-281 (1999).

The trial record reflects that, during the defense's cross-examination of John Sims, after he had surprised defense counsel by testifying that Martin was the driver of the car he had seen before the murder, Sims referred to a statement he gave to the police from which a composite sketch was completed. (T. 621.³) Trial counsel had not been provided with this statement, so they requested it as soon as they learned of it during Sims's cross-examination. The prosecution provided Martin with a statement Sims gave on June 11, 1985, and which had not previously been disclosed to Martin. Sims's June 11 statement, however, contains no mention of a composite. In fact, it was not until the day after he gave that statement, on June 12, that Sims cooperated in the creation of a composite sketch of the driver he saw. (C.R. 3169.⁴) On June 14, 1985, Sims gave a statement at the sheriff's department to Dr. Tom Michel under hypnosis, and the composite was mentioned in this statement. (C.R. 3133.) The June 14, 1985, statement Sims gave while under hypnosis was never provided to trial counsel.

In this Court's opinion, the nondisclosure of the fact that, at the request of law enforcement, Sims underwent hypnosis before he gave one of his statements is a greater violation than the nondisclosure of the statement itself. Regardless of which is the greater violation, however, both the State's failure to disclose to Martin that its key witness was hypnotized and its failure to disclose the statement Sims gave while under hypnosis constitute significant violations of the spirit and the letter of the law regarding disclosure of exculpatory and impeachment evidence.⁵ *Page 671

1. The Nondisclosure of the HypnosisMartin contends that if the prosecution had disclosed that its primary witness had given a statement under hypnosis, he would have challenged Sims's in-court identification and argued that it was unreliable and inadmissible because of the previous hypnosis. He further argues that nondisclosure of the fact that Sims had been hypnotized deprived the jurors of their right to fully evaluate the credibility of Sims's in-court identification and deprived him of valuable impeachment evidence. The State contends that the hypnosis was "unproductive" because, it argues, the statement Sims gave under hypnosis was similar to the June 11 statement he made, and "therefore, the failure to inform the trial court or the jury that [hypnosis] was attempted does not undermine the confidence in the outcome of the trial." (State's brief at p. 15.) We disagree.

In Rock v. Arkansas, 483 U.S. 44 (1987), the United States Supreme Court reviewed Arkansas's per se rule excluding a defendant's hypnotically refreshed testimony and limiting testimony to matters the defendant could prove were remembered before hypnosis. The Court found the per se rule unconstitutional. 483 U.S. at 62. In reaching its decision, the Court discussed the value and risks associated with hypnosis:

"Responses of individuals to hypnosis vary greatly. The popular belief that hypnosis guarantees the accuracy of recall is as yet without established foundation and, in fact, hypnosis often has no effect at all on memory. The most common response to hypnosis, however, appears to be an increase in both correct and incorrect recollection. Three general characteristics of hypnosis may lead to the introduction of inaccurate memories: the subject becomes `suggestible' and may try to please the hypnotist with answers the subject thinks will be met with approval; the subject is likely to `confabulate,' that is, to fill in details from the imagination in order to make an answer more coherent and complete; and, the subject experiences `memory hardening,' which gives him great confidence in both true and false memories, making effective cross-examination more difficult. . . .

"The inaccuracies the process introduces can be reduced, although perhaps not eliminated, by the use of procedural safeguards. One set of suggested guidelines calls for hypnosis to be performed only by a psychologist or psychiatrist with special training in its use and who is independent of the investigation. These procedures reduce the possibility that biases will be communicated to the hypersuggestive subject by the hypnotist. Suggestion will be less likely also if *Page 672 the hypnosis is conducted in a neutral setting with no one present but the hypnotist and the subject. . . .

"The more traditional means of assessing accuracy of testimony also remain applicable in the case of a previously hypnotized defendant. Certain information recalled as a result of hypnosis may be verified as highly accurate by corroborating evidence. Cross-examination, even in the face of a confident defendant, is an effective tool for revealing inconsistencies. Moreover, a jury can be educated to the risks of hypnosis through expert testimony and cautionary instructions."

483 U.S. at 59-61 (citations and footnotes omitted).

By failing to disclose that Sims underwent hypnosis at the request of law enforcement, the trial court and the jury were completely deprived of any consideration of the risks and effects of hypnosis discussed in Rockv. Arkansas and were precluded from considering this important factor in evaluating Sims's vital in-court identification of Martin. Martin was deprived of his right to challenge the in-court identification on this basis and was precluded from cross-examining Sims about the substance of the statement and its inconsistencies with other statements.

This Court, too, has recognized the importance of considering the possible impact of hypnosis on a witness's identification of an appellant. In Chamblee v. State, 527 So.2d 173 (Ala.Crim.App. 1988), the appellant challenged the victim's in-court identification of him because the victim was hypnotized after the crimes had been committed. He also argued that the victim had been given post-hypnotic suggestions that so hardened her recollection that he was denied his right of cross-examination. We found that, because the record disclosed that the victim had observed the defendant for hours during the commission of the crimes, and because she provided detailed descriptions to the officers immediately after the crimes, her identification of the appellant in a subsequent lineup and in court resulted entirely from her contact with him rather than from the effects of the hypnosis. 527 So.2d at 176. We held that the fact that a witness had been hypnotized before testifying did not, per se, disqualify the witness. We stated:

"In considering hypnotically enhanced testimony, a court must conduct a balanced inquiry to determine if the testimony had a basis that was independent of the dangers associated with hypnosis. The inquiry should determine whether the witness's memory and ability to testify from it were distorted by the earlier hypnosis. The admissibility of such testimony is to be evaluated on a case-by-case basis, and the probative value of the testimony is to be weighed against its possible prejudicial effect."

527 So.2d at 177 (citations omitted).

We held in Chamblee that the appellant was not denied his right to confrontation in that case and noted:

"The jury was aware of the hypnotic session, heard the testimony of the expert witnesses from both sides on the subject, and was in a position to fairly evaluate the effect of the hypnotic session on the testimony of the witness. . . . We find in the case sub judice that the hypnotic session did not impair defense counsel's ability to cross-examine the victim, because the manner in which she testified, detailing the facts and circumstances surrounding the crimes, including the identification of appellant, indicates that her recollection and judgment were unimpaired by the hypnosis."

527 So.2d at 177.

A similar crucial evaluation of the effect of hypnosis on Sims's in-court identification of Martin was entirely foreclosed by *Page 673 the State's failure to reveal that Sims had been hypnotized. Martin was unable to cite Sims's hypnosis as a basis for a motion to suppress the in-court identification. At the Rule 32 hearing, trial counsel testified that, if he had been aware that Sims had been hypnotized, he would have considered hiring an expert in hypnosis and witness identification. (R. 296-97⁶.) Information about the hypnosis should have been available to the trial court so that it could fully and fairly evaluate the reliability of Sims's testimony, particularly his in-court identification of Martin. We note that the transcript of the statement discovered in these proceedings does not contain the induction procedure used by the hypnotist. Therefore, the court was unable to evaluate whether the procedure used by the hypnotist was overly suggestive.

Not only was the trial court unfairly limited by the nondisclosure, Martin's jury was also unfairly limited. The jury was aware that Sims had not identified Martin until he testified at trial. Sims acknowledged that he had seen Martin's photographs in circumstances in which Martin was identified as a suspect, but asserted that his recently found ability to identify Martin was unrelated to those viewings. However, the jury should have also been informed that, at the request of law enforcement, Sims had been questioned under hypnosis; it should have been provided with the details of the statement Sims gave under hypnosis so that the jurors could fairly evaluate the credibility of his trial testimony; and it should have been provided with any expert testimony the parties would have offered regarding the effects of hypnosis on recall and memory.

Finally, and most importantly, Martin was foreclosed from cross-examining Sims about the statement he gave under hypnosis and about the effect the hypnosis had on his recall. Because the prosecution failed to disclose the hypnotically induced statement, Martin was unable to question Sims about his inability, even under hypnosis, to identify the men he saw in the car. He was unable to question Sims about discrepancies between that statement and others Sims gave. Because the prosecution failed to disclose the hypnotized statement, Martin had no reason to present expert witnesses who could have evaluated the hypnotic session and explained to the jury the effects of hypnosis on a witness's memory and recall. In sum, Martin was completely deprived of the ability to challenge Sims's in-court identification on the basis that it was affected by hypnosis.

Sims's identification of Martin was of paramount importance to the State's case. As the Alabama Supreme Court stated when it reversed a judgment for a Brady violation, "[T]he failure to disclose the evidence totally prevented [the appellant's] counsel from preparing portions of the defense." Ex parte Williams, 642 So.2d 391, 393 (Ala. 1993). That defense counsel cross-examined Sims as vigorously as they could with the information they had does not eliminate the error caused by the prosecution's failure to disclose this evidence. Id. The undisclosed evidence rendered Martin much like a fighter with one hand tied behind his back — the fact that he was able to land a few punches in cross-examination with one fist did not make the match a fair one. Only by cross-examining Sims about the hypnosis and its effects on his memory and identification would Martin have received all that he was due underBrady.

2. Discrepancies not DisclosedThe State contends that the nondisclosure of the statement given under hypnosis does not constitute a Brady violation *Page 674 because, it says, the hypnosis was not productive and it yielded no new information. The record, however, reveals otherwise.

In the June 11, 1985, statement the prosecution provided to Martin at trial, Sims stated that he had previously told the police that the suspects drove a Chevrolet automobile, a 1965 or 1966 model; he thought it was a Malibu. (C.R. 3140-42.) He said he could not remember whether the car had wheel covers. (C.R. 3140.) He said he believed the car was a two-door model, but he could not remember. (C.R. 3141.) Sims said he did not remember if the car had a license plate. (C.R. 3142.) Sims also said, "And I know the driver didn't have a shirt on and I didn't think the other guy did." (C.R. 3137.)

In the undisclosed June 14, 1985, statement made under hypnosis, Sims said he thought the car had four doors and he did not think it had wheel covers. (C.R. 3122-23.) Although he first said he did not see the tag that closely and he wished he had looked, Sims then said that it seemed like a white tag with a red border, "Probably is a State tag, yeah." (C.R. 3133.) Also Sims said he thought the driver was wearing a white shirt, like an undershirt, and the passenger was not wearing a shirt. (C.R. 3134-35.) He then said that perhaps neither man was wearing a shirt. (C.R. 3135.)

These discrepancies, combined with the fact that Sims was hypnotized on June 14, would have been the proper subject for cross-examination by Martin. Because Sims had already acknowledged that he had seen photographs of Martin before trial, and that the viewings might have influenced his identification, revelation of these discrepancies would have further called into question the reliability of the in-court identification. In addition to the discrepancies identified here, we note that the hypnotist told Sims that, when Sims later saw the composite he had helped create previously, if he saw or felt something that was not quite right, he would be able to pick it out. (C.R. 3133.) Thus, the hypnotist suggested that Sims's memory of the suspect might change in the future and that he should have faith in his feeling about any later changes. Had the prosecution disclosed this statement, the court, the jury, and the defense would have been aware of this suggestion and would have been able to evaluate it in terms of the reliability of Sims's surprising in-court identification of Martin as the driver of the car.

We recognize that hypnosis can be a useful forensic tool. However, we are also aware that the hypnosis of a witness carries identifiable risks. The prosecution's failure to disclose to Martin that the only eyewitness who identified him had undergone hypnosis, and the prosecution's failure to provide the statement the witness made under hypnosis, constituted a significant deprivation to Martin, to the court, and to the jury. Without this information, the court could not fairly evaluate the admissibility of the in-court identification. Without this information, Martin was unable to prepare a significant defense based on a challenge to the post-hypnotic testimony. Without this information, the jury was unable to knowledgeably evaluate the reliability of the in-court identification, if the trial court determined it was admissible. Based on the foregoing, we are forced to conclude that the prosecution violatedBrady principles because the undisclosed evidence constituted impeachment evidence and because Martin was prejudiced by the nondisclosure.

C. Kathy Ellison's Tentative Identification of a Third Person.Martin next argues that the trial court erred when it denied postconviction relief because the State failed to disclose the fact *Page 675 that Kathy Ellison had tentatively identified someone other than Martin at a post-arrest lineup as the man she had seen. The court determined that this was not suppression of evidence, because defense counsel could have asked Ellison before trial whether she had identified someone else. The court stated, alternatively, that, even if the prosecution suppressed any information, Ellison's deposition indicates that any identification she made was equivocal at best. (C.R. 2148-49.)

The trial court's factual findings are not supported by the record, and we disagree with its conclusions of law. The record reflects that, on August 17, 1993, Kathy Ellison was deposed. Martin's counsel asked her about viewing the lineup, and she acknowledged that she had been unable to "positively pick out the person who shot Allen Powell." (C.R. 2221-22.) She said that she had tentatively pointed to someone else as the shooter. (C.R. 2222.) When asked if a member of law enforcement had thereafter told her that she had picked "the wrong person," Ellison testified, "I knew I picked out the wrong person, but I don't know how I heard or who told me." (C.R. 2223.) When counsel sought to confirm that Ellison learned from a member of law enforcement that the person she tentatively identified was not "the right person," Ellison acknowledged that that was correct. (C.R. 2223.) Review of more than the portion of the deposition quoted by the trial court reveals that Ellison identified someone other than Martin as the man who shot Allen Powell. This information was not disclosed to Martin. We find that this constituted another violation of Brady.

At Martin's trial, Ellison testified that she had been asked to identify the two perpetrators but had never been able to identify either of them. She explained that her inability to identify them was due to the fact that her focus during the crime had been on the gun, not on the perpetrators. (T. 572, 586.) The jury was not told that, soon after the crime, Ellison apparently believed that she had focused enough on the perpetrators because she selected someone other than Martin from a police lineup. The jury was left with the impression that Ellison had not been able to identify anyone from the lineup, while the facts were otherwise. The fact that Ellison never identified Martin represented only a portion of the truth, and the facts not disclosed to Martin constituted exculpatory evidence.

"Brady requires the prosecution to produce evidence that someone else may have committed the crime." Nicks v. State, 783 So.2d 895, 916 (Ala.Crim.App. 1999), quoting Jarrell v. Balkcom, 735 F.2d 1242, 1258 (11th Cir. 1984). Nondisclosure of exculpatory evidence bearing on the identity of the perpetrator has often led courts to hold that the principles of Brady were violated. See, e.g., Patton v. State,530 So.2d 886, 890 (Ala.Crim.App. 1988) (a telephone call to police from alleged eyewitness who implicated someone other than appellant was not disclosed); Hall v. State, 625 So.2d 1162, 1164 (Ala.Crim.App.) (remand ordered because crime-scene prints were not processed), aff'd on returnto remand, 625 So.2d 1165 (Ala.Crim.App. 1993); Jefferson v. State,645 So.2d 313 (Ala.Crim.App. 1994) (pretrial statements of witnesses implicating someone other than appellant or containing details inconsistent with witnesses' trial testimony).

While Ellison did not identify Martin at trial, this alone does not preclude a finding of error under the facts of this case. Because the State's witness sought to explain her inability to identify Martin as the killer by claiming that her attention was focused on the gun, the State made Ellison's lack of identification an issue. The fact that *Page 676 Ellison had previously viewed a lineup and had selected someone other than Martin was material evidence, and it should have been disclosed to Martin. This evidence is particularly important, because, in a pretrial statement Ellison gave to the police the day after the murder, she said, "I just think that if I saw the guy that shot him I, I probably could recognize him. Because, I mean, there's just something in the back of my mind that keeps popping up. Then the more I think about it the more . . . he gets a little clearer." (S.R. 49.⁷) If Martin had been aware that Ellison had identified someone else in the lineup, and especially if he had been aware that this misidentification came after she told police that she was reasonably certain that she would recognize the killer if she saw him again, Martin could have cross-examined Ellison about her previous statement and about the misidentification. This information suggested that someone else might have committed the crime, and it should have been disclosed.

Combined with the evidence about Sims's statement under hypnosis, we believe that had this evidence been disclosed, there is a reasonable probability that the result would have been different.

D. Fingerprint EvidenceMartin also contends that the circuit court erred when it denied relief based on the State's failure to disclose fingerprint evidence. Martin claims that the presence of unidentified fingerprints on evidence related to the murder was exculpatory because, he says, it suggested someone else had committed the crime, and the evidence should have been disclosed to him. The trial court found that Martin had failed to establish that the prosecution had suppressed the information. The court found, in the alternative, that even if the prosecution had suppressed the fingerprint reports, Martin would not be entitled to relief because his trial counsel should have filed a motion to compel production of the reports.⁸ Finally, the court found that Martin failed to establish that the evidence was material. Martin contends that a forensic report indicating that an unidentified print had been recovered from the ATM receipt discarded after Powell was murdered suggested the existence of another perpetrator. Martin contends that the State's failure to disclose this information prevented him from arguing to the jury that someone else was responsible for Powell's death. Martin also contends that the court should have granted him relief because the State failed to disclose that prints other than his were found on a rifle scope recovered in connection with this crime. The prosecution presented testimony at trial that Martin had a scope in his possession approximately one month before the shooting. A rifle and the scope attached to it had been stolen before trial, and ballistics evidence established that the shot that killed Powell had been fired from the stolen rifle. Martin argues that, if the State had disclosed the fact that unidentified prints had been lifted from the scope, he would have been able to use the evidence to argue to the jury that Powell was murdered by someone other than him. The evidence was exculpatory and material, he asserts, because it weakened the link between Martin and the murder weapon. The State contends that Martin certainly knew that his prints were not on the rifle scope and that he could have argued that *Page 677 this lack of evidence created a reasonable doubt. It argues, further, that there was no proof that the prints on the scope matched the prints on the ATM receipt, and that, if this proof had existed, then Martin's claim that an unidentified person had committed the murder would have been stronger.⁹

Again, we are forced to conclude that the trial court's findings are not supported by the evidence presented or by the relevant law. Martin's trial counsel testified at the Rule 32 hearing that they did not receive any fingerprint reports until, during trial, they became aware that some reports existed. The court ordered the prosecution to disclose all fingerprint reports, but trial counsel testified at the Rule 32 hearing that the reports regarding the unidentified prints on the rifle scope and on the ATM receipt were in the police files but were never disclosed at trial. (R. 301-02, 359.) The circuit court's finding that Martin failed to prove any suppression is simply not supported by the Rule 32 record. The court's assertion that, if the prosecution had failed to disclose these particular fingerprint reports, defense counsel should have moved to compel the prosecution to disclose them is not well taken. Martin's counsel could not be expected to know that the prosecution disclosed only some of the reports, especially after the trial court ordered the prosecution to disclose all fingerprint reports and the prosecution agreed to do so.

We further find that the circuit court erred when it determined that Martin had failed to establish that the undisclosed fingerprint evidence was material. The fatal shots were fired from a rifle that had been stolen approximately one month before the murder. The only piece of demonstrative evidence linking Martin to the murder was the scope attached to the rifle when it was stolen. A former girlfriend of Martin's codefendant testified at trial that she saw Martin with a rifle scope a day or two before the murder. (T. 488-89.) The fact that three latent prints were recovered from the scope, but that they did not match Martin's or his codefendant's, was material because it weakened the connection the State attempted to make between Martin, the murder weapon, and the fatal shooting.¹⁰ Moreover, the evidence suggested that someone other than Martin had been in possession of the scope. The evidence should have been provided to Martin, and the failure to do so violated Brady principles.See, e.g., Hamilton v. State, 677 So.2d 1254, 1259-61 (Ala.Crim.App. 1996) (reversible error found when undisclosed evidence weakened a significant link in the prosecution's evidence).

On the day after the shooting, credit cards and an ATM receipt belonging to the victim were found on the roadside several miles from the murder scene. The evidence suggested that the perpetrators had taken the items after the shooting and had discarded them on the roadside. A latent print was recovered from the ATM receipt, but a forensic report indicated that the print did not match Martin's or his codefendant's print, nor did it match the prints *Page 678 taken from Powell, his wife, or the investigator who collected the evidence. (C.R. 3109, 3111.) In making its determination that the evidence was not material, the circuit court noted that the receipt was apparently found by a member of a road crew, and it was unknown how many people had handled the receipt. (C.R. 2151.) While this is certainly true, we believe that this is precisely the type of factual determinationa jury is to make in a trial, after it is presented with all of the relevant facts. Martin was deprived of the opportunity to even present to the jury the evidence indicating that an unidentified print was found on an item apparently stolen from the victim during the robbery-murder. We believe that evidence of the unidentified print on the receipt should have been disclosed to Martin, who then could have argued that it constituted evidence that a third person committed the crime. Again, nondisclosure of this fingerprint evidence constituted a violation ofBrady principles.

The prosecution's failure to disclose evidence that prints other than Martin's and his codefendant's were recovered from items related to the crime strengthened the State's case unfairly. This is particularly true with regard to the evidence that three prints, none of which was the defendant's, were lifted from the rifle scope. The fingerprint evidence was material, and it should have been disclosed to Martin. When this nondisclosure is combined with the others discussed in this Court's opinion, we are compelled to find that the prosecution's evidentiary suppression undermines the confidence in the outcome of Martin's trial.Kyles v. Whitley, 514 U.S. 419, 434 (1995).

E. Identification of the VehicleMartin also argues that the trial court erred when it determined that a photographic array of automobiles shown to witness Stanley Wright and Wright's pretrial statement were not material evidence. Wright was the only witness who, at trial, identified Martin's car as the one he saw near the scene of the murder moments before the shooting occurred. (T. 553-54.)

Stanley Wright had been working in a yard near the murder scene. Approximately two weeks after the murder, Wright gave a statement to the police. (S.R. 52.) In that statement, Wright described the car as a dark brown 1974 Chevrolet Impala. The statement was not disclosed to Martin. (R. 305.) Approximately two months later, after Martin had been arrested, Wright was shown photographs of six Chevrolet automobiles. (C.R. 3182-90.) Wright gave a taped statement and indicated that, from the photographs, he had identified a 1967 Chevrolet Impala as the car he had seen on the day of the murder. (S.R. 51.) Martin was not provided with a copy of that statement, and counsel testified at the Rule 32 hearing that he could not recall having seen the photographic array displayed to Wright. (R. 310.) At Martin's trial, Wright testified that, on the day of the murder, he had seen a 1968 Impala in the neighborhood. He said the car he saw on the day of the murder was the same as that depicted in State's Exhibits 1-4, which were photographs of Martin's car. (T. 552-53.)

The circuit court determined that the prosecution's failure to disclose Wright's pretrial statement, in which he identified the car as a "1974 Impala," caused no prejudice to Martin. The court found that Wright had explained at trial that he was able to identify the car as a 1968 Impala because his neighbor owned the same type of car. (C.R. 2159.) Martin correctly argues to this Court that cross-examination of Wright regarding his earlier identification of the car as being a much later model would have been important *Page 679 impeachment evidence. Contrary to the determination of the circuit court, we believe that Wright's explanation regarding his neighbor owning a 1968 Impala did not sufficiently explain Wright's earlier identification of the car he saw as a 1974 Impala. To the contrary, if Wright was familiar with his neighbor's car, it would have been more reasonable to assume that Wright would have correctly identified the car he saw on the day of the murder. The statement should have been disclosed to Martin, so that he could have impeached Wright's testimony.

We find the use of the photographic array even more troubling. The circuit court found that the array was not unduly suggestive, that all of the cars were late 1960s model Chevrolets, and that the outcome of the trial would not have been different if Martin had been able to impeach Wright with the array. (C.R. 2154.) Our viewing of the photographs in Defendant's Exhibit 20 and our reading of the testimony about the photographs compels a different determination. Although Wright described the car he saw as brown, the first car in the array is white, and the sixth car is orange. Of the six cars depicted in the photographs, Martin's car is the only 4-door Impala. The remaining cars are 2-door Chevelle Malibus. (R. 430-31.) Martin's car is the only one with the front wheels on blocks, and it is the only car with a marked patrol car immediately behind it in the photograph. These factors caused Martin's car to stand out so much from the others in the array that it was unduly suggestive. If the prosecution had disclosed the array to Martin, he could have moved for suppression of Wright's in-court identification of Martin's car or, at a minimum, used the suggestive array to impeach Wright's identification of Martin's car.

Wright was the only witness at trial to identify Martin's car. Therefore, opportunities for impeachment of his testimony were crucial to Martin. If the evidence had been disclosed, Martin could have presented it to the trial court to aid in determining whether to allow Wright's testimony, and he could have presented it to the jury so it could evaluate the reliability of Wright's testimony. As was true with the nondisclosure of witness Sims's prior statement and the fact that Sims was hypnotized, these nondisclosures unfairly limited Martin's development of his defense on these extremely important matters. The confidence in the result reached in Martin's trial has been further undermined by the prosecution's failure to disclose evidence regarding Stanley Wright's identification of Martin's car.

F. Promises to Robert Todd ElliotMartin argues that the circuit court erred when it found that no promises of benefits were made to Robert Todd Elliot, a former cell mate of Martin's, other than the promise of a transfer to another prison, in exchange for his testimony against Martin. Elliot testified that while they were cell mates, Martin told him that he had shot Powell accidentally, that he told his brother, codefendant Edward Monteith, to dispose of the gun and the car, and that he (Martin) had been placed in a lineup but not identified. (T. 720-22, 731.)

Martin correctly argues that a prosecutor's failure to disclose the existence of an agreement with an informant witness violates Brady principles, because the agreement might provide evidence regarding the witness's motivation to testify. E.g., Ex parte Womack, 541 So.2d 47,61-62 (Ala. 1988). At Martin's trial, Elliot acknowledged that he hoped that in exchange for testifying he would be moved to a prison closer to his family, that he might get an early release on parole, and that he might have an easier time in prison. (T. *Page 680 727.) The record developed in the Rule 32 proceeding indicates that the only agreement reached between the prosecution and Elliot's lawyer in exchange for his testimony was that Elliot would be transferred to a prison closer to his family and that the prosecutor would inform the parole board of Elliot's cooperation. (R. 405-08.)

Elliot's trial attorney testified in this proceeding that he did all that he could do to gain favorable treatment for his client, but that he and the prosecutor did not come to any agreements other than that he would be transferred to a prison closer to his family. (R. 414-16.) Elliot had already entered a guilty plea and was outside the trial court's jurisdiction at the time he testified at Martin's trial. (R. 413.) Elliot was transferred to a prison closer to his family, but he did not receive an early release on parole. (R. 416.)

Although Martin argues that evidence regarding negotiations for reducing Elliot's sentence and that Elliot's expectations regarding a sentence reduction constituted material evidence that should have been disclosed, we agree with the State that the record does not support an inference that the prosecutor made promises that were not disclosed to Martin. Moreover, in his testimony in Martin's trial, Elliot acknowledged that he hoped to receive several benefits in exchange for his testimony against Martin. Elliot's testimony did not go unimpeached. An argument can be made that Elliot's testimony was impeached to a greater degree than might have been warranted, because, in the end, Elliot received only a transfer, instead of all of the benefits he testified that he hoped to receive.

G. Descriptions by Kathy EllisonMartin argues that the circuit court erred when it found that the composite sketch created with Kathy Ellison's input was produced at his trial. He also contends that the circuit court erred when it found that no Brady violation had occurred as a result of the prosecution's failure to disclose a "be on the lookout" (BOLO) containing Ellison's description of the killers.

Martin's attorney testified at the Rule 32 proceeding that the composite Ellison created was disclosed at trial. (R. 307.) Thus, there was no suppression of the composite.

With regard to the BOLO, the circuit court found that disclosure of the description would not have affected the outcome of the trial, because Ellison was unable to identify Martin at trial. (C.R. 2157.) Martin contends that the BOLO description differed so significantly from his actual appearance that it was material and exculpatory, for the same reasons as those given regarding the failure to disclose Ellison's selection of another person in the lineup. We agree with Martin, and find that the circuit court erred when it found that no Brady violation resulted when the prosecutor failed to disclose the BOLO. We find, for the reasons discussed in Part C, above, that the description in the BOLO constituted impeachment evidence that should have been disclosed. At trial, Ellison contended that she had never been able to identify the killers because her focus had been on the murder weapon. (T. 572, 586.) The fact that she described the killer in some detail after the murder, and that the description was dissimilar from Martin's appearance, were matters that might have affected the jury's assessment of Ellison's credibility. United States v. Bagley, 473 U.S. 667, 676 (1985). Therefore, the prosecution should have disclosed this evidence.

ConclusionAs the United States Supreme Court recognized decades ago:

"Society wins not only when criminals are convicted but when criminal trials are fair; our system of the administration of justice suffers when any accused is treated unfairly. . . . A prosecution that withholds evidence on demand of an accused which, if made available, would tend to exculpate him or reduce the *Page 681 penalty helps shape a trial that bears heavily on the defendant. That casts the prosecutor in the role of an architect of a proceeding that does not comport with the standards of justice."

Brady v. Maryland, 373 U.S. 83, 87-88 (1963).

James Martin's capital-murder conviction and death sentence followed a trial proceeding in which the prosecution failed to disclose, not just one, but several pieces of material evidence. The cumulative effect of these numerous nondisclosures was so serious that we are compelled to find that there is a reasonable probability that disclosure of the suppressed evidence would have produced a different result at Martin's trial. "The question is not whether the defendant would more likely than not have received a different verdict with the evidence, but whether in its absence he received a fair trial, understood as a trial resulting in a verdict worthy of confidence." Kyles v. Whitley, 514 U.S. 419, 434 (1995). The verdict here was so tainted by the prosecutorial nondisclosure of material evidence that it is not "worthy of confidence." Martin was not convicted or sentenced in accordance with fundamental due process.

"Unless, indeed, the adversary system of prosecution is to descend to a gladiatorial level unmitigated by any prosecutorial obligation for the sake of truth, the government simply cannot avoid responsibility for knowing when the suppression of evidence has come to portend such an effect on a trial's outcome as to destroy confidence in its result."

Kyles v. Whitley, 514 U.S. 419, 439 (1995).

For the foregoing reasons, we hold that the trial court abused its discretion when it denied Martin's postconviction petition. We now reverse the judgment of the Montgomery Circuit Court and remand the cause for a new trial.

REVERSED AND REMANDED.

McMILLAN, P.J., and SHAW, J., concur; BASCHAB, J., concurs in the result; WISE, J., dissents, without opinion.