concurring:
“Judge Friendly’s observation a quarter of a century ago that ‘the one thing almost never suggested on collateral attack is that the prisoner was innocent of the crime’ remains largely true today.” Schlup v. Delo, 513 U.S. 298, 321-22, 115 S.Ct. 851, 130 L.Ed.2d 808 (1995) (quoting Henry J. Friendly, Is Innocence Irrelevant? Collateral Attack on Criminal Judgments, 38 U. Chi. L.Rev. 142, 145 (1970)). This appeal from a judgment granting a petition for a writ of habeas corpus is yet another case that confirms Judge Friendly’s observation. While the prisoner here challenges the admissibility of an unnecessarily suggestive eyewitness identification, a challenge that on its face goes to the issue of his guilt of the crime with which he is charged, the record contains compelling independent evidence corroborating the reliability of the eyewitness identification. This, for me, provides the basis for denying the writ.
Unlike my colleagues, I am unable to conclude that the on-the-scene identification of petitioner was not unnecessarily suggestive. While I agree that on-the-scene identifications shortly after the commission of a crime are necessary and appropriate, even though they are inherently suggestive, the manner in which the procedure was conducted in this case rendered it significantly more suggestive than the circumstances necessitated. Similarly, while I agree that such an identification may be admissible, if it is found to be independently reliable, I do not believe that the record is sufficiently developed with respect to this issue. Nevertheless, I concur in the reversal of the judgment granting the writ because the overwhelming evidence of the petitioner’s guilt is sufficient to make up for any deficiencies in the record with respect to the application of the five factor test prescribed in Neil v. Biggers, 409 U.S. 188, 93 S.Ct. 375, 34 L.Ed.2d 401 (1972), and Manson v. Brathwaite, 432 U.S. 98, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977), upon which the majority relies.
Corroborative evidence of guilt goes to the heart of the Supreme Court cases that initially addressed the problem posed by eyewitness identifications. Those cases developed out of concerns that eyewitness identifications (by strangers) may not be reliable, that juries could not be trusted to evaluate the deficiencies of such apparently compelling evidence, and that the admission of suggestive identifications risked the ultimate injustice — the conviction of an innocent person. In short, the reason the Court involved itself in the first place with suggestively obtained identifications was the recognition that they were a “major factor contributing to the high incidence of miscarriage of justice.... ” United States v. Wade, 388 U.S. 218, 228, 87 S.Ct. 1926, 18 L.Ed.2d 1149 (1967); see also Stovall v. Denno, 388 U.S. 293, 297, 87 S.Ct. 1967, 18 L.Ed.2d 1199 (1967) (“A conviction which rests on a mistaken identification is a gross miscarriage of justice.”), overruled on other grounds by Griffith v. Kentucky, 479 U.S. 314, 107 S.Ct. 708, 93 L.Ed.2d 649 (1987). If the reason for ousting the jury from its historic role of evaluating evidence is that suggestively obtained eyewitness identifications can lead to erroneous convictions, then surely independent evidence of guilt provides an essential protection against such “gross misearriage[s] of justice.” Stovall, 388 U.S. at 297, 87 S.Ct. 1967. As the Eighth Circuit put it, “it *97seems unnatural to set such evidence aside when one considers that the ultimate purpose of the Biggers factors is to avoid eyewitness testimony where there is ‘a very substantial likelihood of irreparable misidentification.’ ” Graham v. Solem, 728 F.2d 1533, 1546 (8th Cir.1984) (en banc).
It is true that Brathwaite contains ambiguous language that has been read to oppose the use of other evidence of guilt to establish the reliability of eyewitness identifications. After analyzing the record in light of the five factors articulated in Biggers, Justice Blackmun concluded that “[t]hese indicators of [the witness’s] ability to make an accurate identification are hardly outweighed by the corrupting effect of the challenged identification itself.” Manson v. Brathwaite, 432 U.S. 98, 116, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977). Justice Blackmun then stated that, “[although it plays no part in our analysis, all this assurance as to the reliability of the identification is hardly undermined by the facts that respondent was arrested in the very apartment where the sale had taken place, and that he acknowledged his frequent visits to that apartment.” Id.
The explanation for this less than edifying analysis may be found in Judge Friendly’s opinion in Brathwaite v. Manson, 527 F.2d 363 (2d Cir.1975), which held that the suggestive eyewitness identification was inadmissible under the Biggers standard, and which the Supreme Court reversed. Prior to our decision in Brathwaite, Judge Friendly observed that “the rule in this circuit is that other evidence connecting a defendant with the crime may be considered on the issue whether there was a substantial likelihood of misidentification.” United States v. Reid, 517 F.2d 953, 967 (2d Cir.1975) (citations omitted). Such evidence was present in Brathwaite, corroborating the blatantly suggestive eyewitness identification: An undercover police officer had purchased drugs from Brathwaite at an apartment, and there was other evidence linking the defendant to that same apartment. Judge Friendly concluded, however, that this evidence was not strong enough to sustain the admissibility of the eyewitness identification. As he explained: “Perhaps the strongest bit of evidence to strengthen the identification was Brathwaite’s arrest in the very apartment where the sale was made. But Brathwaite offered an explanation of this which was not implausible, although evidently not credited by the jury....” Brathwaite, 527 F.2d at 372. This may explain Justice Blackmun’s unwillingness to rely on the corroborating evidence other than to say, quite consistently with Judge Friendly’s analysis, that it did not undermine the reliability of the identification in that case, although it did not enhance it. While Justice Blackmun did not offer this explanation, it is clear that the majority of the Supreme Court there did not expressly hold that the presence of corroboration could never be considered. Indeed, the language Justice Blackmun used does not even rise to the level of dictum to this effect.
Nonetheless, the language used by Justice Blackmun does explain Justice Stevens’s separate concurrence in Brathwaite, which was not joined by any of the other six members of the majority. Justice Stevens first commended Justice Blackmun’s opinion for avoiding what he characterized as the “pitfall” of considering other evidence of guilt. Brathwaite, 432 U.S. at 118, 97 S.Ct. 2243 (Stevens, J., concurring). He then went on to expressly articulate the view that corroborating evidence should not “be considered to support the admissibility of eyewitness testimony when applying the criteria identified in Neil v. Biggers.” Id. at 118 n. *, 97 S.Ct. 2243. “Properly analyzed,” he continued, “such facts would be relevant to a question *98whether error, if any, in admitting identification testimony was harmless.” Id. While the Stevens opinion clearly says what Justice Blackmun avoided saying, it certainly does not constitute a holding of the Supreme Court that corroborative evidence should not “be considered to support the admissibility of eyewitness testimony when applying the criteria identified in Neil v. Biggers.” Id.
Significantly, after Brathwaite, many courts have continued to consider other proof of guilt in deciding whether to admit eyewitness identification testimony. An in-court identification was judged reliable by the Eighth Circuit, in part because two other government witnesses identified the defendant, including the driver of the getaway car. United States v. Rogers, 73 F.3d 774, 778 (8th Cir.1996). As the panel explained, the “additional testimony diminishes any likelihood of irreparable misidentification.” Id. Similarly, in evaluating the reliability of eyewitness identifications in a drive-by shooting, the Seventh Circuit explained that “Corroborated eyewitness testimony is much less suspect than one individual’s visual impression.” United States ex rel. Kosik v. Napoli, 814 F.2d 1151, 1156 (7th Cir.1987). The Seventh Circuit then proceeded to consider, along with the Biggers-Brathwaite factors, that a car like the one eyewitnesses described was registered to the defendant, that two other witnesses who did not make identifications gave descriptions of the driver that fit the defendant, and that the shooting took place in the defendant’s neighborhood. Id. at 1156-57, 1161.
A First Circuit case authored by then-Judge Breyer likewise relied in part on corroborative evidence of guilt in holding that an eyewitness identification was properly admitted. United States v. Lau, 828 F.2d 871, 875 (1st Cir.1987). The corroboration in that case included evidence that one defendant had a license to fly the type of plane used in the crime and another was nearby at the time of the incident. Judge Breyer went through the Biggers-Brathwaite factors for and against a correct identification — the witness’s confidence in his identifications, opportunity to observe, accuracy of prior descriptions, and the time elapsed between the incident and the identification. Id. The corroborating evidence appears as part of this same analysis as one of the “factors [that] weigh in favor of a correct identification.” Id.; see also United States v. Wilkerson, 84 F.3d 692, 695 (4th Cir.1996) (“Courts may also consider other evidence of the defendant’s guilt when assessing the reliability of the in-court identification.”). Contra United States v. Rogers, 126 F.3d 655, 659 (5th Cir.1997); United States v. Emanuele, 51 F.3d 1123, 1128 (3d Cir.1995); Green v. Loggins, 614 F.2d 219, 225 (9th Cir.1980).
We too have acknowledged the significance of such evidence before and after Brathwaite. Again, in Reid, Judge Friendly observed that “the rule in this circuit is that other evidence connecting a defendant with the crime may be considered” to determine an identification’s reliability. Reid, 517 F.2d at 967 (citing United States ex rel. Gonzalez v. Zelker, 477 F.2d 797, 803-04 (2d Cir.1973); United States v. Bynum, 485 F.2d 490, 503-04 (2d Cir.1973), vacated on other grounds, 417 U.S. 903, 94 S.Ct. 2598, 41 L.Ed.2d 209 (1974)). Nevertheless, our case law on this issue has been something less than a seamless web. In Raheem v. Kelly, 257 F.3d 122 (2d Cir.2001), without reference to the rule in this circuit established by the foregoing precedent, and without reference to the deferential scope of review mandated by 28 U.S.C. § 2254(d), we rejected an argument that an unnecessarily suggestive identification could be deemed reliable because of evidence of the defendant’s guilt. Id. at 140-41. More recently, in Ken*99naugh v. Miller, 289 F.3d 36, 47-48 (2d Cir.2002), specifically referencing the appropriate deferential standard of review applicable to habeas corpus proceedings, we observed that, “[although we have recently rejected this ‘sixth factor’ approach [which looks to corroborative evidence of guilt] as a matter of federal law in this circuit, choosing instead a harmless error approach, Raheem v. Kelly, 257 F.3d 122, 140 (2d Cir.2001), it seems likely that under the AEDPA, a state court that used a ‘sixth factor’ analysis would be applying the Manson requirements in a perfectly reasonable way.”
In any event, the focus of our decision in this habeas corpus case, as we have observed in response to a recent reminder by the Supreme Court, is not on “warring decisions among the federal circuits,” Rodriguez v. Miller, 537 F.3d 102, 106 (2d Cir.2008) (citing Carey v. Musladin, 549 U.S. 70, 74, 127 S.Ct. 649, 166 L.Ed.2d 482 (2006)), or on inconsistencies in our own case law. Instead, the issue is whether, considering the independent corroborative evidence of the petitioner’s guilt, the state courts’ rejection of petitioner’s challenge to the admissibility of the unnecessarily suggestive eyewitness identification 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). “[C]learly established Federal law ... refers to the holdings, as opposed to the dicta.” Musladin, 549 U.S. at 74, 127 S.Ct. 649 (quoting Williams v. Taylor, 529 U.S. 362, 365, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000)). No holding of the Supreme Court precludes consideration of corroborative evidence in determining the independent reliability of an unnecessarily suggestive eyewitness identification. Because the nature of that evidence in this case is overwhelming, and because it makes up for any deficiencies in the record with respect to the other relevant Biggers factors, I join my colleagues in voting to reverse the judgment of the district court granting the writ.