Appeal from the United States District Court for the Middle District of Alabama. (94-01137-CV-A-S), W. Harold Albritton, III, Chief Judge.

Before COX, BIRCH and BLACK, Circuit Judges.

COX, Circuit Judge:

Clyde Cade, an Alabama inmate under a death sentence for murder, appeals the district court's denial of relief on his 28 U.S.C. 2254 petition. We affirm.

I. Background

Cade was originally convicted and sentenced to death in 1978. His conviction was vacated by the United States Supreme Court because at the time the Alabama death-sentencing procedure did not comply with the Eighth Amendment. See Cade v. State, 375 So.2d 802 (Ala.Crim.App.1978), aff'd, 375 So.2d 828 (Ala.1979), vacated, 448 U.S. 903 , 100 S.Ct. 3043, 65 L.Ed.2d 1133 (1980). After Alabama changed that procedure, see Beck v. State, 396 So.2d 645 (Ala.1980), Cade was again tried, convicted, and sentenced to death in 1982, see Cade v. State, 521 So.2d 80 (Ala.Crim.App.1986), aff'd, 521 So.2d 85 (Ala.1987), cert. denied, 488 U.S. 871 , 109 S.Ct. 184, 102 L.Ed.2d 153 (1988).

Cade's conviction is based on events that took place in 1977. Late in the afternoon on August 3, Cade shot the sheriff of Geneva County, Alabama three times.1 SheriffSizemore had responded to a domestic disturbance call reporting that Cade was threatening his girlfriend and her sister. Diane Butts, the sister, testified that when Sizemore called Cade over to his car, Cade responded by saying, "You ain't taking me nowhere." (State R.1-P2 at 114.)2 He did, however, walk to the car. After Sizemore talked with Cade and patted him down, Butts heard Cade say "I ain't going to jail." (State R.1-P2 at 116.) A struggle then ensued, during which Cade wrestled Sizemore's gun away from him. After the Sheriff hopped back into the car to get away, Cade shot him three times. The coroner testified that the shots had killed Sizemore and forensic analysis indicated that the offending bullets came from a gun with Cade's fingerprints. Various witnesses testified to indications that Cade had been drinking, but they disagreed over how much. Cade took the stand in his own defense and testified that he had been threatened on several occasions by Sizemore. He also confirmed that he had been drinking on the day of the homicide and claimed that although he could remember the struggle, he could not recall anything about the shooting itself.

The jury rejected Cade's voluntary-intoxication defense, found him guilty, and recommended the death penalty. The trial judge, after holding a bench hearing and reviewing a presentence investigation report (PSI), sentenced Cade to death.3 The court found two aggravating factors: first, that Cade murdered Sheriff Sizemore while "he was performing an official or job related act of arresting Clyde Cade," and second, that Cade committed the murder "for the purpose of avoiding or preventing a lawful arrest." (State R.2-P14 at 433.) The court also rejected three statutory mitigators: whether Cade was "under extreme duress," "had the capacity to appreciate the criminality of his conduct ... [and] conform his conduct to the requirements of law," or was "under the influence of extreme mental or emotional disturbance." (State R.2-P14 at 435.)4 The court stated that it weighed the aggravating and mitigating circumstances against each other.

As previously noted, Cade's conviction was affirmed on direct appeal. Cade collaterally attacked his conviction under the procedure then established by Ala. Temp. R.Crim. P. 20. Cade's state-court petition included, in relevant part, a Sixth Amendment claim of ineffective assistance of counsel at the sentencing phase alleging failure to adequately investigate, prepare, and present mitigating evidence, and a due process claim alleging that the evidence supporting the two aggravating factors was insufficient. The petition did not include a claim that the trial court gave insufficient consideration to mitigating factors-an omission of importance to this appeal. After an evidentiary hearing, the court denied Cade's petition, concluding that the sentencing-phase ineffective-assistance claim lacked merit, and that the insufficiency-of-the-evidence claim relating to the aggravators had already been decided against Cade on direct appeal. The Alabama Court of Criminal Appeals affirmed. See Cade v. State, 629 So.2d 38 (Ala.Crim.App.), cert. denied (Ala.1993).

Cade then filed the instant federal petition which, after amendment, asserts twenty-three claims. Included among the claims are ineffective assistance of counsel at the sentencing stage and two claims alleging insufficient consideration of and findings regarding mitigation. The petition does not include a claim that the evidence supporting aggravating factors was insufficient. The district court handled the petition in a two-step process, referring to a magistrate judge for report and recommendation the question of which among Cade's claims were procedurally defaulted and the question of which among any undefaulted claims warranted an evidentiary hearing.

In his brief to the magistrate judge, Cade conceded procedural default of his two claims asserting insufficient treatment of mitigation, but also argued that the default was excused under Sawyer v. Whitley, 505 U.S. 333, 112 S.Ct. 2514, 120 L.Ed.2d 269 (1992). As part of the argument necessary to support his Sawyer theory, Cade reasserted a contention he had pursued as a claim in state court but had not raised in his federal petition: that the evidence substantiating the aggravating factors supporting his sentence had been insufficient. He later amended his federal petition to explicitly include this contention as a separate claim.

The magistrate judge recommended an evidentiary hearing on the claim Cade asserted of ineffective assistance of counsel at the sentencing stage because the state-court evidentiary hearing neither resolved the merits of the factual dispute underlying the claim nor produced a record supporting the state court's fact finding. See Townsend v. Sain, 372 U.S. 293, 313, 83 S.Ct. 745, 757, 9 L.Ed.2d 770 (1963) (articulating these two criteria, among others, as grounds requiring a federal evidentiary hearing to find facts regarding claims about which a state court has already held a hearing), overruled in part on other grounds by Keeney v. Tamayo-Reyes, 504 U.S. 1, 112 S.Ct. 1715, 118 L.Ed.2d 318 (1992). Treating a determination on the merits of the newly added claim regarding aggravation as essential to deciding whether the two claims regarding mitigation were procedurally defaulted, the magistrate judge also relabeled the two mitigation claims as a single claim that Cade's sentencing hearing was an unconstitutional "sham," (R.2-42 at 33), and concluded that that "claim" should be considered on the merits without an evidentiary hearing if the district court found the claim regarding the aggravators to have sufficient merit.

The district court adopted the magistrate judge's recommendation on all of these points and conducted a three-day eviden- tiary hearing to find facts regarding the ineffective-assistance-at-the-sentencing-phase claim. After additional briefing, the district court denied relief. On the ineffective-assistance claim, it concluded that Cade had shown neither that his counsel had performed deficiently nor that he was prejudiced by any deficient performance that might have occurred. The district court also concluded that the procedural default of the claims that the trial judge failed to sufficiently consider and make findings regarding mitigating factors was not excused under Sawyer, and in the alternative, that Cade was not due relief on those claims on their merits.

On appeal, Cade challenges only the district court's decisions on the ineffective-assistance-at-sentencing and insufficient-treatment-of-mitigators claims.

II. Contentions of the Parties on Appeal

Cade makes two arguments. He first contends that his trial counsel was ineffective for failing to investigate and present various mitigating evidence, including expert opinions regarding his history of mental illness and the opinions of family, friends, or community acquaintances regarding not only his history of mental problems, but also his character and generally difficult, impoverished background. His second contention is that the trial judge constitutionally erred by giving mitigating factors insufficient consideration in sentencing. Although he admits that this latter claim is procedurally defaulted and that he cannot show cause for the default, Cade argues that the default is nevertheless excused under Sawyer v. Whitley, 505 U.S. 333, 112 S.Ct. 2514, 120 L.Ed.2d 269 (1992) because, but for constitutional error in the sentencing process, no reasonable jury would have found him eligible for the death penalty.5

III. Ineffective Assistance of Counsel at the Sentencing Stage

A petitioner pursuing an ineffective-assistance claim under Strickland v. Washington6 must show both that "counsel's representation fell below an objective standard of reasonableness," id. at 688, 104 S.Ct. at 2064, and "a reasonable probability [defined as one 'sufficient to undermine confidence in the outcome'] that, but for counsel's unprofessional errors, the result of the proceeding would have been different," id. at 694, 104 S.Ct. at 2068. In an Alabama case arising from acts committed in 1977, this means casting doubt on both the jury's decision to "fix the punishment at death" and the trial judge's decision to uphold it. See supra note 3. Both the former prong, deficient-performance, and the latter, prejudice, present mixed questions of law and fact reviewed de novo on appeal. See id. at 698, 104 S.Ct. at 2070. Because a petitioner must prove both prongs, neither the district court nor this court need consider both or address them in a particular order if one element is missing. See id. at 697, 104 S.Ct. at 2069. Although a district court's ultimate conclusions as to deficient performance and prejudice are subject to plenary review, we subject underlying findings of fact only to clear error review. See Byrd v. Hasty, 142 F.3d 1395, 1396 (11th Cir.1998).

We begin our review in this case by asking whether Cade has shown the requisite prejudice. A petitioner seeking to prevail on a Strickland claim must "affirmatively prove prejudice." Strickland, 466 U.S. at 693, 104 S.Ct. at 2067. Cade attempted to do this in the district court by presenting at the evidentiary hearing three new mental-health experts and five of his community acquaintances or family members whom he claims should have been called by counsel at the sentencing phase of his trial. The presentation of these or similar witnesses, Cade contends, would have resulted in an "accurate life profile" of Cade as a less culpable, mentally ill individual. Jackson v. Herring, 42 F.3d 1350, 1367 (11th Cir.1995). The omission of that profile before the jury is, Cade further argues, sufficiently prejudicial to warrant relief under Strickland.

The district court rejected this argument, making four main findings of fact (supported by subsidiary findings) regarding aspects of Cade's case. It found, first, that the testimony of family and friends would not have overcome the testimony at trial of witnesses who had known Cade for a long time but did not present favorable testimony.7 Second, it found that Cade's own testimony at trial was too damaging to be salvaged by this sort of evidence.8 Third, it found that Cade's mental state was, in any event, already before the jury at sentencing and thus that an additional array of evidence about Cade's background and mental health would not have substantially changed the picture the jury had of him. Finally, it gave some weight to the "brutal nature of the act itself." (R.4-96 at 18.) We discern no clear error in these findings. They support a conclusion that Cade has not shown prejudice. Cade's Strickland claim therefore fails, and we need not consider the deficient-performance prong.

To explain this conclusion, we begin by focusing on what was known about Cade's mental health at the time of sentencing. In 1977, Cade's original counsel9 moved for a mental evaluation of Cade to determine his competence both to stand trial and be held criminally responsible for this homicide. The trial court granted the motion, and Cade was admitted for inpatient evaluation at the Alabama Department of Mental Health's Searcy Hospital. There, various mental health professionals evaluated him for about a month.10 At the end of this period, the hospital issued a report premised, first, on psychological and psychiatric testing and observation, and second, on neurological testing using an electroencephalogram and skull x-ray. The diagnosis was that Cade had an "[a]ntisocial personality with superficial passive-aggressive features." (State R.3-P19 at 44.) The report also concludes that Cade's intelligence is in the "low average" range. (State R.3-P19 at 43.)

When Joseph Hughes took over as counsel for Cade, he moved for an evaluation, at state expense, by a private physician or psychiatrist of Cade's choosing. This motion was denied. He then hired a psychiatrist at his own expense. That psychiatrist, Dr. Lopez, testified at both the first and second trials during the guilt phases. Lopez stated that Cade suffered from "organic brain syndrome" caused by blows to the head and alcohol consumption and that his mental capacity was diminished at the time of the crime. (State R.1-P2 at 271.) Lopez also agreed with the personality-disorder diagnosis by Searcy Hospital professionals and with their assessment of Cade's intelligence, and testified that the personality disorder was a product of Cade's childhood about which Cade could do very little. At the second trial, Hughes also called another expert, Mr. Corbitt, a clinical psychologist (masters-degree level) from the prison where Cade was confined, who testified that Cade was generally a well-behaved inmate, but that he exhibited some bizarre behavior and beliefs, was being medicated with the mild anti-psychotic mellaril and had been the subject of a sanity review board at the prison. In total, Cade's two lawyers sought the expertise of an entire state team of mental-health experts and two individual professionals, one hired at personal expense. Although their evaluations and testimony were primarily directed toward guilt-phase mental-health issues, we discern no clear error in the district court's finding that the testimony "allowed for broad discussion of evidence" that included a history of head injuries, drinking, and allusion to the role that the "culture" in which Cade was raised had on the development and fixity of his personality. (R.4-96 at 9.)

And, as the district court observed, Hughes closed at sentencing with an argument designed to place that broad discussion in the minds of the jurors as they made their sentencing decision:

The evidence is very clear that whether Clyde Cade is insane within the legal definition of that term or not, he is certainly not a normal individual. The reports throughout refer to him as containing certain mental deficits, being unable to conform to standards, being mentally retarded, being mentally ill, and whatever your conclusions are as to exactly what he is suffering from, he's just not like most of us that are walking around out here. I submit to you that it would be a very cruel system that would electrocute a man under those conditions. I'm not excusing what he's done in any way, and I've told you all the time that there is no excuse for that. It's just a tragedy that shouldn't have happened. But committing the further tragedy of killing another human being who is not mentally right is certainly not going to undo it, it's not going to bring Red Sizemore back to us, and all it's going to do in this case is just exercise some sort of vengeance on this man. I ask you not to do that.

(State R.2-P8 at 416.) See Waters v. Thomas, 46 F.3d 1506, 1512 (11th Cir.1995) (en banc) (discerning no ineffective assistance in the mere decision to rely on guilt-phase mental-illness evidence at the sentencing phase without reintroducing the evidence).

It is against this background that we must evaluate Cade's argument that testimony of the kind that he presented at the evidentiary hearing in the district court would have made a difference. See Holladay v. Haley, 209 F.3d 1243, 1250-51 (11th Cir.2000) (concluding that it was not ineffective assistance for trial counsel to rely on the results of a month-long inpatient evaluation at a state hospital for competency to stand trial rather than pursue and present additional experts for sentencing). At the § 2254 hearing, Cade presented three new experts-Drs. Lyman, Halleck, and Kirkland-who offered testimony generally more favorable to Cade than the expert testimony offered at trial.11 Lyman, the chairperson of the psychology department at the University of Alabama, testified, with some "tentativeness" about the diagnosis, that Cade is a "paranoid schizophrenic" or has a "schizoaffective disorder." (R.5 at 40.) He also testified that performance testing indicated some neurological damage. Similarly, Halleck, a retired teaching psychiatrist at the University of North Carolina, referred to Cade's condition as "schizophrenia" combined with a "thought disorder." (R.6 at 272-73.) Both Lyman and Halleck also testified to a relationship between Cade's drinking and his mental illness, construing the drinking as a form of self-medication for mental illness, a causal factor in his mental illness, or both. Cade's final expert, Kirkland, another clinical psychologist, did not make a diagnosis that included schizophrenia. Instead, he described Cade as having "a delusional disorder of the paranoid type," "alcohol dependence that could have affected him in terms of brain behavior functioning," (R.5 at 132), and "an antisocial personality disorder," (R.5 at 154).

In addition to Lyman, Halleck, and Kirkland, Cade also called two younger sisters, a cousin, and two acquaintances from Cade's rural community. Together, these witnesses testified to an over-worked and under-schooled upbringing, and said that Cade had since childhood exhibited a bizarre tendency to laugh inexplicably and inappropriately, a disconnectedness from the world and conversation, and a tendency to ramble. They also expressed concern, pity, and love for Cade and noted that he did have mechanical and artistic talents that he shared with the community.

This additional expert and lay testimony might at some level have helped Cade's case for a lesser sentence. But our task is to determine whether this "missing" testimony is significant enough to "undermine [our] confidence in the outcome" of Cade's sentencing, Strickland, 466 U.S. at 694, 104 S.Ct. at 2068, not to ask whether it would have had "some conceivable effect on the outcome of the proceeding," id. at 693, 104 S.Ct. at 2067. Our confidence in the outcome is not undermined.

First of all, none of Cade's habeas experts examined Cade remotely near the time of the crime or second trial. Although they made use of the records underlying the hospital evaluation and added to them records from Cade's years of imprisonment, none of them saw Cade during the time periods at which the judge and jury would have evaluated him, and their observations are, at best, only weakly probative of the possibility for a favorable and credible diagnosis during those periods.12 More importantly, as the district court found, they painted only a mixed picture even at the evidentiary hearing, and they could have done no better for Cade in 1982. While Lyman and Halleck came to more significant diagnoses than antisocial personality disorder, Kirkland was unable to remove the lesser, personality-disorder diagnosis from his final conclusions. See Clisby v. Alabama, 26 F.3d 1054, 1056 & n. 2 (11th Cir.1994) (noting reasons why antisocial-personality-disorder diagnoses are not mitigating). And Dr. Rivenbark, the State's expert (a clinical psychologist) who examined Cade in 1990, fortified the conclusions of the trial experts by contradicting diagnoses of schizophrenia or delusional disorders and reporting indications that at least some of Cade's symptoms may have been the result of malingering. Finally, this expert testimony from the § 2254 hearing is not only internally mixed in its implications; none of it represents a dramatic departure from the basic point made at trial that Cade exhibited some real mental problems of uncertain nomenclature. The habeas presentation was richer and more complete, but, as the district court found, the debate it provokes was in the air at Cade's sentencing.13

There are similar problems with the testimony of Cade's family and friends. To begin with, the family members who testified saw Cade only infrequently after childhood, and their knowledge about the day-to-day aspects of his adult life was limited. More importantly, while the testimony the laypersons gave provides some corroboration for the symptoms addressed by the experts, it also focuses attention on some negative traits. The witnesses confirmed Cade's history of drinking. Although this history adds credence to the experts' conclusions that Cade's mental condition was created or exacerbated by drinking, it also, as this court has held before in an assessment of trial realities, provides an independent basis for moral judgment by the jury. See Tompkins v. Moore, 193 F.3d 1327, 1338 (11th Cir.1999) ("[A] showing of alcohol and drug abuse ... can harm a capital defendant as easily as it can help him at sentencing."); Waldrop v. Jones, 77 F.3d 1308, 1313 (11th Cir.1996) (holding that a "history of excessive alcohol and drug use ... might have been harmful to [the petitioner's] case" even though offered for mitigation). The two community acquaintances also noted that Cade had been known to go into occasional rages. And most of the witnesses knew of his history of drifting in and out of jail, although none testified to visiting him during the in-jail periods. Finally, although all professed concern and affection for Cade, only one of his sisters (among all the lay witnesses) attended his second trial.

This ambivalent picture is corroborated by the PSI. There, the probation officer recognized Cade's deprived background but also concluded that Cade was not held in esteem or deep concern by the community. And Hughes's testimony during both the state and federal habeas proceedings, summarized by the district court's finding that trial counsel "sought background information on Cade from his family members and other community members, but ... was not given what he considered to be helpful information," (R.4-96 at 15),14 is to the same effect.

Thus, we agree with the district court that Cade has not made a showing of prejudice sufficient to undermine our confidence in the outcome of sentencing hearings before the jury and judge. Cade has not overcome either the inconsistencies in the expert opinions at issue nor the double-edged quality of the lay testimony offered. He has not shown that an "accurate life profile" that Hughes might have painted by combining the lay testimony with that of even the full panoply of his § 2254 experts creates a reasonable probability of a different result: The profile shown is not mitigating to a degree sufficiently greater than the profile the jury actually had at sentencing.15 See Mills v. Singletary, 63 F.3d 999, 1024-26 (11th Cir.1995); Marek v. Singletary, 62 F.3d 1295, 1298-1301 (11th Cir.1995); Francis v. Dugger, 908 F.2d 696, 703-04 (11th Cir.1990) (all holding the omission of testimony regarding difficult childhood experiences nonprejudicial). Cf. Williams v. Taylor, --- U.S. ----, 120 S.Ct. 1495, 1514-15, 146 L.Ed.2d 389 (2000) (affirming the conclusion of both state habeas and federal district courts that prejudice was shown where omitted evidence included the actual child-abuse conviction (under extreme circumstances) of both the petitioner's parents); Tompkins v. Moore, 193 F.3d 1327, 1335-38 (11th Cir.1999) (finding, after testimony of family members to the defendant's good character, that omission of additional evidence indicating childhood abuse was not prejudicial); Dobbs v. Turpin, 142 F.3d 1383, 1389-1390 (11th Cir.1998) (affirming the district court's conclusion that the petitioner satisfied the prejudice prong where a finding of historical fact simply labeled his childhood "unfortunate" and there was no evidence of a potentially negative impact from the omitted evidence); Jackson, 42 F.3d at 1362-69 (prejudice found where omitted testimony from the family was unambiguously positive).

IV. Eighth Amendment Violation

Cade's other claim is that the trial court's consideration of mitigating factors was so limited that it was unconstitutional under Lockett v. Ohio16 and its progeny. Cade concedes that this claim is procedurally defaulted and that he cannot show the cause and prejudice necessary to excuse the default under Wainwright v. Sykes, 433 U.S. 72, 97 S.Ct. 2497, 53 L.Ed.2d 594 (1977). We conclude that he cannot show the "actual innocence" of the death penalty required under Sawyer v. Whitley, 505 U.S. 333, 112 S.Ct. 2514, 120 L.Ed.2d 269 (1992), to maintain this Lockett claim in the face of his conceded procedural default.

Sawyer excuses procedural default, even without cause and prejudice, when a petitioner shows "by clear and convincing evidence that, but for a constitutional error, no reasonable juror would have found the petitioner eligible for the death penalty under the applicable state law," id. at 336, 112 S.Ct. at 2517, and construes such demonstrations as showing " 'actual innocen[ce]' of the death penalty," id. at 335, 112 S.Ct. at 2517, such that a " 'miscarriage of justice' " will result if the federal habeas court refuses to consider the constitutional error, id. at 339, 112 S.Ct. at 2518. The opinion clearly permits "actual innocence" to be proven not only through a demonstration of "innocence of the capital crime itself" but also through "a showing that there was no aggravating circumstance or that some other condition of eligibility [under state law] had not been met." Id. at 345, 112 S.Ct. at 2522. On the other hand, it expressly rejects the notion that "the showing should extend beyond these elements of the capital sentence to the existence of additional mitigating evidence." Id., 112 S.Ct. at 2522. In other words, a showing of actual innocence can only refer to those state-law requirements that must be satisfied to impose the death penalty, i.e., the elements of the capital crime and minimum required aggravating factors.

At the time of Cade's second sentencing, Alabama law, like the Louisiana law considered in Sawyer, required guilt of the capital offense and at least one aggravator for the death penalty to be imposed. Compare Beck v. State, 396 So.2d 645, 663 (Ala.1980) and Joseph A. Colquitt, The Death Penalty Laws of Alabama, 33 Ala. L.Rev. 213, 284-86 (1982) with Sawyer, 505 U.S. at 342, 112 S.Ct. at 2520 (discussing Louisiana's minimum criteria for imposing the death penalty). Because the Lockett error asserted relates to mitigation and does not relate to either of these minima, the Lockett claim is not one whose procedural default can be excused by the actual innocence showing under Sawyer. We thus conclude that Cade is not entitled to relief on his Lockett claim.17

V. Conclusion

For the foregoing reasons, the judgment of the district court denying Cade's petition is AFFIRMED.

NOTES:

1 In Cade's direct appeal from the second conviction and sentencing, the Alabama Court of Criminal Appeals referred to its opinion from Cade's first appeal to summarize the facts of the case. See Cade, 521 So.2d at 81 ("Because the facts of the case are set out in detail in the original appeal of Cade v. State, 375 So.2d 802 (Ala.Crim.App.1978), we will not repeat them in this opinion.") The Alabama Supreme Court found this improper, see Ex parte Cade, 521 So.2d 85, 87-88 (Ala.1987), but affirmed because in its judgment "the Court of Criminal Appeals resolved the issue raised by Cade from the evidence as presented in his second trial," id. at 88. No Alabama court has summarized the evidence at the second trial. We therefore undertake to distill the facts directly from the 1982 trial transcript with an eye only toward providing a sense of the background for Cade's petition.

2 Four volumes of documents from Cade's trial, direct appeal, and state collateral attack were appended as exhibits to his federal habeas corpus checklist. For convenience, this document collection will be abbreviated "State R."

3 For crimes committed before July 1, 1981, see Ala.Code § 13A-5-57 (Michie 1994), Alabama capital sentencing proceeds in two phases. The first phase is before the jury. The jury can "fix the punishment at death," Beck, 396 So.2d at 660 (citing 1975 Ala. Acts 213), but if it is unable to do so, the defendant will be sentenced to life imprisonment without parole, see Baldwin v. Alabama, 472 U.S. 372, 389 n. 9, 105 S.Ct. 2727, 2737 n. 9, 86 L.Ed.2d 300 (1985); Beck, 396 So.2d at 663; Joseph A. Colquitt, The Death Penalty Laws of Alabama, 33 Ala. L.Rev. 213, 282-83 (1982). If the jury recommends the death penalty, the judge holds a hearing following which he can either accept the jury's recommendation or sentence the defendant to life without parole. See Beck, 396 So.2d at 663 (citing 1975 Ala. Acts 213); Colquitt, supra, at 329-30.

4 In its written order, the court also stated that it "considered all evidence touching on any mitigating circumstances." (R.4-95 (emphasis added).)

5 Recent changes in governing law require us to address the manner in which Cade was required to perfect this appeal in January of 1999. Cade applied for and was granted a certificate of probable cause (CPC) under the pre-AEDPA law then thought to govern the appeal of all § 2254 petitions filed in district courts before AEDPA's effective date. See Hardwick v. Singletary, 126 F.3d 1312, 1313 (11th Cir.1997). The Supreme Court has since held that AEDPA governs the perfection of any appeal noticed after its effective date, even if the petition was filed in the district court before that date. See Slack v. McDaniel, --- U.S. ----, 120 S.Ct. 1595, 1600, 146 L.Ed.2d 542 (2000). Thus, Cade should have applied for the certificate of appealability (COA) mandated by the current version of 28 U.S.C. 2253 rather than a CPC. See id. And the district court should have "indicate[d] which specific issue or issues," § 2253(c)(3), showed "denial of a constitutional right," § 2253(c)(2), as the statute requires for issuing a COA.

Because these steps were not taken, we must analyze Cade's application to the district court anew, construed as one made for a COA, just as we would if the CPC had been issued by a single judge of this court. See Franklin v. Hightower, 215 F.3d 1196, 1199 (11th Cir.2000) ("fixing" a CPC granted by a circuit judge after the effective date of AEDPA by reviewing it under the standard governing COAs). For each claim Cade pursues in this appeal, we now grant a COA.

6 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984).

7 As Cade points out, all of the witnesses from Cade's community called during the trial were witnesses to the crime or the events surrounding it. They included Diane Butts, her mother, L.C. Thomas, the local "Auxiliary Deputy Sheriff," (State R.1-P2 at 208), whom the sisters summoned to the scene immediately after the shooting, and David Key, a friend of Cade's who testified to Cade's drinking the day of the homicide. All of these witnesses, however, testified to a history of Cade's drinking to excess. And Key also testified that he heard Cade state some weeks before that he would kill Sizemore if he attempted to take him to jail.

8 The district court did not indicate what in Cade's testimony it found so damaging. We note that Cade confirmed his own history of drinking, indicated his willingness to fight back physically when confronted, attributed various crimes for which he had been jailed to the lies of others, and expressed confidence that the prosecutor could not confuse him into testifying that he remembered certain events when he did not.

9 For reasons unclear to us, Cade's original counsel (Myron H. Thompson, now a United States District Judge in the Middle District of Alabama) was replaced before trial by Joseph Hughes, who served as trial counsel through both trials and whose efforts are the main focus of the ineffective-assistance claims in this appeal.

10 Dr. Rudder, one of the evaluating physicians, testified at trial that the evaluating team included psychiatrists, clinical psychologists, a social worker, registered nurses, and various orderlies, all of whose input was valued in developing a complete clinical picture of Cade under nearly continuous observation.

11 Kirkland was originally retained by the State, but was called by Cade at the hearing.

12 This is not to say that there was no earlier evidence indicating a schizophrenia diagnosis. There is a 1978 annotation of uncertain authorship in Cade's prison record indicating that one observer saw him as a type of schizophrenic even then. Cade also points out a 1985 personality inventory administered in prison that Lyman and Kirkland both testified might be read as indicating schizophrenia, but which Kirkland admitted showed a score of "questionable validity." (R.5-147.) Nevertheless, our basic point holds: Even 1985 was well after the second trial, and, presumably, if the consensus opinion at Holmon had been that Cade was schizophrenic, Corbitt, the prison psychologist, would have so testified.

13 Cade also argues in his briefs that his schizophrenic or delusional disorder extended in a specific way to his perceptions of Sizemore, and that in committing the murder, he was acting only on the basis of his delusions. But this argument rests on diagnoses that are, as we have discussed, debatable.

14 The statement summarizes the basic theme from Hughes's testimony relevant to prejudice: that Hughes-whose judgment was formed by spending his entire legal career in rural, tight-knit Geneva County-testified to finding, on balance, an unpromising picture of Cade validates the mixed picture we see in the family's testimony and the PSI. At the Rule 20 hearing, trial counsel testified as follows:

Q:In preparation for trial, Mr. Hughes, what kind of investigation of the Defendant's background did you undertake?

A:I talked with Mr. Cade himself on numerous occasions. I talked to members of his family. I talked with people from the community where he lived up around Bellwood, Alabama. I talked with law enforcement officers. I got all the information I could and followed up as best I could on anything that was brought out that I though would be helpful.

Q:Were you able to generate any favorable evidence that you could use? For example, in mitigation at the penalty stage?

A:Very little. Clyde just had a general reputation for violence and for drinking and this type of thing. He was not a real reputable character and I just couldn't find much along that line that would be of any help to me.

Q:Would it be fair to say that what you found was used a trial?

A:Yes, if I found any, I used it.

Q:Mr. Hughes, at the penalty phase before the Jury, what was the strategy that you used to attempt to prevent the imposition of the death penalty?

A:I don't remember too clearly, but what I do recall is that I continued somewhat with my legal argument to the Jury that this procedure was not proper and that they should not impose the death penalty for that reason. I also argued simply along the lines that we don't have a right to impose the death penalty and that's not ou[r] prerogative.

Q:Did you argue any inferences from the testimony concerning the Petitioner's mental condition?

A:During the penalty phase?

Q:Yes, sir.

A:I don't really remember.

Q:Were you able to find any sympathetic witnesses in the community that Mr. Cade came from?

A:None other than his immediate family.

(State R.3-P21 at 80-81.)

And at the § 2254 hearing, Hughes testified:

Q:Did you attempt an investigation into mitigating circumstances for Mr. Cade?

A:I'm sure I did. I talked to a lot of people.

Q:Do you recall anybody that you talked to particularly?

A:I can't remember the names. I talked with the woman that ... that was said to be his girlfriend or that he had been living with. I talked with a number of people in that community. And most of the people who testified in the case and testified for the state, I had talked with.

Q:Are you talking about the Bellwood community?

A:Yes.

Q:Do you remember anybody in particular that you talked to in the Bellwood community?

A:I can't remember anybody's name now.

....

Q:What about the people in the Bellwood community; what did you learn from them about Mr. Cade?

A:Generally my recollection now is that I learned that he was known to be pretty wild and had a reputation of being a troublemaker.

Q:Did you get any favorable information about Mr. Cade from the people in the Bellwood community?

A:Not that I recall.

Q:Did you attempt to get information from Mr. Cade?

A:Yes. I talked with him on numerous occasions.

Q:Did you get any names from him of people to talk to?

A:Some. I don't remember any of the names now. But he gave me some names.

Q:Did you talk to any family members?

A:Yes. I believe I talked with his mother and several other people that were related to him, or at least closely associated with him.

Q:Did you think ... Did you receive any information from the investigation that you thought could be used as mitigating evidence?

A:Well, I was of the impression from all that I learned that Clyde's mental condition was not the best in the world. And, also, I received information, I don't know from who all or what sources, that he was drinking heavily that day. And I believe the general testimony in the trial was that he was drinking heavily that day. And, so, for whatever we could, you know, we tried to make something out of that, as far as his mental condition was concerned.

Q:You attempted to present that as far as mitigation at the trial?

A:Yes.

Q:Would it be fair to say that if you found anything that you presented it at trial, as far as mitigation?

A:Yes. I think I presented everything that I had that I though would help.

Q:Were you able to find any sympathetic witnesses for Clyde Cade in the Bellwood community?

A:Not generally, no.

(R.6 at 332-35.)

In response to later questioning directly from the court, Hughes responded in a manner indicating his fading memory regarding his decision-making process (he later completely conceded his inability to recall exactly why he chose not to call family members who might have testified to Cade's mental irregularities) but also showing his persistent tendency to recall doubts about the usefulness of family testimony.

THECOURT: Now, is there any ... Can you tell me any reason that you had for not calling any of the family members to testify during the sentencing phase of the trial?

A:The only thing that I can think of now is that ... My recollection is that most of those family members had testified mostly against Mr. Cade [an incorrect assertion; no family members testified], and their testimony was to the effect that he was drinking and wild and out of hand. That I didn't think that it would really do any good to rehash that at that phase of the trial.

(R.6 at 352-53.)

One of Cade's sisters confirmed during the federal habeas hearing that trial counsel had conferred with several family members before the first trial. Hughes had a vague recollection of meeting again before the second trial, but the testimony of Cade's family does not corroborate that recollection.

15 Cade also argues that, in light of the changes to Alabama's sentencing procedure made between his first and second trials, Hughes's assistance was self-evidently ineffective because Hughes did not visit or investigate new possibilities for mitigation. Even if this argument has some arguable merit with respect to deficient-performance analysis, we need not consider it because it makes no difference to our consideration of prejudice.

16 438 U.S. 586, 597-609, 98 S.Ct. 2954, 2961-67, 57 L.Ed.2d 973 (1978).

17 Cade's argues that he has satisfied Sawyer to excuse default of his Lockett claim by using the evidence that underpins his entirely distinct claim that the evidence supporting the aggravators was insufficient. This argument need not detain us because there clearly was evidence to support the finding of aggravators.