A state-court jury convicted Michael Bigelow of kidnapping, felonious assault and arson for his alleged involvement in an attack on a woman in Toledo, Ohio on June 17, 1993. From his initial arrest to the present, Bigelow has insisted that he did not commit the crime and indeed could not have committed the crime because he was residing and working 150 miles away in Columbus, Ohio on the day of the assault. At each stage in the proceedings — in state court, in his state post-conviction proceedings, and now in his federal habeas corpus proceedings — Bigelow also has claimed that his court-appointed lawyer, Peter Rost, did not adequately investigate this alibi defense, most notably by failing to identify three witnesses who could have placed Bigelow in Columbus on the day of the assault.

In one sense, it is easier to sympathize with Rost than with Bigelow when it comes to this claim. Bigelow lived an itinerant life in Columbus; he did not remember exactly where he was in Columbus on the day of the crime; he did not fully communicate all possible leads to Rost and apparently did not inform him about his own letter-writing investigation efforts from prison; and Rost in fact did pursue many leads, none of which bore fruit. Until four days before Bigelow's trial, it is indeed difficult to second-guess Rost's efforts, frustrating as they were, to advance his client's defense.

On the fourth day before the commencement of the criminal trial, however, Vernon Greenlee, an employee of Orkin Pest Control, called Rost and told him that he could place Bigelow in Columbus on the day of the crime. (Greenlee's call was prompted by a letter that Bigelow had written to Orkin from prison.) Realizing the significance of this testimony, Rost subpoenaed Greenlee and one other Orkin employee to testify at the trial. The testimony was helpful because Greenlee identified Bigelow in court as the man he saw at the home of Gary Chasen in Columbus on June 17th, the day of the assault, but the testimony was vulnerable to impeachment because Greenlee worked at the house at issue on two consecutive days. In convicting Bigelow, the jury apparently was swayed by the two primary pieces of evidence submitted by the State — the testimony of the victim who was able to pick Bigelow out of a lineup (and identify him at trial) based on brief glances at him during the assault and the testimony of an individual who claimed to see Bigelow (from the back and side) running across a field away from the crime scene.

In rejecting Bigelow's ineffective-assistance-of-counsel claim, the state courts and federal district court focused primarily on whether an alibi witness contacted Rost during the week before trial and whether Rost failed to return the phone call. The state courts found as a matter of fact that Rost did not know about any other alibi witnesses before the trial. The district court properly respected this finding in view of the competing evidence on the issue and the rigorous requirements for rejecting such a finding under The Antiterrorism and Effective Death Penalty Act (AEDPA), Pub.L. No. 104-132, 110 Stat. 1214 (1996).

The problem with the district court's decision is that it did not address the other aspect of Bigelow's Sixth Amendment claim: Rost's failure to conduct any additional investigation after the sudden appearance of Greenlee four days before trial. While the State urges us to reject this alternative argument on our own, we refuse to do so in view of the seriousness of the claim. Once Greenlee appeared, Rost had ample reasons to re-commit himself to finding additional alibi witnesses in the Columbus area — whether by asking for a postponement of the trial, by hiring an investigator or by traveling to Columbus himself to talk firsthand to the other people that might have been working at the same house as Greenlee (and apparently Bigelow) on June 17th. Had Rost pursued any of these options, he likely would have identified three other witnesses, all of whom have since come forward to testify that they saw Bigelow in Columbus on the day of the attack and none of whom had a prior relationship with Bigelow (or any other reason to be untruthful).

Whether Rost's failure to take additional action after being contacted by Greenlee constituted ineffective assistance deserves consideration by the district court in the first instance and possibly an evidentiary hearing. As the United States Supreme Court first indicated in Strickland v. Washington, 466 U.S. 668, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984), and reaffirmed just recently in Wiggins v. Smith, 539 U.S. 510, 123 S. Ct. 2527, 156 L. Ed. 2d 471 (2003), the respect that attorneys' strategic decisions in a criminal trial will receive is proportionate to the extent of the investigation they in fact conducted. See Strickland, 466 U.S. at 691, 104 S. Ct. 2052 ("[S]trategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation."). As the case comes to us, there is no indication that Rost performed any further investigation after Greenlee came forward — even though his alibi testimony was sufficiently important that Rost put him on the stand virtually sight unseen. For these reasons and those elaborated below, we vacate the judgment of the district court, remand the case to the district court and allow it to consider in the first instance whether to grant the writ on the basis of this claim.

On the morning of June 17, 1993, Charlotte Schrier, a real estate agent, was sitting in her car behind an apartment complex in Toledo, Ohio, waiting for her next appointment. At some point she felt a tap on her left shoulder, and she heard a man's voice telling her not to move as he entered the back seat of her car. Although she could detect his presence in the back seat, she obeyed his commands to face forward and not turn around.

At some point, the man instructed her to start the car and proceed out of the complex. After Schrier drove a short distance, he asked her to pull over and light his cigarette. Schrier did as asked, then resumed driving. While she was driving, the man threatened her multiple times, saying he wanted to injure her physically and see her bleed.

At some point, the attacker asked Schrier to pull over again. This time, he got out of the car, opened her door, and pulled her out by the hair. He first instructed her to lean into the back seat, but then told her to get up again. Schrier stood up and faced the car, with the attacker behind and to her right. His hand suddenly swung down in front of her face, and she noticed that he was holding a razor blade. Pressing the blade into her hand, he told her that he wanted her to cut her own arm. When she hesitated, his arm swung again and he either cut or forced Charlotte to cut her arm with the blade. She then turned around to face him, kicked him in the groin and managed to escape. After Schrier fled, the assailant apparently set her car on fire.

Schrier gave a statement about the attack to Detective Kulakoski on the day after the incident. She described her attacker's clothing — white T-shirt, brown pants and tennis shoes — and his physical appearance — white male, late 30's to early 40's, 5'9" or 5'10", no facial hair or visible tattoos and a very close, almost Marine-like haircut. JA 654. After giving this statement, Schrier looked through photograph arrays and did so again on several other occasions, but she never recognized any of the men as her attacker. JA 611. An initial attempt by police to create a composite sketch of the assailant failed to produce a passable likeness. A police artist later attempted a free-hand drawing of the man based on Schrier's input, the end result of which looked much like her attacker, Schrier concluded, prompting police to distribute copies of the sketch to patrol officers and to local media on July 8, 1993.

The next day, police brought Bigelow in for questioning based on his resemblance to the man in the drawing. They photographed him, and included his picture in a photo array shown to Schrier. She pointed out Bigelow's photo to Detective Kulakoski but noted that she did not remember the deep lines in his face and could not be certain that this was her attacker unless she saw him in person. Several hours later, Schrier identified Bigelow in a line-up, and he was detained. Bigelow maintained his innocence and rejected a plea offer that included a five-year prison sentence. His case proceeded to trial five months later.

At trial, the State presented Charlotte Schrier's testimony, as well as the testimony of Thomas Mermer. Mermer had been near the scene of the incident and had radioed for help. After doing so, he noticed a man running into the field behind Schrier's burning car. Mermer testified that he could see the man only from the side and behind, JA 628; unlike Schrier, the police never showed Mermer any photo arrays or asked him to attend a line-up. Two weeks before trial, however, Mermer saw Bigelow giving an interview on television and identified him as the man he saw running into the field on the day of the crime. JA 629-30. The State did not introduce any other evidence connecting Bigelow to the crime.

The defense claimed that Bigelow was in Columbus, Ohio, 150 miles southeast of Toledo, on June 17, 1993, the day of the attack. In support of this alibi defense, it relied on two witnesses. John Laughner, the Columbus branch manager of Orkin Pest Control, testified that his office records showed that Vernon Greenlee, an employee of Orkin, had worked at the Columbus home of Gary Chasen over a period of two days, including from 11:00 a.m. to 5:00 p.m. on June 17th. JA 677. Greenlee also testified, and he confirmed that he treated the Chasen home for termites on the 17th. JA 684. He testified that a man had helped him move some objects from the garage so he could perform the treatment, and that the man was present when Greenlee arrived and when he left. JA 687. Greenlee identified Bigelow in court as the man who helped him in the garage on June 17th. JA 690-91. His testimony also acknowledged, however, that he had failed to identify Bigelow's photograph in an array shown to him by police, that he had worked at the Chasen home over two consecutive days, not just on the date of the crime, and that he had performed between fifty and one hundred jobs since June 17th of that year. Bigelow did not testify.

The jury convicted Bigelow of kidnapping, felonious assault and arson, and the trial court sentenced him to consecutive prison terms, which together created a twenty to forty-two year prison term. Bigelow unsuccessfully appealed his conviction. He then filed a petition for state post-conviction relief claiming he had received ineffective assistance of trial counsel because his court-appointed lawyer had not adequately investigated his alibi defense. The state trial court denied Bigelow's petition, but the appeals court remanded for an evidentiary hearing. After hearing testimony from the three alibi witnesses whom Bigelow claimed his attorney should have identified before trial, the court again denied his petition, the appeals court affirmed, and the state supreme court denied review. Bigelow filed this suit for a writ of habeas corpus in federal district court, again claiming ineffective assistance of counsel in violation of his Sixth (and Fourteenth) Amendment rights.

B. Bigelow's Allegations of Ineffective Assistance of Counsel

Bigelow first complained about the ineffectiveness of his lawyer in a letter to the state court before trial. The letter prompted the court to conduct a hearing on November 2, 1993, to determine whether Bigelow's lawyer, Peter Rost, should be replaced. At the hearing, Rost explained the difficulties of representing Bigelow and the investigative work he had undertaken on behalf of his client. Among the difficulties in supporting the alibi defense were that Bigelow could not remember exactly where in Columbus he had been on June 17th and that Bigelow suffered from a then-untreated mental illness, which contributed to his inability to aid the defense. JA 471, 726-27.

Despite these challenges, Rost noted that Bigelow had provided him with a list of names and possible leads — all but one of which Rost had pursued by telephone. In particular, Rost talked on the telephone to Gary Chasen, who besides being the owner of 654 Indian Mound Road where Greenlee had worked on June 17th, also owned rental property in Columbus where Bigelow at one point had lived. Chasen told Rost he could not confirm that Bigelow was in Columbus on the 17th. Similarly, Dan Watson and Deborah Gray, who also owned a home in Columbus where Bigelow had rented a room, were unable to remember whether Bigelow was with them on the 17th. At Rost's request, they checked the records of the telephone line that Bigelow shared with them to see if Bigelow had placed any calls to his friends or family on that date. The records showed only calls for June 20th and June 24th, but not June 17th. Rost followed up on other leads that Watson and Gray provided, but they too proved unsuccessful. Rost next contacted Greyhound Bus to determine whether Bigelow had traveled to Columbus around June 17th; Greyhound informed Rost that they did not have passenger records from June. Rost then called Bigelow's physician and dentist in Columbus to determine whether Bigelow had attended an appointment with either of them on June 17th; their records showed he had not. Finally, Rost searched Bigelow's personal papers for documentation that could establish Bigelow's presence in Columbus on June 17th, but again to no avail. After hearing about these efforts, the trial judge denied Bigelow's motion for a change of counsel.

On the day after this hearing, Bigelow began his own investigation. He wrote two letters from his prison cell — one to Orkin Pest Control, the other to Moonlighting Landscape and Lighting. In both letters, he asked whether anyone at the companies could verify his presence at Gary Chasen's home at 654 Indian Mound Road on June 17th and asked them to contact Rost if they had any such information. At Orkin, Laughner received Bigelow's letter, checked his records of work performed, and learned that Greenlee had in fact been there. He spoke to Greenlee, who remembered seeing (and talking to) Bigelow that day. On the evening of Thursday, November 11th — four days before the trial was to begin on Monday, November 15th — Greenlee telephoned Rost about his recollection of Bigelow at the Chasen home on June 17th. The phone call apparently did not spur Rost into performing any further investigation, though he did subpoena Laughner and Greenlee and did present their testimony at trial.

Bigelow's second letter, to Moonlighting Landscape, also arrived at its intended destination. Christine Patridge (a co-owner) checked the company's records and learned that she, Vic Timler (also a co-owner) and Jay Loyzelle (an employee) had all been at the Chasen home on June 17th. Like Orkin, they had been hired by Chasen to help prepare his home and yard for his daughter's wedding. All three recalled seeing Bigelow on June 17th and either having an extended encounter with Bigelow or noticing an idiosyncratic feature of his behavior or appearance. Patridge recalled that Bigelow talked to her for an extended period of time, and that he was very inquisitive about their landscaping work. JA 756. Timler noticed Bigelow oddly trimming a boxwood shrub with scissors and noticed that he was dressed atypically for yard work. JA 105, 776. Loyzelle saw Bigelow there both when he arrived in the morning and when he left late in the day, JA 805, 810; he spoke with Bigelow for "quite a while"; and he recalled that Bigelow cut his hand and asked for a band aid, JA 803.

The three witnesses agree that one of them (likely Patridge) contacted Gary Chasen to tell Chasen about Bigelow's letter and their memories of him being there. Chasen, however, told Patridge that it was better not to get involved and that she should ignore the letter. JA 757. Nonetheless, Patridge claimed that she called Rost and left him a message, but she could not remember when she did so and whether she left the message on an answering machine or with Rost's assistant. JA 766, 768. Rost claimed that he never received any message from Patridge and did not learn of the existence of the Moonlighting witnesses until he read Bigelow's state post-conviction petition. Bigelow insisted that he told Rost about both the Orkin and Moonlighting letters. JA 824-25.

Bigelow's essential claim, in both the state and federal post-conviction proceedings, is that Rost failed to satisfy the minimal requirements of effective advocacy because he did not adequately investigate Bigelow's alibi defense. As a factual matter, Bigelow claims that Rost knew about the Moonlighting Landscape lead, either because Bigelow told him or because he received the message from Christine Patridge, and accordingly Rost's failure to contact these witnesses and present their testimony at trial constituted ineffective assistance of counsel. Regardless of whether Rost actually knew about the Moonlighting employees before trial, Bigelow adds that Rost's failure to investigate adequately still constitutes ineffective assistance of counsel. Had his investigation been more thorough — i.e., had Rost obtained court funds for and hired an investigator, traveled to the location of Bigelow's alibi or at least investigated further once he learned that Greenlee could place Bigelow at the Chasen residence on the day of the attack — he undoubtedly would have uncovered the Moonlighting employees. The Ohio courts rejected these arguments, as did the federal district court, determining that Rost's assistance was not ineffective under Strickland v. Washington.

In reviewing the denial of a habeas petition, we consider the district court's legal conclusions anew, applying the same standard of review to the state court decision that the district court applied. See Smith v. Hofbauer, 312 F.3d 809, 813 (6th Cir.2002). That standard of review is supplied by AEDPA — The Antiterrorism and Effective Death Penalty Act, Pub.L. No. 104-132, 110 Stat. 1214 (1996). When a state court has already adjudicated a federal constitutional claim, AEDPA establishes that the writ of habeas corpus may issue in just two instances: (1) if the state court decision "was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States," 28 U.S.C. § 2254(d)(1); or (2) if the state court decision was "based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding," id. § 2254(d)(2).

To succeed on an ineffective assistance of counsel claim, a petitioner must show (1) that his lawyer's performance was deficient and (2) that the deficiency prejudiced the defense. Strickland, 466 U.S. at 687, 104 S. Ct. 2052. In establishing the first requirement, the petitioner must demonstrate that his lawyer's performance "fell below an objective standard of reasonableness" as measured by "prevailing professional norms." Id. at 687-88, 104 S. Ct. 2052. Judicial review of the lawyer's performance must be "highly deferential," and "indulge a strong presumption" that a lawyer's conduct in discharging his duties "falls within the wide range of reasonable professional assistance," since reasonable lawyers may disagree on the appropriate strategy for defending a client. Id. at 689, 104 S. Ct. 2052. While "strategic choices made after thorough investigation of law and facts ... are virtually unchallengeable[,][] strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation. In other words, counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary." Id. at 690-91, 104 S. Ct. 2052; see also O'Hara v. Wigginton, 24 F.3d 823, 828 (6th Cir.1994) ("[A] failure to investigate, especially as to key evidence, must be supported by a reasoned and deliberate determination that investigation was not warranted."); cf. ABA Standards for Criminal Justice 4-4.1(a) (3d ed. 1993) ("Defense counsel should conduct a prompt investigation of the circumstances of the case and explore all avenues leading to facts relevant to the merits of the case and the penalty in the event of conviction.").

In establishing prejudice, Bigelow must demonstrate a "reasonable probability" that the result of his trial would have been different but for Rost's mistakes. Strickland, 466 U.S. at 694, 104 S. Ct. 2052. A "reasonable probability" is a probability "sufficient to undermine confidence in the outcome," id., but something less than a showing that the outcome more likely than not would have been different, id. at 693, 104 S. Ct. 2052. While the petitioner need not conclusively demonstrate his "actual innocence," compare Schlup v. Delo, 513 U.S. 298, 327, 115 S. Ct. 851, 130 L. Ed. 2d 808 (1995) (requiring petitioner to establish more likely than not that a reasonable juror would not have convicted him), with Strickland, 466 U.S. at 693, 104 S. Ct. 2052 ("we believe that a defendant need not show that counsel's deficient conduct more likely than not altered the outcome in the case"), the focus should be on whether the result of the trial was "fundamentally unfair or unreliable," Lockhart v. Fretwell, 506 U.S. 364, 369, 113 S. Ct. 838, 122 L. Ed. 2d 180 (1993).

Bigelow first claims that Rost knew about the Moonlighting Landscape employees but refused to communicate with them about testifying as alibi witnesses. In the state courts and in the district court, no one has debated whether this allegation, if true, would present a serious Sixth Amendment claim. For the failure to call a known alibi witness generally would constitute ineffective assistance of counsel. See, e.g., Matthews v. Abramajtys, 319 F.3d 780, 789-90 (6th Cir.2003); Blackburn v. Foltz, 828 F.2d 1177, 1182-83 (6th Cir.1987). Rather, the debate has been joined on the question whether the facts support the claim, an issue that the state courts resolved against Bigelow.

In bringing this claim, Bigelow thus must overcome the state court's factual finding that "Pete Rost was not aware of Christine Patridge, Victor Timler, or Jay Loyzelle as additional witnesses who would support Bigelow's alibi defense." JA 279. That is no small task. Under AEDPA, we presume that the state court's factual findings are correct, and the petitioner bears the burden "of rebutting the presumption of correctness by clear and convincing evidence." 28 U.S.C. § 2254(e)(1); see Mitchell v. Mason, 325 F.3d 732, 737-38 (6th Cir.2003).

Although the record reveals some support for his position, Bigelow has not rebutted this presumption of correctness. At the state-court evidentiary hearing, Rost testified that Bigelow did not tell him about the letter to the landscape company and that he did not know about these potential witnesses. Bigelow, however, gave inconsistent testimony. He first stated that he told Rost about the letters, and that he remembered doing so because he had put Rost's home and work phone numbers in the letter (presumably so Rost would expect any resulting calls). JA 824-25, 827. But on cross-examination, Bigelow admitted that he "didn't tell Pete about [the letters] because the [other] leads, like the people said earlier, were no good that I gave to Pete Rost." JA 836.

Patridge's testimony that she left a message for Rost also does not undermine the state court's finding. Her testimony at the hearing, as an initial matter, conflicted with her affidavit in which she stated that she ignored the letter from Bigelow after talking to Chasen, who discouraged her from getting involved in the case. Patridge also could not remember when she left the message, and given the proximity between the date of Bigelow's letter to Moonlighting (November 3rd) and the date of the trial (November 15th), she may well have called after, possibly well after, the trial. And even if she left the message before trial, it requires few inferences to believe that Rost did not get the message, whether Patridge left it on an answering machine or with Rost's assistant. In view of the deferential standard of review that applies in this setting, the district court correctly rejected Bigelow's challenge to this factual finding and correctly rejected this ground for granting the writ.

Bigelow next argues that Rost's failure to investigate after he became aware of the Greenlee evidence constituted ineffective assistance of counsel. In the words of Bigelow's appellate brief:

Once defense counsel learned of the Orkin witnesses, he had an obligation to follow-up on the Orkin information and pursue this lead. From Orkin, defense counsel learned that Mr. Bigelow was on the Chasen property on June 17, 1993. From Orkin, defense counsel knew that Mr. Chasen had been wrong about Mr. Bigelow being on the property on June 17, 1993. He now had the tools to refresh Mr. Chasen's memory with this new information. Had defense counsel followed up on this information, Moonlighting's presence at the Chasen home, which corroborated Orkin's testimony, could have been easily discovered.

Appellant's Br. at 32. Bigelow raised the same argument in the district court, Traverse to Respondent's Return of Writ at 22-24, and in state court, Mem. for Pet'r, JA 238. In some contrast to his challenge to the state court's factual finding, Bigelow may prevail on this claim if he can show that the state court's application of Supreme Court precedent in this area was "objectively unreasonable." Wiggins, 123 S.Ct. at 2534-35.

Even though Bigelow raised this issue below and in state court, the district court did not address it — perhaps because the issue was obscured by the understandable focus in state court and in the district court on whether Rost in fact knew about these other alibi witnesses. In response, the State asks us to affirm the judgment nonetheless, arguing that the claim is meritless. We disagree. The claim is sufficiently serious that it warrants consideration by the district court in the first instance and may even warrant an evidentiary hearing. Even after looking at the issue through the prism of AEDPA, the State has not shown that Bigelow's claim under Strickland prong one (the adequacy of counsel's performance) or prong two (prejudice to the defendant) deserves plenary rejection by us. To the ends of facilitating the district court's consideration of these issues, we offer some explanation for rejecting the State's invitation to reject this claim at this stage of the case.

The record regarding the adequacy of counsel's investigation raises as many questions as answers. In spite of Rost's initial lack of success in investigating Bigelow's defense — or perhaps because of that lack of success — it is difficult on this record to understand why the surfacing of the Greenlee evidence did not prompt Rost to investigate further. Greenlee was the first person to come forward who could corroborate Bigelow's claim that he had been in Columbus, not Toledo, on the day (and at the time) of the crime. Rost's own actions, moreover, prove he understood the significance of the evidence, as he put Greenlee on the stand virtually sight unseen and without any further investigation. Indeed, Rost admitted on the morning of trial that he had not yet spoken to Bigelow's only alibi witness, namely Greenlee. JA 487. At a minimum, it would seem that this evidentiary breakthrough would have prompted additional inquiry either by Rost or by a publicly-funded investigator. Cf. Workman v. Tate, 957 F.2d 1339, 1345 (6th Cir.1992) (concluding that "reasonable prudence" should have prompted the lawyer to recognize the importance of potential witness testimony to the defense); Griffin v. Warden, Maryland Corr. Adjustment Ctr., 970 F.2d 1355, 1359 (4th Cir.1992) (recognizing the significance of alibi evidence in countering an eyewitness case assembled by the State).

Nor, at this point, can we readily agree with the State that Rost made "a reasonable decision that ma[de] [this] particular investigation[] unnecessary." Strickland, 466 U.S. at 691, 104 S. Ct. 2052. To our knowledge, the State has not even attempted to offer a strategic explanation for Rost's failure to investigate further once he learned of Greenlee's evidence. The information provided by Greenlee did nothing to suggest that further investigation would be futile or damaging to his client, but in point of fact suggested just the opposite. See Wiggins, 123 S.Ct. at 2537, 123 S. Ct. 2527 (noting that the fact the lawyers "uncovered no evidence in their investigation to suggest that ... further investigation would have been fruitless" differentiated Wiggins' case from those in which limited investigations were reasonable); see also Workman, 957 F.2d at 1345 ("Where counsel fails to investigate and interview promising witnesses, and therefore has no reason to believe they would not be valuable in securing defendant's release, counsel's inaction constitutes negligence, not trial strategy.") (quotation and citation omitted); id. (noting that the case was not one where further investigation would be unlikely to bear fruit, which could excuse a lawyer's failure to investigate).

Bigelow himself also did not supply any reasons why further inquiry would be unproductive. Instead, Greenlee's evidence amounted to the first evidentiary breakthrough in the case, which suggested that (1) Bigelow was in Columbus on the date of the assault, (2) Bigelow at the very least was in Columbus on the day before or the day after the assault (since Greenlee worked at the Chasen home for two consecutive days) and (3) Gary Chasen, the owner of the house where Greenlee and Bigelow had worked, erred in telling Rost he had no relevant information about Bigelow's whereabouts on June 17th.

Given that Greenlee's testimony was the only evidence then available for Bigelow's defense and given that his testimony would be vulnerable on cross-examination since he had been at Chasen's house for two days, Rost's failure to do anything at this point remains unexplained, if not inexplicable. While Bigelow was "not entitled to an attorney who will leave not the smallest stone unturned," since he had "but one stone, it should at least [have been] nudged." Coleman v. Brown, 802 F.2d 1227, 1234 (10th Cir.1986) (internal quotation, citation omitted). Rost appeared to have a host of options available to him for nudging that stone: he could have requested a continuance in order to investigate this new lead in the case; he could have hired an investigator (with court funds); or at the very least he could have spent the four days before trial following up on the information that Greenlee provided him. But as in Wiggins, it appears on this record that Rost "chose to abandon [his] investigation at an unreasonable juncture, making a fully informed decision" with respect to trial strategy very difficult, if not impossible. 123 S.Ct. at 2538.

Like the state-court proceedings, the State's contrary arguments suffer from a mistaken emphasis on (1) whether Rost knew about the Moonlighting employees and (2) what Rost did earlier in the case rather than what Rost failed to do once he learned of Greenlee. As to the first point, it makes no difference whether Rost knew about these witnesses if a reasonable investigation (after Greenlee came forward) would have uncovered their identities anyway. See Wiggins, 123 S.Ct. at 2537 (noting that if counsel had performed a reasonable investigation, they likely would have discovered evidence of sexual abuse that could have been presented at sentencing hearing); Lindstadt v. Keane, 239 F.3d 191, 200-01 (2d Cir.2001) (determining that counsel was ineffective for failing to investigate the case, as a reasonable investigation undoubtedly would have uncovered the error in the time frame of plaintiff's accusations, which would have led to alibi defense); Sims v. Livesay, 970 F.2d 1575, 1580 (6th Cir.1992) ("Although defense counsel might not have been told about [potentially exculpatory evidence] by Sims, prior to trial [defense counsel] was in possession of an FBI report that should have alerted him to the significance of [this evidence] for the defense.").

Once Rost learned of Greenlee, had he taken even minimal additional investigative steps — e.g., by contacting the initially-reluctant Chasen and confronting him with the new information about Greenlee, asking Chasen for records of the companies that helped with wedding preparations on the 17th, or talking to Chasen's neighbors — he likely would have uncovered the Moonlighting Landscape employees. The same is true of a follow-up conversation with his client. Had he discussed Orkin with Bigelow, he assuredly would have learned of Moonlighting, since Bigelow's letters demonstrate that he had already developed a link between the two companies and his whereabouts on June 17th.

As to the second point — the adequacy of Rost's investigative efforts before Greenlee came forward — we agree with the State that Rost's performance surpassed the "objective standard of reasonableness" described in Strickland. But Rost's commendable efforts before Greenlee arrived on the scene cannot shield from scrutiny his efforts after this evidence surfaced. Wiggins demonstrates that it does not invariably suffice that a lawyer make some efforts to investigate a case; the proper inquiry is "whether the known evidence would lead a reasonable attorney to investigate further." 123 S.Ct. at 2538; see also, e.g., Montgomery v. Petersen, 846 F.2d 407, 414 (7th Cir.1988) (defense counsel's failure to investigate a promising lead that would have uncovered a disinterested alibi witness constituted ineffective assistance even though the lawyer had interviewed and put on the testimony of twelve other witnesses). Neither does the late arrival of the Greenlee evidence necessarily excuse Rost's failure to act. See Bryant v. Scott, 28 F.3d 1411, 1417 (5th Cir.1994) (determining that counsel was ineffective, because it was "incumbent upon [him] to at least try to contact" a potential alibi witness that he learned about seventy-two hours before the start of trial); cf. id. (noting that even if the lawyer had learned of the alibi witnesses on the first day of trial, he "nevertheless should have contacted the witnesses and made his record to the trial court as to the significance of the alibi and the fact that it was newly discovered") (quotation and citation omitted).

In the end, given what Rost learned from Greenlee four days before trial (that Bigelow was at the Chasen home on the day of the attack) and given what Rost knew about other alibi evidence up to that point (nothing), his apparent decision to do no further investigation deserves fresh consideration by the district court. Thus far, the State has given no indication that Rost did any further investigation after hearing from Greenlee and has offered no reasoned explanation — strategic or otherwise — why Rost should not have investigated further at that point. To the ends of answering these questions (and potentially developing a record in support of those answers), we ask the district court to take an initial look at the adequacy of Bigelow's counsel in these respects.

Nor, on this record, can one say that any errors in Rost's investigation were not prejudicial. Although we leave the question open for the district court to decide, it seems on this record that Rost would have uncovered the Moonlighting witnesses had he investigated further after learning of Greenlee. And Rost acknowledges that he would have subpoenaed the three witnesses to testify at trial if he had known about them. JA 747.

The key question is whether the testimony of the Moonlighting employees would have made a persuasive case that Bigelow was at the Chasen home in Columbus on June 17th, 1993. We leave this issue for the district court to consider in the first instance because, contrary to the State's contention, its resolution is not obvious. At the state court evidentiary hearing, Moonlighting's records showed that Patridge, Timler and Loyzelle all worked at the Chasen home on June 17th, JA 105B, 107, 753, 795-96, and at least one of them (Timler) did not work at the property on any other day, JA 778. All three witnesses identified Michael Bigelow, sitting before them at the evidentiary hearing, as the man they had seen and interacted with at the Chasen home on the 17th. JA 754-55, 781-84, 804. Timler initially testified that he was "80%" sure Bigelow was at the Chasen home that day, but then said he was "100%" certain it was Bigelow. JA 783-84. Patridge's in-court identification of Bigelow went unchallenged by the State, and she stated in her affidavit that Chasen "introduced" Bigelow (presumably by name). JA 105B. Loyzelle identified Bigelow in court as well, JA 804, and his affidavit noted that Chasen had addressed Bigelow on the 17th as "Mike," JA 107. All three witnesses, moreover, recalled idiosyncratic details about their encounters with Bigelow (i.e., his inquisitiveness, the fact that he was pruning hedges with a pair of scissors and that he cut his hand and asked for a band aid), adding credence to their memories of seeing him.

Loyzelle testified that he saw Bigelow at various points both in the morning and the afternoon of June 17th, JA 805, 811, making it impossible for Bigelow to have slipped away to make the six-hour round-trip to Toledo and back. This testimony, if believed, would have precluded the possibility that Bigelow attacked Schrier at midday on the 17th.

All three witnesses were completely disinterested, as none of them had any previous connection to Bigelow. JA 753-54, 774-75, 802. In fact, Timler testified that it was quite costly for the three to attend the evidentiary hearing in view of the money they could have been earning on another landscaping job. JA 780.

This evidence plainly would have bolstered Bigelow's defense and was anything but cumulative. See Washington v. Smith, 219 F.3d 620, 634 (7th Cir.2000); Montgomery, 846 F.2d at 413 (introduction of alibi witness would not have been cumulative despite testimony by twelve other witnesses where the new witness did not bear the same weakness as the others). Bigelow had just one witness at trial (Greenlee) who could support his alibi defense. Doubtless, three other witnesses, who like Greenlee did not previously know Bigelow and accordingly had no axe to grind in testifying on his behalf, would have aided the defense. This testimony also would have shored up the weaknesses in Greenlee's testimony — that he had worked at Chasen's house over a period of two days, which cast doubt on his certainty as to the date that he actually witnessed Bigelow, and that he had been unable to select Bigelow from a photo lineup, which cast doubt on his in-court identification.

The testimony of the Moonlighting witnesses also would have facilitated Bigelow's efforts to undermine the State's case. The State presented two eyewitnesses, both of whom identified Bigelow in court, but did not introduce any forensic or other evidence in the case. Schrier, for one, testified that she had a limited opportunity to view her attacker. Because the assailant remained behind her and she obeyed his instructions not to turn around, JA 498-501, she had two opportunities to view his face: (1) when he told her to light his cigarette; and (2) when she turned around to face him just before kicking him and escaping. JA 510, 525, 528, 530. Both opportunities, however, were fleeting. See JA 511, 534-35. There also was a disparity between the exacting details about her attacker's face that she supplied at trial and her initial description of him. At trial, she testified about his hairline, Adam's apple, cheekbones, eyes and lips. But in the original description she gave to the police, she focused on the attacker's clothing, height and age.

The testimony of the other eyewitness, Thomas Mermer, was even weaker. He admitted (1) that he saw the assailant only from the back and side at a distance running away from the scene of the crime, JA 628, and (2) that police never showed him a lineup or photo display, JA 629. Mermer first identified Bigelow as the defendant after seeing Bigelow's face on television many months after the incident, JA 629-30, a classically suggestive setting because Bigelow was giving an interview to local media about his impending trial.

The state trial court's conclusion to the contrary does not alter this analysis. It described the Moonlighting witnesses' testimony as "vague and unconvincing," pointing to inconsistencies between their affidavits and their hearing testimony. The inconsistencies, however, are just three, and they are insignificant to boot. They were: (1) typographical mistakes as to the year that Bigelow was at the Chasen home (which were corrected and initialed by the notary public); (2) Patridge's testimony that she contacted Rost versus her affidavit statement that she ignored his letter (which relates to the separate question whether Rost knew about these witnesses before trial); and (3) Loyzelle's testimony that he left the property between 5:00 p.m. and 5:30 p.m., not at 4:15 p.m. as he indicated in the affidavit (which would not have made a difference as to whether Bigelow committed a midday attack). None of the inconsistencies undermine the pivotal facts established by the affidavits: that Patridge checked Moonlighting's records and verified that they were at Chasen's on the 17th; that they saw Bigelow there both in the morning and in the afternoon; and that Chasen had introduced Bigelow to Patridge and referred to him as "Mike" within earshot of Loyzelle. Each of these statements was consistent with the evidentiary hearing testimony of the witnesses.

In the final analysis, the addition of the three Moonlighting witnesses would have presented the jury with (1) four witnesses on the one hand who could identify Bigelow in court as the man they saw in Columbus on June 17th and (2) two witnesses on the other hand who could identify him in court as the assailant in one instance and as the man running from the crime scene in the other. In a case involving identification and identification alone, it is not easy to imagine a defense lawyer who would pass on the chance to bolster the defense with evidence of this sort — particularly since eyewitness evidence is "precisely the sort of evidence that an alibi defense refutes best," Griffin, 970 F.2d at 1359.

But because these issues were not addressed by the district court and because some of them may benefit from additional evidence, we remand the case to the district court to determine whether Bigelow has shown that the writ should be granted on this alternative theory of ineffective assistance of counsel. Consideration of this theory will require the district court to focus on two questions that we have addressed but ultimately leave open for consideration on remand: (1) Was it objectively unreasonable for Rost to fail to conduct further investigation after learning of the Orkin employees?; and (2) If Rost's representation was ineffective, was it likely the three alibi witnesses would have been identified had he conducted a reasonable investigation?

For the foregoing reasons, we vacate the judgment and remand the case to the district court for a determination whether the writ should be granted.