David Scott DETRICH, Petitioner-Appellant, v. Charles L. RYAN, Director of Arizona Department of Corrections, Respondent-Appellee, 740 F.3d 1237

OPINION W. FLETCHER, Circuit Judge: Judges Pregerson and Reinhardt concur in the entirety of the following opinion. Judge Christen concurs in

Part II and in the result. Judges Nguyen and Watford concur in the result. David Scott Detrich appeals from the district court’s denial of his habeas petition. An Arizona judge sentenced Detrich to death after a jury convicted him of murder, kidnapping, and sexual abuse. The district court held that several of Det-rich’s claims of ineffective assistance of counsel (“LAC”) by his trial counsel were procedurally defaulted because he had failed to raise them during his state post-conviction relief (“PCR”) proceedings.

Applying then-governing law, the district court rejected Detrieh’s argument that ineffective assistance of his PCR counsel could excuse his procedural default. While Detrich’s appeal from the district court decision was pending in this court, the Supreme Court decided Martinez v. Ryan,—U.S.-, 132 S.Ct. 1309, 182 L.Ed.2d 272 (2012). The Court held in Martinez that a state PCR counsel’s ineffective assistance in failing to raise trial-counsel IAC claims can excuse a procedural default. Detrich moved for a remand to the district court to allow that court to rule on his Martinez motion in the first instance.

We grant the motion and remand to the district court. I. Background Detrich was charged with first-degree murder, kidnapping, and sexual assault in connection with the killing of Elizabeth Souter. State v. Detrich (Detrich I), 178 Ariz.380, 873 P.2d 1302, 1304 (1994). Alan Charlton, who participated in the events that culminated in Souter’s murder, pled guilty to kidnapping. He then testified against Detrich in exchange for a ten-and-a-half-year sentence.

Detrich’s first trial ended in a mistrial after a prosecution witness testified that Detrich had invoked his rights under the Fifth Amendment during the investigation. Id. The Arizona Supreme Court reversed the conviction in Detrich’s second trial because of a defective jury instruction. Id. at 1306. After a third trial, the jury convicted Detrich of kidnapping and first-degree murder. State v. Detrich (Detrich II), 188 Ariz.57, 932 P.2d 1328, 1331 (1997). The jurors could not agree on a basis to support the first-degree murder conviction. Nine jurors found Detrich guilty of premeditated murder.

Three jurors found him guilty of only felony murder. That is, it appears that only nine jurors were convinced that Detrich, rather than Charlton, was the actual killer. The trial judge concluded beyond a reasonable doubt that Detrich was the killer. Based on that conclusion, he sentenced Detrich to death for the murder and to twenty-one years in prison for the kidnapping. Id. The Arizona Supreme Court affirmed Detrich’s convictions and sentence. Id. With the assistance of new counsel, Det-rich filed a PCR petition in Pima County Superior Court.

Detrich alleged in his PCR petition that counsel at his third trial had been ineffective for failing to (1) present mitigating evidence during sentencing; (2) present an expert witness to rebut the aggravating factors presented by the state; (3) retain an expert witness to examine certain pieces of forensic evidence at trial; (4) present live testimony from exculpatory witness William Shell instead of relying on Shell’s recorded testimony from a prior trial; (5) object to testimony that Charlton’s plea agreement required that he testify truthfully; and (6) preserve other constitutional challenges for appeal.

The superior court rejected Detrich’s claims on the merits, holding that “neither prong of the Strickland v. Washington[, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984),] test has been met as to any claims of ineffective assistance of counsel.” The Arizona Supreme Court denied review, leaving the superior court’s four-page order as the only reasoned state-court PCR decision. Detrich then filed a habeas petition in federal district court.

The petition alleged some of the claims that had been rejected in the state PCR proceeding, including the trial-counsel IAC claims for failure to present mitigating evidence and failure to present an expert witness to rebut the state’s aggravation case. The petition also raised trial-counsel IAC claims that had not been presented in the state PCR proceedings. Before the district court ruled, Detrich filed a second PCR petition in the superior court. In this petition, Detrich raised many of the trial-counsel IAC claims he had alleged for the first time in his federal petition.

The superior court held that these new claims were procedurally barred under Arizona Rule of Criminal Procedure 32.2(

b) because they could have been raised in Detrich’s first PCR petition. The district court then ruled that Det-rich’s new trial-counsel IAC claims were procedurally defaulted for purposes of federal habeas review. The court rejected Detrich’s argument that the ineffectiveness of his first PCR counsel excused his procedural default, noting that there was no constitutional right to counsel in PCR proceedings. The district court rejected all of Det-rich’s non-defaulted claims on the merits.

After an evidentiary hearing, the district court held that Detrich’s counsel had performed deficiently by failing to investigate and present mitigating evidence at sentencing, contrary to the holding of the state PCR court. Detrich v. Schriro, No. CV-03-229-TUC-DCB, 2007 WL 4024551, at * 3-10 (D.Ariz. Nov. 15, 2007). The district court concluded, however, that Detrich had failed to show prejudice resulting from that deficient performance as required under Strickland. Detrich, 2007 WL 4024551, at * 10-24. A three-judge panel of this court reversed, vacating Detrich’s death sentence. Detrich v.

Ryan, 619 F.3d 1038 (9th Cir.2010). The panel agreed with the district court that the Arizona PCR court had unreasonably applied Strickland when it concluded that Detrich’s sentencing counsel had not performed deficiently. Id. at 1052-57. However, it disagreed with the district court on the prejudice prong of Strickland, holding that the PCR court’s conclusion that Detrich was not prejudiced by trial counsel’s failure to investigate and present mitigating evidence was based on an unreasonable determination of the facts. Id. at 1057-69.

The Supreme Court vacated our decision and remanded in light of its decision in Cullen v. Pinholster,—U.S.-, 131 S.Ct. 1388, 179 L.Ed.2d 557 (2011). Ryan v. Detrich,—U.S.-, 131 S.Ct. 2449, 179 L.Ed.2d 1206 (2011) (mem.). On remand, the three-judge panel again reversed the district court and vacated the death sentence. Detrich v. Ryan, 677 F.3d 958 (9th Cir.2012). After the panel issued its second opinion, Detrich moved to remand under Martinez. We granted rehearing en banc. 696 F.3d 1265 (9th Cir.2012). Detrich’s Martinez motion is now before our en banc panel. II. Martinez v. Ryan and Trevino v.

Thaler The district court properly concluded under then-governing law that Detrich’s trial-counsel IAC claims raised for the first time in his federal habeas petition had been procedurally defaulted, and that it therefore could not hear them. A federal court sitting in habeas ordinarily cannot hear a petitioner’s procedurally defaulted federal claims absent a showing of cause and prejudice, or a showing that failing to review the claim will result in a fundamental “miscarriage of justice.” Wainwright v.

Sykes, 433 U.S. 72, 88, 90-91, 97 S.Ct. 2497, 53 L.Ed.2d 594 (1977) (applying the rule in the context of failure to make contemporaneous objection); Coleman v. Thompson, 501 U.S. 722, 750, 111 S.Ct. 2546, 115 L.Ed.2d 640 (1991) (making “explicit” that Wainwright and its progeny apply in “all eases in which a state prisoner has defaulted his federal claims in state court pursuant to an independent and adequate state procedural rule”).

In Coleman, the Court held that ineffective assistance of counsel in a state PCR proceeding cannot constitute cause to excuse a procedural default because there is no constitutional right to an attorney in state PCR proceedings. 501 U.S. at 752-53, 111 S.Ct. 2546. Applying Coleman, the district court correctly held, based on the law as it then stood, that the ineffectiveness of Detrich’s state PCR counsel in failing to raise trial-counsel IAC claims could not constitute cause. While the district court’s decision was on appeal, the Supreme Court changed the law. See Martinez, 132 S.Ct. at 1315.

The Court held in Martinez that “[inadequate assistance of counsel at initial-review collateral proceedings may establish cause for a prisoner’s procedural default of a claim of ineffective assistance at trial.” Id. The Court addressed the situation in Arizona, where a prisoner is forbidden to raise a trial-counsel IAC claim on direct review. Such a claim may be brought only in state PCR proceedings. Id. The Court wrote that in such cases, “the collateral proceeding is in many ways the equivalent of a prisoner’s direct appeal as to the ineffective-assistance claim.” Id. at 1317.

The Court recognized that “if counsel’s errors in an initial-review collateral proceeding do not establish cause to excuse the procedural default in a federal habeas proceeding, no court will review the prisoner’s claims.” Id. at 1316.

The Court therefore held, Where, under state law, claims of ineffective assistance of trial counsel must be raised in an initial-review collateral proceeding, a procedural default will not bar a federal habeas court from hearing a substantial claim of ineffective assistance at trial if, in the initial-review collateral proceeding, there was no counsel or counsel in that proceeding was ineffective. Id. at 1320. The Court did not reach the question of whether there is a constitutional right to effective assistance of counsel during a state PCR proceeding. Id. at 1319-20.

Rather, the Court established an equitable rule that IAC during initial-review PCR proceedings may constitute “cause” to excuse a state-court procedural default. Id. In Trevino v. Thaler,—U.S.-, 133 S.Ct. 1911, 185 L.Ed.2d 1044 (2013), the Court reaffirmed and slightly expanded the Martinez rule. The question in Trevino was whether the Martinez rule applies in states, such as Texas, where an appellate counsel is legally permitted to assert a claim of trial-counsel IAC on direct review, but where it is “highly unlikely” as a practical matter that appellate counsel will have a “meaningful opportunity” to do so.

Id. at 1921. The Court noted that Texas, unlike Arizona, “appears at first glance to permit ... the defendant initially to raise a claim of ineffective assistance of trial counsel on direct appeal.” Id. at 1915. But “in actual operation,” the “structure and design of the Texas system ... make it ‘virtually impossible’ for an ineffective assistance claim to be presented on direct review.” Id. (quoting Robinson v. State, 16 S.W.3d 808, 810-11 (Tex.Crim.App.2000)). The Court therefore found “no significant difference between this case and Martinez.” Id. at 1921.

The Court concluded that “where, as here, [the] state procedural framework, by reason of its design and operation, makes it highly unlikely in a typical case that a defendant will have a meaningful opportunity to raise a claim of ineffective assistance of trial counsel on direct appeal, our holding in Martinez applies.” Id. This is the first occasion for an en banc panel of our court to deal with what we will call, for ease of reference, a Martinez motion. For the guidance of the district court on remand, we take the opportunity to explicate three aspects of the Court’s decisions in Martinez and Trevino.

A. “Cause” under Martinez and Trevino In Martinez and Trevino, the Court created an exception to the normally applicable “cause” and “prejudice” rule for excusing state-court procedural default on federal habeas. Under the usual rule, “ ‘cause’ ... must be something external to the petitioner, something that cannot fairly be attributed to him.” Coleman, 501 U.S. at 753, 111 S.Ct. 2546. State PCR counsel’s “ignorance or inadvertence” cannot constitute “cause” under this rule. Id.

PCR counsel acts as “the petitioner’s agent ..., and the petitioner must ‘bear the risk of attorney error’ ” because there is no constitutional right to counsel in state PCR proceedings. Id. at 753-54, 111 S.Ct. 2546 (quoting Murray v. Carrier, 477 U.S. 478, 488, 106 S.Ct. 2639, 91 L.Ed.2d 397 (1986)).

To show “prejudice” under the usual rule, the “habeas petitioner must show ‘not merely that the errors at ... trial created a possibility of prejudice, but that they worked to his actual and substantial disadvantage, infecting his entire trial with error of constitutional dimensions.’ ” Murray, 477 U.S. at 494, 106 S.Ct. 2678 (quoting) United States v. Frady, 456 U.S. 152, 170, 102 S.Ct. 1584, 71 L.Ed.2d 816 (1982) (omission and emphasis in original). The strict “cause” and “prejudice” rule, first articulated by the Court in 1977 in Wainwright v.

Sykes, replaced the more lenient “deliberate bypass” rule that had been established in Fay v. Noia, 372 U.S. 391, 438-39, 83 S.Ct. 822, 9 L.Ed.2d 837 (1963). The Court in Wainwright justified the strictness of the new rule as necessary to prevent competent defense counsel from “sandbagging” the prosecution at trial. 433 U.S. at 89, 97 S.Ct. 2497.

The Court explained that the rule discouraged “ ‘sandbagging’ on the part of defense lawyers, who may take their chances on a verdict of not guilty in a state trial court with the intent to raise their constitutional claims in a federal habeas court if their initial gamble does not pay off.” Id. Sandbagging might consist, for example, of competent defense counsel deliberately failing to make a constitutional objection to testimony of a key prosecution witness, with the result that neither the court nor the prosecutor takes corrective action during the trial.

Then, in the event that the defendant is convicted, defense counsel could raise for the first time on federal habeas the constitutional objection he deliberately failed to make during trial, with the result that the conviction would be set aside. The concern that gave rise to the strict “cause” and “prejudice” rule is not at issue in a Martinez motion. There is no concern about competent defense counsel who might “sandbag” at trial. The premise of Martinez is incompetent counsel. Indeed, the premise is two incompetent counsel — trial counsel and state PCR counsel.

This quite different circumstance is reflected in the Court’s more lenient rule in Martinez for excusing procedural default. The Court justified the rule by emphasizing the importance of the right to effective assistance of trial counsel, which the Court called “a bedrock principle in our justice system.” Martinez, 132 S.Ct. at 1317. “[T]he limited nature of the qualification to Coleman adopted here reflects the importance of the right to the effective assistance of trial counsel and Arizona’s decision to bar defendants from raising ineffective-assistance claims on direct appeal.” Id. at 1320.

Under the new Martinez rule, a procedural default by state PCR counsel in failing to raise trial-counsel IAC is excused if there is “cause” for the default.

The Court wrote in Trevino, summarizing its holding in Martinez: We consequently read Coleman as containing an exception, allowing a federal habeas court to find “cause,” thereby excusing a defendant’s procedural default, where (1) the claim of “ineffective assistance of trial counsel” was a “substantial” claim; (2) the “cause” consisted of there being “no counsel” or only “ineffective” counsel during the state collateral review proceeding; (3) the state collateral review proceeding was the “initial” review proceeding in respect to the “ineffective-assistance-of-trial-counsel claim”; and (4) state law requires that an “ineffective assistance of trial counsel [claim] ... be raised in an initial-review collateral proceeding.” Martinez, [132 S.Ct. at 1318-19, 1320-21], Trevino, 133 S.Ct. at 1918.

The Court thus clarified that these are the only four requirements to overcome a procedural default under Martinez. Of the four requirements for “cause,” we need not pause over the third or fourth. The third is self explanatory. The fourth was modified in Trevino, as explained above. The first and second requirements, however, merit attention. The first requirement, that the prisoner show a “substantial” underlying trial-counsel IAC claim, may be seen as the Martinez equivalent of the “prejudice” requirement under the ordinary “cause” and “prejudice” rule from Wainwright.

The second requirement, that there have been “ ‘no counsel’ or only ‘ineffective’ counsel,” may be seen as the Martinez equivalent of the “cause” requirement of the rule from Wainwright. With respect to the first requirement, that there be a “substantial” claim, the Court wrote that a prisoner must demonstrate that the underlying ineffective-assistance-of-trial-counsel claim is a substantial one, which is to say that the prisoner must demonstrate that the claim has some merit. Cf. Miller-El v.

Cockrell, 537 U.S. 322 [123 S.Ct. 1029, 154 L.Ed.2d 931] (2003) (describing standards for certificates of appealability to issue). Martinez, 132 S.Ct. at 1318-19.

Under the standard for issuing a certificate of appealability, which the Court incorporated in its definition of substantiality, “a petitioner must show that reasonable jurists could debate whether (or, for that matter, agree that) the petition should have been resolved in a different manner or that the issues presented were adequate to deserve encouragement to proceed further.” Miller-El, 537 U.S. at 336, 123 S.Ct. 1029 (internal quotation marks and alterations omitted).

Stated otherwise, a claim is “insubstantial” if “it does not have any merit or ... is wholly without factual support.” Martinez, 132 S.Ct. at 1319. The second requirement, that there have been “no counsel” or “ineffective” counsel, does not demand a showing of prejudice beyond that demanded under the first requirement. The Comb described the two cases to which the second requirement applies: The first is where the state courts did not appoint counsel in the initial-review collateral proceeding for a claim of ineffective assistance at trial.

The second is where appointed counsel in the initial-review collateral proceeding, where the claim should have been raised, was ineffective under the standards of Strickland v. Washington. Id. at 1318. In the first of the two posited cases, the second requirement is satisfied simply by showing that the prisoner was not represented by counsel during state PCR proceedings.

There is no need to show “prejudice” resulting from the failure of the pro se prisoner during the state PCR proceeding to raise a claim of trial-counsel IAC, over and above the need to satisfy the first Martinez requirement that the underlying trial-court IAC claim be “substantial.” In the second of the two posited cases, the Court did not specify the manner in which Strickland should be applied. We conclude, for the narrow purpose of satisfying the second Martinez requirement to establish “cause,” that a prisoner need show only that his PCR counsel performed in a deficient manner.

A prisoner need not show actual prejudice resulting from his PCR counsel’s deficient performance, over and above his required showing that the trial-counsel IAC claim be “substantial” under the first Martinez requirement. This reading of the requirements in the second-posited case, where the prisoner had PCR counsel, is necessary to harmonize the second Martinez requirement with the rest of the Martinez framework.

If a prisoner who had PCR counsel were required to show prejudice, in the ordinary Strickland sense, resulting from his PCR counsel’s deficient performance in order to satisfy the second Martinez requirement, the prisoner would have to show, as a condition for excusing his procedural default of a claim, that he would succeed on the merits of that same claim.

But if a prisoner were required to show that the defaulted trial-counsel IAC claims fully satisfied Strickland in order to satisfy the second Martinez requirement, this would render superfluous the first Martinez requirement of showing that the underlying Strickland claims were “substantial” — that is, that they merely had “some merit.” See Martinez, 132 S.Ct. at 1318-19. Our conclusion is reinforced by Justice Breyer’s recent statement in Gallow v. Cooper, 570 U.S.——, 133 S.Ct. 2730, 186 L.Ed.2d 935 (2013) (statement “respecting the denial of the petition for writ of certio-rari”).

Justice Breyer, the author of Trevino, indicated in Gallow that once a finding of “cause” under Martinez has been made, a federal habeas court may proceed to the merits of the trial-counsel IAC claim under Strickland. He wrote, “The ineffective assistance of state habeas counsel might provide cause to excuse the default of the claim, thereby allowing the federal habeas court to consider the full contours of Gallow’s ineffective-assistance claim.” Id. at 2731. That is, cause and prejudice under Strickland are determined separately from, and after, a determination of “cause” under Martinez.

We therefore read the Court’s reference to Strickland in the second-posited case of the second requirement (where the prisoner had PCR counsel) to mean the same thing as in the first-posited case (where the prisoner was pro se in PCR proceedings). That is, in both of the posited cases, no showing of prejudice from PCR counsel’s deficient performance is required, over and above a showing that PCR counsel defaulted a “substantial” claim of trial-counsel IAC, in order to establish “cause” for the procedural default. B.

Discovery and Evidentiary Hearing Martinez does not apply to claims that were not procedurally defaulted, but were, rather, adjudicated on the merits in state court. For procedurally defaulted claims, to which Martinez is applicable, the district court should allow discovery and hold an evidentiary hearing where appropriate to determine whether there was “cause” under Martinez for the state-court procedural default and to determine, if the default is excused, whether there has been trial-counsel IAC.

The Court recognized in Martinez that determining whether there has been IAC often requires factual development in a collateral proceeding. The Court emphasized that IAC claims can require “investigative work” and development of an “evidentiary basis” that “often turns on evidence outside the trial record.” Martinez, 132 S.Ct. at 1317; cf. id. at 1318 (explaining that some states require delaying trial-counsel IAC claims until post-conviction proceedings because “[djirect appeals, without evi-dentiary hearings, may not be as effective as other proceedings for developing the factual basis for the claim.”).

For example, to determine whether an attorney’s performance was deficient, it is often necessary to ask the attorney to state the strategic or tactical reasons for his or her actions. To determine prejudice, it is often necessary to authorize discovery and conduct an evidentiary hearing to assess the effect of the attorney’s deficient performance. The Supreme Court held in Cullen v. Pinholster,—U.S.-, 131 S.Ct. 1388, 179 L.Ed.2d 557 (2011), that a federal ha-beas court is ordinarily confined to the evidentiary record made in state court.

However, Pinholster does not prevent a district court from holding an evidentiary hearing in a Martinez case. Pinholster applies when a “claim” has been “ ‘adjudicated on the merits in State court proceedings.’ ” Id. at 1398 (quoting 28 U.S.C. § 2254(d)). But Pinholster’s predicates are absent in the context of a procedurally defaulted claim in a Martinez case in which a habeas petitioner seeks to excuse his default.

First, “cause” to excuse a procedural default under Martinez is not a “claim.” A finding of IAC by the PCR counsel under Martinez is only an “equitable” ruling that there is “cause” excusing the state-court procedural default. Martinez, 132 S.Ct. at 1319-20. Second, in a Martinez case, neither the underlying IAC claim nor the question of PCR-counsel ineffectiveness has been adjudicated on the merits in a state-court proceeding.

Martinez would be a dead letter if a prisoner’s only opportunity to develop the factual record of his state PCR counsel’s ineffectiveness had been in state PCR proceedings, where the same ineffective counsel represented him. See Strickland, 466 U.S. at 694, 104 S.Ct. 2052 (noting the unfairness of applying the restrictive “newly discovered evidence standard” where ineffective assistance of counsel was the reason the evidence was not discovered earlier). The same is true of the factual record of his trial-counsel’s ineffectiveness.

In deciding whether to excuse the state-court procedural default, the district court thus should, in appropriate circumstances, allow the development of evidence relevant to answering the linked Martinez questions of whether there was deficient performance by PCR counsel and whether the underlying trial-counsel IAC claims are substantial. If the district court holds an evidentiary hearing before ruling on the Martinez motion, evidence received at that hearing is not subject to the usual habeas restrictions on newly developed evidence.

Evidentiary hearings to develop the factual basis of a “claim” are ordinarily governed by 28 U.S.C. § 2254(e)(2). But as we have already noted, a prisoner making a Martinez motion is not asserting a “claim” for relief but instead is seeking, on an equitable basis, to excuse a procedural default. See Martinez, 132 S.Ct. at 1319-20; cf. id. at 1320 (finding that § 2254® does not apply to a Martinez motion because “cause” to overcome a procedural default “is not synonymous with a ‘ground for relief ”).

Indeed, even with respect to the underlying trial-counsel IAC “claim,” given that the reason for the hearing is the alleged ineffectiveness of both trial and PCR counsel, it makes little sense to apply § 2254(e)(2). The Court made the nature of the problem clear in Strickland: Even when the specified attorney error results in the omission of certain evidence, the newly discovered evidence standard is not an apt source from which to draw a prejudice standard for ineffectiveness claims.

The high standard for newly discovered evidence claims presupposes that all the essential elements of a presumptively accurate and fair proceeding were present in the proceeding whose result is challenged. An ineffective assistance claim asserts the absence of one of the critical assurances that the result of the proceeding is reliable, so finality concerns are somewhat weaker and the appropriate standard should be somewhat lower. 466 U.S. at 694,104 S.Ct. 2052. C.

New Trial-Counsel IAC Claims The fact that some trial-counsel IAC claims may have been properly raised by the allegedly ineffective state PCR counsel does not prevent a prisoner from making a Martinez motion with respect to trial-counsel claims that were not raised by that counsel. Nothing in Martinez suggests that a finding of “cause” excuses procedural default only when state PCR counsel raised no claims of trial-counsel IAC whatsoever.

Rather, Martinez authorizes a finding of “cause” excusing procedural default of any substantial trial-counsel IAC claim that was not raised by an ineffective PCR counsel, even if some trial-counsel IAC claims were raised. The Court wrote in Martinez: Where ... the initial-review collateral proceeding is the first designated proceeding for a prisoner to raise a claim of ineffective assistance at trial, the collateral proceeding is in many ways the equivalent of a prisoner’s direct appeal as to the ineffective-assistance claim....

As Coleman recognized, an attorney’s errors during an appeal on direct review may provide cause to excuse a procedural default; for if the attorney appointed by the State to pursue the direct appeal is ineffective, the prisoner has been denied fair process and the opportunity to comply with the State’s procedures and obtain an adjudication on the merits of his claims. Id. at 1317. The concerns expressed by the Court apply equally to all claims of trial-counsel IAC. As the Court recognized, an ineffective PCR counsel’s failure to raise a valid claim of trial-counsel IAC is a denial of fair procedure.

It is no less a denial of fair procedure if the ineffective PCR counsel happened to raise other, less viable, claims of trial-counsel IAC. We therefore read Martinez to allow new claims of trial-counsel IAC, asserted for the first time on federal habeas, even if state PCR counsel properly raised other claims of trial-counsel IAC. The Court implicitly confirmed this reading in Trevino, where it held that Martinez applied to Trevino’s procedurally defaulted trial-counsel IAC claims even though Trevino’s state PCR counsel had presented other trial-counsel IAC claims during the initial-review collateral proceeding.

See Trevino, 133 S.Ct. at 1915-16. III. Remand to the District Court We remand to the district court under Martinez to determine, in the first instance, whether there is “cause” to excuse state PCR counsel’s procedural default. If the district court finds that there was “cause,” it should then address on the merits the substantial trial-counsel IAC claims that it previously held procedurally defaulted under pre-Marimez law. As to these claims, the two-step cause and prejudice test of Strickland applies.

Depending on its ruling on the merits of these new trial-counsel IAC claims, the district court may have occasion to revisit its earlier conclusion in this case that the deficient performance of trial counsel did not cause prejudice within the meaning of Strickland. A standard practice, in habeas and non-habeas cases alike, is to remand to the district court for a decision in the first instance without requiring any special justification for so doing. In cases where there is little doubt about the correct answer, we will sometimes decide an issue in the first instance rather than remand to the district court.

But our general assumption is that we operate more effectively as a reviewing court than as a court of first instance. We see no reason why a remand to the district court in a Martinez case should be treated differently from a remand in other cases. Indeed, we have remanded Martinez motions for initial decision by the district court on prior occasions, including in Martinez itself. See, e.g., Martinez v. Ryan, 680 F.3d 1160 (9th Cir.2012).

The dissenters in this case believe that Detrich’s Martinez motion has so little merit that we can confidently decide it ourselves, with little risk that we will misunderstand the trial-court record or that we will mistakenly conclude that evidentia-ry development will not alter our view of the case. When we remand to the district court for a decision in the first instance, we usually do not provide a preliminary analysis of the relevant evidence or law. We simply leave it to the district judge to decide the remanded matter.

Here, however, we feel compelled to respond to the dissent’s conclusion that Detrich’s defaulted claims of trial-counsel IAC in this capital case are so obviously meritless that we can safely decide his motion. We therefore describe the state trial record and provide an analysis of why we believe that Detrich’s new trial-counsel IAC claims are sufficiently plausible that we should remand to the district court to decide in the first instance whether they are “substantial” within the meaning of Martinez.

The central question is whether any of Detrich’s newly presented trial-counsel IAC claims prejudiced him at sentencing. Even if, based on new evidence, the jury were unwilling to find beyond a reasonable doubt that Detrich was the actual killer, it would almost certainly convict him of felony murder. Under the circumstances of this case, a felony-murder conviction would likely still make Detrich death-eligible. However, the trial judge did not sentence Detrich to death for felony murder.

Instead, the judge sentenced Detrich to death based on his own conclusion, beyond a reasonable doubt, that Detrich rather than Charlton killed Souter. Thus, in practical effect, Detrich’s trial-counsel IAC claims are primarily directed to his sentence rather than his conviction. The question is whether, if the evidence that Charlton was the actual killer were stronger — and the evidence against Detrich therefore weaker — Detrich would nonetheless have been sentenced to death. To evaluate substantiality under Martinez, it is necessary to assess the evidence at trial.

If the evidence that Detrich killed Souter were overwhelming and unassailable, as the dissent contends, then any deficient performance by Detrich’s counsel in failing to present additional evidence would be very unlikely to have altered the sentencing judge’s conclusion beyond a reasonable doubt that Detrich was the killer.

In that event, we could safely conclude that Detrich has no real chance of showing that his new trial-counsel IAC claims are “substantial.” But if the trial evidence were close, then it would not take too much new exculpatory evidence to call into question the trial judge’s sentencing decision. We need look no further than the jury verdict to get a general sense of the strength of the evidence. Three out of the twelve jurors refused to convict Detrich of premeditated murder. They were willing to convict Detrich of only felony murder.

The dissent contends that these three jurors could have concluded that Detrich was the actual killer, but that he lacked the premeditation necessary for a first-degree murder conviction. See Dissent at 1269 n.7. However, neither the prosecution nor Detrich advocated such a theory at trial.

In fact, the prosecution strenuously argued against finding either second-degree murder or manslaughter, stating during closing argument that “[tjhere is no way ... that one can argue it is not premeditated.” None of the witnesses supported a theory of non-premeditated killing either, so there is little reason to believe that the jurors adopted this view of events. A description of the evidence at trial shows why the jurors may have been hesitant to find that Detrich killed Souter. We first describe the evidence. We then discuss four of Detrich’s trial-counsel IAC claims that may be “substantial” under Martinez.

A. Evidence at Trial Alan Charlton and David Detrich, two white men, were driving in Charlton’s car one evening after drinking heavily. They picked up a black woman, Elizabeth Souter. They drove with Souter to a bar where Souter knew to get drugs. They then drove to Souter’s house. In the house were Souter’s two adult daughters and an adult white woman. The drugs were bad, and Detrich threatened Souter with a small knife because they were bad. Detrich and Charlton then drove away from the house with Souter. Souter’s body was discovered in the desert two days later. She had been stabbed numerous times.

The critical question in the case was who killed Souter — Detrich or Charlton. The following evidence was presented at trial. We describe the evidence that Det-rich killed Souter, including all of the evidence described by the dissent. Unlike the dissent, we also describe the evidence that Charlton killed her. 1. Events Inside the House Three adult women were in Souter’s house when she came home with Detrich and Charlton — Gwen and Caprice Souter, Souter’s daughters; and Tami Winsett, a white woman. Gwen had a baby and never came out of the bedroom. She never saw Detrich or Charlton.

Caprice and Winsett testified that they, Souter, and the two men were together in the living room. They testified that Det-rich became angry when he and Souter discovered that the drugs were bad. They testified that Souter took off her dress and lay down on a mattress on the floor. Win-sett testified that Detrich then lay down next to Souter and told her that she would have to have sex with him because the drugs were bad. Caprice and Winsett testified that Detrich held a small knife to Souter’s throat. No one testified that Det-rich in fact had sex with Souter on the mattress.

Winsett testified that she sat on the couch in the living room with Charlton. She testified that Charlton told her he had previously killed someone.

Winsett added: “He told me if I lied to him, if I stabbed him in the back, or he said if I fucked him over, he would kill me, too.” Gwen Souter, who was in the bedroom and who never saw Detrich or Charlton, testified that she heard Detrich say, “I will kill you.” She identified Detrich as the speaker based solely on the sound of his voice, after hearing Charlton speak in court months later. (Detrich did not testify.) Neither Caprice nor Winsett, who were in the living room with Detrich and the victim, testified that they heard Det-rich say “I will kill you.” The only witness in the room who testified that she heard a man say he would kill someone was Win-sett, who testified that Charlton had said he would kill her. 2.

Events Outside the House Souter was taken outside and put into Charlton’s car. It is not clear who forced her into the car. Winsett was the only one of the three women who testified about events outside. She had left the house to call the police and was half a block away. It was late at night, and there were no street lights. Winsett testified that there was some moonlight. “[I]t was kind of dark but you could see a little bit.” Win-sett could not determine who forced Souter into the car. She testified, “I saw them. I don’t know who put her in the car or what.

They were all on the passenger side and then they put her in the car.” Charlton testified that Detrich forced Souter into the car at knife-point. But Winsett never testified that she saw a knife. It is unlikely that Detrich would have used the knife he had held while still in the house. After the murder, a small knife was found in the living room. That knife was almost certainly the knife Det-rich had held to Souter’s throat when he lay next to her on the mattress.

An expert testified that Charlton’s fingerprints were found on the inside of the driver’s side window while Detrich’s prints were found on the exterior of the car on the front passenger-side fender. There was no testimony regarding when these prints were left, whether in the days leading up to the murder, on the night of the murder itself, or sometime thereafter. 3. Charlton’s Testimony Regarding Events in the Car Charlton testified that he drove the car away from Souter’s house, with Detrich in the middle of the front seat and Souter in the front seat next to the passenger-side door.

Charlton did not explain how Det-rich ended up sitting in the middle if Det-rich had forced Souter into the car from the passenger side. He also did not explain why they would have allowed Souter, a kidnapping victim, to sit by the passenger-side door where she could escape by simply opening the door. Charlton testified that while driving he looked over and saw that Detrich was “on top of’ Souter, “humping her.” Charlton stated that he “couldn’t tell” whether or not Detrich was having actual intercourse. In a previous trial, Charlton had testified that he saw Detrich “raping” Souter.

He had stated multiple times at that trial that he saw actual intercourse occurring. Both of these accounts differed from Charlton’s initial statement to police, which was introduced in the final trial. In that statement, Charlton said that he could not tell whether Detrich was having either oral or vaginal sex with Souter because he was “paying attention to the road.” When confronted with this inconsistency, Charl-ton testified that he had “remembered some things” since the time of his initial statement. The state’s expert pathologist testified that he had found no physical evidence of sexual assault.

Charlton testified that he looked over later and saw that Souter’s throat had been slit. He testified that Detrich hit Souter and asked her several times who had provided the bad drugs; Souter “just gurgled something” in response. Charlton testified that Detrich then “asked me if I wanted a shot of it, it is dead but it is warm.” Charlton testified that he declined, and that they then drove to the desert where Detrich deposited Souter’s body. As the dissent notes, some of the forensic evidence regarding the manner of death was consistent with Charlton’s account.

For example, the state’s pathologist testified that Souter received forty cutting or stabbing wounds as well as additional blunt-force injuries. Charlton’s knowledge of the manner of death indicates that he was present at Souter’s murder, but it does nothing to establish whether he or Detrich killed her. 4. Knives Found on Detrich and Charlton The police found a knife on Detrich when he was arrested. Charlton testified that the knife was his, and that he had not given it to Detrich. Charlton claimed that the knife must have fallen out of his pocket and that Detrich must have picked it up.

Charlton testified that the knife found on Detrich was the only knife he saw Detrich holding on the night of the murder. He testified that he later saw the knife covered in blood. The prosecutor suggested at the outset of the trial that Detrich killed Souter with this knife. In its narrative of the evidence, the Arizona Supreme Court also suggested that Det-rieh killed Souter with this knife. See Detrich II, 932 P.2d at 1332; Dissent at 56-57. The knife found on Detrich was not the murder weapon.

The state’s expert who performed the autopsy testified that Souter’s wounds were 0.9 centimeters wide on the skin’s surface. The blade of the knife found on Detrich was 1.7 centimeters wide. The expert testified that Detrich’s knife could not have caused Souter’s stab wounds because it was much too wide. The police found a knife on Charlton when they arrested him. The knife was engraved with the letters “FTW,” which stood for “Fuck The World.” Charlton testified that this knife was in his pocket on the night of the murder.

The state’s expert testified that this knife was more consistent with Souter’s wounds than the knife found on Detrich. 5. Testimony of Phillip Shell Phillip Shell testified that Charlton confessed to him in jail that he had stabbed Souter. Shell testified that Charlton had bragged to him about having lied in court, and that he had bragged about getting Detrich the death penalty. According to Shell, Charlton told him that after leaving Souter’s house he parked the car behind a bar. Detrich started kissing Souter in the front seat.

Charlton grew angry that Detrich was kissing a black woman, especially after she had short-changed him on the drugs. Charlton pulled out a knife and stabbed Souter. Souter “went crazy,” and Charlton cut her throat. After killing Souter, Charlton climbed into the back seat and passed out. Some of the evidence at trial corroborated Shell’s story. Charlton testified that they were driving when Detrich killed Souter; that they then drove to where Detrich left Souter’s body; and that they then drove straight to his friend William Carbonell’s house.

The entire trip described by Charlton at trial would have taken about forty-five minutes. But according to other evidence at trial, several hours elapsed between the time they left Souter’s home and the time they arrived at Carbonell’s house. Charlton never explained how they spent the rest of the time. Shell explained that Charlton told him that he spent this time sleeping in the back seat of the car. Charlton also never explained why the police found a pair of bloodstained jeans in the back seat. The stains were consistent with Charlton having wiped his hands on the jeans before going to sleep in the back seat.

Unlike other jailhouse snitches, Shell testified against the interest of the prosecution. He received no benefit from testifying. In fact, he was kept in custody after he was acquitted in his own case in order to testify in Detrich’s. Charlton admitted that he had spoken to Shell in jail, and that he had never had any fights or disagreements with him. 6.Testimony of William Carbonell William Carbonell worked with Charlton at Ocotillo Motors in Benson, Arizona. Carbonell and Charlton were friends. Carbonell had sold Charlton the car that Charlton drove on the night of the murder.

A day after the murder, Carbonell drove Charlton to work from Tucson to Benson, a distance of about 45 miles. Detrich also worked at Ocotillo Motors. Showing his degree of friendship with Det-rich, Carbonell testified that Detrich had worked at Ocotillo Motors for six months, and that the two of them had gone drinking together on multiple occasions. However, the owner of Ocotillo Motors testified that Detrich had been an employee for only a month. Carbonell testified at trial that Charlton and Detrich arrived at his house the morning after the murder. He testified that Detrich was covered in blood.

He testified that Charlton had blood on only his right side. Carbonell had initially told police that Detrich and Charlton were both covered with blood without specifying that Charlton had blood on only one side. His earlier statement to police was introduced at trial. The prosecutor used Carbonell’s trial testimony to emphasize his theory of the case — that Charlton had been driving while Detrich killed Souter, which accounted for Charlton having blood on only his right side. The evolution of Carbonell’s story was obviously helpful to the prosecution.

The prosecutor who elicited Carbo-nell’s testimony was later disbarred for suborning perjury in another capital case. See In re Peasley, 208 Ariz. 27, 90 P.3d 764 (2004) (en banc). Carbonell testified that Detrich confessed to him on the morning after the murder that he had stabbed Souter. Carbonell’s testimony that Detrich had confessed to him was critical to the state’s case. As we discuss below, however, Carbonell had earlier made an inconsistent statement to investigators.

Detrich’s attorney failed to introduce that statement into evidence. 7.Scratches on Charlton There were scratches or cuts on Charl-ton’s arm at his arrest. Charlton never explained where they came from. There were also wounds on Souter’s hands. The state’s expert testified that Souter’s wounds were consistent with “defensive injuries” from trying to fight off an attacker. There were no scratches or cuts on Detrich. 8.Charlton’s Racism Two witnesses testified that Charlton was prejudiced against black people.

First, Phillip Shell testified that Charlton referred to Souter as “that black bitch” and referred to Detrich as a “nigger lover.” Second, Charlton’s wife, Deborah Charl-ton, testified that Charlton did not like black people “at all” and referred to them as “mud ducks.” Deborah’s testimony was undermined by the fact that she and Charlton were in the midst of an acrimonious divorce. Deborah’s testimony was further undermined by a support letter she had written to the court some time earlier, during Charlton’s sentencing for his participation in the kidnapping and death of Souter.

In that letter, she made no mention of any racism. 9.Other Evidence of Charlton’s Violent Tendencies Charlton testified at trial that he wore Grim Reaper earrings — the “sign of death” — on the night of the murder. He testified that he was very upset at the time because his marriage had fallen apart and his wife had given their kids up for adoption without his knowledge. He testified, “I didn’t care if I lived or died.” Deborah Charlton testified' that Charl-ton had an “obsession” with Bruce Lee.

She testified that he “would go to an import store and constantly look at the swords” and other martial arts paraphernalia. The police found a third knife Charlton owned at his house when they arrested him: a “[d]ouble-edged hunting knife.” Deborah testified that Charlton sometimes carried this knife in a sheath down his leg. She also testified that Charlton threatened to kill her with it the night she left him. B. New Trial-Counsel IAC Claims The dissent rejects several of Detrich’s trial-counsel IAC claims for reasons unrelated to their substance.

First, the dissent points out that some of Detrich’s trial-counsel IAC claims were adjudicated on the merits by the first state PCR court and were thus never procedurally defaulted. We agree with the dissent that Martinez does not apply to such claims, and that the following three claims were' adjudicated on the merits: (1) trial counsel failed to bring witness Shell to court to testify; (2) trial counsel failed to rehabilitate jurors regarding death-penalty qualification; and (3) trial counsel failed to object to prosecutorial vouching during cross-examination of Charlton.

However, we do not agree with the dissent as to one claim. Detrich claims that three pieces of evidence should have been subjected to further testing. We agree with the district court that this claim was not adjudicated but was, rather, procedurally defaulted. Second, the dissent would hold that two of Detrich’s trial-counsel IAC claims were waived for purposes of Martinez because he did not raise them with sufficient specificity in his motion to remand to the district court.

Those claims are: (1) trial counsel failed to interview two key witnesses, Darci and Donald Bell; and (2) trial counsel failed to cross-examine William Carbonell by introducing a prior inconsistent statement. We do not agree with - the dissent that these claims are waived. Detrich moved in this court for a remand to the district court for a determination of “cause” under Martinez after supplemental briefing in that court. His motion did not ask us to make the initial decision whether there was “cause.” Det-rich’s then-pending appeal before our en banc court involved other issues.

None of the briefs or excerpts of record had been prepared with Martinez in mind. Indeed, they could not have been, since the Court decided Martinez after briefing on appeal was complete. Because Detrich moved in our court for a remand and not for a ruling under Martinez, and because he was subject to our relatively stringent page limits for a motion to this court, he explicitly stated that he was providing only a “summary” of his underlying trial-counsel IAC claims. He referred us to his amended habeas petition for a fuller description of those claims.

To the extent the dissent would reach the merits of Detrich’s contention that he can show “cause” under Martinez, it would hold that none of the procedurally defaulted trial-counsel IAC claims is “substantial.” Unlike the dissent, we do not decide whether Detrich’s procedurally defaulted trial-counsel IAC claims are “substantial.” We remand for the district court to decide that question in the first instance. However, we feel compelled, given the dissent, to show that some of Detrich’s trial-counsel IAC claims are sufficiently plausible to warrant remanding to the district court.

The following four procedurally defaulted claims support our decision to remand to the district court for a decision under Martinez in the first instance. 1. Failure to Interview Darci and Donald Bell Detrich alleges that his trial counsel was ineffective for failing to interview Detrich’s sister and brother-in-law, Darci and Donald Bell. Neither Darci nor Donald testified at trial. In an affidavit presented to the district court, Darci states that she saw Detrich the morning after the murder. She states that she washed Detrich’s bib overalls and that they had no blood on them.

She states, further, that the police officers investigating the crime were extremely aggressive, searching her house with guns drawn and without a warrant. Finally, she states that the officer who interviewed her “stopped the tape a few times to tell me what I was supposed to say.” Donald describes similar police conduct in his affidavit. According to Darci and Donald, Detrich’s attorneys never contacted them. Detrich’s trial counsel was on notice in numerous ways that the Bells had useful information. Darci states that she “left many messages” for Detrich’s attorney, but he never responded.

Darci reports that Detrich gave his attorney her contact information as well. Charlton testified in one of the earlier trials that Detrich had asked him for a ride to Darci’s house the morning after the murder. Charlton testified that Detrich said Darci “was going to wash his clothes for him.” Charlton’s testimony indicated that there was blood on Detrich’s clothes that needed to be washed off. It also made clear that Darci could have been a key witness. Trial counsel’s failure to contact Darci and call her as a witness may have prejudiced Detrich in several ways.

First, Dar-ci’s eyewitness statements that there was no blood on Detrich’s clothes would have directly rebutted Charlton and Carbonell’s accounts. Darci did not see Detrich immediately after the murder, but she saw him later the same morning, after he left Car-bonell’s house. Detrich did not return home in the interim, and there is no evidence that he washed his clothes before arriving at Darci’s. To the contrary, Charlton testified that Detrich asked Charlton to take him to Darci’s house specifically so that she could wash his clothes for him.

Darci states unambiguously that she washed Detrich’s clothes, including his “bib overalls,” and that they did not have blood on them. ■ Charlton testified that Detrich was wearing bib overalls the night of the murder and that they were “saturated” with blood. Detrich was wearing the overalls when he arrived at Carbonell’s house allegedly covered in blood, as Carbo-nell recounted to investigators: Q: What were they wearing, do you recall? [Carbonell]: [Detrich] had a set of bib overalls on that he had on the day before, and [Charlton], he just had a regular pair of Levis and a shirt on, I think.

And, they was both all covered with blood. (Emphasis added.) Darci’s testimony would have directly rebutted Charlton and Carbonell’s accounts that the same overalls were saturated with blood. The dissent speculates, somewhat oddly and entirely without evidence, that Detrich might have been carrying multiple pairs of “bib overalls.” Second, Darci’s testimony that the police instructed her to lie would have suggested that other witnesses were not testifying truthfully. For example, her testimony would have raised additional doubts about why Carbonell’s story about-the blood on Charlton had improved at trial.

Finally, Darci stated in her affidavit that Charl-ton’s earring was “an Aryan Nation symbol.” As we discuss next, testimony about Charlton’s connection to the Aryan Nation (or Brotherhood) would have strengthened the evidence of Charlton’s motive for the killing. 2. Failure to Investigate and Introduce Evidence Against Charlton Detrich alleges that his trial counsel was ineffective for failing to interview or investigate Charlton and for failing to introduce evidence of Charlton’s connections to the Aryan Brotherhood. Detrich’s trial counsel never tried to interview Charlton before trial.

He also failed to investigate Charlton’s connections to the Aryan Brotherhood, a violent white supremacist organization. James Williams, defense counsel’s investigator, stated in a post-trial affidavit: I was not asked by defense counsel to pursue or conduct interviews of witnesses pertaining to the issue of [Charl-ton’s] involvement with the Aryan Brotherhood; I believe this was important since [Charlton] was known to be involved with the Aryan Brotherhood, and have a hatred of African-Americans, whereas Mr. Detrich did not have such a hatred; the victim in this matter was African-American.

Deborah Charlton told investigator Williams before trial that her husband was involved with the Aryan Brotherhood while in jail. Charlton had been in jail for three years at the time. There were numerous people in the jail who might have known about Charlton’s Aryan Brotherhood affiliations. Further, Charlton might also have discussed with other inmates his involvement in Souter’s murder.

Williams stated in a post-trial interview that this was common in his experience, noting that “they can’t keep their mouths shut for a month over in Pima County jail so after three years they’re all bragging there.” Phillip Shell testified that this was precisely what Charlton had done in confessing to him. Williams suggested to defense counsel that he go to the jail, request Charlton’s records, and interview officers and Charl-ton’s former cell mates. Williams thought it was “really critical” to perform this investigation. Defense counsel rejected Williams’ suggestion.

Defense counsel’s failure to develop and introduce evidence of Charlton’s Aryan Brotherhood connection, as well as additional evidence of his racism, may have prejudiced Detrich. Even though Charl-ton’s trial testimony was impeached in some respects, the jury would have found it difficult to accept that he was the actual killer without a coherent explanation for why he would have killed Souter. Charlton’s racist attitudes were his alleged motivation for killing Souter, but evidence at trial of Charlton’s racism was weak. Only two witnesses testified to his racism: Phillip Shell and Deborah Charl-ton.

Deborah had credibility problems since she and Charlton were in the midst of a contentious divorce. The prosecutor also impeached Deborah with her prior letter to the court, in which she had described Charlton in entirely positive and non-racist terms. In closing arguments, the prosecutor specifically commented on the weakness of the evidence of Charlton’s racism: [Defense counsel] suggested ... Alan Charlton doesn’t like blacks and had a pet name for them. That came from but one witness, Mr. Charlton’s soon-to-be ex-wife. Number one, I would submit to you that the Defendant doesn’t have any burden in this case.

He doesn’t have to call a witness. He doesn’t have to do anything. But what he does have is the subpoena power. I can assure you that if any of that were true, that Charlton didn’t like blacks and has a grudge against them, that would somehow or another explain what happened here, you would have seen a line of witnesses come up to the witness stand. They weren’t called because it simply isn’t true. You had a chance to hear the testimony of the soon-to-be ex-wife of Alan Charlton and she simply lied.

Exactly why, I don’t know, and I don’t— you can use your common sense as well. (Emphasis added.) Additional evidence of Charlton’s racism would have foreclosed the prosecutor’s argument. The prosecutor may also have committed misconduct— to which defense counsel did not object— by giving his personal assurance that there were no other witnesses to testify to Charlton’s racist beliefs and by stating his opinion that a defense witness had lied. See United States v. Necoechea, 986 F.2d 1273 (9th Cir.1993) (prosecutorial misconduct to comment on facts not in evidence); United States v.

Garcia-Guizar, 160 F.3d 511 (9th Cir.1998) (prosecutor impermissi-bly vouched by calling defendant a “liar”). There was no evidence whatsoever introduced at trial about Charlton’s connections to the Aryan Brotherhood. The Aryan Brotherhood is a powerful white supremacist group responsible for a number of race-based murders. Even if the jury believed that Charlton did not like black people, it would have been qualitatively different to know that he was involved in the Aryan Brotherhood, whose members commit racially motivated murders. 3.

Failure to Conduct Forensic Testing Detrich alleges that his counsel was ineffective for failing to conduct independent forensic testing of three pieces of evidence: human hairs found on Souter, the knives alleged to be possible murder weapons, and a needle found at Souter’s house. In his first PCR petition, Detrich alleged that his trial counsel was ineffective for failing adequately to test “the physical evidence.” He specifically named certain pieces of evidence, including the front-seat covers of Charlton’s car, a jacket recovered with the seat covers, and the jeans found in the back seat.

He did not mention the hairs found on Souter, the knives alleged as murder weapons, or the needle. The PCR court denied the eviden-tiary-testing claim raised in that court, specifically discussing the seat covers, the car, and the victim’s fingernail. The district court found that the forensic-testing claim had been properly raised before the first state PCR court as to the pieces of evidence listed in the PCR petition. But it held that the claim had been procedurally defaulted in state court as to the hairs, the knives, and the needle.

The dissent contends that the district court erred in its determination that the trial-counsel IAC claim as to these three pieces of evidence was defaulted. We disagree. Detrich’s non-specific reference to all of “the physical evidence” in his first PCR petition did not sufficiently present the claim to the state court as to each piece of evidence.

Indeed, we have little doubt that, in a non -Martinez setting, the dissenters would agree that such a catchall reference did not exhaust the specific claims later asserted, any more than a PCR petition asserting trial-counsel IAC for violations of “all amendments to the U.S. Constitution” would exhaust all possible constitutional challenges. On the merits, it is easy to see how testing of this evidence could have been useful. For example, the state had begun to test the hairs found on Souter. But the state’s analyst had insufficient time to complete her testing.

As a result, the analyst could not identify who had left the hairs. The analyst specifically admits in her affidavit that the step she skipped might have allowed her to identify the source of the hairs. If it turns out that Charlton’s hairs were found on the victim, it would provide additional evidence that he killed Souter while she struggled against him. Similarly, any additional evidence that the knife found on Charlton was the murder weapon would create further doubt that Detrich was the killer.

The dissent contends that any new forensic evidence implicating Charlton could not have made a difference because the existing evidence at trial was so strong. We have already noted that three jurors do not appear to have believed that Det-rich actually killed Souter. Those jurors convicted him of only felony murder. More specifically, the dissent contends that the forensic evidence at trial corroborated the evidence against Detrich.

The dissent points to the incomplete test results of the hairs found on the victim, as well as to the testing of fingernails and the “examinations for fingerprints, blood, and semen in the car.” Dissent at 1272. We take in turn the four pieces of evidence upon which the dissent relies. None of this evidence demonstrated that Detrich, as opposed to Charlton, was the killer. If anything, it pointed to Charlton as the actual killer. First, one of Souter’s fingernails was found in the car. Souter may well have lost the nail while trying to fight off her attacker with her hands.

This would be consistent with the state pathologist’s testimony that Souter had “defensive injuries.” But the only physical evidence that either Charlton or Detrich had engaged in a struggle was the scratches on Charlton’s arms. Second, Charlton’s fingerprints were found on the driver’s side of the car. This proved little since Charlton owned the car. It was uncontested that he had been the driver on the night of the murder. Det-rich’s fingerprints were found only on the front passenger-side fender.

Third, blood was found on the front-seat covers, including the seat cover on the driver’s side of the car, and on the jeans in the back seat. The blood suggested strongly that Souter had been killed inside the car, but it did not answer the question of who killed her. The bloody seat cover on the driver’s side of the car, as well as the jeans in the back seat, were consistent with Charlton having killed Souter. Fourth, there was no semen found, in the car or elsewhere. The state’s pathologist specifically testified that he found no semen during the autopsy.

Not only did the pathologist testify that he found no semen, he also testified that he had found no physical evidence of sexual assault. The pathologist’s testimony cast doubt on Charlton’s claims about Detrich “humping” or “raping” Souter in the car. It also reinforced Shell’s testimony that Charlton had killed Souter when Detrich was only kissing her. 4. Failure to Cross-Examine William Carbonell Detrich alleges that defense counsel failed to cross-examine Carbonell with a prior inconsistent statement.

Carbonell testified at trial that Detrich had confessed to him that he had killed Souter: Q: Tell the jury what it was [Detrich] told you the second time. [Carbonell]: [Detrich] told me he killed a girl Q: Did he say how he killed her? A: With a knife. Q: Did he say he cut her throat? A: Yes, sir. This testimony was devastating to Detrich. It was also inconsistent with what Carbo-nell had stated in a pre-trial interview with investigators. Carbonell had stated in his pre-trial interview that Detrich had not told him that he had killed Souter.

Carbonell had stated that he had merely “surmised” that Detrich had been the killer. He had stated that Charlton told him that Detrich had killed Souter: [Carbonell]: [Detrich] didn’t say that he killed the girl. But, he said [Charlton] was driving the car. So, I guess — I just sum — surmised that — that— Q: Yeah. A: — he had — he had been the one that cut the girl. ‘Cause [Charlton] told me later on[] that he did cut the girl up. (Emphasis added.) Carbonell’s prior statement thus specifically contradicted his repeated trial testimony that Detrich had confessed to killing Souter.

The jury never heard Carbonell’s prior statement because Detrich’s trial counsel failed to introduce it. Carbonell’s testimony about Detrich’s confession was one of the most damaging pieces of evidence at trial. In sentencing Detrich to death, the trial judge specifically cited Detrich’s alleged confession to Carbonell that he had “ ‘[k]illed some chick for getting bad drugs.’ ” Carbonell’s prior statement would have directly contradicted his own testimony.

Without this testimony, the state’s only direct evidence of guilt came from Charlton, who had an obvious motive to lie and whose story was contradicted by other evidence in the record. The dissent writes that Carbonell’s prior statement was that Detrich had told him that “Charlton was driving at the time” Souter was killed. Dissent at 68. If this were true, Detrich’s reported statement would have implied that he had killed Souter. But Carbonell did not say that. In his prior statement, Carbonell said only that Detrich had told him that Charlton “was driving” the car.

It was uncontested that Charlton was driving the car on the night of the murder. Carbonell never stated that Detrich had told him that Souter had been killed while Charlton was driving. The dissent also asserts that any difference between Carbonell’s testimony and his prior statement was “negligible.” Dissent at 68. We disagree. Carbonell testified at trial that a few hours after the murder, Detrich told him that he had killed Souter by slitting her throat with a knife.

In his prior statement, Carbonell stated that Detrich “didn’t say that he killed the girl.” Carbonell stated that he had merely “surmised” that Detrich was the killer based in part on what Charlton told him. No reasonable trier of fact could find the difference between these statements to be “negligible.” IV. Conclusion We grant Detrich’s motion to remand for the district court to rule, in the first instance, on his Martinez motion. We do not reach Detrich’s non-defaulted sentencing-phase IAC claims.

We cannot properly evaluate those claims at this time because Detrich’s Martinez motion raises additional claims that his trial counsel’s ineffectiveness prejudiced him at sentencing. It would be premature to evaluate prejudice from his non-defaulted claims before we know what additional prejudice might have resulted from the defaulted ones. We will reach the non-defaulted claims, if appropriate, after the district court has decided Detrich’s Martinez motion and any trial-counsel IAC claims for which Detrich’s state-court procedural default is excused.

The en banc panel will retain jurisdiction over any subsequent appeal. VACATED IN PART and REMANDED.