Rompilla v. Horn

Opinions of the United
2004 Decisions States Court of Appeals
for the Third Circuit

1-13-2004

Rompilla v. Horn
Precedential or Non-Precedential: Precedential

Docket No. 00-9005

Follow this and additional works at: http://digitalcommons.law.villanova.edu/thirdcircuit_2004

Recommended Citation
"Rompilla v. Horn" (2004). 2004 Decisions. Paper 1047.
http://digitalcommons.law.villanova.edu/thirdcircuit_2004/1047

This decision is brought to you for free and open access by the Opinions of the United States Court of Appeals for the Third Circuit at Villanova
University School of Law Digital Repository. It has been accepted for inclusion in 2004 Decisions by an authorized administrator of Villanova
University School of Law Digital Repository. For more information, please contact Benjamin.Carlson@law.villanova.edu.
 Volume 1 of 2

PRECEDENTIAL

Filed January 13, 2004

UNITED STATES COURT OF APPEALS
FOR THE THIRD CIRCUIT

Nos. 00-9005, 00-9006

RONALD ROMPILLA
v.
MARTIN HORN, COMMISSIONER,
PENNSYLVANIA DEPARTMENT OF
CORRECTIONS
Martin Horn,
Appellant/Cross-Appellee

On Appeal From the United States District Court
For the Eastern District of Pennsylvania
(D.C. Civ. No. 99-cv-00737)
District Judge: Honorable Ronald L. Buckwalter

Argued: May 22, 2002
Before: SLOVITER, ALITO, AND STAPLETON,
Circuit Judges.

(Filed: January 13, 2004)
AMY ZAPP (Argued)
Senior Deputy Attorney General
Office of Attorney General
16th Floor, Strawberry Square
Harrisburg, PA 17120
Counsel for Appellant/Cross-Appellee
 2

BILLY H. NOLAS (Argued)
DAVID W. WYCOFF
MICHAEL WISEMAN
Defender Association of Philadelphia
Federal Court Division
The Curtis Center, Suite 545 West
Independence Square West
Philadelphia, PA 19106
Counsel for Appellee/Cross-Appellant

OPINION OF THE COURT

ALITO, Circuit Judge:
The Commissioner of the Pennsylvania Department of
Corrections (hereinafter “the Commonwealth”) appeals from
a District Court order granting the petition for a writ of
habeas corpus that was filed by Ronald Rompilla, a
Pennsylvania prisoner who was sentenced to death. The
District Court ordered that Rompilla be released unless he
is either resentenced to life imprisonment or a new penalty
phase trial is held. Rompilla cross-appeals from the denial
of his petition insofar as it challenged his conviction. We
conclude that the Pennsylvania Supreme Court’s decision
regarding Rompilla’s sentencing proceeding was not
contrary to and did not involve an unreasonable application
of clearly established Supreme Court precedent, and
therefore we reverse the decision of the District Court with
respect to Rompilla’s sentence. We affirm the decision of
the District Court with respect to his conviction. By
separate order, however, we have granted Rompilla’s
application to file a successive petition for a writ of habeas
corpus so that he will be able to assert his claim that,
under Atkins v. Virginia, 536 U.S. 304 (2002), he may not
be executed because of mental retardation.

I.
In 1988, Rompilla was tried for the murder of James
Scanlon. Scanlon’s body was found lying in a pool of blood
in his bar, the Cozy Corner Café in Allentown,
 3

Pennsylvania. Scanlon had been stabbed repeatedly and set
on fire. There were no eyewitnesses to the killing, but the
Commonwealth introduced substantial circumstantial
evidence of Rompilla’s guilt. In its opinion on direct appeal,
the state supreme court summarized the prosecution’s
evidence as follows:
Appellant was seen in the Cozy Corner Café on
January 14, 1988, from approximately 1:00 a.m. to
2:00 a.m. During that time, he was observed going to
the bathroom approximately ten times. A subsequent
police investigation determined that the window in the
men’s bathroom was used as the point of entry into the
bar after it had closed.
When questioned by an investigating detective from
the Allentown Police Department, Appellant stated that
he had been in the Cozy Corner Café on the night of
the murder and left between 2:00 a.m. and 2:30 a.m.
because he had no money. He stated that he had only
$2.00 to buy breakfast at a local diner. A cab driver
testified that he picked up Appellant at the diner and
drove him to two different hotels where Appellant was
unable to rent a room. The driver then took Appellant
to the George Washington Motor Lodge where he was
able to rent a room. Appellant paid the cab fare of
$9.10.
Appellant rented a room for two nights at the George
Washington Motor Lodge. In doing so, he paid $121.00
in cash and flashed a large amount of cash to the desk
clerks. Appellant also used a false name when he
checked in.
The police secured a search warrant for Appellant’s
motel room and seized several items, including
Appellant’s sneakers. These sneakers matched a
footprint in blood that was discovered near the victim’s
body. In addition, the blood found on the sneakers
matched the victim’s blood type.
The Commonwealth also presented other
circumstantial evidence that linked Appellant with the
robbery and murder of James Scanlon. First, Mr.
Scanlon’s wallet was found by a groundskeeper in the
 4

bushes, six to eight feet outside the room that
Appellant had rented at the George Washington Motor
Lodge. Second, Appellant’s fingerprint was found on
one of the two knives used to commit the murder.
Finally, there were numerous inconsistencies between
what Appellant had told the police concerning his
activities on January 14 and 15, 1988, and the
testimony of other witnesses.
Commonwealth v. Rompilla, 653 A.2d 626, 629 (Pa. 1995).
The jury found Rompilla guilty of first degree murder and
other related offenses.
At the penalty phase of the trial, the prosecution
attempted to establish three aggravating factors: (1) that
Rompilla committed the murder while perpetrating a felony,
42 Pa. C.S. § 9711(d)(6), namely, the burglary and robbery
of the bar; (2) that he committed the murder by means of
torture, 42 Pa. C.S. § 9711(d)(8); and (3) that he had a
significant history of felony convictions involving the use or
threat of violence to the person, 42 Pa. C.S. § 9711(d)(9). To
establish torture, the Commonwealth called Dr. Isidore
Mihalakis, a forensic pathologist, who testified to the
multiple wounds inflicted on Scanlon and opined that
Scanlon was conscious and alive when many of those
wounds were inflicted. App. 698-707. Based on the nature
of the wounds, Dr. Mihalakis also opined that Scanlon’s
killer had deliberately attempted to inflict pain before
Scanlon died. Id. at 707-08. To show the defendant’s prior
history of violent felonies, the Commonwealth proved that
he had been convicted in 1976 of rape, burglary, and theft.
App. 651-52. Commonwealth v. Rompilla, 378 A.2d 865 (Pa.
1977). The testimony of the rape victim, which was read
into the record, showed that Rompilla had burglarized a bar
after closing and had raped the bar owner and slashed her
with a knife. App. 662-696.
The defense presented the testimony of five members of
the Rompilla family. Rompilla’s older brother, Nicholas, and
his wife, Darlene, testified that Rompilla had lived in their
home with their children for the three and one-half months
before the killing. Nicholas testified that he and the
defendant had grown up together, that the defendant had
worked for him as a house painter before his arrest, and
 5

that the defendant had also helped out around the house.
App. 738-41. Nicholas said that he had visited his brother
frequently in prison and that they had a good relationship.
Id. at 739. Nicholas added that he did not think that his
brother had killed Scanlon, whom he had known for about
10 or 11 years, and that his heart went out to the Scanlon
family. Id. at 740. He concluded his testimony by asking
the jury to have mercy on his brother. Id. at 741.
Darlene Scanlon testified that her children were “very
attached” to the defendant and that he “was very good in
our house.” App. 734-35. Darlene said that he helped out
in the home and that “he was a good family member” who
felt “strongly about family” and “respected the family very
well.” Id. at 735. While the defendant was in prison,
Darlene testified, he frequently wrote to her and spoke
about “[f]amily, his son, his wife, his brothers and sisters.”
Id. She stated that the defendant’s relationship with his son
was good. Id. at 736. Like her husband, Darlene testified
that she had known Scanlon, that she did not believe that
the defendant had killed him, and that she felt for the
Scanlon family “[v]ery, very much.” Id. Weeping, she
concluded her testimony by telling the jury: “We go to bed
crying, we wake up crying, it’s been very hard on my
children . . . . [W]e want Ron alive even if it’s in jail, we
want him alive.” Id.
Another brother, Bobby, also took the stand. Bobby
testified that during the time when the defendant was out
of prison they had an “[e]xcellent relationship” and were
“[v]ery close.” App. 745. Before then, according to Bobby, he
had visited the defendant in prison and the defendant had
written him letters in which he expressed great concern for
his son and other family members. Id.
The defendant’s sister, Sandra Whitby, testified that she
had grown up with the defendant. App. 754. Crying, she
said that she loved him very much and that she thought
that he was a “good person.” Id. at 755. Asked what things
were important for the defendant, she answered: “Family,
his son, his wife, things my children were doing, my
brothers.” Id. She testified that she was praying for the
Scanlon family and for her brother’s life. Id. at 756. She
added: “[W]e’re not God, and we can’t take people’s lives
 6

(crying) . . . I love my brother. Taking one life is never going
to replace another life.” Id.
Finally, Aaron Rompilla, the defendant’s 14-year old son
testified. He said that after his father’s release from prison
they had regular visits, which he enjoyed, that his father
was proud of him, that he loved his father, and that he
would visit him if he was sentenced to prison. Id. at 757-
59. He said that he did not think it would be “right” to
sentence his father to death, and when he was asked
whether there was anything else he wanted to tell the jury,
he simply cried. Id. at 759.
In her closing argument, defense counsel made an
impassioned plea for Rompilla’s life. Her closing appears to
have had three major themes. First, she repeatedly stressed
that, although the jury had found the defendant guilty, they
must have had at least some lingering doubt about what
had happened, and therefore they should not sentence him
to death, which “is final, irreversible.” Id. Second, she
reminded the jury of the good qualities mentioned by
Rompilla’s family members. She argued that Rompilla was
“more than this act that you have found him to have
committed,” id. at 769, and she emphasized the love of
Rompilla’s family members. Id. at 773. She particularly
asked the jury to keep in mind the defendant’s 14 year old
son, who had come to court to ask the jury “not to kill his
father.” Id. Third, she pled for mercy and warned the jury:
“I tell you this from the bottom of my heart, if you order
death, it will take a part of your life away. Don’t stain your
hands and your souls with this man’s blood, don’t do it
please.” Id. at 772.
After deliberating, the jury unanimously found all three
of the aggravating circumstances alleged by the
prosecution, and the jury stated that one or more members
found two mitigating factors under the “catchall” provision
of 42 Pa. C.S. § 9711(e)(8), i.e., “Rompilla’s son being
present and testifying” and the possibility of rehabilitation.
The jury found that the aggravating circumstances
outweighed the mitigating factors and sentenced Rompilla
to death.
The Pennsylvania Supreme Court affirmed the conviction
 7

and sentence. Commonwealth v. Rompilla, 653 A.2d 626
(Pa. 1995) (hereinafter “Rompilla-1”).1 In December 1995,
Rompilla filed a petition under the Pennsylvania Post-
Conviction Relief Act (“PCRA”). After an evidentiary hearing,
the PCRA denied the petition, and the Pennsylvania
Supreme Court affirmed. Commonwealth v. Rompilla, 721
A.2d 786 (Pa. 1998) (hereinafter “Rompilla-2”).
Rompilla then filed a petition for a writ of habeas corpus
pursuant to 28 U.S.C. § 2254 in the United States District
Court for the Eastern District of Pennsylvania. Rompilla’s
petition raised 11 claims. Rompilla v. Horn, No. Civ.A.99-
737, 2000 WL 964750 (E.D. Pa. July 11, 2000).2 The

1. While the case was pending on direct appeal, Rompilla filed a petition
for a writ of habeas corpus in the United States District Court for the
Eastern District of Pennsylvania. The District Court dismissed the
petition without prejudice for lack of exhaustion. Rompilla v. Love, No.
94-cv-4196 (E.D. Pa.).
2. The claims as identified by the District Court were as follows:
(1) Trial counsel were ineffective at the capital sentencing phase for
failing to investigate, develop and present significant mitigating
evidence related to petitioner’s traumatic childhood, alcoholism,
mental retardation, cognitive impairment and organic brain damage;
(2) Petitioner is entitled to relief from his conviction and death
sentence because of the trial court’s improper instruction on
accomplice liability;
(3) The trial court’s failure to instruct the jury that “life
imprisonment” means life without possibility of parole, even after
the jury repeatedly asked about parole eligibility; and the trial
court’s provision, instead, of inaccurate and misleading information,
violated petitioner’s Sixth, Eighth and Fourteenth Amendment
rights;
(4) Petitioner is entitled to relief from his death sentence because
the (d)(8) aggravating circumstance was unconstitutionally obtained;
(5) Petitioner is entitled to relief from his death sentence because
the (d)(9) “significant history” of felony convictions aggravating
circumstance is unconstitutionally vague; the jury instructions did
not cure this vagueness; and the jury instructions, to the extent
they provided guidance at all, directed a finding of this aggravating
circumstance;
 8

District Court denied relief as to the guilt phase but
granted relief as to the penalty phase based on its
conclusion that counsel had been ineffective in failing to
investigate and present mitigating evidence. Id. at *14. The
District Court thus ordered that a writ of habeas corpus
would be granted unless the Commonwealth conducted a
new sentencing hearing or resentenced Rompilla to life
imprisonment. Id. at *21. The Commonwealth appealed,
and Rompilla cross-appealed. Three questions are
presented on appeal: (1) whether Rompilla’s trial counsel
was constitutionally ineffective during the penalty phase;
(2) whether the trial court committed constitutional error in
giving an accomplice liability instruction; and (3) whether
the trial court erred in failing to instruct the jury that “life

(6) Petitioner is entitled to relief from his death sentence because
of improper prosecutorial argument at the penalty phase;
(7) Petitioner is entitled to relief from his death sentence because
the trial court improperly allowed the prosecutor to read to the jury
the inflammatory and prejudicial testimony of the victim of a prior
rape and counsel were ineffective for failing to cite controlling
authority that would have prevented the introduction of the
inflammatory and prejudicial evidence;
(8) The seating of the juror who visited the scene of the crime ten
times, including, during the trial proceedings, who knew the victim
of the offense and the victim’s son, who knew an employee of the
prosecutor’s office and who expressed substantial doubts regarding
the presumption of innocence; and trial counsel’s failure to
challenge this juror for cause, violated petitioner’s sixth, eighth and
fourteenth amendment rights;
(9) Petitioner’s death sentence should be vacated because the
arbitrary “proportionality review” performed by the Pennsylvania
Supreme Court violated his Eighth Amendment and Fourteenth
Amendment rights;
(10) Petitioner is entitled to an evidentiary hearing on his claim
that the prosecution violated his due process by introducing false
and misleading evidence; and
(11) Petitioner is entitled to relief because of cumulative prejudicial
effects of errors in this case.
Rompilla, 2000 WL 964750.
 9

imprisonment” under Pennsylvania law meant life without
the possibility of parole.3

II.
Because the District Court did not conduct an
evidentiary hearing, our review of that Court’s decision is
plenary. Duncan v. Morton, 256 F.3d 189, 196 (3d Cir.),
cert. denied, 122 S. Ct. 269 (2001). However, our review of
the decision of the Pennsylvania Supreme Court is
circumscribed by the Antiterrorism and Effective Death
Penalty Act of 1996 (“AEDPA”). See Hartey v. Vaughn, 186
F.3d 367, 371 (3d Cir. 1999). Under AEDPA, a federal court
may not grant habeas relief on any claim adjudicated on
the merits in state court unless the adjudication
(1) resulted in a decision that was contrary to, or
involved an unreasonable application of, clearly
established Federal law, as determined by the Supreme
Court of the United States; or
(2) resulted in a decision that was based on an
unreasonable determination of the facts in light of the
evidence presented in the State court proceeding.
28 U.S.C. § 2254(d). Under the “contrary to” clause of
§ 2254(d)(1), relief may be ordered if the state court arrived
at “a conclusion opposite to that reached by [the Supreme]
Court on a question of law” or if the state court decided “a
case differently than [the Supreme] Court has on a set of
materially indistinguishable facts.” Williams v. Taylor, 529
U.S. 362, 412-13 (2000). Under the “unreasonable
application” clause of § 2254(d)(1), relief may be awarded if
the state court identified the correct governing legal
principle from Supreme Court decisions but unreasonably
applied that principle to the facts of the case, id., or was
“unreasonable in refusing to extend the governing legal

3. A certificate of appealability was not required for the Commonwealth’s
appeal of the ineffective assistance of counsel claim. Fed. R. App. P.
22(b)(3). The District Court granted a certificate of appealability as to the
parole ineligibility claim, Rompilla, 2000 WL 964750, at *21, and we
granted a certificate of appealability as to the accomplice liability
instruction claim. 28 U.S.C. § 2253(c).
 10

principle to a context in which the principle should have
controlled.” Ramdass v. Angelone, 530 U.S. 156, 166
(2000). Relief may not be granted under the “unreasonable
application” clause merely because the federal court
“concludes in its independent judgment that the relevant
state-court decision applied clearly established federal law
erroneously or incorrectly. Rather, that application must
also be unreasonable.” Williams, 529 U.S. at 411. See also
Bell v. Cone, 122 S. Ct. 1843, 1850 (2002) (“The focus . . .
is on whether the state court’s application of clearly
established federal law is objectively unreasonable, and we
stressed in Williams that an unreasonable application is
different from an incorrect one.”). Furthermore, a federal
court “must presume that the factual findings of both state
trial and appellate courts are correct.” Everett v. Beard, 290
F.3d 500, 508 (3d Cir. 2002) (citing 28 U.S.C. § 2254(e)(1)).
This presumption may be rebutted only by clear and
convincing evidence. 28 U.S.C. § 2254(e)(1).

III.
Rompilla claims that his trial attorneys were ineffective at
the penalty phase. First, Rompilla alleges that his attorneys
were derelict in failing to obtain school, hospital, court, and
prison records that reveal a number of IQ test results in the
mentally retarded range, low achievement scores,
placement in special education classes, childhood neglect,
problems with alcohol, and an alcoholic mother. Second, he
faults his attorneys for failing to provide such records to
the psychologist and the two psychiatrists whom they
retained to examine him and for allegedly failing to
communicate adequately with these experts. Third, he
alleges that his attorneys were negligent in interviewing
members of his family. He criticizes trial counsel because
they did not interview two sisters who did not testify at the
penalty phase, and because they did not ask sufficiently
specific questions when they interviewed other family
members. Fourth, Rompilla charges that his attorneys were
derelict in failing to investigate the possible effects on
Rompilla of childhood trauma and alcoholism.
 11

A.
Rompilla was represented by two public defenders,
Frederick Charles, who was then the Chief Public Defender
for Lehigh County, and Maria Dantos, a full-time assistant
public defender. App. 1055-56, 1059. Both testified at
length at the PCRA hearing. For present purposes, two
aspects of these attorneys’ preparation for the penalty
phase are most pertinent: first, their efforts to obtain
information about Rompilla’s childhood and schooling and,
second, their communications with the three mental health
professionals whom they consulted.
Both Charles and Dantos explained that the defense had
questioned Rompilla and numerous members of his family
about his childhood, schooling, and background and had
asked for any information that might be helpful at
sentencing. According to their testimony, however, neither
Rompilla himself nor any family member even hinted at the
problems on which Rompilla’s ineffective assistance claim is
based. Dantos testified that she had developed a good
relationship of trust with Rompilla, that she felt that she
had gotten to know him, and that she had a lot of
discussions with Rompilla “about who he was and his life.”
App. 1073, 1163. When she asked Rompilla about school,
Dantos recounted, he told her “[t]hat there was nothing
unusual about it.” Id. at 1197. Dantos also said that she
specifically asked Rompilla about drinking and that he
responded that he sometimes drank but “could handle it”
and was not an alcoholic. Id. at 100-01. On the night of the
killing, Rompilla told her, he had consumed three or four
beers over the course of the entire evening. Id. at 1101.
Charles added that Rompilla had responded to questions
about his background by saying that nothing was wrong:
“Is there anything that happened? What was it like
growing up? Is there anything you can tell us that
could help us?” And he said, “No, there was nothing
wrong.” He was very, very, smooth about it. It wasn’t
that he was reluctant to talk about anything. He said,
“Your conversations about the possibility of the death
penalty bore me.” . . .
 12

There was no indicator from anything he told us that
would send us searching for . . . any kind of records.
He said everything was fine. He had a normal
childhood. There was nothing there. . . .
. . . I remember [Dantos] specifically going one by one
and talking to him. ‘Is there anything you can tell me?
Tell me about yourself. Tell by about your background.’
She was, you know, meticulous to cover points.
App. 1303. Both Dantos and Charles said that nothing in
their discussions with Rompilla ever suggested that he was
mentally retarded, id. at 1181, 1393, and Charles
elaborated that Rompilla did not have difficulty in
understanding what was said to him or in expressing his
feelings. Id. at 1393.
Dantos and Charles also testified that members of
Rompilla’s family provided no hint that Rompilla had
mental problems, had suffered child abuse, or was an
alcoholic. Dantos stated that she spoke with three of
Rompilla’s siblings, his sister-in-law Darlene Rompilla, and
his ex-wife and that she had formed a “very close”
relationship with the family. App. 1092, 1065. She said
that they had discussed the importance of mitigation
evidence but that no one had provided any useful
information about Rompilla’s background. She testified that
she had “spent hours with these family members,” that
they had “discussed the family dynamics and what
[Rompilla’s] family relationship was with his parents,” and
that there was no indication “that there was any sort of
abuse within the family.” Id. at 1097 Likewise, she stated
that “there was nothing exceptional presented to [her]
about drinking within the family.” Id. Charles added that
although the family members said that they did not know
Rompilla well because he had spent so many years behind
bars, the family was a “constant source of information.” Id.
at 1303, 1384.
Charles was questioned repeatedly and extensively by
Rompilla’s attorneys about the failure of the defense to look
for school, prison, and medical records, and Charles
explained his strategy as follows:
 13

I would investigate by asking my client, “how was your
childhood? Were there any problems that you suffered?
Any kind of abuse? . . . [I]s there anything that sticks
out? Don’t think whether it’s important or not. You just
tell us, and then we’ll determine whether or not we can
use it.” Investigate it that way. Would I send somebody
to the person’s elementary school to talk to the teacher
to see if they remember him from 25 years or 40 years
before. No. I didn’t have those resources in the office.
. . . I had two investigators and 2,000 cases. . . . I will
talk to the client, talk to the family, and see if anything
developed from there.
Id. at 1293-94. If these inquiries provided any hint that
records would be helpful, he added, he would go anywhere
to get the records. Id. at 1307.
With respect to the development of mitigation evidence
regarding Rompilla’s mental condition, Charles said that he
had sent Rompilla to “the best forensic psychiatrist around
here, to [another] tremendous psychiatrist and a fabulous
forensic psychologist” and that he relied on them to detect
any mental problems that might be useful to the defense
and to request that the defense provide them with any
records that they might need. App. 1307-08. He elaborated
that when he sent Rompilla to the psychologist, he expected
the psychologist to administer “a Reitan battery test to
determine if there was any brain damage,”4 a personality
test such as “the Minnesota Multiphasic Personality
Inventory (“MMPI”)”,5 a Rorschach test,6 and an IQ test. Id.

4. “The Reitan Battery measures organic impairment by means of a
variety of exercises testing concentration, coordination, memory, motor
control, abstract abilities and other cognitive and physical functions.”
Barker v. Secy of Health and Human Services, 882 F.2d 1474, 1476 (9th
Cir. 1989).
5. The Minnesota Multiphasic Personality Inventory or “MMPI” is “one of
the best known and widely used personality assessment tests.” Richard
Sloane, THE SLOANE-DORLAND ANNOTATED MEDICAL-LEGAL DICTIONARY — 1992
SUPPLEMENT 522 (1992).
6. The Rorschach test is “a projective psychological [test] in which the
subject reveals his or her attitudes, emotions, and personality by
reporting what is seen in each of 10 inkblot pictures.” STEDMAN’S MEDICAL
DICTIONARY 1808 (27th ed. 2000).
 14

at 1323. He stated that he also expected that the results of
this testing would be available to the psychiatrists. Id. If the
mental health professionals needed any records or other
information, he said, he expected them to ask for them, and
he would have tried to provide whatever they sought. Id. He
said that he also thought that the mental health
professionals would interview Rompilla and that they would
detect whether Rompilla’s denial of anything unusual in his
background seemed suspicious. Id. at 1308-09.
All three of the mental health professionals to whom
Rompilla had been sent by the defense testified at the PCRA
hearing either in person or by deposition. All three stated
that they had examined Rompilla and had found nothing
useful to the defense, but as Rompilla now stresses, all
three also testified that, if they had been provided with
Rompilla’s school and other records, they would have done
additional testing.
Dr. Gerald Cooke, an experienced clinical and forensic
psychologist, no longer possessed records regarding his
examination of Rompilla other than the letter that he had
sent to the public defender’s office after completing the
examination, but he was able to state with “99 percent”
certainty what he would have done in a case of that type.
App. at 1797. In a death penalty case, he testified that one
of the questions in his mind would have been whether the
defendant “showed any mental illness, emotional
disturbance, or other sorts of problems that might be a
psychological mitigating circumstance.” App. at 1808. He
would have administered an MMPI, “an Incomplete
Sentence Blank,7 the Rorschach inkblot technique, and two
or three subtests of the verbal subscale of the Wexler Adult
Intelligence WAIS Revised.”8 Id. at 1797. Dr. Cooke testified
that he knew that the IQ test that he would have
administered to Rompilla had not produced a score in the

7. The Rotter Incomplete Sentences Blank Test tests maladjustment by
providing the beginning of sentences (e.g., “I feel . . .”, “My nerves . . .”)
that are completed by the subject. See www.cps.nova.edu/
~cpphelp/ROTTER.hrml.
8. The Wechsler Adult Intelligence WAIS Revised is a test that measures
general intelligence in adults. STEDMAN’S MEDICAL DICTIONARY, supra, 1596.
 15

mentally retarded range because, in that event, he would
have given further tests. Id. at 1810. He explained that he
had not prepared a report regarding his examination
because it was his practice to discuss the results of an
examination orally with the referring attorney and to leave
it to the attorney to decide whether he should prepare a
report. Id. at 1816. He interpreted the letter that he wrote
in this case to mean that he “didn’t have anything that [he]
felt could be helpful.” Id.
Dr. Robert Sadoff, an experienced board-certified forensic
psychiatrist, also had no records about Rompilla but
testified that in a case of that type he would have examined
the defendant for competency to stand trial, criminal
responsibility, and mitigating circumstances. App. 1841,
1859. He would have questioned and observed the
defendant and would have requested psychological testing
if he thought it was necessary. Id. From the letter that he
sent to the Public Defender’s Office, he said, it was fair to
infer that he had found no mitigating evidence. Id. at 1859.
A second psychiatrist, Paul K. Gross, also testified that
he had examined Rompilla prior to his trial at the request
of the Public Defender’s office to determine his mental
status at the time but was not asked to look for mitigating
factors. App. 1504, 1506, 1549. Dr. Gross stated that
Rompilla “denied any abuse as a child, by either parent.”
Id. at 1517. According to Dr. Gross, Rompilla said that he
had “a good relationship with his father” and a “fairly
normal childhood except for the fact that he didn’t like
school, which he left in the ninth grade.” Id. at 1517. Dr.
Gross’s conclusion at the time was that, although there was
some evidence of antisocial behavior, “[t]here was no other
evidence for underlying psychiatric or mental disorder.” Id.
at 1540. He added that he did not see anything in the
materials shown to him by Rompilla’s attorneys at the
PCRA hearing that would have changed his opinion. Id. at
1539-40. Dr. Gross also reported that although Rompilla
denied it, there was a possibility that he could become
violent while under the influence of alcohol.
In the PCRA proceedings, Rompilla presented testimony
of three family members: two sisters, Barbara Harris and
Randi Rompilla, who testified that they were not
 16

interviewed before sentencing, and Nicholas Rompilla, who
had testified at the sentencing hearing. These family
members stated, among other things, that their parents
were alcoholics; that their mother drank while pregnant
with Rompilla; that their father was physically abusive to
the children and their mother; that Rompilla was locked in
an outdoor dog pen; and that Rompilla was told he was
stupid and would not amount to anything and was a very
nervous child who kept everything inside. Nicholas testified
that during the pre-sentencing interview he was asked only
about the three months prior to the offense and was not
asked about Rompilla’s childhood. App. 1462-63, 1467-71,
1477-78.
Rompilla also presented evidence from two psychologists,
Carol L. Armstrong and Barry Crown, both of whom had
evaluated and tested Rompilla after he was convicted and
sentenced. These psychologists’ evaluations included
neuropsychological testing, review of Rompilla’s school,
medical, and prison records, and review of post-sentencing
declarations by Barbara Harris, Darlene Rompilla, and
Nicholas Rompilla. The psychologists stated that the low IQ
and achievement test results documented in Rompilla’s
school records, his medical history, and his abusive
background were all “red flags” indicating that further
objective evaluation was necessary. App. at 1692, 1739,
1743.
Drs. Armstrong and Crown opined that Rompilla suffers
from organic brain damage, an extreme mental disturbance
significantly impairing several of his cognitive functions.
They expressed the view that Rompilla’s problems relate
back to his childhood and were likely caused by fetal
alcohol syndrome, and they concluded that Rompilla’s
capacity to appreciate the criminality of his conduct or to
conform his conduct to the law was substantially impaired
at the time of the offense. See Rompilla Br. at 58-61.

B.
In evaluating Rompilla’s ineffective assistance of counsel
claim, the PCRA court noted that Strickland v. Washington,
466 U.S. 668 (1984), requires proof of deficient performance
 17

and prejudice. App. 2027. The court then observed that
Pennsylvania cases employ a three-pronged test:
First, a defendant must demonstrate that his claim is
of arguable merit. In the event this threshold
requirement is satisfied, a defendant must show that
counsel had no reasonable basis for the act or
omission in question. Finally a defendant must
establish that but for counsel’s act or omission, the
outcome of the proceedings would have been different.
App. 2028 (quoting Commonwealth v. Buehl, 658 A.2d 771
(Pa. 1995).
Applying this standard, the PCRA court concluded that
Rompilla satisfied the first prong because he was “entitled
to have relevant information of mental infirmity” presented
to the jury, but the court held that the second prong was
not met because “counsel had a reasonable basis for
proceeding as they did during the penalty phase.” App.
2028. The court made the following findings:
• Drs. Cooke and Sadoff, are “recognized experts in the
field of psychiatry and psychology.”
• “These experts administered tests, evaluated Mr.
Rompilla, and reported back to defense counsel.”
• They found “no organic brain damage” and “nothing that
could be used in mitigation.”
• They diagnosed Rompilla as a “sociopath,” and this
evidence “would not have been of benefit to Mr.
Rompilla’s case.”
• The defense attorneys provided Drs. Cooke and Sadoff
“with whatever they asked for,” and Drs. Cooke and
Sadoff did not request the records later unearthed by
PCRA counsel.
• Trial counsel “also obtained an evaluation by Dr. Paul
Gross, a well respected Lehigh Valley psychiatrist,” and
Dr. Gross “found nothing that would have been
beneficial in the penalty phase.”
• Rompilla did not provide trial counsel with “any
indication of mental problems or alcoholic blackouts” or
anything else that was “particularly useful.”
 18

• Trial counsel “spoke with members of the family in a
detailed manner,” and the family did not reveal the
family background information adduced in the PCRA
proceeding.
• The family members’ testimony at the PCRA hearing was
not credible insofar as it “attempted to contradict what
defense counsel indicated was asked of them during
numerous communications prior to trial.”
App. 2028-2029.
Based on these findings, the PCRA concluded:
• “Given the fact that [the] three health care professionals
[retained by trial counsel], all of whom were experienced
forensic experts, had provided opinions . . . , and none
of them asked for more information, it was hardly
unreasonable or ineffective for defense counsel to have
relied upon their opinions.”
• Trial counsel was not ineffective in questioning family
members because family members were questioned “in a
detailed manner,” and contrary testimony at the PCRA
hearing was rejected.
• “Defense counsel was reasonable in believing that the
only avenue available to them was to ask the jury to
show mercy upon Mr. Rompilla.” “Under the
circumstances of this case, this strategy was not only an
appropriate one, but the only one reasonably available to
counsel.”
On appeal, the Pennsylvania Supreme Court also held
that Rompilla’s ineffective assistance claim lacked merit.
The Court agreed with the PCRA court “that trial counsel
was effective with respect to their investigation and
presentation of mitigating evidence.” Rompilla-2, 721 A.2d
at 790. Noting that trial counsel employed three experts to
evaluate Appellant” and that “the experts found nothing
helpful to Appellant’s case,” the Court added: “[W]e agree
with the PCRA court that under the facts of this case,
counsel reasonably relied upon their discussion with
Appellant and upon their experts to determine the records
needed to evaluate his mental health and other potential
mitigating circumstances.” Id.
 19

In the habeas proceeding, the District Court not only
disagreed with the decisions of the PCRA court and the
state supreme court on the ineffective assistance issue, but
the District Court found that those decisions were
unreasonable. Rompilla, 2000 WL 964750, at *14. However,
the Court stated that its decision was “a very close call . . .
because trial counsel performed so admirably according to
my review of the record” and further commented that its
review of the record revealed that “trial counsel were
intelligent, diligent and devoted to their task of representing
[Rompilla].” Id. at *8, *12.
The District Court found two omissions in the opinions of
the PCRA court and the state supreme court. First, the
District Court stated, the prior opinions lacked “an in depth
analysis of what the duty to investigate consists of in a case
of this nature.” Id. at *8. In particular, the District Court
faulted the state courts for failing to “include a reference to
the ABA Standards for Criminal Justice.” Id. Second, the
District Court found the state court opinions deficient for
failing to discuss alcoholism. Id. at *8. Observing that it
appeared that trial counsel “thought they could rely on the
experts to determine [Rompilla’s] general mental ability or
capacity,” the Court opined: “[I]t seems to me on balance
that [trial counsel] were obliged to go a bit farther to fulfill
their duty to investigate.” Id. at *9, *12. Concluding that
trial counsel’s performance at the penalty phase had been
inadequate and that Rompilla had been prejudiced, the
Court held that Rompilla was entitled to habeas relief with
respect to his sentence. Id. at *14.

IV.
Strickland is the Supreme Court precedent governing
ineffective assistance of counsel claims, Williams, 529 U.S.
at 391, and in order to show ineffective assistance under
Strickland, a defendant must demonstrate (1) that counsel’s
“representation fell below an objective standard of
reasonableness,” Strickland, 466 U.S. at 688; and (2) that
“there is a reasonable probability that, but for counsel’s
unprofessional errors, the result of the proceeding would
have been different.” Id. at 694. “Judicial scrutiny of a
counsel’s performance must be highly deferential,” and
 20

“every effort [must] be made to eliminate the distorting
effects of hindsight, to reconstruct the circumstances of
counsel’s challenged conduct, and to evaluate the conduct
from counsel’s perspective at the time.” Id. at 689. “There
is a ‘strong presumption’ that counsel’s performance was
reasonable.” Jermyn v. Horn, 266 F.3d 257, 282 (3d Cir.
2001). Thus, “a defendant must overcome the presumption
that, under the circumstances, the challenged action might
be considered sound trial strategy.” Bell, 122 S. Ct. at 1852
(quotations omitted).
“The test for ineffectiveness is not whether counsel could
have done more; perfection is not required. Nor is the test
whether the best criminal defense attorneys might have
done more. Instead the test is . . . whether what [counsel]
did was within the ‘wide range of reasonable professional
assistance.’ ” Waters v. Thomas, 46 F.3d 1506, 1518 (11th
Cir. 1995) (en banc) (quoting Strickland, 466 U.S. at 689).
Ultimately, the issue is not what conduct is “prudent or
appropriate, but only what is constitutionally compelled,”
Burger v. Kemp, 483 U.S. 776, 794 (1987) (quoting United
States v. Cronic, 466 U.S. 648, 665 n.38 (1984)), and the
Sixth Amendment “simply . . . ensure[s] that criminal
defendants receive a fair trial,” Strickland, 466 U.S. at 689.
See also Kokoralies v. Gilmore, 131 F.3d 692, 696 (7th Cir.
1997) (“The sixth amendment does not guarantee success
or entitle defendants to the best available counsel or the
most prudent strategies. . . . [T]he Constitution is satisfied
when the lawyer chooses a professionally competent
strategy that secures for the accused the benefit of an
adversarial trial.”)

V.
In this appeal, the Commonwealth contends that the
state courts reasonably applied the Strickland ineffective
assistance standard to the facts of this case. According to
the Commonwealth, the District Court erred by essentially
considering the ineffective assistance issue as if it were
conducting a de novo review — that is, by making its own
independent application of the Strickland test to the facts
rather than considering only whether the state supreme
court’s application of that test was reasonable. Rompilla of
 21

course disagrees, but he also goes further and argues that
we need not reach the unreasonable application issue at all
for two reasons: first, because the Pennsylvania Supreme
Court did not render a “decision” within the meaning of 28
U.S.C. § 2254(d)(1), and, second, because the state supreme
court’s decision is “contrary to” Strickland. We discuss each
of these arguments below.

A.
Rompilla argues that the Pennsylvania Supreme Court
did not render a “decision” on his Sixth Amendment claim
because “[t]he Pennsylvania Court’s opinion lacks any real
Sixth Amendment analysis.” Rompilla Br. at 98. Rompilla
criticizes the state supreme court for applying the three-
pronged test set out in its own cases rather than the two-
pronged Strickland test. Id. He contends that the
relationship between the three-part Pennsylvania
formulation and the Strickland test is “not clear,” and he
faults that state supreme court for supposedly failing to
address either prong of the Strickland test. Id. at 98-99. He
asserts that the state supreme court mistakenly thought
that his federal claim was that trial counsel did not conduct
any investigation at all, whereas his actual claim was that
trial counsel did not conduct a sufficiently “thorough”
investigation. Id. at 100. Finally, he criticizes the state
supreme court for failing “to render a ‘decision’ at all on
critical aspects of this claim,” viz., the prejudice prong of
Strickland and his arguments regarding alcoholism and
intoxication. Id.
In order to address these arguments, we must first
explain the critical difference under the habeas statute
between, on the one hand, the failure of a state court to
adjudicate a federal claim on the merits (something that
may occur if the state court misconstrues the federal claim)
and, on the other, the failure of a state court to hand down
an opinion that discusses every argument, sub-argument,
and legal authority offered by the habeas petitioner. For
purposes of the habeas statute, a failure to decide affects
the standard of review; a failure to discuss (either at all or
to the satisfaction of the habeas petitioner or the federal
court) is irrelevant.
 22

Under the habeas statute, as previously noted, if a claim
was “adjudicated on the merits” in a state court proceeding,
relief may not be granted “unless the adjudication of the
claim . . . resulted in a decision that was contrary to, or
involved an unreasonable application of, clearly established
Federal law.” 28 U.S.C. § 2254(d)(1)(emphasis added). An
“adjudication on the merits” “has a well settled meaning: a
decision finally resolving the parties’ claims, with res
judicata effect, that is based on the substance of the claim
advanced, rather than on a procedural, or other, ground.”
Sellan v. Kuhlman, 261 F.3d 303, 311 (2d Cir. 2001). See
also Neal v. Puckett, 286 F.3d 230, 235 (5th Cir. 2001)
(“adjudication ‘on the merits’ is a term of art that refers to
whether a court’s disposition of the case was substantive as
opposed to procedural”), cert. denied, 123 S. Ct. 963 (2003).
In Chadwick v. Janecka, 312 F.3d 597, 605-07 (3d Cir.
2002), cert. denied, 123 S.Ct. 1914 (2003), we discussed
this issue in detail and reviewed both pertinent Supreme
Court precedent and our own prior decisions in Everett v.
Baird, 290 F.3d 500, 507-08 (3d Cir. 2002), cert. denied,
537 U.S. 1107 (2003); Appel v. Horn, 250 F.3d 203, 210 (3d
Cir. 2001); and Hameen v. Delaware, 212 F.3d 226, 248
(3d Cir. 2000). We noted that under Weeks v. Angelone,
528 U.S. 225, 237 (2000), a state court may render an
adjudication or decision on the merits of a federal claim by
rejecting the claim without any discussion whatsoever.
Chadwick, 238 F.3d at 606. We explained that if an
examination of a state court opinion reveals that the state
court did not decide a federal claim on the merits, the
deferential standards of review set out in § 2254(d)(1) do not
apply. Id. at 605-07. But if the state court decided the
claim, the § 2254(d)(1) standards govern — regardless of the
length, comprehensiveness, or quality of the state court’s
discussion. Id. See also Sellan, 261 F.3d at 312 (a state
court adjudicates a claim on the merits when it “(1)
disposes of the claim ‘on the merits,’ and (2) reduces its
disposition to judgment. . . . [e]ven if the state court does
not explicitly refer to . . . relevant federal case law.”); Bell v.
Jarvis, 236 F.3d 149, 160 (4th Cir. 2000) (en banc) (section
2254(d) “does not require that a state court cite to federal
law in order for a federal court to determine whether the
state court decision is an objectively reasonable one.”), cert.
 23

denied sub nom., Bell v. Beck, 122 S. Ct. 74 (2001); Aycox
v. Lytle, 196 F.3d 1174, 1177 (10th Cir. 1999) (“The focus
is on the state court’s decision or resolution of the case.”);
Wright v. Secretary for Dep’t of Corrections, 278 F.3d 1245,
1255 (11th Cir. 2002) (section 2254(d)(1) “focuses on the
result, not on the reasoning that led to the result”) cert.
denied, 123 S.Ct. 1511 (2003).
Here, it is abundantly clear that the state supreme court
adjudicated Rompilla’s Sixth Amendment claim on the
merits. Although the state supreme court referred to its
own three-pronged ineffective assistance test rather than
the two-pronged Strickland test, the Pennsylvania Supreme
Court has explicitly held that the state standard is “the
same” as Strickland’s and that Pennsylvania law does not
provide “any greater or lesser protection” than the Sixth
Amendment. Commonwealth v. Pierce, 527 A.2d 973, 976-
77 (Pa. 1987). As we put it in Werts v. Vaughn, 228 F.3d
178, 203 (3d Cir. 2000) cert. denied, 532 U.S. 980 (2001),
the Pennsylvania Supreme Court has “opined that the
Pennsylvania standard judging ineffectiveness claims [is]
identical to the ineffectiveness standard enunciated by the
United States Supreme Court in Strickland.” Since the brief
filed in the state supreme court on Rompilla’s behalf in the
PCRA appeal left no doubt that the ineffectiveness claim
asserted was based on federal law, it is perfectly clear that
the state supreme court adjudicated that federal claim on
the merits but simply chose to address the claim within the
framework of its own, familiar three-part test, which it
regards as “the same” as Strickland.
Moreover, it is apparent that this essentially stylistic
choice on the part of the state supreme court had no effect
on its decision. The state supreme court held that the
second prong of the state formulation — requiring a
showing that “counsel had no reasonable basis for the act
or omission in question,” Rompilla-2, 721 A.2d at 789 —
had not been met because Rompilla’s trial “counsel acted
reasonably” “with respect to their investigation and
presentation of mitigation evidence.” Id. at 790. The second
prong of the state formulation is substantively
indistinguishable from the first prong of Strickland — which
requires a showing that counsel’s “representation fell below
 24

an objective standard of reasonableness”. Strickland, 466
U.S. at 688. Rompilla argues that the significance of the
first Pennsylvania prong is unclear9 and that the third
Pennsylvania prong differs from Strickland’s prejudice
prong,10 Rompilla Br. at 98-99, but because the state
supreme court held that Rompilla failed the second state
prong, which is indistinguishable from the first Strickland
prong, these arguments are entirely beside the point.
Rompilla’s remaining arguments on the question whether
the state supreme rendered an “adjudication” or “decision”
on the merits of his federal ineffectiveness claim require
little additional discussion. There is plainly no merit to
Rompilla’s argument that the state supreme court
mistakenly thought that his Sixth Amendment claim was
that counsel failed to conduct any investigation at all
regarding the matters at issue and did not appreciate that
his claim was that counsel did not conduct a sufficiently
thorough investigation. Although the state supreme court
did use the phrase “failed to investigate,” 721 A.2d at 790,
it seems clear that this was simply a shorthand way of
referring to the claim. The court’s discussion of the various
steps that counsel took and its conclusion that counsel
acted “reasonably” make it clear that the court did not
think that any investigation at all would suffice but instead
understood that the extent of the investigation had to be
“reasonabl[e].”

9. Rompilla takes exception with the first prong of the Pennsylvania
formulations — whether the underlying claim has arguable merit.
Although the question of the merit of an underlying claim is not an
explicit step under Strickland, we have held that it is a determinative
factor in the “deficient performance” prong of the Strickland analysis in
at least some contexts. See Parrish v. Fulcomer, 150 F.3d 326, 328 (3d
Cir. 1998) (counsel not ineffective for failing to raise meritless claims).
10. Rompilla notes that the third Pennsylvania prong, as stated by the
Pennsylvania Supreme Court’s opinion, asks whether counsel’s act or
omission “would have” produced a different outcome, not whether, as
stated in Strickland, there is a reasonable probability of a different
outcome. Strickland, 466 U.S. at 694. Since neither the decision of the
state supreme court nor our decision in this appeal turns on the issue
of prejudice, we have no occasion to explore the question whether in
practice prejudice is analyzed differently under these two standards.
 25

What we have said about the relationship between the
Pennsylvania formulation of the test for ineffective
assistance and the Strickland formulation is sufficient to
dispose of Rompilla’s criticism of the state supreme court
for failing to discuss Strickland in so many words.
Rompilla’s criticism of the state supreme court for failing to
decide whether the prejudice prong of Strickland was
satisfied is further flawed because, in light of the state
supreme court’s holding that counsel’s performance was
reasonable, the court had no need to address the issue of
prejudice. See Strickland, 466 U.S. at 697 (“[T]here is no
reason for a court deciding an ineffective assistance claim
to . . . address both components of the inquiry if the
defendant makes an insufficient showing on one.”); Sistrunk
v. Vaughn, 96 F.3d 666, 673 n.8 (3d Cir. 1996) (if counsel’s
performance is objectively reasonable, there is no need to
decide prejudice).
Finally, Rompilla’s complaint that the state supreme
court “failed to discuss any aspect” of his arguments
regarding evidence of alcoholism and intoxication goes to
the style of the state court’s opinion, not the question
whether the state court rendered an adjudication on the
merits. There are many different theories about how judicial
opinions should be written. While some opinions make a
point of specifically addressing every argument and every
significant legal authority offered by counsel, others favor
brevity and comment on only those points that the court
finds most important. Because the state supreme court in
this case rendered an “adjudication on the merits” of
Rompilla’s Sixth Amendment claim, Rompilla’s criticisms of
the state court’s opinion cannot free him from the
restrictive standards of review set out in 28 U.S.C.
§ 2254(d)(1).

B.
There is similarly no merit in Rompilla’s contention that
the decision of the state supreme court was “contrary to”
Strickland because it did not “identify and apply actual
Sixth Amendment standards.” Rompilla Br. at 101. A
decision is “contrary to” a Supreme Court holding if the
state court “contradicts the governing law set forth in [the
 26

Supreme Court’s] cases” or if it “confronts a set of facts that
are materially indistinguishable from a decision of th[e
Supreme] Court and nevertheless arrives at a [different]
result.” Williams, 529 U.S. at 405-06.
In this case, the critical standard applied by the state
supreme court — whether trial counsel had a “reasonable
basis for the act[s] or omission[s] in question” — did not
contradict Strickland but rather, as we have explained, was
entirely consistent with Strickland. In Wertz v. Vaughn, 228
F.3d at 202-04, as noted, we compared Pennsylvania’s
ineffective-assistance test with the test enunciated in
Strickland, and we found that a state court decision that
applied the Pennsylvania test did not apply a rule of law
that contradicted Strickland and was thus not “contrary to”
established Supreme Court precedent. Id. at 204. In the
instant case, the state court applied the same state test.
Accordingly, here, as in Wertz, the state court’s application
of that test does not mean that its decision is “contrary to”
established Supreme Court precedent.
Rompilla also argues that the state court decision is
contrary to Strickland because counsel’s failure to obtain
records, without knowing the contents of the records,
“cannot be deemed a reasonable strategy/tactic, as a matter
of Sixth Amendment law,” and because “Sixth Amendment
law does not condition [the] obligation [to obtain records] in
any way on counsel’s duties with regard to expert
evaluations.” Rompilla Br. at 106, 108 (emphasis in
original). Rompilla is essentially challenging the manner in
which the state court applied the law to the facts, and thus
this argument must be analyzed under the “unreasonable
application” clause of § 2254(d)(1). See Williams, 529 U.S.
at 406 (a “run-of-the-mill state court decision” that is “in
accord with . . . Strickland as to the legal prerequisites for
establishing an ineffective-assistance claim . . . may be
contrary to the federal court’s conception of how Strickland
ought to be applied in that particular case” but is not
contrary to Strickland itself). For these reasons, we reject
Rompilla’s arguments under the “contrary to” clause of
§ 2254(d)(1).
 27

VI.
We now turn to the ‘unreasonable application’ clause of
§ 2254(d)(1). To obtain relief under this clause, Rompilla
must do more than show that he would have satisfied
Strickland’s test if his claim were being analyzed in the
first instance, because under § 2254(d)(1), it is not
enough to convince a federal habeas court that, in its
independent judgment, the state-court decision applied
Strickland incorrectly. Rather, he must show that the
[state court] applied Strickland to the facts of his case
in an objectively unreasonable manner.
Bell, 122 S. Ct. at 1852 (citation omitted). In other words,
Rompilla must demonstrate that “the state court decision,
evaluated objectively and on the merits, resulted in an
outcome that cannot reasonably be justified under
Strickland.” Wertz, 228 F.3d at 204.
We hold that the Pennsylvania Supreme Court’s
conclusion that trial counsel acted reasonably and rendered
effective assistance was not an unreasonable application of
Strickland. The findings of the PCRA court and
uncontradicted testimony at the PCRA hearing establish
that trial counsel conducted an extensive investigation for
mitigating evidence. According to their testimony, trial
counsel got to know Rompilla well during the course of
their representation and established a good relationship
with him. Rompilla was questioned about his background
but provided no useful information or leads. Trial counsel
also spoke to three of Rompilla’s siblings, as well as a
sister-in-law and Rompilla’s ex-wife. Family members were
questioned “in a detailed manner,” but they did not allude
to any of the new evidence adduced at the PCRA
proceeding.
Trial counsel retained three well-qualified mental health
experts to examine Rompilla. Dr. Cooke, a psychologist,
testified that in a case of this type he would have looked for
mitigating evidence and would have performed a battery of
tests, including an IQ test. Based on his letter to the Public
Defender’s office, he concluded that Rompilla’s IQ test must
not have shown mental retardation and that his evaluation
must not have revealed any abnormalities that would have
 28

been useful in showing mitigation. Dr. Sadoff, a
psychiatrist, interviewed Rompilla and evaluated him for
the presence of mitigating factors, but Dr. Sadoff inferred
that he also found nothing useful. A second psychiatrist,
Dr. Gross, was also retained but similarly found nothing
useful. In view of this record, we cannot say that the
Pennsylvania Supreme Court unreasonably applied
Strickland in concluding that trial counsel’s investigation
regarding mitigating evidence relating to Rompilla’s family
background and mental condition did not fall below the
Sixth Amendment floor.
Rompilla criticizes many aspects of trial counsel’s
performance at the penalty phase, but we see no ground for
relief under the habeas statute. Rompilla faults his trial
attorneys for failing to interview two sisters who testified at
the PCRA hearing — Barbara Harris, the oldest of
Rompilla’s siblings, and Randi Rompilla, the second
youngest. But trial counsel did interview three other
siblings, including two who were a few years older than
Rompilla (Nicholas Rompilla and Sandra Whitby) and one
Robert Rompilla, who was younger. At least some of the
siblings who were interviewed must have been aware of the
lurid conditions in the family home that were portrayed at
the PCRA hearing, but they never mentioned anything
about these matters to trial counsel despite being
interviewed “in a detailed manner.” It was thus not
constitutionally ineffective for trial counsel to fail to
anticipate that interviewing Barbara Harris or Randi
Rompilla would have yielded important new information
about the family home.
Rompilla also criticizes his trial attorneys for allegedly
failing to ask sufficiently specific questions when
interviewing family members and for allegedly failing to ask
Nicholas Rompilla about any time period other than the
three and one-half months just before the murder. This
argument is defeated by findings of fact made by the state
courts. The PCRA found that trial counsel “spoke with
members of the family in a detailed manner.” App. 2028.
The PCRA court also rejected Nicholas Rompilla’s testimony
insofar as it contradicted “what defense counsel indicated
was asked . . . during numerous conversations prior to
 29

trial.” Id. at 2030. Testimony at the PCRA hearing supports
these findings. See, e.g., id. at 1303 (“I remember [Dantos]
specifically going one by one and talking to him. ‘Is there
anything you can tell me? Tell me about yourself. Tell by
about your background.’ She was, you know, meticulous to
cover points.”). The state courts’ findings must be
presumed to be correct, and Rompilla has not rebutted that
presumption “by clear and convincing evidence.” 28 U.S.C.
§ 2254(e)(1). We accordingly reject Rompilla’s argument
regarding the interviewing of family members.
Rompilla contends that trial counsel should have sought
out school, medical, police, and prison records, but this
argument also falls short of meeting the AEDPA standard.
With the benefit of hindsight, we know that these records
contain useful information about Rompilla’s childhood
home environment, his mental problems, and his problems
with alcohol, but trial counsel had grounds for believing
that if there was any mitigating evidence of this sort to be
found, at least a hint of its availability would be disclosed
in the interviews with Rompilla and his family members or
in the testing and evaluations performed by the three
mental health experts whom they retained.
Charles, the lead trial attorney who had the final say on
every issue in the case, App. 1258, explained why these
records were not sought out. He testified that he would
have done whatever was needed to get records if there had
been any indication from the interviews or from the
information provided by the mental health experts that
such records would be helpful. Id. at 1307. But since these
interviews and evaluations suggested that Rompilla’s home
environment, schooling, and mental condition were not
promising avenues of investigation in the search for
mitigating evidence, he did not think that obtaining those
records would have represented a sound allocation of his
office’s resources. Id. at 1293-94.
This explanation provides a reasonable basis for
counsel’s decision not to seek the records. Defense counsel
was permitted to rely on statements made by their client in
deciding on the extent of the investigation that should be
conducted in particular areas. See Strickland, 466 U.S. at
691 (“The reasonableness of counsel’s actions may be
 30

determined or substantially influenced by the defendant’s
own statements or actions.”); id. (“[W]hen a defendant has
given counsel reason to believe that pursuing certain
investigations would be fruitless or even harmful, counsel’s
failure to pursue those investigations may not later be
challenged as unreasonable.”). As noted counsel’s
interviews with Rompilla himself and with family members
provided no indication of Rompilla’s abusive childhood or
mental problems. See Thomas, 144 F.3d at 515 (“if the
client and his family and friends throw the lawyer off the
scent . . . by denying the existence of psychological
problems that might have provided an alternative theory of
mitigation, the lawyer cannot be faulted for failing to go
down the path thus closed off ”). “The right to counsel does
not require that a criminal defense attorney leave no stone
unturned and no witness unpursued. [Although] it does
require a reasoned judgment as to the amount of
investigation the particular circumstances of a given case
require[,] [a]n attorney need not fully investigate every
potential avenue if he or she has reasonable grounds for
not doing so.” Berryman v. Morton, 100 F.3d 1089, 1101
(3d Cir. 1996). See also Thomas v. Gilmore, 144 F.3d 513,
515 (7th Cir. 1998) (“A reasonable investigation is not . . .
the investigation that the best criminal defense lawyer in
the world, blessed not only with unlimited time and
resources but also with the inestimable benefit of hindsight,
would conduct.”) cert. denied, 525 U.S. 1123 (1999).
It was likewise not unreasonable for trial counsel to rely
on its mental health experts to detect whether there was
any basis for further pursuit of mitigating evidence relating
to their client’s mental condition. Trial counsel retained no
fewer than three highly qualified experts. Dr. Cooke and Dr.
Sadoff looked for any evidence that could be used as
mitigation but found none. A battery of tests was performed
but yielded no indication of mental retardation or anything
else that would have been useful for mitigation. Although
all three of the experts testified that the records that PCRA
counsel subsequently obtained would have caused them to
do further investigation, none of the experts asked for
records or suggested that any further testing be done.
In view of these circumstances, it was not unreasonable
 31

for the state courts to conclude that trial counsel did not
fall below the constitutionally mandated level of
representation by failing to search out the records at issue
and by failing to provide those records to their mental
health experts. This was not a case where counsel knew of
the defendant’s mental health or other problems and failed
to inform or provide the experts with the information. See
Caro v. Calderon, 165 F.3d 1223, 1228 (9th Cir. 1999) (“A
lawyer who knows of but does not inform his expert
witnesses about . . . essential pieces of information going to
the heart of the case for mitigation does not function as
‘counsel’ under the Sixth Amendment.”) cert. denied, sub
nom, Woodford v. Caro, 527 U.S. 1049 (1999). Despite
counsel’s attempts to find out otherwise, counsel had no
knowledge of Rompilla’s problems and/or childhood abuse.11
In Card v. Dugger, 911 F.2d 1494 (11th Cir. 1990), the
court addressed an argument very similar to Rompilla’s
argument regarding trial counsel’s failure to provide records
to their mental health experts. In Card, a capital case, the
mental health expert spoke to members of the defendant’s

11. For this reason, the authority cited by Rompilla is inapposite. See
Jermyn, 266 F.3d at 275 (counsel knew that expert was willing to testify
regarding defendant’s abusive childhood and its mitigating effects on the
defendant’s adult life); Profitt v. Waldron, 831 F.2d 1245, 1246-49 (5th
Cir. 1987) (counsel knew defendant escaped from mental hospital shortly
before the crime); Beavers v. Balkcom, 636 F.2d 114, 115-16 (5th Cir.
1981) (counsel knew of defendant’s prior hospitalization in mental
institution); Glenn v. Tate, 71 F.3d 1204,1208 (6th Cir. 1995) (counsel
knew defendant had received mental health counseling) cert. denied, 519
U.S. 910 (1996); Antwine v. Delo, 54 F.3d 1357, 1365-66 (8th Cir. 1995)
(counsel knew that defendant was acting “odd” prior to the offense) cert.
denied sub nom, Bowersox v. Antwine, 516 U.S. 1067 (1996); Hill v.
Lockhart, 28 F.3d 832, 838-40, 845 (8th Cir. 1994) (counsel knew of
defendant’s prior hospitalization); Kenley, 937 F.2d at 1300, 1304-06
(counsel knew of but failed to contact a doctor who had previously
treated the defendant); Clabourne v. Lewis, 64 F.3d 1373, 1377, 1388
(9th Cir. 1995) (counsel knew defendant was mentally impaired at the
time of the offense, and had “display[ed] a broad spectrum of psychiatric
and emotional disorders”); Wallace v. Steward, 184 F.3d 1112, 1114,
1116 (9th Cir. 1999) (counsel knew from first sentencing hearing of
defendant’s “chaotic upbringing,” substance abuse, and mother’s mental
illness) cert. denied, 528 U.S. 1105 (2000).
 32

family and was provided copies of pre-trial reports
indicating that the defendant was competent to stand trial.
Id. at 1512. The expert testified at the sentencing hearing
that the defendant suffered from a sociopathic personality
disorder and that his abusive childhood was consistent
with the development of the disorder. Id. at 1508. The
defendant claimed that counsel was ineffective in failing to
provide the expert with various school, incarceration, army,
and medical records and information from family members.
Id. at 1512. This information, the defendant argued, would
have enabled the expert to diagnose the defendant as
suffering from organic brain damage and schizophrenia,
rather than merely a sociopathic personality. Id. at 1512-
13. The Eleventh Circuit rejected this argument observing
that
[t]here is no indication that the experts felt incapable of
basing their conclusions on the information they
obtained through their own testing and examinations.
Nor is there any reason that, after receiving the
experts’ reports, counsel was obligated to track down
every record that might possibly relate to [the
defendant’s] mental health and could affect a
diagnosis.
Id. at 1512.
We find this reasoning persuasive. As in Card, there was
no indication in the present case that Drs. Cooke, Sadoff,
and Gross “felt incapable of basing their conclusions on the
information they obtained through their own testing and
examinations.” Card, 911 F.2d at 1512. Thus, it was not
unreasonable for counsel to rely on the expertise and
experience of the mental health professionals in
determining the need for any records. Such deference to the
legitimate role of experts is well within the wide range of
reasonable professional assistance.
Furthermore, it was permissible for Charles to consider
his office’s limited investigative resources in determining
the extent of the investigation that should be conducted
with respect to Rompilla’s childhood, family, and mental
condition. See McWee v. Weldon, 283 F.3d 179, 188 (4th
Cir. 2002) (“the reasonableness of an investigation . . .
 33

must be considered in light of the scarcity of counsel’s time
and resources in preparing for a sentencing hearing”), cert.
denied, 537 U.S. 893 (2002); Mahaffey v. Page, 151 F.3d
671, 685 (7th Cir.) (“A ‘reasonable investigation’ does not
mandate a ‘scorch-the-earth strategy,’ a requirement that
would fail to consider the limited time and resources that
defense lawyers have in preparing for a sentencing
hearing.”) (citations omitted), vacated in part on other
grounds, 162 F.3d 481 (7th Cir. 1998); Rogers v. Zant, 13
F.3d 384, 387 (11th Cir. 1994) (the focus on whether a
decision not to conduct a particular investigation was
reasonable “reflects the reality that lawyers do not enjoy the
benefit of endless time, energy or financial resources”).
Rompilla argues that his trial attorneys were deficient
because they did not provide adequate information or
instructions to Drs. Cooke, Sadoff, and Gross. Rompilla
maintains that his attorneys should have instructed Dr.
Cooke to test for brain damage, but Charles testified that
he relied on the experts to do whatever testing was
appropriate. App. 1307-08. Dr. Cooke was aware of the
need for mitigating evidence. He testified that in a capital
case he would have looked for indications of “any mental
illness, emotional disturbance, or other sorts of problems
that might be a psychological mitigating circumstance,” and
he performed the battery of tests that he felt was called for.
Id. at 1808-09. He stated that if the results of some of
those tests had provided a reason to do a “neuropsyche
test,” he would have done so. Id. at 1809. Thus, Rompilla
is in effect faulting his trial attorneys for failing to instruct
an expert to do a test that the expert apparently did not
think was warranted under the circumstances. This
argument demands much more than the Sixth Amendment
requires.
Rompilla criticizes his trial attorneys for allegedly failing
to educate Dr. Sadoff about the meaning of mitigating
evidence. Based on a snippet of Dr. Sadoff ’s testimony,
Rompilla suggests that Dr. Sadoff thought that “mitigating
evidence was limited to evidence about the offense itself,”
Rompilla Br. at 87 (citing App. 1874), whereas under the
law, “mitigation is anything that ‘might serve as a basis for
a sentence less than death’ and need ‘not relate specifically
 34

to [the defendant’s] culpability for the crime he
committed.’ ” Rompilla Br. at 88 (quoting Skipper v. South
Carolina, 476 U.S. 1, 4-5 (1986)) (emphasis and brackets in
original). There are at least three weaknesses in this
argument.
First, in light of Dr. Sadoff ’s background — which
included extensive experience relating to the criminal
justice system (see App. 1876-1892) — it is far from clear
that trial counsel had a constitutional obligation to brief
him on the meaning of exculpatory evidence.12 Second, it
also is not clear that Rompilla’s argument accurately
represents what Dr. Sadoff said. Asked by Rompilla’s PCRA
attorney whether he thought that evidence was mitigating
only “if it relates to the defendant’s state of mind at the
time of the commission of the offense,” Dr. Sadoff
answered: “It relates to that and the kind of person he is
. . . .” App. 1874. (emphasis added). He continued that in
the past he had not “used irrelevant material that had
nothing to do with the crime itself as a mitigating factor. It
had to be something that would have impacted on [the
defendant’s] state of mind at the time.” Id. Thus, Dr. Sadoff
did not say that he thought that mitigating evidence had to
be evidence “about the offense.” Rompilla Br. at 87
(emphasis in original removed). Rather, he said that the
mitigating evidence that he had used in the past was
evidence of something that had an impact on the
defendant’s mind at the time. App. at 1874. It is not
apparent that mental health evidence is likely to be
persuasive for mitigation purposes if it cannot have had
any impact on a murderer’s mind at the time of offense.

12. Both Drs. Sadoff and Cooke had previously testified in criminal
proceedings. See, e.g., United States v. Green, 544 F.2d 138, 144 (3d Cir.
1976) (Sadoff and Cooke); New Jersey v. Zola, 548 A.2d 1022, 1030 (N.J.
1988) (Sadoff and Cooke); Commonwealth v. Comitz, 530 A.2d 473, 603-
04 (Pa. Super. Ct. 1987) (Sadoff); Commonwealth v. Stark, 526 A.2d 383
(Pa. Super. Ct. 1987) (Sadoff). They had also previously testified in
capital cases. See, e.g., New Jersey v. Pitts, 562 A.2d 1320, 1325-27
(N.J. 1989) (Sadoff and Cooke); Commonwealth v. Aulisio, 522 A.2d 1075,
1087 (Pa. 1987) (Cooke); Commonwealth v. Terry, 521 A.2d 398, 405,
409 (Pa. 1987) (Cooke); Commonwealth v. Whitney, 512 A.2d 1152, 1157
(Pa. 1986) (Cooke); Commonwealth v. Terry, 462 A.2d 676 (Pa. 1983)
(Sadoff and Cooke).
 35

Third and most important, even assuming for the sake of
argument that Dr. Sadoff ’s understanding of the range of
mental health evidence that can be mitigating was unduly
narrow, we see absolutely no reason to believe that Dr.
Sadoff ’s understanding contributed in any way to his
failure to detect the conditions that Rompilla now claims.
Rompilla argues that he “suffers from organic brain
damage, including damage to the frontal lobes and parietal
area” and that these impairments, among other things,
prevent him from “adequately control[ling] and guid[ing] his
behavior,” “emotions,” and “impulses.” Rompilla Br. at 58.
He argues that these conditions “substantially impair[ ] his
capacity to appreciate criminality of conduct or conform
conduct to the requirements of law.” Id. at 96. Accordingly,
these impairments surely would have “impacted” his mind
at the time of the Scanlon murder and thus fall squarely
within the scope of the sort of mitigating evidence to which
Dr. Sadoff referred.
Rompilla argues that his trial attorneys were ineffective
because they did not ask Dr. Gross to look for mitigating
evidence. Rompilla Br. at 65 (citing App. 1535-36, 1545).
However, Drs. Cooke and Sadoff did look specifically for
mitigating evidence, and the scope of Dr. Gross’s evaluation
encompassed at least some mitigating evidence. He stated:
“I was looking to see if he knew what he was doing, whether
he was suffering from any kind of psychiatric disorder,
whether he may have been intoxicated, exactly what his
emotions and cognitive processes were.” App. 1506. In light
of what Drs. Cooke and Sadoff were asked to do and in fact
did, trial counsel’s instructions to Dr. Gross do not show
that trial counsel were ineffective.
Finally, Rompilla argues that trial counsel were
ineffective in failing to do further investigation relating to
Rompilla’s alcohol use. Rompilla never explains exactly
what alcohol-related mitigation argument he believes his
trial attorneys should have made, but he refers to
information suggesting that there is a history of alcohol-
abuse in the Rompilla family, that he is an alcoholic, that
his alcoholism was caused by his traumatic childhood, that
he suffers from alcohol-induced blackouts, and that he was
intoxicated on the night of the murder. While these
 36

elements can be woven into an attractive mitigation
argument, most of the elements are based on information
in the records that trial counsel was not, for the reasons we
have explained, constitutionally obligated to obtain. This
includes evidence of drinking in the Rompilla family,
Rompilla’s abusive childhood home, blackouts, and
Rompilla’s alcoholism.
The only alcohol-related mitigation argument that trial
counsel could have made without obtaining the previously
discussed records — namely, that Rompilla was so
intoxicated at the time of the Scanlon murder that his
judgment was impaired — would not have been particularly
compelling. Moreover, while evidence to support this
argument was available,13 the argument would have to have
been made without the support of testimony from Rompilla,
who did not take the stand at either phase of the trial and
denied intoxication or alcohol problems. We cannot say that
it was objectively unreasonable for trial counsel not to put
on such a defense.
Finally, Rompilla argues that the Supreme Court’s recent
decision in Wiggins v. Smith, 123 S.Ct. 2527 (2003), shows
that his trial attorneys conducted an unreasonably
abbreviated investigation of potential mitigating factors, but
there is a sharp contrast between the conduct of the
attorneys in Wiggins and those in this case. In Wiggins, the
trial attorneys claimed that they made a tactical decision
not to present mitigating evidence but instead to pursue
the alternative strategy of attempting to prove that someone
other than the defendant actually killed the victim. Id. at
2533. Accordingly, they made little effort to investigate the
defendant’s background although they had reason to
believe that such an investigation would have been fruitful.
The Supreme Court found that defense counsel’s
investigation of their client’s background was based entirely
on the presentence investigation report (PSI) and
documents from the Baltimore City Department of Social
Services (DSS). Id. at 2540. Although funds were available

13. As Rompilla points out, police reports on the Scanlon case indicated
that “Rompilla was drinking heavily and was intoxicated and incoherent
around the time of the offense.” Rompilla Br. at 92.
 37

for the preparation of a social history of the defendant,
counsel did not commission such a report. Id. at 2536. But
even the limited documents that counsel reviewed
contained information that should have alerted counsel
that an expanded investigation into their client’s
background was likely to yield mitigation evidence. The PSI
reported that the defendant had spent most of his
childhood in foster care and quoted the defendant as
describing his background as “disgusting.” Id. at 2536. The
DSS records disclosed that the defendant’s “mother was a
chronic alcoholic; Wiggins was shuttled from foster home to
foster home and displayed some emotional difficulties while
there; he had frequent, lengthy absences from school; and,
on at least one occasion, his mother left him and his
siblings alone for days without food.” Id. at 2537.
Wiggins’s trial counsel nevertheless failed to follow these
leads. Counsel moved for bifurcation of the penalty phase
trial so that they could first attempt to show that their
client did not kill the victim by his own hand and then, if
that defense was rejected, offer evidence in mitigation. Id. at
2532. When bifurcation was denied, defense counsel made
a brief reference in her penalty phase opening to the
defendant’s “difficult life” but “introduced no evidence of
[his] life history.” Id. And although counsel proffered some
evidence that would have been introduced if the bifurcation
had been allowed, no evidence of the defendant’s life history
or family background was included. Id. After the defendant
was sentenced to death, new counsel commissioned the
preparation of a social history. Id. at 2532. A social worker
interviewed Wiggins and family members, obtained
additional documents, and compiled evidence of severe
physical and sexual abuse by Wiggins’s mother. Id. at
2532, 2542.
Faced with these facts, the Supreme Court held that it
was objectively unreasonable for counsel not to conduct a
further investigation for mitigating evidence regarding
Wiggins’s background. Id. at 2541. The Court stated that
“any reasonably competent attorney” would have realized
that it was necessary to follow the leads in the PSI and DSS
documents before making a choice among potential
defenses. Id. at 2537. The Court stressed that “counsel
 38

uncovered no evidence in their investigation to suggest that
. . . further investigation would have been fruitless.” Id. The
Court also emphasized that Strickland does not require
counsel to investigate every conceivable line of mitigating
evidence no matter how unlikely the effort would be to
assist the defendant at sentencing. Id. Instead, the Court
reiterated that “[a] decision not to investigate . . . ‘must be
directly assessed for reasonableness in all the
circumstances.’ ” Id. (quoting Strickland, 466 U.S. at 691).
There are critical differences between the conduct of
Wiggins’s and Rompilla’s trial attorneys. Wiggins’s attorneys
were presented with leads that “any reasonably competent
attorney” would have realized were promising. Rompilla’s
attorneys had no comparable leads. Wiggins described his
childhood as “disgusting”; Rompilla insisted that his was
“normal.” In Wiggins, postconviction interviews with
Wiggins himself and with members of his family produced
evidence of severe abuse. 123 S.Ct. at 2542. In the present
case, Rompilla’s attorneys interviewed their client and
numerous family members in detail before the penalty
proceeding but discovered no mitigating evidence. Wiggins’s
attorneys “uncovered no evidence in their investigation to
suggest that . . . further investigation would have been
fruitless.” Id. at 2537. Rompilla’s trial attorneys had a body
of evidence that suggested that a further investigation into
Rompilla’s family background would not have been
productive. Not only did Rompilla’s attorneys probe their
client for mitigating evidence, but they interviewed
numerous members of his family, including siblings who
bracketed him in age. At least one of these siblings,
Nicholas Rompilla, later revealed that he knew about the
conditions in the home on which Rompilla now relies, but
he never provided that information to trial counsel, and
although he now claims that trial counsel never asked him
about any period of time other than the months
immediately preceding the murder, the PCRA court found
that this testimony was not believable.
In short, the attorneys in Wiggins did little to investigate
their client’s background although they possessed
information that should have prompted them to do so.
Rompilla’s attorneys conducted a much greater
 39

investigation, but their interviews with their client and his
family provided a reasonable basis for concluding that
additional investigation would not have represented a wise
allocation of limited resources. In our view, Wiggins is
critically different from the present case.
Rompilla asserts that Dr. Gross’s report “explicitly
recommend[ed] that counsel investigate alcoholism,”
Rompilla Br. at 66, but this is not exactly what Dr. Gross
wrote. Dr. Gross said that “[t]he possibility does remain . . .
that Mr. Rompilla while under the influence of alcohol, can
become prone to violent behavior although he himself
strongly denies this.” Dr. Gross added: “My
recommendation is that this area should be further
evaluated before any definitive conclusions are drawn.”
(emphasis added). Thus, Dr. Gross did not say that further
evaluation was likely to produce evidence that Rompilla
becomes violent when drunk, only that this was a
“possibility.” Nor did he recommend further evaluation as a
promising strategy for the defense. Instead, he merely said
that “no definitive conclusions” should be drawn about this
possibility unless there was a further evaluation. And
finally, it should be noted that he referred to a further
“evaluation” (a term that suggests a psychiatric
examination), not an “investigation.” Rompilla’s brief
repeatedly uses the term “investigate” in an apparent
attempt to suggest that Dr. Gross recommended that trial
counsel search for evidence such as the records discussed
above. The use of this term obscures the fact that trial
counsel did precisely what Dr. Gross recommended: they
had Rompilla “evaluated” further, i.e., examined by another
psychiatrist, Dr. Sadoff.
In conclusion, it appears to us that Rompilla is now
arguing that his trial counsel were constitutionally derelict
in failing to take all the steps that might have been pursued
by the most resourceful defense attorneys with bountiful
investigative support. But while we may hope for the day
when every criminal defendant receives that level of
representation, that is more than the Sixth Amendment
demands. See Strickland, 466 U.S. at 689 (“the purpose of
the effective assistance guarantee of the Sixth Amendment
is not to improve the quality of legal representation,
 40

although that is a goal of considerable importance to the
legal system”); Burger, 483 U.S. at 794 (“The record at the
habeas corpus hearing does suggest that [counsel] could
well have made a more thorough investigation than he did.
Nevertheless, in considering claims of ineffective assistance
of counsel, ‘[w]e address not what is prudent or
appropriate, but only what is constitutionally compelled.’ ”)
(quoting Cronic, 466 U.S. at 665 n.38) (alteration in
original). The Sixth Amendment is satisfied when “counsel’s
conduct falls within the wide range of reasonable
professional assistance,” thereby “ensur[ing] that criminal
defendants receive a fair trial.” Strickland, 466 U.S. at 689.14

14. In finding the state courts’ analysis unreasonable, the District Court
relied in part on the absence of any “in depth analysis” of the duty to
investigate, including the lack of any reference to the ABA Standards for
Criminal Justice. Rompilla, 2000 WL 964750, at *9. Three important
observations must be made about the District Court’s treatment of this
point. First, the state court’s failure to discuss these standards does not
mean that the standards themselves or the principles they embody were
not considered. Second, we see nothing in the quoted portions of the
ABA standards that dictates that records of the sort at issue here must
always be sought. Third, although the ABA standards are important
guides, they cannot be viewed as a codification of the requirements of
the Sixth Amendment. In Strickland, the Supreme Court recognized that
“[p]revailing norms of practice as reflected in American Bar Association
standards and the like . . . are guides to determining what is
reasonable.” Strickland, 466 U.S. at 688 (emphasis added). See also
Williams, 529 U.S. at 396. However, the Supreme Court has made it
clear that the ABA standards
are only guides. No particular set of detailed rules for counsel’s
conduct can satisfactorily take account of the variety of
circumstances faced by defense counsel or the range of legitimate
decisions regarding how best to represent a criminal defendant. Any
such set of rules would interfere with the constitutionally protected
independence of counsel and restrict the wide latitude counsel must
have in making tactical decisions. . . . [T]he purpose of the effective
assistance guarantee of the Sixth Amendment is not to improve the
quality of legal representation, although that is a goal of
considerable importance to the legal system. The purpose is simply
to ensure that criminal defendants receive a fair trial.
Strickland, 466 U.S. at 688-89 (citation omitted). See also Roe v. Flores-
Ortega, 528 U.S. 470, 481 (2000) (the Court has “consistently declined
to impose mechanical rules on counsel — even when those rules might
lead to better representation”).
 41

See also Waters, 46 F.3d at 1512 (“The test has nothing to
do with what the best lawyers would have done. Nor is the
test even what most good lawyers would have done . . . . We
are not interested in grading lawyers’ performances; we are
interested in whether the adversarial process at trial, in
fact, worked adequately.”) (quoting White v. Singletary, 972
F.2d 1218, 1220-21 (11th Cir. 1992)). Here, the state
court’s determination that counsel acted reasonably was
not “contrary to” or an “unreasonable application” of
Strickland.

VII.
Rompilla advances three claims related to the trial court’s
response to the jury’s questions about accomplice liability.
First, Rompilla argues that his federal constitutional right
to due process was violated because the court instructed
the jury on accomplice liability without providing “notice or
fair warning that he might be subjected to liability as an
accomplice.” Rompilla Br. at 14. Second, he contends that
his federal constitutional right to due process was abridged
because the accomplice-liability comments did not state
that “accomplice liability must be proven beyond a
reasonable doubt and that conviction for murder, even
based on accomplice liability, required proof of a specific
intent to kill.” Id. at 20. Third, he maintains that his rights
under the Sixth, Eighth, and Fourteenth Amendments were
violated because the trial judge’s remarks prevented the
jury “from considering and giving exculpatory and
mitigating effect to evidence that someone other than
Rompilla was the killer.” Id. at 22 (emphasis in original
removed).

A.
Under 28 U.S.C. § 2254(b)(1), a state prisoner’s
application for a writ of habeas corpus “shall not be granted
unless it appears that . . . the applicant has exhausted the
remedies available in the courts of the State” or the case
falls within one of two exceptions to the exhaustion
requirement, i.e., “there is an absence of available State
corrective process” or “circumstances exist that render such
 42

process ineffective to protect the rights of the applicant.”
Moreover, under 28 U.S.C. § 2254(b)(3), a state may not be
deemed to have waived exhaustion or be estopped from
relying on exhaustion “unless the State, through counsel,
expressly waives the requirement.” In order to exhaust state
remedies, a habeas petitioner must present the petitioner’s
federal claim in the state courts. See Keller v. Larkins, 251
F.3d 408, 413 (3d Cir. 2001). It is not enough for a
petitioner to advance a similar state-law claim. Duncan v.
Henry, 513 U.S. 364, 366 (1995). Rather, a petitioner must
“present a federal claim’s factual and legal substance to the
state courts in a manner that puts them on notice that a
federal claim is being asserted.” McCandless v. Vaughn, 172
F.3d 255, 261 (3d Cir. 1999).
Three sections of Rompilla’s brief on direct appeal to the
state supreme court raised claims concerning the
accomplice liability instruction. First, Rompilla argued that
the judge erred in giving the instruction because the
defense had no prior notice that this theory of liability was
in the case and because the judge refused a defense
request that “the jury be specifically instructed that under
an accomplice theory every element must be proven beyond
a reasonable doubt.” Appellant’s Br. at 11, Commonwealth
v. Rompilla, Supreme Court of Pennsylvania, Philadelphia
District, No. 53 Capital Appeal Docket (hereinafter
“Rompilla Direct Appeal Br.”). Although this section of
Rompilla’s brief did refer to the requirement of proof beyond
a reasonable doubt — something demanded by both the
federal Constitution and Pennsylvania law15 — the brief
made no reference to the federal Constitution and did not
cite any federal authorities. Second, Rompilla argued that
the judge erred in refusing to poll the jury as to whether
the guilty verdict was based on the theory of accomplice
liability. Rompilla Direct Appeal Br. at 15-17. This
contention was based exclusively on state law. Third,
Rompilla repeated the argument that the trial judge erred
in refusing to give the requested defense instruction to the
effect that conviction as an accomplice required proof
beyond a reasonable doubt. Id. at 17-18. Only state
authorities grounded on state law were cited.

15. See Commonwealth v. Bonomo, 151 A2d 441, 446 (Pa. 1959).
 43

Based on our examination of Rompilla’s brief to the
Pennsylvania Supreme Court, it is not clear that Rompilla
properly presented any federal constitutional claim to that
court.16 Moreover, even if Rompilla’s brief to the state
supreme court is read as having raised the federal claim
that the trial judge erred in refusing to give the requested
instruction on the requirement of proof beyond a
reasonable doubt, it seems clear that the remaining federal
constitutional claims advanced by Rompilla in this appeal
were not presented to the state supreme court. Rompilla’s
argument in the state supreme court concerning the lack of
prior notice was based solely on state authorities, and the
argument that Rompilla has made to us that the trial
judge’s instruction effectively prevented the jury from
considering exculpatory and mitigating evidence was never
made, either as a state- or federal-law claim, in Rompilla’s
direct appeal brief.
In view of these deficiencies, if Rompilla could still return
to the state courts and present the federal constitutional
claims that he has made here, we would probably be
obligated to regard those claims as unexhausted. It is clear,
however, that Rompilla can no longer raise those claims in
state court, 42 Pa. Cons. Stat. Ann. § 9545(b), and therefore
it appears that those claims are exhausted but are probably
procedurally defaulted. See Fahy v. Horn, 240 F.3d 239,
245 (3d Cir. 2001). The Commonwealth has not raised the
defense of procedural default, but we have the discretion to
do so in appropriate circumstances. See Sweger v.
Chesney, 294 F.3d 506, 520 (3d Cir. 2002) cert. denied,
123 S.Ct. 1902 (2003). Nevertheless, we do not rely on the
doctrine of procedural default in deciding this appeal.
Assuming for the sake of argument that our consideration
of those claims on the merits is not barred, we find that the
claims do not provide a basis for habeas relief.

16. On direct appeal, the state supreme court described Rompilla’s
argument as follows:
by being instructed on an accomplice theory and then being told to
disregard it, the jury might have been confused and convicted
[Rompilla] as an accomplice.
Rompilla-1, 653 A.2d at 631.
 44

B.
When language in jury instructions is challenged, the
language in question “must be considered in the context of
the instructions as a whole and the trial record.” Estelle v.
McGuire, 502 U.S. 62, 72 (1991). We then consider
“ ‘whether there is a reasonable likelihood that the jury has
applied the challenged instructions in a way’ that violates
the Constitution.” Smith v. Horn, 120 F.3d 400, 411 (3d Cir.
1997) (quoting McGuire, 502 U.S. at 72.)) Under these
standards, the trial judge’s comments about accomplice
liability did not violate Rompilla’s constitutional rights.

1.
Since there were no eyewitnesses to the killing, the
Commonwealth’s case was based almost entirely on the
previously noted circumstantial evidence. See Rompilla-1,
653 A.2d at 629. The defense argued that someone other
than Rompilla was the actual assailant.
During jury deliberations in the guilt phase, the jury
asked:
If Defendant was an accomplice to the Charge of
Criminal Homicide, can he be charged with Murder in
the First Degree?
App. at 585. In response, the trial judge began by informing
the jury that the Commonwealth had not proceeded on the
theory that Rompilla was liable for the murder as an
accomplice and that no instruction on accomplice liability
had been given. The judge stated:
Well, preliminarily, let me say this. If you recall, the
Court, in [its] Charge, at no time referred to any theory
of accomplice in this case. It is not the Commonwealth’s
theory in this case that the Defendant was an
accomplice. I’m not quite sure what prompted the
question. If it arose because of something that was
argued by Counsel in the closing speeches, if you
recall, I indicated that you are to give — and you
should be guided by each Lawyer’s arguments to the
extent they’re supported by the evidence and in so far
as they aid you in applying your own reason and
 45

common sense. However, you are not required to
accept the arguments of either Lawyer. It is for you
alone — you and you alone to decide this case based
on the evidence as it was presented from this witness
stand and in accordance with the instruction which the
Court gave you. Now, again, as I say, I’m not quite sure
what prompted the question, but you did ask the
question, and I assume you want an answer, is that
correct?
Id. at 585-86 (emphasis added). The jury foreman nodded
in the affirmative, and the trial court then attempted to
explain when an accomplice could be held liable for a crime
committed by another. Id. at 586-87. The court concluded,
however, with the following comments:
I can only say this to you, that you should determine
this case based on the Charge of the Court as it had
given you originally. There was no evidence in this case
with regard to the question as posed but a simple clear
answer to the question is yes, you could be charged
with, and you could be convicted of Murder in the First
Degree if, indeed, you’re an accomplice, but it requires
other proof or findings on your part that were not a part
of this case nor were they made a part of this case, all
right.
Id. at 587-88 (emphasis added).
On direct appeal, the state supreme court did not decide
whether the trial judge’s comments about accomplice
liability were flawed but instead held that any error was
cured by the judge’s admonition to the jury that it should
not consider that theory of liability. Id.17 The District Court

17. In the PCRA appeal, Rompilla again attempted to challenge the trial
judge’s instruction on accomplice liability, but the state supreme court
held that relief was unavailable because the claim had already been
litigated on direct appeal. Commonwealth v. Rompilla, 721 A.2d 786, 793
(Pa. 1999).
In the federal habeas proceeding, the District Court also found that the
instruction did not prevent consideration of exculpatory and mitigating
evidence and that Rompilla’s claim that there was evidence of an
accomplice was not supported by the record. Id.
 46

— presented with arguments notably different from those
advanced in the state supreme court — went beyond the
state supreme court in holding that the judge’s explanation
of accomplice liability was incomplete, but the District
Court agreed with the state supreme court that any flaw
was cured by the trial court’s admonition that the theory of
accomplice liability was not before the jury in this case.
Rompilla, 2000 WL 964750, at *14. Similarly, we conclude
that, even assuming arguendo that the challenged
comments were constitutionally flawed, there is no basis for
habeas relief. The state supreme court’s decision that any
defect in the challenged comments was cured by the judge’s
admonition that the jury was not to consider the theory of
accomplice liability did not “involve an unreasonable
application of Federal law,” 28 U.S.C. § 2254(d)(1), namely,
the rule that jury instructions do not violate the
Constitution unless there is a reasonable probability that
they were applied in a way that abridged constitutional
rights. Moreover, even under a standard of plenary review,
we would find no constitutional violation.
Here, the clear thrust of the trial court’s response to the
jury’s question was that the Commonwealth was not relying
on the theory that Rompilla was liable as an accomplice
and that the jury should not consider that theory as a
possible basis for conviction. As noted, when the judge first
responded to the jury’s question, the judge began by
reminding the jury that the court’s instructions “at no time
referred to any theory of accomplice in this case. It is not
the Commonwealth’s theory in this case that the Defendant
was an accomplice.” App. 585-86. In addition, the judge
told that jury that they were required to decide the case “in
accordance with the instruction which the Court gave you.”
Id. Since “the instruction which the Court gave [the jury]”
“at no time referred to any theory of accomplice in this
case,” id. at 585-86, the judge’s comments meant that the
jury was to decide the case without relying on the theory of
accomplice liability. The judge’s final statements regarding
accomplice liability made exactly the same point. The judge
stated: “I can only say this to you, that you should
determine this case based on the Charge of the Court as it
had given you originally.” Id. at 587. Jurors are presumed
to follow the court’s instructions. Richardson v. Marsh, 481
 47

U.S. 200, 211 (1987). In view of the instructions given here,
it is not reasonably likely that the jury interpreted the
judge’s comments to mean anything other than that they
were not to consider accomplice liability as a possible
ground for conviction.
Once the meaning of the judge’s comments is
understood, it is apparent that there is no merit in
Rompilla’s argument that his right to due process was
violated because he did receive prior notice that he might
be subjected to liability as an accomplice or in his
argument that the judge’s explanation of accomplice
liability was constitutionally flawed. Since the trial judge
told the jury that it was to decide the case without relying
on the theory of accomplice liability, these arguments fail.
Rompilla contends that the trial court told the jury that
it should not consider accomplice liability because no
evidence was presented during the trial to support such a
finding. Rompilla Br. at 16. Rompilla then maintains that
the trial judge gave contradictory instructions in that, on
the one hand, he told the jury not to consider accomplice
liability because there was no evidence to support such a
finding and, on the other hand, he told the jury that it was
the sole factfinder and its recollection of the evidence
should therefore control. According to Rompilla, the jury
was left with two irreconcilable options — either to
disregard the theory of accomplice liability because the trial
court believed there was no evidence to support the theory
or to consider the theory if the jury, as sole factfinder,
believed, as suggested by its question, that there was
evidence of an accomplice.
This elaborate argument rests on the faulty premise that
the trial judge instructed the jury that the reason why the
theory of accomplice liability was not in the case was
because there was no supporting evidence. In fact, however,
the judge never made such a statement. The trial judge did
state that “[t]here was no evidence in this case with regard
to the question as posed,” but he did not state that this
was why the theory was not in the case. Rather, as noted,
the judge twice told the jury that they were not to consider
accomplice liability because it was not in the original
charge, and the judge also noted that “[i]t [was] not the
 48

Commonwealth’s theory in this case that the Defendant
was an accomplice.”
Rompilla’s remaining argument (that the trial judge’s
statements prevented the jury from considering exculpatory
and mitigating evidence) entirely misconstrues the judge’s
words. The judge did not, as Rompilla contends, say that
there was no evidence that another person was involved.
Rather, the judge stated that “[t]here was no evidence in
this case with regard to the question as posed.” The
“question as posed” was: “If Defendant was an accomplice
to the Charge of Criminal Homicide, can he be charged with
Murder in the First Degree?” Thus, the judge essentially
said that there was no evidence to support a conviction
under an accomplice liability theory; the judge did not say
that there was no evidence that might justify reasonable
doubt, or residual doubt for purposes of sentencing
mitigation, as to whether Rompilla committed the killing
with the requisite intent. Accordingly, nothing that the
judge said foreclosed consideration of exculpatory or
mitigating evidence.
In sum, after considering all of Rompilla’s arguments
regarding accomplice liability, we see no ground for habeas
relief.

VIII.
Rompilla contends, finally, that his future dangerousness
was put at issue during the sentencing phase of his trial
and that therefore, under Simmons v. South Carolina, 512
U.S. 154 (1994), the trial court was required to inform the
jury that “life” imprisonment under Pennsylvania law
means “life without parole.” We agree with the District
Court, however, that the Pennsylvania Supreme Court’s
rejection of this claim cannot be disturbed under the
narrow standard of review prescribed by AEDPA.

A.
Rompilla maintains that the state supreme court’s
decision was “contrary to” and represented “an
unreasonable application” of Simmons.18 The actual holding
in Simmons, however, as set out in Justice O’Connor’s

18. Simmons applies here because it was decided before Rompilla’s
conviction became final upon conclusion of his direct appeal in 1995. Cf.
O’Dell v. Netherland, 521 U.S. 151, 156-57 (1997).
 49

controlling opinion, is narrower than Rompilla
acknowledges.
In Simmons, the defendant was convicted of capital
murder for the killing of an elderly woman. Simmons, 512
U.S. at 156. The defendant had a history of assaulting
elderly women, and both defense and state witnesses
agreed that the defendant posed a continuing danger to
elderly women. Id. at 157. During the penalty phase, the
state argued that the question for the jury was “what to do
with [the defendant] now that he is in our midst.” Id. The
state also urged that a death sentence would be “a
response of society to someone who is a threat. Your verdict
will be an act of self-defense.” Id. The defendant sought to
rebut the state’s argument by presenting evidence that his
dangerousness was limited to elderly women and that there
was no reason to expect acts of violence once he was in a
prison setting. Id. The trial court denied defense counsel’s
request for a parole ineligibility instruction. Id. at 158-60.
During deliberations, the jury asked if a life sentence
included the possibility of parole. Id. at 160. The trial court
instructed the jury not to consider parole or parole
eligibility and told the jury that life imprisonment and
death were to be understood in their plain and ordinary
meaning. Id. Shortly thereafter, the jury returned with a
sentence of death. Id.
The United States Supreme Court held that, under these
circumstances, due process required the trial judge to
inform the jury that the defendant would not have been
eligible for parole if sentenced to imprisonment for life. The
plurality opinion endorsed by four Justices reasoned that
the jury reasonably may have believed that petitioner
could be released on parole if he were not executed. To
the extent this misunderstanding pervaded the jury’s
deliberations, it had the effect of creating a false choice
between sentencing petitioner to death and sentencing
him to a limited period of incarceration. This grievous
misperception was encouraged by the trial court’s
refusal to provide the jury with accurate information
regarding petitioner’s parole ineligibility, and by the
State’s repeated suggestion that petitioner would pose a
future danger to society if he were not executed.
 50

Id. at 161-62 (plurality)(emphasis added). The plurality
found that the jury was then “left to speculate about . . .
parole eligibility when evaluating petitioner’s future
dangerousness, and was denied a straight answer . . . even
when it was requested.” Id. at 165-66. Consequently, the
plurality stated that “where the defendant’s future
dangerousness is at issue, and state law prohibits the
defendant’s release on parole, due process requires that the
sentencing jury be informed that the defendant is parole
ineligible.” 512 U.S. at 156 (plurality)(emphasis added). Id.
at 161-62.19
The controlling opinion in Simmons — Justice O’Connor’s
concurrence in the judgment, in which the Chief Justice
and Justice Kennedy joined — seemed to phrase the
holding more narrowly. Justice O’Connor stated that a
“defendant should be allowed to bring his parole ineligibility
to the jury’s attention . . . in cases in which the only
available alternative sentence to death is life imprisonment
without possibility of parole and the prosecution argues that
the defendant will pose a threat to society in the future.” Id.
at 177 (O’Connor, J., concurring in the judgement)
(emphasis added). Other passages in Justice O’Connor’s
opinion pointed in the same direction. She referred to cases
“ ‘[w]here the prosecution specifically relies on a prediction of
future dangerousness in asking for the death penalty,’ ” id.
at 175 (emphasis added)(citation omitted)(brackets in
original) and cases where “the State seeks to show the
defendant’s future dangerousness.” Id. 177. See also id. at
176 (“the State sought to show that petitioner is a vicious
predator who would pose a continuing threat to the
community”); id. at 177 (“The prosecutor in this case put
petitioner’s future dangerousness in issue”).
In short, whereas the Simmons plurality opinion may be
read to state that the Simmons rule applies whenever a

19. Justice Souter’s concurring opinion, in which Justice Stevens joined,
was phrased in seemingly broader terms. See 512 U.S. at 172 (Souter,
J., concurring)(“at least when future dangerousness is an issue . . . .”).
The Simmons dissenters interpreted the plurality opinion as apparently
requiring admission of parole ineligibility even when the prosecution
does not argue future dangerousness. Id. at 183 (Scalia, J., joined by
Thomas, J., dissenting).
 51

defendant’s future dangerousness is “at issue” (whether or
not the prosecution makes that argument), Justice
O’Connor’s controlling opinion may be read more narrowly
to hold that the rule governs only where the prosecution
actually “argue[s]” that a defendant poses a future threat.
Neither position, of course, has perfectly clear boundaries.
There is a sense in which future dangerousness is “at
issue” at the penalty phase of virtually all capital cases. The
possibility that a first-degree murderer will kill again is
likely to be present in the jurors’ mind in all cases except
those where the defendant committed the murder for which
he or she was convicted only because of unusual
circumstances that are very unlikely to occur again even if
the defendant is released from prison. As the Chief Justice
has put it, “[i]t is difficult to envision a capital sentencing
hearing where the State presents no evidence from which a
juror might make [an inference of future dangerousness to
society].” Kelly v. South Carolina, 122 S.Ct. 726, 735
(2002)(Rehnquist, C.J., dissenting). Thus, read liberally, the
plurality’s position in Simmons would mandate that the
jury be informed about parole ineligibility in almost all
capital cases in which the only possible sentences are
death or life without parole.
On the other hand, the fine line that Justice O’Connor’s
opinion seemed to draw — between cases in which the
prosecution “argue[s]” future dangerousness and cases in
which future dangerousness is inferred by the jury from the
evidence that is brought to its attention — is difficult to
police and arguably superficial. A prosecutor may
encourage a jury to think about future dangerousness
without expressly referring to that concept.
Last year, in Kelly v. South Carolina, 122 S. Ct. 726
(2002),20 the Court arguably broadened the holding in
Simmons. In Kelly, the state told the jurors in its opening
at the penalty phase: “I hope you never in your lives again

20. Kelly was decided after briefing was completed and before oral
argument in this case. Rompilla submitted a Notice of Supplemental
Authority, pursuant to Fed. R. App. P. 28(j). After oral argument both
parties submitted supplemental memoranda addressing the issue of
whether Kelly was applicable to the instant matter.
 52

have to experience what you are experiencing right now.
Being some thirty feet away from such a person. Murderer.”
Kelly, 122 S. Ct. at 729. The state then presented evidence
that while in prison, Kelly had made a knife, had attempted
to escape from prison, and had planned to hold a female
guard as a hostage. Id. The state also brought out evidence
of “Kelly’s sadism at an early age, and his inclination to kill
anyone who rubbed him the wrong way.” Id. (citation
omitted). During closing arguments, the state referred to
Kelly as “the butcher of Batesburg,” “Bloody Billy,” and
“Billy the Kid.” Id. The state also told the jurors that “Kelly
doesn’t have any mental illness. He’s intelligent . . . He’s
quick-witted. Doesn’t that make somebody a little more
dangerous . . . . for this lady . . . doesn’t that make him
more unpredictable for [the victim] . . . . murderers will be
murderers. And he is the cold-blooded one right over there.”
Id. at 729-30. The trial court did not give a parole
ineligibility instruction. Id. at 730.
The Supreme Court held that Kelly’s future
dangerousness was sufficiently put at issue that he was
entitled to a parole ineligibility instruction. Id. at 731, 733-
34. In the body of its opinion, the Court stated that
“[e]vidence of future dangerousness under Simmons is
evidence with a tendency to prove dangerousness in the
future; its relevance to that point does not disappear merely
because it might support other inferences or be described
in other terms.” Id. at 732. The Court recognized that “it
may well be that the evidence in a substantial proportion,
if not all, capital cases will show a defendant likely to be
dangerous in the future,” but the Court declined to address
the issue of whether a defendant is entitled to a parole
ineligibility instruction when the state’s evidence shows
future dangerousness but the prosecutor does not argue it.
Id. at 732 n.4. The Court concluded that “[t]he prosecutor
accentuated the clear inference of future dangerousness
raised by the evidence and placed the case within the four
corners of Simmons.” Id. at 732.
The Kelly dissenters, including two of the Justices who
had joined Justice O’Connor’s opinion in Simmons, argued
that the Court had improperly extended Simmons’s reach.
See id. at 735 (Rehnquist, C.J., joined by Kennedy, J.,
 53

dissenting) (“the test is no longer whether the State argues
future dangerousness to society[,] [as the Court found in
Simmons]; the test is now whether evidence was introduced
at trial that raises an ‘implication’ of future dangerousness
to society”) and id. at 737 (Thomas, J., joined by Scalia, J.,
dissenting) (“the Court dilutes the Simmons test, now
requiring that a parole ineligibility instruction be given
where the prosecution makes arguments that have a
‘tendency to prove dangerousness in the future.’ ”).
Even if Kelly broadened Simmons, however, Kelly cannot
aid Rompilla here. Under 28 U.S.C. § 2254(d)(1), our review
is limited to deciding whether a state court decision is
“contrary to” or an “unreasonable application” of Supreme
Court precedent “as of the time of the relevant state-court
decision.” Williams, 529 U.S. at 412 (emphasis added). The
state court decisions in this case preceded Kelly.
Accordingly, regardless of whether Kelly expands the
circumstances when future dangerousness is at issue for
purposes of requiring a parole ineligibility instruction, the
case is not applicable to the instant matter. We are limited
to deciding whether the state court decisions in this case
were contrary to or an unreasonable application of the
holding in Simmons itself — which is to say Justice
O’Connor’s controlling concurrence. We must thus
determine whether the Pennsylvania Supreme Court was
reasonable in concluding that the prosecutor in the present
case did not “argue[ ]” that Rompilla would present a future
danger if not sentenced to death. We therefore turn to the
relevant events at the penalty phase of Rompilla’s trial.
Volume 2 of 2
 56

B.
During the penalty phase, the Commonwealth attempted
to establish the aggravating factor that Rompilla had a
“significant history of felony convictions” as specified in 423
Pa. C.S. § 9711(d)(9). In doing this, the Commonwealth read
into the record the testimony of J.M., a female bar owner
whom Rompilla had previously robbed, slashed with a
knife, and raped. During closing arguments, the
Commonwealth argued:
[T]he woman that was raped, was raped pretty brutally.
She was raped at knife point. . . . [I]sn’t it frightening,
the similarity between that case and this case. I mean,
it is absolutely astounding. Both take place around the
bar. The Defendant gets in after closing or right before
closing. . . . On both occasions, a knife was used.
Steals money both times. Isn’t it frightening the
similarities in those crimes. Takes a taxi away from
[J.M.’s] Bar, takes a taxi the night of this crime. He
slashes [J.M.] in the breast with a knife. He uses a
knife on Jimmy Scanlon. It’s absolutely frightening to
think of the similarities in those two crimes. But there
is one difference, one major difference, [J.M.] lived
through her experience. Jimmy Scanlon didn’t. . . . I
keep wondering, why did the Defendant wait till Jimmy
Scanlon left the Bar? He knew he was closing up. Why
didn’t he just wait until he had left to break in and
steal the money, no problem, nobody gets hurt? You
have to ask yourselves, why didn’t he wait for Jimmy
Scanlon to leave? Was it his intent to kill right from the
start? Was it his intent to do serious bodily injury right
from the start? Because if he had waited maybe a half
hour, an hour later, Jimmy Scanlon would have been
gone. This wouldn’t be a homicide case. Jimmy
Scanlon would still be living. I think the Defendant
learned a lesson from [J.M.] in that case, that Rape
case. That lesson was, don’t leave any witnesses. Don’t
leave anybody behind that can testify against you.
App. at 779-80.
During deliberations the jury asked:
 57

If a life sentence is imposed, is there any possibility of
the Defendant ever being paroled?
Id. at 802. The trial judge answered:
I’m sorry to say, I can’t answer that question. That’s
not before you as such. The only matter that you can
consider in the Sentencing Hearing is the evidence that
was brought out in the course of the Hearing and the
Law with respect to the Court’s Charge. That’s the only
consideration you have, I’m sorry to say. I — if there
were other alternatives that you should consider, we
would have outlined them in the Charge, all right.
Id. at 802-03.
Several hours later, the jury asked to see the docket
entries for Rompilla’s prior convictions for the purpose of
ascertaining Rompilla’s prior sentence. The judge denied
the request, and the following exchange took place:
Juror No. 3: . . . we want to know if it was — the
sentence was — if he got released on
behavioral —
Foreman: It was commuted in any way, the
original sentence.
Court: Well, that we can’t give you.
Prosecutor: You can’t tell them.
Id. at 823-24.
The next day, after two more hours of deliberation, the
jury asked:
Was the Defendant offered any type of rehabilitation
either while in prison or after his release from prison?
Id. at 842. The following colloquy then occurred:
Court: Well, I’m sorry to say, I can’t answer
that. I can only tell you that you’re going
to have to make your decision based
upon the evidence that was presented
and in accordance with the Law with
respect to Sentencing Hearing. First of
 58

all, I couldn’t even answer it if I wanted
to or if I could, I don’t know.
Foreman: Could I change the question to the point
that is — isn’t rehabilitation available in
prison?
Court: Well, again, I would like to even answer
that, and I can’t. You’re going to have to
rely upon your own knowledge of that
aspect if, indeed, that is a part that
troubles the area that you’re interested
in. The penology system, I’ll be quite
frank with you, is not an issue before —
before you with respect to the Law that
it’s a decision that you must make. I can
understand your interest, however, as I
say, we’re constrained to, you know,
comply with whatever evidence that was
put on in the hearing and then your
decision must be based upon whatever
the Law says and whatever you may
find. All right.
Id. at 842-43. Three hours later the jury returned a
sentence of death.
The PCRA court found that Simmons did not apply
because the prosecutor did not argue future
dangerousness. App. at 2025. The District Court similarly
found that a “fair reading of [the prosecutor’s summation]
leads to the conclusion that the state’s reasoning for the
death penalty was not based upon future dangerousness,”
but rather upon Rompilla’s “despicable, savage and
cowardly beating” of the victim. Rompilla, 2000 WL 964750,
at *15. The District Court concluded that although a “close
issue,” the state supreme court’s decision was not an
unreasonable application of federal law. Id.

C.
As noted, Rompilla relies on both the “contrary to” and
“unreasonable application” prongs of 28 U.S.C. § 2254(d)(1).
We will discuss each prong separately.
 59

1.
A result is “contrary to” a Supreme Court holding if the
state court “contradicts the governing law set forth in [the
Supreme Court’s] cases” or if it “confronts a set of facts that
are materially indistinguishable from a decision of [the]
Supreme] Court and nevertheless arrives at a [different]
result.” Id. at 405—06. Williams v. Taylor, 529 U.S. 362,
412 (2000). This prong is not met here because the
Pennsylvania Supreme Court applied the holding of
Simmons and because the facts of Simmons and the present
case are materially distinguishable.
Rompilla argues that the state supreme court’s decision
is “contrary to” Simmons because the state supreme court
did not inquire whether there was “a ‘reasonable likelihood’
that parole and future dangerousness affected the
sentencing decision.” Rompilla Br. at 42 (citing Simmons,
512 U.S. at 170). We reject this argument. Neither the
Simmons plurality nor Justice O’Connor’s concurrence
stated that the test for entitlement to an instruction on
parole ineligibility hinges on whether there is a reasonable
likelihood that future dangerous might figure in the jury’s
verdict. Rather, the Simmons plurality opinion referred to
the concept of “reasonable likelihood” in considering a
different question — whether, once future dangerousness
was put at issue, there was a “reasonable likelihood” that
a jury instruction “directing juries that life imprisonment
should be understood in its ‘plain and ordinary’ meaning”
dispelled any misunderstanding that the jurors might have
had as to the meaning of life imprisonment. Simmons, 512
U.S. at 169-70 (citing Boyde v. California, 494 U.S. 370,
380 (1990)).
Rompilla contends that the prosecution necessarily put
his future dangerousness at issue by presenting evidence of
his prior criminal conduct to establish the aggravating
factor that he had a significant history of felony convictions.
According to Rompilla, the state supreme court rendered a
decision that was “contrary to” Skipper v. South Carolina,
476 U.S. 1 (1986) and Ramdass v. Angelone, 530 U.S. 156
(2000), when it rejected this argument on the “formalistic”
ground that this aggravator concerns past conduct rather
than future behavior. Rompilla Br. at 42. This argument is
 60

also meritless. It overlooks the scope of Justice O’Connor’s
Simmons concurrence, which, as noted, can reasonably be
read as focusing on the prosecution’s arguments, not the
inferences that arise from the facts that are proven. On this
view, proving past crimes that may cause jurors to worry
about future dangerousness is quite different from arguing
that a defendant presents a future threat.
In addition, Skipper does not hold or even state that
proving prior felony convictions is tantamount to arguing
that a defendant presents a future threat for Simmons
purposes. In Skipper, the Court held that the capital
defendant’s right to present all relevant mitigating evidence
at the sentencing stage of the trial was violated by the state
court’s refusal to admit evidence that the defendant had
made a good adjustment to jail during the time between his
arrest and the trial. Thus, Skipper had nothing to do with
an instruction on parole ineligibility.21
Rompilla’s reliance on Ramdass is also misplaced. Not
only does Ramdass post-date the decision of the
Pennsylvania Supreme Court, but Ramdass does not
address the question of when future dangerousness is
sufficiently put in issue to require an instruction on parole
ineligibility. In Ramdass, the prosecution expressly argued
future dangerousness as an aggravating circumstance, and
thus that point was not disputed. Ramdass, 530 U.S. at
161. Instead, the issue in Ramdass was whether the
defendant was eligible for parole at the time that was
relevant for Simmons purposes. See id. at 166-68.

21. Rompilla refers to the Skipper Court’s statement that “[c]onsideration
of a defendant’s past conduct as indicative of his probable future
behavior is an inevitable and not undesirable element of criminal
sentencing.” 476 U.S. at 5. If this dictum is true — that people will
inevitably draw inferences of probable future behavior from evidence of
past conduct — it may argue against drawing the line suggested in
Justice O’Connor’s Simmons concurrence between cases in which the
prosecution argues that a defendant presents a future danger and cases
in which the facts give rise to an inference of future dangerousness. But
that plainly does not make the decision of the Pennsylvania Supreme
Court in this case “contrary to” Skipper, which was not even a Simmons
case.
 61

Third, Rompilla argues that his case is “materially
indistinguishable” from Simmons, but we disagree. In
Simmons, as noted, the prosecution expressly argued that
the defendant posed a future threat, stating that “a verdict
for death would be ‘a response of society to someone who
is a threat. Your verdict will be an act of self-defense.’ ”
Simmons, 512 U.S. at 157. In this case, by contrast, the
Commonwealth made no specific references to any possible
future conduct by the defendant. We thus hold that the
decision of the Pennsylvania Supreme Court was not
“contrary to” Simmons.

2.
Rompilla next contends that the state supreme court’s
decision represents an “unreasonable application” of
Simmons, but while we agree with the District Court that
the application of Simmons to the present case would be
“close” if we were exercising plenary review, we are
convinced that the state supreme court’s decision was
reasonable.
Rompilla maintains that the state supreme court’s
decision was unreasonable because its opinion did not
discuss all of evidence on which he relies in arguing that
his future dangerousness was sufficiently put at issue
during the penalty phase of his trial. Rompilla Br. at 45. As
we have already explained, however, 28 U.S.C. § 2254(d)(1)
calls upon us to decide whether the adjudication of a claim
on the merits in state court “resulted in a decision” that
“involved an unreasonable application of . . . clearly
established Federal law.” This standard applies even when
a claim is adjudicated on the merits without any discussion
at all. See Weeks, 528 U.S. at 237; Chadwick v. Janecka,
312 F.3d at 606. Accordingly, we must look to the
reasonableness of what the state supreme court decided,
not the detail included in its opinion.
Rompilla argues that the prosecution raised the issue of
future dangerousness in its closing by stating that Rompilla
had “learned a lesson” from the prior rape, namely, that he
should not leave any witnesses and by repeatedly asking
the jury whether the similarities between the rape and his
 62

murder of Scanlon were not “frightening.” Rompilla
contends that these comments painted a picture of him as
a “hardened, frightening, violent recidivist who, if ever
released, would commit additional violent crimes and would
not ‘leave any witnesses.’ ” Rompilla Br. at 34-35.
We have given this argument careful consideration, but
in the end it does not persuade us that the difficult
standard of § 2254(d)(1) is met. Reading the
Commonwealth’s remarks in the context of the entire
sentencing proceedings, it appears to us that the
Commonwealth did not argue future dangerousness. In her
emotional closing argument at the penalty phase, defense
counsel argued over and over that, although the jury had
found the defendant guilty, they must have had some
doubts about his guilt in view of the evidence and that they
should therefore not impose a sentence of death. She
stated:
You, as a Jury, convicted him of First Degree Murder.
You convicted him, in your minds, beyond a reasonable
doubt, but you must have some doubt. There has got
to be some doubt there with no eye witness, with
circumstantial evidence . . . .
So I know that you have to have some doubt as to
what actually happened that night. We may never
know what actually happened that night . . . .
Don’t you want to be sure before you condemn that
man to die because don’t fool yourselves. A death
sentence is death. . . . No more questions after that.
. . .
I saw you all struggling with this. I saw it Monday
night at 10:00 o’clock when we let you go for the
evening. You looked tired, you looked nervous, and you
looked like you’ve been struggling. . . . [I]f you’re
struggling with it, the fact that you had some doubt
should also stay in your mind, now. If there was any
doubt, and I submit to you that there has to be some
doubt. There has to be. You don’t know what
happened. You’ll never know exactly what happened.
There’s got to be some doubt here. . . .
 63

What if you’re wrong? Can you live with that? . . . .
You’ve rendered what you felt was a just verdict. I have
no quarrel with you, that was your job, you came back
with a verdict but none of you can tell me that you
have no doubts and if you have a doubt, then you
better think very, very, carefully about what you’re
going to do here. . . .
Don’t take a chance that you may be making a wrong
decision that you can’t live with. Please, spare his life,
thanks.
App. 767-774.
The prosecutor’s remarks upon which Rompilla now
relies came in response to these defense comments
suggesting that the jury should still have doubts about the
defendant’s guilt. Seeking to dispel any such doubts, the
prosecutor’s obvious point in stressing the similarities
between the circumstances of the rape for which Rompilla
had previously been convicted and the Scanlon murder was
to convince the jury that the same man had committed
both crimes. Although the prosecutor at times termed the
similarities between the two crimes “frightening,” it seems
clear that he did not use this term to suggest that Rompilla
was “frightening” in the sense that he posed a future threat,
but simply that the similarities were, as he more aptly put
it at another point, “astounding.” That his argument did not
go to future dangerousness is illustrated by the fact that
many of the features of the two crimes that he highlighted
— both occurred at a bar, both took place at around closing
time, and in both instances the perpetrator used a taxi —
are not features that naturally suggest future
dangerousness. The prosecutor’s further comment that
Rompilla had learned a lesson from the rape, i.e., that he
should not leave any witnesses, came immediately after this
litany of similarities, and the comment seems to have had
two likely purposes: to explain why there was no eyewitness
to the most recent crime and to explain why the two crimes
differed in the important respect that one involved a killing
and the other did not.
In any event, even if this interpretation of the
prosecutor’s comments is incorrect and even if they were
 64

meant to imply that Rompilla would present a future
danger if he was ever released from prison, the fact remains
that the prosecutor never actually argued that Rompilla
presented a future threat. Concluding that these facts did
not bring the case within the holding of Simmons, as set
out in the controlling concurrence, is by no means
“unreasonable.” Particularly in view of the fact that the
Justices who endorsed the controlling concurrence in
Simmons apparently felt that it was important to draw the
line where they did, rather than approving the arguably
broader reach of the plurality, the state court’s failure to
extend Simmons to situations in which the prosecution
does not argue future dangerousness was not an
“unreasonable application” of Simmons.22
Rompilla also argues that the state supreme court’s
decision is an unreasonable application of Supreme Court
law because it “flies in the face of ” the state court’s prior
decisions regarding future dangerousness and prior
criminal history and thus violates the Eighth and
Fourteenth Amendment requirements that capital
sentencing be applied consistently and with an even hand
(citing Eddings v. Oklahoma, 455 U.S. 349, 361 (1977), and
Gardner v. Florida, 430 U.S. at 349, 361 (1977). Rompilla
Br. at 44. In making this argument, Rompilla points to the
passage in the state supreme court opinion addressing his
argument that his future dangerousness was put at issue
for Simmons purposes by the prosecution’s effort to prove
as an aggravating circumstance that he had a significant
history of violent felony convictions. Rompilla-2, 721 A.2d
at 795. Rejecting this contention, the state supreme court

22. Rompilla points to the testimony regarding his parole status and the
lack of rehabilitative services as indicating that his future dangerousness
was an issue for the jury. This evidence was brought out by defense
counsel on direct examination of Rompilla’s witnesses and was not
argued by the Commonwealth. Rompilla also points to the jury’s
questions regarding parole and future dangerousness. Although the
questions may demonstrate that the jury did not know the meaning of
“life” imprisonment, see Simmons, 512 U.S. at 170 n.10; id. at 178
(O’Connor, J., concurring); Rompilla-2, 721 A.2d at 795 (Flaherty, C.J.,
dissenting), a jury’s uncertainty as to parole ineligibility is insufficient to
require an instruction under Simmons.
 65

stated that “this aggravating circumstance only addresses
Appellant’s past conduct, not his future dangerousness. Id.
Rompilla attacks this reasoning on the ground that the
underlying rationale for this aggravating circumstance is
that a history of violent felony convictions is a “barometer
of future danger.” Rompilla Br. at 44. He therefore contends
that the state supreme court’s holding in the present case
violates the principle that a state must administer its death
penalty statute evenhandedly. There is no merit in this
argument.
The constitutional principle on which Rompilla relies
demands consistency in a state’s treatment of its capital
cases, and we have not been presented with any evidence
that Pennsylvania does not apply the Simmons standard
evenhandedly. On the contrary, it appears to us that
Pennsylvania has uniformly required a Simmons instruction
when the prosecutor specifically raised the issue of future
dangerousness, see, e.g., Commonwealth v. Trivigno, 750
A.2d 243, 252-54 (Pa. 2000) (prosecutor asked jury to use
prior convictions as a “weather vane looking into the
future” and a “determinant of where [the defendant] is
going, not just where he’s been”); Commonwealth v.
Chandler, 721 A.2d 1040, 1046-47 (Pa. 1998) (prosecutor
asked jury to “stop [the defendant] from ever killing another
woman again”), and has not required the instruction when
the prosecutor only presented evidence of prior convictions.
E.g., Commonwealth v. Robinson, 721 A.2d 344, 355 (Pa.
1999) (prosecutor only made references to defendant’s past
dangerousness); Commonwealth v. King, 721 A.2d 763, 779
(Pa. 1998) (prosecutor only made reference to defendant’s
past violent acts); May, 710 A.2d at 47 (prosecutor did not
argue future dangerousness). See also Robinson, 721 A.2d
at 355 (recognizing that the court requires a Simmons
instruction only when future dangerousness is “expressly
implicated”).
Rompilla does not dispute the fact that Pennsylvania
applies the Simmons rule evenhandedly. Instead, his
complaint is that there is, in his view, a logical
inconsistency between the rationale for the aggravator in
question and the Pennsylvania Supreme Court’s
interpretation of Simmons. This argument, however, is
 66

invalid for numerous reasons. First, the constitutional
principle on which Rompilla relies concerns consistency in
the treatment of cases, not logical consistency in all aspects
of a state’s death penalty jurisprudence. Second, Rompilla
has not identified a logical inconsistency. The Legislature
may have adopted the aggravating circumstance in question
wholly or partially for the purpose of retribution, not to
protect the public from those defendants who are likely to
pose a threat. Moreover, even if this aggravating
circumstance was intended solely to serve as a “barometer
of future danger,” there is nothing illogical about taking the
position that, while proof of a defendant’s past history of
violence may tend to suggest that the defendant may pose
a future danger if released, the jury should be instructed
about parole only in those cases in which there is a
particularly strong reason to think that concern about
future danger will decisively influence the jury’s sentence.
Informing a jury about the correctional consequences of a
verdict is an exception from usual practice, and it is not
illogical to confine this exception to narrow circumstances.
Third, if there is a logical inconsistency, it flows from the
fine line drawn by the controlling opinion in Simmons, not
from the Pennsylvania Supreme Court. For all these
reasons, Rompilla’s argument is rejected.

IX.
For the reasons set out above, the decision of the District
Court is reversed with respect to the ineffective assistance
of counsel claim and affirmed as to the accomplice liability
instruction and parole ineligibility claims.
 67

SLOVITER, Circuit Judge, dissenting.
I respectfully dissent from the well-crafted (albeit, in my
view, flawed) decision of the Majority that, in effect,
reinstates the death penalty for appellant Ronald Rompilla.
There are two claims that I believe warrant the grant of a
writ of habeas corpus. One, which was the basis for the
District Court’s grant of the writ requiring a new sentencing
hearing, is the shocking ineffective assistance of counsel at
the sentencing phase. Rompilla’s trial counsel failed to
obtain Rompilla’s school, medical, court and prison records
as part of their investigation and, as a result, failed to
present to the jury any mitigating evidence regarding
Rompilla’s “childhood, alcoholism, mental retardation, or
possible organic brain damage.” Rompilla v. Horn, 2000 WL
964750, at *9. Counsel also failed to communicate with two
of Rompilla’s siblings who lived nearby and would have
advised counsel of evidence that Rompilla was raised by
alcoholic parents in a cold, violent, frightening and abusive
home.
The standard for establishing ineffective assistance of
counsel was set forth in the Supreme Court’s seminal
decision in Strickland v. Washington, 466 U.S. 668 (1984).
Strickland was applied in Williams v. Taylor, 529 U.S. 362
(2000), to the issue of the investigation required of counsel
at the penalty phase. Holding, in “a straightforward
application of Strickland,” that trial counsel was ineffective
because he “failed to discover or failed to offer” certain
mitigating evidence, id. at 393, the Court held that the
Virginia Supreme Court decision denying the requested writ
of habeas corpus was both contrary to and involved an
unreasonable application of Strickland. Id. at 391.
As I will discuss, the Pennsylvania Supreme Court’s
failure to grant relief to Rompilla because of trial counsel’s
grossly inadequate investigation also was both contrary to
and involved an unreasonable application of Strickland and
Williams. The Majority’s decision overturning the District
Court’s grant of a writ of habeas corpus and rejecting
Rompilla’s claim of ineffective assistance of counsel is
inexplicable in light of the Supreme Court’s most recent
application of Strickland in Wiggins v. Smith, 123 S. Ct.
 68

2527 (2003), under circumstances remarkably similar to
those presented here.
The second ground on which I would grant a writ of
habeas corpus is the refusal of the trial judge to advise the
jury as to the meaning of a life sentence in Pennsylvania,
notwithstanding the jury’s questions on that issue on three
different occasions during its sentencing deliberations. As a
result, the jury imposed a death sentence, rather than a life
sentence, without ever being told by the trial court in
response to the jury’s question that a life sentence in
Pennsylvania means life without parole. The Pennsylvania
courts’ rule in this respect is, in my judgment, an
unreasonable application of the Supreme Court’s decisions
in Simmons v. South Carolina, 512 U.S. 154 (1994), and
Kelly v. South Carolina, 534 U.S. 246 (2002). Because this
appeal is literally, not figuratively, a matter of life or death,
I elaborate on each of these grounds.

I.

Ineffective Assistance of Counsel
There is no need to restate the facts of this case and its
procedural history, both of which are fully and accurately
set forth in the majority opinion. The ineffective assistance
of counsel claim is before us on the Commonwealth’s
appeal.
The applicable legal principles are not in dispute. Under
the Antiterrorism and Effective Death Penalty Act of 1996
(AEDPA), 28 U.S.C. § 2254(d)(1), the federal court may
grant a petition for habeas corpus only if the state’s
adjudication resulted in a decision that is “contrary to, or
involved an unreasonable application of, clearly established
Federal law, as determined by the Supreme Court,” or
§ 2254(d)(2), the state decision was based on an
unreasonable determination of the facts. Both parties agree
that it is section 2254(d)(1) that is at issue here. Strickland
supplies the standard for addressing a claim of ineffective
assistance of counsel:
A convicted defendant’s claim that counsel’s assistance
was so defective as to require reversal of a conviction or
 69

death sentence has two components. First, the
defendant must show that counsel’s performance was
deficient. This requires showing that counsel made
errors so serious that counsel was not functioning as
the “counsel” guaranteed the defendant by the Sixth
Amendment. Second, the defendant must show that
the deficient performance prejudiced the defense. This
requires showing that counsel’s errors were so serious
as to deprive the defendant of a fair trial, a trial whose
result is reliable.
466 U.S. at 687.
In Williams, the Supreme Court held that “the rule set
forth in Strickland qualifies as clearly established federal
law, as determined by the Supreme Court.” 529 U.S. at 391
(internal quotation marks omitted). The Court further held
that a state decision is “contrary to . . . clearly established”
federal law if, inter alia, “the state court confronts a set of
facts that are materially indistinguishable from a decision
of this Court and nevertheless arrives at a result different
from our precedent.” Id. at 406. As the decision in Williams
made clear, relief should be granted under the
“unreasonable application” clause “if the state court
identifies the correct governing legal principle from [the
Supreme Court’s] decisions but unreasonably applies that
principle to the facts of the prisoner’s case.” Id. at 413.
Williams instructs that “clearly established Federal law”
refers “to the holdings, as opposed to the dicta, of this
Court’s decisions as of the time of the relevant state-court
decision.” 529 U.S. at 412. The Court in Wiggins decided
the ineffectiveness of counsel claim before it using the
standards articulated in Strickland, noting that Williams did
not create new law, but illustrated “the proper application”
of these standards. 123 S. Ct. at 2535. Because Williams
fell squarely within Strickland, Williams, 529 U.S. at 390,
and Wiggins was decided pursuant to the same standards
of Strickland, these two later cases demonstrate how
Strickland should be applied.
Under Strickland, a petitioner seeking to prove ineffective
assistance of counsel must show that counsel’s
performance was deficient and that the deficiency
prejudiced the defense. 466 U.S. at 687. To make the
 70

requisite showing of deficiency, the petitioner must show
that counsel’s representation “fell below an objective
standard of reasonableness.” Id. at 688. Among counsel’s
responsibilities with regard to the sentencing phase is the
“obligation to conduct a thorough investigation of the
defendant’s background.” Williams, 529 U.S. at 396.
Coincidentally, in both Williams and Wiggins, two of the
principal Supreme Court cases dealing with ineffective
assistance of counsel, the Court found that a writ of habeas
corpus was appropriate because the failure of counsel to
conduct the requisite thorough investigation was objectively
unreasonable and prejudicial.
In Williams, the Court noted that counsel “failed to
conduct an investigation that would have uncovered
extensive records graphically describing Williams’
nightmarish childhood, not because of any strategic
calculation but because they incorrectly thought that state
law barred access to such records.” Id. at 395. The Court
continued,
Had they done so, the jury would have learned that
Williams’ parents had been imprisoned for the criminal
neglect of Williams and his siblings, that Williams had
been severely and repeatedly beaten by his father, that
he had been committed to the custody of the social
services bureau for two years during his parents’
incarceration (including one stint in an abuse foster
home), and then, after his parents were released from
prison, had been returned to his parents’ custody.
Id. (footnotes omitted). The Court, noting that mitigating
evidence “may alter the jury’s selection of penalty, even if it
does not undermine or rebut the prosecution’s death-
eligibility case,” id. at 398, concluded that “the Virginia
Supreme Court rendered a ‘decision that was contrary to,
or involved an unreasonable application of, clearly
established Federal law,’ thereby violating Williams’
constitutional right to the effective assistance of counsel as
defined in Strickland.” Id. at 399.
Our court in Jermyn v. Horn, 266 F.3d 257 (3rd Cir.
2001), a decision following Williams, concluded that
Jermyn’s trial counsel was ineffective “because he failed to
 71

conduct an investigation, failed to prepare adequately for
the penalty phase of Jermyn’s trial, and consequently,
failed to present substantial mitigating evidence that would
have directly undercut the state’s penalty-phase case.” Id.
at 306. We explained,
Counsel failed to investigate the circumstances
surrounding Jermyn’s childhood, even though counsel
admitted at the PCRA hearing that he was aware that
Jermyn had claimed that he was abused as a child. Dr.
Phillips told counsel before the original trial that
Jermyn had been abused as a child, and told counsel
that the abuse was a critical component to
understanding Jermyn’s mental illness. Nonetheless,
counsel did not attempt to locate any fact witness who
witnessed the abuse and could testify about it
specifically.
Id. at 306 (internal citation omitted).
We stated that if counsel had investigated further “he
would have obtained powerful and, as the PCRA court
noted, ‘credible,’ testimony” from witnesses who presented
testimony at the PCRA hearing that
was replete with first-hand accounts of instances of
mental and physical abuse that Jermyn suffered at the
hands of his father. . . . Counsel could have used that
testimony to provide the jury with critical insight into
the root of Jermyn’s mental illness. The [witnesses’]
testimony also offered valuable insight into Jermyn’s
mother’s role in the household, and how she failed to
intervene on Jermyn’s behalf.
Id. at 306-07. Counsel in Jermyn “also did not seek to
obtain records from the school which corroborate the fact
that Jermyn was abused as a child, and that his mother
did not intervene on his behalf.” Id. at 307. Based on the
ineffective assistance of counsel for this and other reasons
set forth in that opinion, we affirmed the District Court’s
grant of a writ of habeas corpus.
With these cases as a background, we turn to the
conduct of Rompilla’s trial counsel, informed by the
Supreme Court’s recent decision in Wiggins.
 72

Counsel in this case were two public defenders. Frederick
Charles, the senior of the two, was a veteran attorney with
significant criminal trial experience as a defense lawyer,
who never before tried a capital case. His role here was as
the principal counsel for Rompilla at the guilt phase. His
colleague, Maria Dantos, who was two and a half years out
of law school at the time of the trial, App. at 1060-61,
1065-67, was given the responsibility of handling the
penalty phase. She had the role of making the arguments
and presenting witnesses at the penalty phase, under
Charles’ supervision. This was her first capital case and
first homicide trial. Counsel’s devotion to Rompilla’s cause
has not been challenged. They impressed the District Court
as “intelligent, diligent and devoted to their task of
representing [Rompilla].” 2000 WL 964750, at *9. It is their
competence and effectiveness that are at issue.
The PCRA court determined that “counsel had a
reasonable basis for proceeding as they did during the
penalty phase,” Maj. Op. at 17 (quoting App. at 2028),
because counsel retained three health professionals, two of
whom, Drs. Cooke and Sadoff, are “recognized experts in
the fields of psychiatry and psychology,” id., who
administered tests, evaluated Rompilla and reported back
to defense counsel that they found that nothing could be
used in mitigation. Counsel also obtained an evaluation by
a local psychiatrist, Dr. Paul Gross, who found nothing that
would have been beneficial in the penalty phase.
In affirming the PCRA court’s determination rejecting,
inter alia, Rompilla’s claim of ineffective assistance of
counsel, the Supreme Court of Pennsylvania concluded that
counsel “reasonably relied upon their discussions with
[Rompilla] and upon their experts to determine the records
needed to evaluate mental health and other potential
mitigating circumstances.” Commonwealth v. Rompilla, 721
A.2d 786, 790 (Pa. 1998).
In accepting the Pennsylvania courts’ conclusions, the
Majority also concludes that “trial counsel conducted an
extensive investigation for mitigation evidence,” Maj. Op. at
27, because trial counsel retained three well-qualified
mental health experts to examine Rompilla. The Majority
recognizes that counsel failed to seek out school, medical,
 73

police, and prison records, which they recognize “contain
useful information about Rompilla’s childhood home
environment, his mental problems, and his problems with
alcohol,” Maj. Op. at 29. Nevertheless, the Majority
characterizes counsel’s decision as “reasonable” because
Charles, who “had the final say on every issue in the case,”
Maj. Op. at 29, “did not think that obtaining those records
would have represented a sound allocation of his office’s
resources.” Maj. Op. at 29. The question before us is not
whether we believe counsel’s explanation for his failure to
obtain the relevant records, some of which were available
across the street from the trial and others in the same
building as the trial, was “reasonable” but whether his
failure to take such action to save money was objectively
reasonable and consistent with his obligation to conduct a
thorough investigation.
A comparison of counsel’s actions in this case with those
of counsel in Wiggins, where the Supreme Court concluded
that counsel was ineffective, is instructive.
Wiggins, who was represented by two public defenders,
was convicted of capital murder in 1989 by a Maryland
judge, and the jury sentenced him to death the same
afternoon. A divided Maryland Court of Appeals affirmed.
He filed a motion for post-conviction relief in which he
challenged the adequacy of his representation at
sentencing, “arguing that his attorneys had rendered
constitutionally defective assistance by failing to investigate
and present mitigating evidence of his dysfunctional
background.” Wiggins, 123 S. Ct. at 2532. At the PCRA
hearing, a licensed social worker testified regarding an
elaborate social history in which he chronicled Wiggins’
“bleak life history” based on state social services, medical
and school records and interviews with Wiggins and
numerous family members. Id. at 2533. His mother, a
chronic alcoholic, frequently left him and his siblings alone
for days, forcing them to beg for food and to eat paint chips
and garbage. Her abusive behavior included beating the
children and having the children present while she engaged
in her active sexual life. Wiggins suffered severe physical
and sexual abuse at the hands of his mother and father
and while in the care of a series of foster parents. The
 74

father in his second foster home repeatedly molested and
raped him. In one foster home, he was gang raped on more
than one occasion and was sexually abused by his
supervisor on a job corps program. Id.
Wiggins’ counsel decided to focus their request for post-
conviction relief on retrying the factual case and disputing
Wiggins’ direct responsibility for the murder.
Notwithstanding that counsel, as a practical matter, did not
compile a social history of Wiggins, the state post-
conviction court concluded that when the decision not to
investigate is a matter of trial tactics, there is no ineffective
assistance of counsel. Wiggins, 123 S. Ct. at 2533. The
Maryland Court of Appeals affirmed the denial of relief on
the ground that trial counsel, although they knew of
Wiggins’ unfortunate childhood and had available both the
presentence investigation report and the more detailed
social service records, “made a reasoned choice to proceed
with what they thought was their best defense.” Id. at
2533-34.
In holding that Wiggins’ trial counsel provided
constitutionally ineffective assistance of counsel and that
the Maryland Court of Appeals unreasonably applied
Strickland in failing to so conclude, the Supreme Court
specified why counsel’s conduct was defective. The Court
noted that trial counsel had limited their investigation to
two sources. One was the presentence investigation report
(PSI) prepared by the Division of Parole and Probation,
“which included a one-page account of Wiggins’ ‘personal
history’ noting his ‘misery as a youth,’ quoting his
description of his own background as ‘disgusting,’ and
observing that he spent most of his life in foster care,”123
S. Ct. at 2536 (quoting PSI). The other source counsel
examined was the records kept by the Baltimore City
Department of Social Services (DSS) documenting Wiggins’
various placements in the state foster care system. The
Supreme Court noted that counsel chose not to expand
their investigation beyond those records even though the
Public Defender’s Office made funds available for the
retention of a forensic social worker. Id. The Supreme Court
thus concluded that, “counsel abandoned their
investigation of [Wiggins’] background after having acquired
 75

only rudimentary knowledge of his history from a narrow
set of sources.” Id. at 2537.
The Supreme Court characterized the Maryland Court of
Appeals’ application of Strickland’s governing legal
principles as “objectively unreasonable.” Id. at 2538. That
court merely assumed that counsel’s investigation was
adequate and failed to focus on counsel’s failure to engage
in further investigation, despite the information in the DSS
records with respect to Wiggins’ mother’s alcoholism, his
shuttling from foster home to foster home, and his lengthy
absences from school. Further investigation would have
discovered the sexual abuse that was later revealed during
state post-conviction proceedings. Id.
The investigation by Rompilla’s lawyers, albeit different
from that conducted by Wiggins’ lawyers, was similarly
defective. They did not present at the penalty phase
evidence that was available, had they investigated
thoroughly, and which was presented for the first time at
Rompilla’s PCRA hearing. Rompilla’s parents were both
severe alcoholics who drank constantly. His mother drank
during her pregnancy with Rompilla, and he and his
brothers eventually developed serious drinking problems.
His father, who had a vicious temper, frequently beat
Rompilla’s mother, leaving her bruised and black-eyed, and
bragged about his cheating on her. His parents fought
violently, and on at least one occasion his mother stabbed
his father. App. at 1408, 1416-17, 1450, 1460, 1487, 1492,
1495-96. He was abused by his father who beat him when
he was young with his hands, fists, leather straps, belts
and sticks. All of the children lived in terror. There were no
expressions of parental love, affection or approval. Instead,
he was subjected to yelling and verbal abuse. His father
locked Rompilla and his brother Richard in a small wire
mesh dog pen that was filthy and excrement filled. He had
an isolated background, and was not allowed to visit other
children or to speak to anyone on the phone. They had no
indoor plumbing in the house, he slept in the attic with no
heat, and the children were not given clothes and attended
school in rags. These facts were not presented at the
sentencing hearing because counsel did not know them
and, more important for our purposes, failed to make the
reasonable investigation that would have uncovered them.
 76

Rompilla’s counsel presented the testimony of three of
Rompilla’s siblings at the sentencing hearing but did not
interview two of Rompilla’s sisters, Barbara Harris and
Randi Rompilla, both of whom lived nearby, prior to
sentencing, App. at 1422, 1436-37, 1489-90, and did not
present their testimony. Further, Rompilla’s brother,
Nicholas Rompilla, Jr., who had briefly testified at the
sentencing hearing, claimed at the post-conviction hearing
that Rompilla’s counsel only asked him about the three
months prior to the offense and did not ask him about
Rompilla’s childhood. App. at 1462-63. We must then ask
whether it was unreasonable for counsel not to have
interviewed all of Rompilla’s siblings. It was these family
members who testified at the PCRA hearing that, among
other things, their parents were alcoholics and their mother
drank while pregnant with Rompilla. They detailed the
physical abuse referred to above. Rompilla was told he was
stupid and would not amount to anything; they also
testified that Rompilla was a “very nervous child,” who kept
everything inside. App. at 1407-13, 1424, 1451, 1480-84,
1487-88.
It was Rompilla’s PCRA counsel, not trial counsel, who
presented evidence from two mental health experts, Drs.
Armstrong and Crown, both of whom evaluated and tested
Rompilla post-conviction. Their evaluations included
neuropsychological testing, review of Rompilla’s school,
medical, and prison records (none of which had been
examined by trial counsel), and review of declarations by
Barbara Harris, Darlene Rompilla, and Nicholas Rompilla,
Jr., App. at 1562-64, 1567, 1704, 1736, 1741, 1743, 1745;
Commonwealth Response to Habeas Pet., Exh. B, C, & Pet.,
Exh. B, C, & D. The doctors stated that Rompilla’s low IQ
and achievement test results documented in his school
records, his medical history, and his abusive background
were all “red flags” indicating that further objective
evaluation was necessary. App. at 1614, 1686, 1692-93,
1739, 1743, 1745-46.
These doctors determined that Rompilla suffers from
organic brain damage, an extreme mental disturbance
significantly impairing several of his cognitive functions:
impulsivity, reasoning and judgment, ability to make sense
 77

out of experience, ability to draw conclusions, ability to
accept long-term consequences of immediate behavior,
emotional liability, concentration, mental flexibility,
recalling and integrating information, controlling behavior,
and controlling motor movements (hands). App. at 1572-74,
1577-79, 1581-84, 1617-18, 1707, 1717-18, 1721-22,
1726, 1728-33. The doctors believe Rompilla’s problems
relate back to his childhood, and were likely caused by fetal
alcohol syndrome. App. at 1601-02, 1606, 1615. They
concluded that Rompilla’s capacity to appreciate the
criminality of his conduct or to conform his conduct to the
law was substantially impaired at the time of the offense.
App. at 1616, 1687-88, 1735-36.
The PCRA court, applying Pennsylvania’s three-prong
standard for ineffective assistance of counsel claims, stated
that although Rompilla’s claim arguably had merit because
he was “entitled to have relevant information of mental
infirmity” presented to the jury, App. at 2028, “[g]iven the
fact that three health care professionals, all of whom were
experienced forensic experts, had provided opinions to
defense counsel, and none of them asked for more
information, it was hardly unreasonable or ineffective for
defense counsel to have relied upon their opinions.” App. at
2030.
The PCRA court accepted Rompilla’s trial counsel’s
testimony that they had spoken to family members in detail
and that the family had not revealed the information that
was claimed in the PCRA hearing. App. at 2029-30. The
court also noted that Rompilla “made contradictory
statements” to counsel during their representation, and
that he had not indicated he had any mental problems or
alcoholic blackouts. App. at 2029. The PCRA court
concluded that counsel was reasonable in believing that
seeking mercy was the only available strategy, and found
that there was a reasonable basis for counsel’s actions.
App. at 2029-30.
The Pennsylvania Supreme Court, also citing the state’s
three-prong test for ineffective assistance of counsel claims,
agreed with the PCRA court’s conclusion that counsel was
effective, concluding without discussion that counsel acted
reasonably. Rompilla, 721 A.2d at 789-90. The Court,
 78

relying on the conclusions of the PCRA court, noted that
counsel had investigated Rompilla’s mental health by
retaining three experts and reasonably relied on the experts
and on their own discussions with Rompilla to determine
the records needed to evaluate Rompilla, and it effectively
adopted the PCRA court’s credibility determination of trial
counsel’s conversations with family members, leading it to
conclude that counsel had not failed to investigate. The
Court rejected Rompilla’s argument that the experts
received inadequate information because of counsel’s
failure to investigate his background. Id.
An examination of the record shows that Rompilla’s
lawyers did less investigation into mitigating evidence than
did counsel for Wiggins. The Pennsylvania Supreme Court,
in affirming the decision of the PCRA court denying post-
conviction relief, stated that there was no “arguable merit”
in Rompilla’s claim that trial counsel failed to investigate
Rompilla’s family background, Rompilla, 721 A.2d at 790.
The Court never referred to counsel’s failure to speak to two
of Rompilla’s siblings, Randi Rompilla and Barbara Harris,
who lived nearby and who attended the trial. Dantos, when
questioned at the post-conviction hearing, said that she
didn’t remember whether she spoke to those sisters or not,
App. at 1099-1100, but those sisters testified at the PCRA
hearing that they would have testified at the penalty phase,
if asked, and would have told counsel about Rompilla’s
dysfunctional background, if they had been asked.
Counsel certainly had reason to inquire further as to the
availability of other family members. Counsel was aware
that the family members whom she interviewed did not
know a great deal about Rompilla. Charles testified that the
family said “they hardly know him.” One said, “He was in
a reformatory. He’s been away the whole time. We didn’t
know him well.” Even Dantos testified that “[T]he
overwhelming response from the family was that they didn’t
really feel as though they knew him all that well since he
had spent the majority of his adult years and some of his
childhood years in custody. . . ,” App. at 1094, and that the
family had “limited knowledge of their brother.” App. at
1098. Dantos stated that “it seemed pretty clear that [the
family members she interviewed] didn’t feel as though they
 79

knew Ron very well,” App. at 1166. It is thus apparent that
trial counsel had sufficient leads, as in Wiggins, to
investigate further to find family members who did know
more about Rompilla’s youth. Counsel never explained why
she did not interview the other sisters.
The Pennsylvania Supreme Court, in rejecting Rompilla’s
PCRA claim that counsel did not obtain records that would
have aided the mental health experts who evaluated him,
agreed with the PCRA court that counsel reasonably relied
on their experts to determine the record needed to evaluate
his mental health. Rompilla, 721 A.2d at 790. Counsel
cannot so easily shed their constitutional obligations.
Moreover, it appears that counsel directed the experts to
the guilt phase, giving no or little attention to the penalty
phase.
Rompilla’s lawyers sought opinions from the mental
health experts they hired primarily about Rompilla’s
“mental infirmity or mental insanity for the guilt phase.”
App. at 1069-71. Although Dantos stated that she also
instructed the experts to see if there was any issue “to
possibly use in mitigation any mental infirmity,” id. at
1067, Dr. Gross’ report says that counsel’s instruction as to
the evaluation was “to determine Mr. Rompilla’s mental
state during the time of the alleged charges.” Id. at 1069.
That is supported by Dantos’ own testimony that the
mental health professionals employed by the defense were
asked to look into Rompilla’s “mental state at the time of
the commission of the offense.” App. at 1071. Counsel did
not themselves investigate Rompilla’s medical history, drug
use, birth trauma or developmental delays; did not request
or instruct the medical experts they retained to investigate
those issues, and the medical experts did not make any
such investigation. Counsel did not provide Drs. Sadoff and
Cooke, who saw Rompilla, with any of the records that
would have shown Rompilla’s long history of alcoholism
and never themselves investigated the records that would
have shown that history.
Rompilla’s counsel did not seek or obtain any of
Rompilla’s school records and therefore did not learn that
Rompilla was in special education, left school in the 9th
grade, and that his abilities had not advanced beyond the
 80

third grade, suggesting mental retardation because of a
possible organic brain disorder. Rompilla, 2000 WL 964750,
at *4-*7. Because they did not obtain these records, they
did not provide them to the health experts they retained.
The PCRA court, on which the Pennsylvania Supreme
Court relied, did not explain its finding that the school
records were “not entirely helpful” beyond noting that
Rompilla’s IQ can “simply be part of the Bell Curve and a
learning disability is not necessarily caused by an organic
defect,”a statement made by Dr. Cooke, one of the experts
retained by Rompilla’s trial counsel. App. at 2029. Trial
counsel failed to investigate further into Rompilla’s retarded
level IQ scores. Indeed, Dr. Cooke further testified at the
PCRA hearing that had he been provided with Rompilla’s IQ
scores at trial he would have done testing for brain damage,
looked for evidence of prenatal damage to Rompilla’s brain,
and looked at the family situation, whether it was an
abusive situation, or dysfunctional situation. App. at 1800-
01.
Counsel did not investigate records from the
Pennsylvania Department of Corrections where Rompilla
was incarcerated for 14 years and therefore did not learn
that his adult scores on achievement tests were very low,
his spelling and arithmetic achievement scores were below
96% of the population, App. at 1009, and that
psychological tests performed showed serious abnormalities
on the schizophrenia, paranoia, neurosis and
obsessive/compulsive scales. See App. at 1595-99. The
court records, which counsel did not investigate, were used
by the Commonwealth as an aggravating circumstance and
were in the same courthouse where Rompilla’s case was
tried. Because counsel failed to investigate into Rompilla’s
prior correctional experience, they did not provide that
information to the health experts they retained.
In Wiggins, the Supreme Court quoted from the American
Bar Association’s Guidelines for the Appointment and
Performance of Counsel in Death Penalty Cases 11.4.1(C),
p. 93 (1989), which provide that investigations into
mitigating evidence “ ‘should comprise efforts to discover all
reasonably available mitigating evidence and evidence to
rebut any aggravating evidence that may be introduced by
 81

the prosecutor.’ ” 123 S. Ct. at 2537 (emphasis in original).
Because Wiggins’ counsel abandoned the investigation of
his background by failing to pursue evidentiary leads
provided in the DSS records, they fell short of these “well-
defined norms.” Id.
As shown by the above, the investigations by Rompilla’s
counsel were no more thorough, perhaps less, than those
found inadequate by the Supreme Court in Wiggins. The
Majority’s attempt to reconcile its conclusion that
Rompilla’s counsel provided effective assistance of counsel
with the conclusion in Wiggins that defendant’s counsel
were ineffective is nothing short of astonishing. The
Majority states that “[t]here are critical differences between
the conduct of Wiggins’s and Rompilla’s trial attorneys.”
Maj. Op. at 38. It continues, “Wiggins’s attorneys were
presented with leads that ‘any reasonably competent
attorney’ would have realized were promising. Rompilla’s
attorneys had no comparable leads.” Id.
With due respect to my colleagues on the Majority, the
distinction entirely misses the point. If it was ineffective for
Wiggins’ counsel to fail to follow up the leads they had, was
it not even more ineffective, indeed inexcusable, for
Rompilla’s attorneys to fail to investigate to find the leads
that could have been used by the experts they retained or
to retain experts to testify at the penalty phase and present
a viable case for mitigation? At the brief sentencing hearing,
trial counsel called only five witnesses, Rompilla’s sister
Darlene, older brother Nicholas, Jr., younger brother
Robert, sister Sandy Whitby and Aaron, Rompilla’s
fourteen-year old son. The total examination covered about
26 pages of notes of testimony. The witnesses testified in
general that Rompilla was a good family member and never
had a problem. As described by the District Court, the
“testimony was apparently presented to engender sympathy
for [Rompilla].” 2000 WL 964750, at *4. “It seemed to be
designed primarily as an emotional appeal to the jury to
show mercy to [Rompilla] — he wasn’t as bad as he seemed
and his family loved him.” Id. In lieu of the case trial
counsel presented in mitigation, a case that the District
Court described as “unreasonably brief and lacking in real
substance,” 2000 WL 964750, at *4, had they investigated
 82

they could have presented a case such as that presented by
the PCRA counsel. Rompilla’s counsel shifted the
responsibility for finding leads to the medical experts, and
never even told their experts that they had the
responsibility for uncovering that information.
In Wiggins, trial counsel made the tactical decision to
forego mitigating evidence of Wiggins’ dysfunctional
background and mental health problems because they
believed that Wiggins’ “best hope of escaping the death
penalty was for one or more jurors to entertain a
reasonable doubt as to his criminal agency.” Wiggins v.
State, 724 A.2d 1, 15 (Md. 1999), a decision the Maryland
Supreme Court termed a “deliberate, tactical decision”
which, under Strickland, should not be second guessed. Id.
at 15, 17-18. When the United States Supreme Court
reviewed that decision under § 2254(d), it held that because
counsel’s investigation was inadequate, the state court’s
“subsequent deference to counsel’s strategic decision . . .
despite the fact that counsel based this alleged choice on
what we have made clear was an unreasonable
investigation, was also objectively unreasonable.” Wiggins,
123 S. Ct. at 2538.
The PCRA court found that Rompilla’s counsel “had a
reasonable basis for proceeding as they did,” App. at 2028,
because they employed two experts and obtained an
evaluation by another psychiatrist, who also advised
counsel that he found nothing that would be beneficial in
the penalty phase. The Pennsylvania Supreme Court agreed
that “trial counsel was effective with respect to their
investigation and presentation of mitigation evidence.”
Rompilla, 721 A.2d at 790. Notwithstanding the decision in
Wiggins, the Majority, applying § 2254(d)(1), holds that “the
state court’s determination that counsel acted reasonably
was not ‘contrary to’ or an ‘unreasonable application’ of
Strickland.” Maj. Op. at 41.
As noted above, I believe the Majority seriously errs. It is
clearly established by both Williams and Wiggins that
counsel or counsel’s experts cannot make a reasonable
decision at the penalty phase if they do not investigate the
relevant facts that could be used in mitigation. In holding
that the Maryland Court of Appeals’ application of
 83

Strickland’s governing legal principles was objectively
unreasonable, Justice O’Connor, in Wiggins, stated that
counsel’s failure to thoroughly investigate made “a fully
informed decision with respect to sentencing strategy
impossible.” Wiggins, 123 S. Ct at 2538. Therefore the state
Supreme Court’s holding “reflected an unreasonable
application of Strickland.” Id. What was true in Wiggins and
before that in Williams is equally applicable here. It follows
that the District Court properly determined that Rompilla’s
trial counsel did not meet the performance standards
required under Strickland.
I believe that under the circumstances in which counsel
presented an inadequate case for mitigation at the penalty
phase of a capital sentencing hearing, the prejudice prong
of constitutionally ineffective assistance of counsel is clearly
met. The Pennsylvania Supreme Court did not reach this
issue and therefore we may consider it de novo. A
reasonable attorney, if aware of the evidence presented at
the PCRA hearing following a thorough investigation, would
have done more at sentencing than plead for mercy. As the
Court stated in Wiggins, “had the jury been confronted with
[the] considerable mitigating evidence, there is a reasonable
probability that it would have returned with a different
sentence.” 123 S. Ct. at 2543. Quoting from Williams, the
Court stated, “we must evaluate the totality of the evidence,
both that adduced at trial, and the evidence adduced in the
habeas proceeding.” Wiggins, 123 S. Ct. at 2543, citing
Williams, 529 U.S. at 397-98. In Williams, the Court
recognized that the graphic description of the defendant’s
childhood “filled with abuse and privation, or the reality
that he was ‘borderline mentally retarded’ might well have
influenced the jury’s appraisal of his morale culpability.”
529 U.S. at 398 (citation omitted). The Court further stated
that “[m]itigating evidence unrelated to dangerousness may
alter the jury’s selection of penalty, even if it does not
undermine or rebut the prosecution’s death-eligibility case.”
Id.
In considering prejudice to Rompilla from his trial
counsel’s performance, we should look at the totality of
evidence adduced at trial as well as that adduced at the
state post-conviction hearing where counsel, after a
 84

thorough investigation, found the evidence of Rompilla’s
abusive background, his disfunctional family situation, his
low IQ, his meager reading and understanding ability found
in the prison records, and the medical evidence of brain
disfunction. The jury could certainly have considered this
matter as sufficiently mitigating to warrant a different
sentence. Therefore, I believe that the Supreme Court of
Pennsylvania failed to reasonably apply the Strickland
standard when it held that Rompilla had not shown
ineffective assistance of counsel.
I therefore disagree with the majority and would affirm
the District Court’s grant of a writ of habeas corpus
because of trial counsel’s ineffective assistance of counsel.

II.

Failure to Give a Simmons Instruction
Rompilla has cross-appealed from the District Court’s
denial of a writ of habeas corpus on Rompilla’s claim that
his right to due process was violated by the state trial
court’s refusal to inform the jury in response to the jury’s
inquiries that Rompilla was parole ineligible if sentenced to
life imprisonment.
While the jury was deliberating Rompilla’s penalty, life
imprisonment or death, the jury asked in succession, “If a
life sentence is imposed, is there any possibility of the
Defendant ever being paroled?” App. at 802. The trial court
responded:
I’m sorry to say, I can’t answer that question. That’s
not before you as such. The only matter that you can
consider in the Sentencing Hearing is the evidence that
was brought out in the course of the Hearing and the
Law with respect to the Court’s Charge. That’s the only
consideration you have, I’m sorry to say. I—if there
were other alternatives that you should consider, we
would have outlined them in the Charge, all right. Are
there any other questions?
App. at 802-03.
 85

The jury later requested to examine information with
respect to Rompilla’s prior sentence. The trial court refused
because it was not entered into evidence. The jury foreman
clarified that the jury wanted to know the sentence from
Rompilla’s prior conviction. The trial court stated that he
cannot give that. A juror then asked “if he got released on
behavioral . . .” and the foreman added, “It was commuted
in any way, the original sentence.” App. at 823. As before,
the trial court refused to give the requested information.
App. at 824.
Finally, the next day, after more deliberation the jury
asked, “Was the Defendant offered any type of
rehabilitation either while in prison or after his release from
prison?” App. at 842. Once again the trial court declined to
answer, even after the foreman changed the question to ask
“isn’t rehabilitation available in prison?” App. at 842. The
trial court again refused to provide the information sought
by the jury. It was only after the trial court declined to
provide the information requested by the jury time after
time that the jury returned a sentence of death.
On Rompilla’s appeal from the denial of his PCRA
petition, the Supreme Court of Pennsylvania rejected
Rompilla’s contention that due process required that the
jury be instructed that in Pennsylvania “life imprisonment
means life,” ruling as follows:
Under the current state of the law in Pennsylvania,
the jury must be told that life means life without parole
only when the defendant’s future dangerousness is at
issue. Commonwealth v. Clark, 551 Pa. 258, 710 A.2d
31, 35-36 (Pa. 1998). Appellant argues that his future
dangerousness was at issue because the
Commonwealth argued the aggravating circumstance
that he has a significant history of felony convictions
involving the use or threat of violence. The Court
rejected this argument in Commonwealth v. May, 551
Pa. 286, 710 A.2d 44, 47 (Pa. 1998). As stated in that
case, this aggravating circumstance only addresses
Appellant’s past conduct, not his future
dangerousness. See id. Thus, no relief is due.10
10
This author [Justice Nigro] agrees with the dissent’s
position that a Simmons instruction should be given in
 86

all cases and has previously so stated. See Clark, 710
A.2d at 43-44 (Nigro, J., concurring, joined by Flaherty,
J. and joined in relevant part by Zappala, J.); May, 710
A.2d at 49 (Nigro, J., concurring, joined by Zappala,
J.). Under the current state of the law, however,
Appellant’s argument that he was entitled to the
instruction because the Commonwealth argued the
aggravating circumstance identified above, is without
merit. See May, 710 A.2d at 47.
Commonwealth v. Rompilla, 721 A.2d 786, 795 (Pa. 1998).
In his dissent, Chief Justice Flaherty wrote:
I believe the majority is in error in its treatment of the
issue pertaining to the jury’s question about the
defendant’s parole eligibility. Under Simmons v. South
Carolina, 512 U.S. 154, 114 S.Ct. 2187, 129 L.Ed.2d
133 (1994) and Commonwealth v. Clark, 551 Pa. 258,
710 A.2d 31, 35-36 (Pa. 1998), the court must tell a
jury that a life sentence means life without parole, if
the defendant requests the instruction and his future
dangerousness is at issue. Here, during deliberations
in the penalty phase, the jury asked, “If a life sentence
is imposed, is there any possibility of the Defendant
ever being paroled?” I view this question as a clear
expression of the jury’s concern about the defendant’s
future dangerousness. I would therefore hold that the
trial court’s refusal to explain the meaning of life
without parole constituted error under Commonwealth
v. Clark, supra. I would therefore reverse the order of
the court of common pleas and remand for proceedings
consistent with this opinion.
I would go further and require an explanation of the
meaning of a life sentence in all capital cases. There
can be no harm in instructing juries that in
Pennsylvania appellant would be statutorily ineligible
for parole if sentenced to life in prison, but that a life
sentence might nonetheless be commuted by the
governor. On the other hand, if we do not so instruct,
a jury, erroneously believing that a prisoner sentenced
to life may be paroled within a period of years, may
impose the death penalty for reasons which are not
based in law.
 87

Id. at 795-96 (Flaherty, C.J., dissenting).1
I believe that the Pennsylvania Supreme Court’s analysis
of Rompilla’s Simmons claim was an “unreasonable
application” of Supreme Court precedent, specifically
Simmons v. South Carolina, 512 U.S. 154 (1994), and Kelly
v. South Carolina, 534 U.S. 246 (2002), and therefore that
it must be reversed under the standards of AEDPA.
In Simmons, the Supreme Court held that the defendant’s
due process right to answer an allegation against him
requires the trial court to instruct the jury that the
alternative to the death penalty is life without parole (in
states where that is the alternative) if the prosecutor argues
that the defendant will pose a danger to others. As Justice
Blackmun, who authored the plurality opinion, wrote: “The
State may not create a false dilemma by advancing
generalized arguments regarding the defendant’s future
dangerousness while, at the same time, preventing the jury
from learning that the defendant never will be released on
parole.” Simmons, 512 U.S. at 171. In Shafer v. South
Carolina, 532 U.S. 36 (2001), the Supreme Court described
Simmons as holding that when “a capital defendant’s future
dangerousness is at issue, and the only sentencing
alternative to death available to the jury is life
imprisonment without possibility of parole, due process
entitles the defendant ‘to inform the jury of [his] parole
ineligibility, either by a jury instruction or in arguments by
counsel.’ ” Id. at 39 (quoting Ramdass v. Angelone, 530 U.S.
156, 165 (2000) (plurality opinion)).

1. I find it difficult to glean much enlightenment from the two cases cited
by the Pennsylvania Supreme Court. In Commonwealth v. Clark, 710
A.2d 31 (Pa. 1998), the Court did not define what it meant to put future
dangerousness “at issue” because the question did not arise. (In Clark it
was not the Commonwealth that argued future dangerousness; rather,
the defense argued the opposite, noting that chances of a commuted
sentence in Pennsylvania were close to zero; the court gave an
instruction defining life imprisonment.) As for Commonwealth v. May,
710 A.2d 44 (Pa. 1998), its proposition that “[t]he aggravating
circumstance of appellant’s prior record for violent felonies addressed
only appellant’s past conduct, not his future dangerousness,” is a mere
unsupported assertion. Id. at 47.
 88

The Simmons plurality and Justice O’Connor, concurring,
believed there was no question that the prosecutor had
made an issue of Simmons’s future dangerousness. As a
result, the Court had no occasion to define what
constitutes making an issue of future dangerousness. That
issue was directly addressed eight years later in Kelly, an
opinion of the Court authored by Justice Souter.
In Kelly, the prosecutor stated that he would not argue
future dangerousness and “that takes it out of Simmons
anyhow.” Kelly, 534 U.S. at 249. The trial court then denied
Kelly’s counsel’s request for a Simmons instruction, saying
that the State’s evidence went to Kelly’s character and
characteristics, not to future dangerousness. The South
Carolina Supreme Court affirmed the conviction and death
sentence, holding that the Simmons instruction was not
required because future dangerousness was not at issue.
In reversing, Justice Souter stated that the South
Carolina Supreme Court’s statement that Kelly’s future
dangerousness was not at issue “is unsupportable on the
record before us.” Id. at 252. He continued, “It is not that
the state court failed to pose the legal issue accurately, for
in considering the applicability of Simmons it asked
whether Kelly’s future dangerousness was ‘a logical
inference from the evidence,’ or was ‘injected into the case
through the State’s closing argument.’ ” Id. at 252 (internal
citations omitted). In support of the appropriateness of this
statement of the legal issue, Justice Souter included the
following citations and explanatory parentheses: “Shafer,
[532 U.S.] at 54-55 (whether prosecutor’s evidence or
argument placed future dangerousness in issue) [and]
Simmons, 512 U.S. at 165, 171, (plurality opinion) (future
dangerousness in issue because ‘State raised the specter of
. . . future dangerousness generally’ and ‘advanc[ed]
generalized arguments regarding the [same]’).” Kelly, 534
U.S. at 252.
In addressing the trial court’s interpretation of the
evidence, the Kelly Court explained why the trial court
erred:
To the extent that it thought that “[e]vidence that Kelly
took part in escape attempts and carried a shank . . .
 89

is not the type of future dangerousness evidence
contemplated by Simmons,” . . . it overlooked that
evidence of violent behavior in prison can raise a
strong implication of “generalized . . . future
dangerousness.” Simmons, [512 U.S.] at 171. (And, of
course, the state court’s reasoning says nothing about
the evidence of the crime, or of Kelly’s sadism
generally, and his mercurial thirst for vengeance.) A
jury hearing evidence of a defendant’s demonstrated
propensity for violence reasonably will conclude that he
presents a risk of violent behavior, whether locked up
or free, and whether free as a fugitive or as a parolee.
Id. at 253-54 (emphasis added).
In language particularly relevant to this case, the Kelly
Court stated, “The fallacy of the State Supreme Court’s
attempt to portray the thrust of the evidence as so
unrealistically limited harks back to a comparable mistake
by the trial judge, who spoke of the evidence as going, not
to future dangerousness, but ‘to [Kelly’s] character and
characteristics.’ ” Id. at 254. The Court continued,
The error in trying to distinguish Simmons this way lies
in failing to recognize that evidence of dangerous
“character” may show “characteristic” future
dangerousness, as it did here. This, indeed, is the fault
of the State’s more general argument before us, that
evidence of future dangerousness counts under
Simmons only when the State “introduc[es] evidence for
which there is no other possible inference but future
dangerousness to society.” . . . Evidence of future
dangerousness under Simmons is evidence with a
tendency to prove dangerousness in the future; its
relevance to that point does not disappear merely
because it might support other inferences or be
described in other terms.
Id. (emphasis in original).
The Kelly Court thus made explicit what was implicit in
Simmons: the prosecutor need not expressly and separately
argue future dangerousness; rather, future dangerousness
can be made an issue through implication by or inference
 90

from arguments addressing such independent matters as
the defendant’s character.
At Rompilla’s sentencing hearing, the prosecutor, who
was seeking the death penalty, focused on Rompilla’s
“significant history of felony convictions,” as that is one of
the statutory factors that can serve as “aggravators” in the
balancing between mitigating factors and aggravating
factors required under Pennsylvania’s death penalty law. 42
Pa. Cons. Stat. Ann. § 9711(d)(9). During his closing
argument the prosecutor referred to Rompilla’s prior
offense, which involved the robbery, rape and slashing with
a knife of a female bar owner, in the following terms:
Joe Macrenna, the woman that was raped, was raped
pretty brutally. She was raped at knife point. . . . [I]sn’t
it frightening, the similarity between that case and this
case. I mean, it is absolutely astounding. Both take
place around the bar. The Defendant gets in after
closing or right before closing. . . . On both occasions,
a knife was used. Steals money both times. Isn’t it
frightening the similarities in those crimes. Takes a taxi
away from Joe’s Bar, takes a taxi the night of this
crime. He slashes Joe in the breast with a knife. He
uses a knife on Jimmy Scanlon. It’s absolutely
frightening to think of the similarities in those two
crimes. But there is one difference, one major
difference, Joe Macrenna lived through her experience.
Jimmy Scanlon didn’t. . . . I think the Defendant
learned a lesson from Joe Macrenna in that case, that
Rape case. That lesson was, don’t leave any witnesses.
Don’t leave anybody behind that can testify against
you.
App. at 779-80 (emphasis added).
At sentencing the jury also learned via testimony elicited
by both parties that the instant crime occurred shortly after
Rompilla had been released from prison, that there is a lack
of rehabilitation services for prisoners, that Rompilla’s
children were initially frightened of him when he was
released on parole, and that he had been paroled from his
previous sentence only three weeks before the instant
crime.
 91

Rompilla argues that because the prosecutor placed
Rompilla’s future dangerousness at issue, the trial court
was required under Simmons to inform the jury that
Rompilla is parole ineligible. The PCRA court disagreed,
reasoning that Simmons does not apply because the
prosecutor did not argue future dangerousness. The court
stated, “The prosecution merely argued that Mr. Rompilla
had a significant history of felony convictions which was
proper argument as an aggravating factor under
Pennsylvania law.” App. at 2025. The court did not find the
jury’s question concerning whether he had received
rehabilitation supportive of Rompilla’s argument: “This
question . . . relates more to the issues raised by the
defense witnesses during the penalty phase who
complained that Mr. Rompilla had not received
rehabilitation while in prison and while on parole after his
1974 rape and burglary convictions. In fact, that was a
mitigating factor found by the jury rather than an
aggravating factor.” App. at 2026. The court did not
address, perhaps did not recognize, the relevance of the
jury’s question concerning whether rehabilitation is
available in prison to Rompilla’s request for a Simmons
instruction.
On appeal, the Pennsylvania Supreme Court summarized
the facts and issue, and analyzed Rompilla’s Simmons claim
in one paragraph which I quoted in full near the beginning
of this dissent. The Pennsylvania Supreme Court
interpreted Pennsylvania law as requiring a Simmons
instruction “only when the defendant’s future
dangerousness is at issue,” and expressly rejected
Rompilla’s argument that in presenting Rompilla’s extensive
criminal history the prosecutor ipso facto made future
dangerousness an issue. The Pennsylvania Supreme Court
thus appears to have determined that future
dangerousness must be argued expressly and separately
and may not be implied by or inferred from other
arguments; consequently, the prosecutor’s recitation of
defendant’s past criminal conduct cannot, without more,
also amount to making an issue of future dangerousness.
It is unclear whether the Pennsylvania Supreme Court
believes this principle to be consistent with or required by
Simmons.
 92

When the Simmons issue was before the District Court on
Rompilla’s habeas petition, that court also gave the issue
only brief analysis and concluded:
The prosecutor’s summation in this case covers 16
pages of the notes of testimony2 and a fair reading of it
leads to the conclusion that the state’s reasoning for
the death penalty was not based upon future
dangerousness but on the despicable, savage and
cowardly beating the Petitioner inflicted upon his
victim. This is a close issue, however, but the Supreme
Court of Pennsylvania’s decision in the PCRA case was
not an unreasonable application of federal law.
Rompilla v. Horn, 2000 WL 964750, at *15 (E.D. Pa. July
11, 2000). The District Court appears to have agreed with
the Pennsylvania Supreme Court that if the prosecutor
“based his argument on” the despicable nature of the
defendant’s actions, he could not at the same time have
made an issue of Rompilla’s future dangerousness for
Simmons purposes without providing a separate, express
argument to that effect. The Majority appears to agree, as
its opinion states: “even if [the prosecutor’s comments] were
meant to imply that Rompilla would present a future
danger if he was ever released from prison, the fact remains
that the prosecutor never expressly argued that Rompilla
presented a future threat.” Maj. Op. at 63-64.
I do not read Simmons as requiring an express argument
of future dangerousness. In Simmons, the prosecutor put
future dangerousness “at issue” in a few brief comments in
his closing argument. As the Simmons plurality opinion
describes it:
In its closing argument the prosecution argued that
petitioner’s future dangerousness was a factor for the
jury to consider when fixing the appropriate
punishment. The question for the jury, said the
prosecution, was “what to do with [petitioner] now that
he is in our midst.” . . . The prosecution further urged
that a verdict for death would be “a response of society
to someone who is a threat. Your verdict will be an act
of self-defense.”

2. That is not accurate: it covers eight pages. See App. at 774-782.
 93

512 U.S. at 157.
In his dissent, Justice Scalia takes issue with the
conclusion of those Justices who formed the majority that
the above statements constituted an argument for future
dangerousness.3 The language used in the various opinions

3. The dissent states:
Both Justice Blackmun and Justice O’Connor focus on two portions
of the prosecutor’s final argument to the jury in the sentencing
phase. First, they stress that the prosecutor asked the jury to
answer the question of “what to do with [petitioner] now that he is
in our midst.” That statement, however, was not made (as they
imply) in the course of an argument about future dangerousness,
but was a response to petitioner’s mitigating evidence. Read in
context, the statement is not even relevant to the issue in this case:
“The defense in this case as to sentence . . . [is] a diversion. It’s
putting the blame on society, on his father, on his grandmother,
on whoever else he can, spreading it out to avoid that personal
responsibility. That he came from a deprived background. That he
didn’t have all of the breaks in life and certainly that helps shape
someone. But we are not concerned about how he got shaped. We
are concerned about what to do with him now that he is in our
midst.”. . . .
Both opinions also seize upon the prosecutor’s comment that the
jury’s verdict would be “an act of self-defense.” That statement came
at the end of admonition of the jury to avoid emotional responses
and enter a rational verdict:
“Your verdict shouldn’t be returned in anger. Your verdict
shouldn’t be an emotional catharsis. Your verdict shouldn’t be . . .
a response to that eight-year-old kid [testifying in mitigation] and
really shouldn’t be a response to the gruesome grotesque
handiwork of [petitioner]. Your verdict should be a response of
society to someone who is a threat. Your verdict will be an act of
self-defense.” . . . .
This reference to “self-defense” obviously alluded, neither to defense
of the jurors’ own persons, nor specifically to defense of persons
outside the prison walls, but to defense of all members of society
against this individual, wherever he or they might be. Thus, as I
read the record (and bear in mind that the trial judge was on the
lookout with respect to this point), the prosecutor did not invite the
jury to believe that petitioner would be eligible for parole — he did
not mislead the jury.
Simmons, 512 U.S. at 181-182 (Scalia, J., dissenting) (emphasis in
original).
 94

of the Justices who formed the Simmons majority does not
suggest that the prosecutor must expressly argue that
defendant will be a serious threat if not sentenced to death
before the Court will hold that future dangerousness has
been placed before the jury. For example, Justice
Blackmun’s opinion uses the following phrases (emphases
added, passim): “where the defendant’s future
dangerousness is at issue,” id. at 156; “[t]he Due Process
Clause does not allow the execution of a person ‘on the
basis of information which he had no opportunity to deny or
explain.’ ” Id. at 161; “the State’s repeated suggestion that
petitioner would pose a future danger to society if he were
not executed,” id. at 162; “[t]he State raised the specter of
petitioner’s future dangerousness generally . . . ,” id. at
165; “if the State rests its case for imposing the death
penalty at least in part on the premise that the defendant
will be dangerous in the future . . . ,” id. at 168-69; “[t]he
State may not create a false dilemma by advancing
generalized arguments regarding the defendant’s future
dangerousness . . . ,” id. at 171.
In his concurring opinion, Justice Souter writes, “when
future dangerousness is an issue in a capital sentencing
determination . . . .” Id. at 172 (emphasis added). Similarly,
Justice O’Connor states, “When the State seeks to show the
defendant’s future dangerousness,” and “a means of
responding to the State’s showing of future dangerousness.”
Id. at 177. “Moreover, the prosecutor, by referring to a
verdict of death as an act of ‘self-defense,’ strongly implied
that petitioner would be let out eventually if the jury did
not recommend a death sentence.” Id. at 178 (emphasis
added).
There are indeed some statements that can be read to
refer to arguments expressly made by the State as to
Simmons’s future dangerousness, particularly in the
concurring opinion of Justice O’Connor, but even she uses
the phrase “strongly implied.” I therefore repeat my
observation that neither Justice O’Connor nor the other
Justices forming the majority suggest that the instruction
must be given only if the prosecutor makes express,
explicit, exclusive arguments for future dangerousness.
Rather, the Court was willing to infer a future
 95

dangerousness argument from what was, on Justice
Scalia’s account, an argument about a different topic
entirely (the merit of one of Simmons’s mitigating factors).
Nothing in the Justices’ language supports the premise on
which the Pennsylvania Supreme Court’s decision turns.
Any doubts on the matter were dispelled by the Supreme
Court’s subsequent decision in Kelly where the prosecutor
did not expressly argue future dangerousness.
The Majority concedes that Kelly “arguably broadened the
holding in Simmons,” Maj. Op. at 51, and notes that two of
the Justices who joined Justice O’Connor’s concurring
opinion in Simmons dissented in Kelly because “the test is
no longer whether the State argues future dangerousness
to society . . . [but] whether evidence was introduced at trial
that raises an ‘implication’ of future dangerousness to
society.” Kelly, 534 U.S. at 261 (Rehnquist, C.J., joined by
Kennedy, J., dissenting).
The Majority declines to apply Kelly in Rompilla’s favor,
stating, “Even if Kelly broadened Simmons . . . Kelly cannot
aid Rompilla here” because Kelly was decided after the
Pennsylvania Supreme Court’s decision in Rompilla’s case.
Maj. Op. at 53. The Majority notes that under Williams v.
Taylor, 529 U.S. 362 (2000), federal review pursuant to
section 2254(d)(1) is limited to the state court’s application
of federal law “as of the time of the relevant state-court
decision.” Id. at 412.
The Supreme Court does not interpret the unreasonable
application of Supreme Court precedent prong of
§ 2254(d)(1) as narrowly as does the Majority. In Wiggins,
the Court viewed its opinion in Williams as “illustrative of
the proper application” of the Strickland standard,
notwithstanding that “Williams had not yet been decided at
the time the Maryland Court of Appeals rendered the
decision at issue in this case.” Wiggins, 123 S. Ct. at 2535.
Just as the Williams opinion noted that “the merits of
[Williams’s] claim are squarely governed by our holding in
Strickland v. Washington,” Williams, 529 U.S. at 390, so
also the Court’s opinion in Kelly represented an application
of Simmons and did not make new law. There is no
indication anywhere in the opinion that the Kelly majority
thought that they were doing anything other than applying
 96

Simmons. Rather, at every step of the way the Kelly
majority cites Simmons. At one point, the Kelly majority
states that the prosecutor “accentuated the clear
implication of future dangerousness raised by the evidence
and [thereby] placed the case within the four corners of
Simmons.” Kelly, 534 U.S. at 255. Therefore, even if Kelly
“arguably” broadened Simmons, it is dispositive for present
purposes that the majority of the Supreme Court believed
otherwise.
To recapitulate, the Pennsylvania Supreme Court applied
an unjustifiedly narrow test for determining whether the
prosecutor made an issue of Rompilla’s future
dangerousness. That test represented an unreasonable
application of Simmons, especially as Simmons was applied
in Kelly.
In any event, I believe that the record shows that the
prosecutor did make an issue of Rompilla’s future
dangerousness. It is noteworthy that the prosecutor
referred to the similarities between this crime and
Rompilla’s previous crime as “frightening” no less than
three times. The similarities can be “frightening” only if the
prosecutor was sending the message that there is a
possibility that Rompilla will repeat the crime. Similarly,
the prosecutor’s emphasis on the one major difference
between this crime and his previous crime — the murder of
the victim — and his statement to the jury that Rompilla
had learned not to leave any witnesses were tantamount to
a warning about what would happen if Rompilla were
allowed to commit another crime, i.e., that Rompilla would
be dangerous. Moreover, as the Supreme Court has
recognized, evidence of past criminal conduct may be
indicative of future dangerousness. See Skipper v. South
Carolina, 476 U.S. 1, 5 (1986).
The Majority urges us to consider the prosecutor’s
comments in context, claiming he was merely responding to
counsel’s argument that the jury must have some doubt
concerning Rompilla’s guilt and that his comments were a
rebuttal of that argument, not an argument for future
dangerousness. The Majority states, “Seeking to dispel any
such doubts, the prosecutor’s obvious point in stressing the
similarities between the circumstances of the rape for
 97

which Rompilla had previously been convicted and the
Scanlon murder was to convince the jury that the same
man had committed both crimes.” Maj. Op. at 63.
The Majority’s attempt to explain the remark is sheer
conjecture. The prosecutor made his remarks in the course
of outlining why three aggravating factors apply against
Rompilla. After addressing the first two — torture and
murder while committing other crimes — the prosecutor
said:
Commonwealth has more. Commonwealth submits to
you that the Defendant has a significant history of
felony convictions involving violence to the . . . person
or the threat of violence to the person.
App. at 779. The prosecutor then described the previous
rape with his above-quoted use of “frightening” and gave
the State’s explanation that Rompilla did not wait at the
bar until Scanlon closed up before breaking in to steal
because Rompilla did not want to leave any witnesses. App.
at 780. This suggests dangerousness, not rebuttal.
It is plain from Simmons and Kelly that the reasons for
the prosecutor’s statements are not dispositive of whether
they put future dangerousness at issue for the jury. They
could have done so even if their principal thrust were to
dispel doubt raised by the defense. After all, in Simmons
the prosecutor’s remarks were made in response to defense
counsel’s mitigating evidence (see Justice Scalia’s
dissenting opinion, quoted above). Thus, even if the
prosecutor in Rompilla’s case did not make an express
argument, Simmons is still applicable because, as explained
above, Simmons does not require “express arguments.”
The Majority also asserts that the prosecutor merely used
the word “frightening” as a synonym for “astounding,” a
word that the prosecutor also used (“more aptly,” as the
Majority puts it). Maj. Op. at 63. Accepting arguendo the
Majority’s hypothesis, it does not negate the implication of
future dangerousness. The meaning of a word or phrase
depends in part on the reasonable understanding of the
listener, not merely the intent of the speaker.
The Majority also attempts to discount any future
 98

dangerousness implications in the prosecutor’s statement
that by the time of this crime Rompilla had learned to kill
any eyewitnesses by explaining that “the comment seems to
have had two likely purposes: to explain why there was no
eyewitness to the most recent crime and to explain why the
two crimes differed in the important respect that one
involved a killing and the other did not.” Maj. Op. at 63.
Again, even if those were the prosecutor’s purposes, it does
not follow that he did not also thereby inject future
dangerousness into the proceedings. The more immediate
message learned by the jury was surely that if they put
Rompilla in a position from which he could commit another
crime, the jurors had better not be among the witnesses.
Their repeated questions to the judge concerning parole
and rehabilitation suggests that they had learned that
lesson.4

4. This court is limited to reviewing the Pennsylvania Supreme Court’s
opinion for its consistency with federal law, and has no role in its
application of state law. It is of interest, however, that there is
disagreement among the Pennsylvania Supreme Court justices
themselves as to whether the jury need be instructed that in
Pennsylvania the jury may not be informed that life imprisonment means
life without parole. Pennsylvania is one of only two states out of the 50
United States where life imprisonment means life without parole that
hold that the jury may not be so informed. South Carolina, the state that
was the subject of the Supreme Court decisions in Simmons, Kelly, and
Skipper, is the other such state.
The Pennsylvania rule was enunciated in Commonwealth v. Mills, 39
A.2d 572 (Pa. 1944), where the Pennsylvania Supreme Court stated that
in a capital case the jury’s only function was to decide whether the
penalty should be life imprisonment or the death penalty. More recently,
the Pennsylvania Supreme Court, in addressing the effect of Simmons on
the rule announced in Mills, stated that “the per se rule announced in
Mills [that any reference to the possibility of parole was an improper
consideration for the jury in their deliberation of the defendant’s guilt]
has been superseded [by Simmons].” Clark, 710 A.2d at 36. The
Pennsylvania Court then stated “that due process requires the court to
instruct the jury on the law as it relates to the possibility of parole where
that issue [the defendant’s future dangerousness] clearly arises from the
arguments of either counsel in the penalty phase.” Id. Justice Nigro
concurred, suggesting “that the better practice and policy is to require
trial courts to give a Simmons instruction in all death penalty
proceedings, regardless of whether counsel raises the issue of a
defendant’s potential future dangerousness during the penalty phase.”
Id. at 43. Justice Zappala concurred separately, agreeing with Justice
Nigro that a Simmons charge should be mandated. Id.
 99

Justice Souter, in his concurring opinion in Simmons,
joined by Justice Stevens, wrote:
The Eighth Amendment entitles a defendant to a jury
capable of a reasoned moral judgment about whether
death, rather than some lesser sentence, ought to be
imposed. The Court has explained that the Amendment
imposes a heightened standard “for reliability in the
determination that death is the appropriate
punishment in a specific case,” Woodson v. North
Carolina, 428 U.S. 280, 305 (1976) (plurality opinion of
Stewart, Powell, and Stevens, JJ.); see also, e.g.,
Godfrey v. Georgia, 446 U.S. 420, 427-428 (1980); Mills
v. Maryland, 486 U.S. 367, 383-384 (1988). Thus, it
requires provision of “accurate sentencing information
[as] an indispensable prerequisite to a reasoned
determination of whether a defendant shall live or die,”
Gregg v. Georgia, 428 U.S. 153, 190 (1976) (joint
opinion of Stewart, Powell, and Stevens, JJ.), and
invalidates “procedural rules that ten[d] to diminish the
reliability of the sentencing determination,” Beck v.
Alabama, 447 U.S. 625, 638 (1980).
That same need for heightened reliability also
mandates recognition of a capital defendant’s right to
require instructions on the meaning of the legal terms
used to describe the sentences (or sentencing
recommendations) a jury is required to consider, in
making the reasoned moral choice between sentencing
alternatives. Thus, whenever there is a reasonable
likelihood that a juror will misunderstand a sentencing
term, a defendant may demand instruction on its
meaning, and a death sentence following the refusal of
such a request should be vacated as having been
“arbitrarily or discriminatorily” and “wantonly and . . .
freakishly imposed.” Furman v. Georgia, 408 U.S. 238,
249 (1972) (Douglas, J., concurring) (internal quotation
marks omitted); id., at 310 (Stewart, J., concurring).
Simmons, 512 U.S. at 172-73 (Souter, J., concurring).
In this case, the repeated questions by the jury as to the
effect of a sentence by them of life imprisonment
demonstrate unequivocally that the jury did not understand
 100

that under Pennsylvania law a life prison term means life
without parole. I believe the rationale for requiring that the
jury be instructed about parole ineligibility when the
prosecutor puts future dangerousness in issue — i.e., to
ensure that the jurors have accurate information as to the
effect of their sentence — is similarly applicable in a case,
such as this one, where the jurors have requested accurate
information. I do not understand the State to be arguing
that the requested instruction was not legally accurate. The
Majority has provided no rationale why the jury should not
be informed of the applicable Pennsylvania law. Truth in
advertising is now the byword of this generation. Truth in
instructing the jury as to the effect of the sentence in a
capital case is at least as important.
I would grant a writ of habeas corpus on this issue as
well as on the ineffective assistance of counsel issue.

A True Copy:
Teste:

Clerk of the United States Court of Appeals
for the Third Circuit