PUBLISHED

UNITED STATES COURT OF APPEALS

FOR THE FOURTH CIRCUIT

ARTHUR MARTIN BOYD, JR.,
Petitioner-Appellant,

v.

JAMES B. FRENCH, Warden, Central
                                                                      No. 97-23
Prison, Raleigh, North Carolina;
MICHAEL F. EASLEY, Attorney
General of North Carolina,
Respondents-Appellees.

Appeal from the United States District Court
for the Middle District of North Carolina, at Winston-Salem.
William L. Osteen, Sr., District Judge.
(CA-89-127-WS-C)

Argued: March 4, 1998

Decided: June 19, 1998

Before MURNAGHAN, ERVIN, and WILKINS, Circuit Judges.

_________________________________________________________________

Affirmed by published opinion. Judge Wilkins wrote the opinion, in
which Judge Ervin joined. Judge Murnaghan wrote a concurring opin-
ion.

_________________________________________________________________

COUNSEL

ARGUED: James Richard Glover, GLOVER & PETERSEN, P.A.,
Chapel Hill, North Carolina; Thomas Kieran Maher, RUDOLPH &
MAHER, P.A., Chapel Hill, North Carolina, for Appellant. Barry Ste-
ven McNeill, Special Deputy Attorney General, NORTH CARO-
LINA DEPARTMENT OF JUSTICE, Raleigh, North Carolina, for
Appellees. ON BRIEF: Michael F. Easley, Attorney General of
North Carolina, NORTH CAROLINA DEPARTMENT OF JUS-
TICE, Raleigh, North Carolina, for Appellees.

_________________________________________________________________

OPINION

WILKINS, Circuit Judge:

Appellant Arthur Martin Boyd, Jr. filed this petition for habeas cor-
pus relief1 from his North Carolina capital conviction and death sen-
tence for the murder of his former girlfriend, 32-year-old Wanda Mae
Phillips Hartman. See 28 U.S.C.A. § 2254 (West 1994).2 The district
court denied the petition, holding inter alia that the state trial court
committed harmless error in failing to permit Boyd to present mitigat-
ing expert testimony at sentencing. Finding no reversible error in any
of Boyd's numerous arguments, we affirm.

I.

Boyd met Hartman in November 1978 while the two were
employed by the same company. Within days Boyd moved in with
Hartman, and the two resided together for approximately three and
one-half years. In April 1982, Hartman decided to move into her par-
ents' residence with her daughter. Boyd was not supportive of this
_________________________________________________________________
1 Boyd named James B. French, Warden of Central Prison, where Boyd
was then incarcerated, and Michael F. Easley, Attorney General of North
Carolina, as Respondents in this action. For ease of reference, we refer
to Respondents as "the State" throughout this opinion.
2 Because Boyd's petition for a writ of habeas corpus was filed on Feb-
ruary 16, 1989, prior to the April 24, 1996 enactment of the Antiterro-
rism and Effective Death Penalty Act (AEDPA) of 1996, Pub. L. No.
104-132, 110 Stat. 1214, amendments to 28 U.S.C.A.§ 2254 effected by
§ 104 of the AEDPA do not govern our resolution of this appeal. See
Lindh v. Murphy, 

, 2067 (1997). The State does not main-
tain that it has satisfied the opt-in requirements of § 107 such that those
provisions of the AEDPA apply.

                    2
decision and persistently attempted to reconcile with Hartman. Ulti-
mately, on Friday, July 30, 1982, eight days before the murder, Boyd
attempted to visit with Hartman in the front yard of her parents' resi-
dence, but Hartman's father, Lawrence Phillips, instructed Boyd "to
get off of [his] property and stay off of it." S.J.A. 102. Boyd then
threatened Hartman saying, "I'll see you like a German submarine,
when you are not expecting it." S.J.A. 103 (internal quotation marks
omitted). And, Boyd also told Phillips, "I'll meet you in heaven or
hell one day." 

marks omitted). Following this
encounter, Phillips sought a warrant for Boyd's arrest for trespassing,
and the warrant was served on Boyd on Monday, August 2.

On the morning of Saturday, August 7, following a night of drink-
ing and drug use, Boyd called Hartman at 8:00 a.m. and talked to her
for approximately two hours. During this conversation, Boyd learned
that Hartman planned to go to the Mayberry Mall in Mount Airy,
North Carolina to shop and attend a church-sponsored car wash. Boyd
then went to a bar and began drinking and using drugs again. At
approximately 12:00 noon, when the bartender declined to serve him
any more alcohol, Boyd hailed a taxi to take him to the mall.

Upon arriving at the mall, Boyd entered a store that sold knives and
asked the salesman for a lock-blade knife. The owner of the store tes-
tified that "[a] lock-blade knife is a knife that once it's opened it is
locked in an open position. It cannot come back against your hands
or fingers or cut you in any way. It's locked in." S.J.A. 9. Boyd pur-
chased the knife and left the store.

Boyd then saw Hartman and her mother, approached them, and
asked Hartman if she would go outside with him. Boyd and Hartman
sat together on a curb outside the mall in close proximity to the ongo-
ing car wash, apparently discussing again the possibility of a reconcil-
iation. After some period of time had passed, at approximately 2:00
p.m., Hartman's mother approached them and indicated that it was
time to leave. Hartman stood up, but Boyd attempted to prevent her
from leaving, asking her repeatedly to stay with him a few more min-
utes. Hartman responded to Boyd "that she had lived in hell for three
months, that if he was going to kill her just go ahead and kill her and
get it over with." S.J.A. 36.

                     3
Boyd brandished the knife he had just purchased, but offered Hart-
man assurances that he meant her no harm. Despite these assurances,
Boyd began to stab Hartman. As Boyd attacked, Hartman screamed
for help and her mother attempted to intervene, trying to pull Boyd
away from Hartman. Boyd, however, threw the 76-year-old woman
to the ground and resumed his onslaught on Hartman. Forcing Hart-
man to the ground on her stomach and holding her by her hair, Boyd
stabbed her repeatedly. Throughout the attack, numerous witnesses
looked on powerless to stop it, including Hartman's shrieking eight-
year-old daughter. After stabbing Hartman 37 times, Boyd calmly
walked away. He was apprehended quickly as he hid between two
parked vehicles; the murder weapon was recovered from where Boyd
had thrown it under a nearby automobile.

Emergency medical personnel were summoned and arrived on the
scene at approximately 2:20 p.m. These technicians characterized
Hartman's condition as requiring advanced life support treatment and
explained that they were unable to transport Hartman until they could
control her bleeding. They described the extreme difficulty Hartman
was having breathing and the severe pain she was experiencing,
recounting how Hartman moaned and "rak[ed her hands] back and
forth in the dirt" where she was lying. S.J.A. 165. The examining
pathologist later identified wounds to Hartman's throat, chest, left
thigh, and back. Among these were two wounds that punctured Hart-
man's right lung, three that pierced her left lung, one that entered her
stomach, and one that penetrated her sternum. Additionally, several
defensive wounds to Hartman's hands and left arm were present. Loss
of blood from these wounds led to hypovolemic shock, and Hartman
died of exsanguination while being transported to a hospital.

Boyd was charged with first-degree murder. In light of the numer-
ous witnesses to the murder, Boyd did not dispute that he had
inflicted the fatal wounds. However, Boyd presented the testimony of
two friends with whom he had been drinking on the morning of the
murder and of the bartender who had declined to serve him to support
his argument that he was intoxicated at the time of the murder. The
jury convicted Boyd of first-degree murder in violation of N.C. Gen.
Stat. § 14-17 (1993).

At sentencing, Boyd testified concerning his relationship with
Hartman, their break-up, and his attempts at reconciliation. Boyd also

                    4
professed his love for Hartman, saying, "[It was the m]ost beautiful
thing that's ever happened to me. It's the best thing that ever hap-
pened in my life. I loved her, more than anybody, I guess, could ever
love anybody." J.A. 583. Boyd related that when Hartman ended their
relationship, he began to seek mental health assistance because he
was having thoughts of killing people, including himself and Hart-
man. Boyd recounted his almost daily attempts to reunite with Hart-
man. Further, Boyd explained the difficulties he was experiencing in
sleeping and his heavy use of alcohol and illegal drugs.

Boyd also testified concerning various emotional losses he had
experienced as a child. Boyd's father deserted their family when
Boyd was very young, and his grandfather, with whom he was very
close, died when Boyd was five years old. Boyd's mother corrobo-
rated the losses of his father and grandfather.

Boyd then called Dr. Jack Humphrey, a professor of criminology
at the University of North Carolina.3 The State objected, and Dr.
Humphrey was examined outside the presence of the jury. Dr. Hum-
phrey testified about a study he had performed over a two-year period
in conjunction with the North Carolina Department of Corrections.
The study had two elements. First, researchers compared prison
records, social histories, and psychiatric histories of North Carolina
prisoners convicted of homicide with those convicted of property
offenses. He concluded that prisoners convicted of homicide had suf-
fered over the course of their lives more stressful events than nonvio-
lent offenders. The second aspect of the study dealt with whether
there was a difference between individuals who had killed strangers
and individuals who had killed family members or those close to
them. Dr. Humphrey concluded that individuals whose victims were
close to them tended to have experienced more loss in their lives than
those who had killed strangers:
_________________________________________________________________
3 Dr. Humphrey earned a Ph.D. in sociology with a concentration in
criminology from the University of New Hampshire. Employed by the
University of North Carolina since 1972, Dr. Humphrey taught classes in
criminology, criminal justice, juvenile delinquency, and deviant behav-
ior. He had conducted a number of studies and published extensively in
the areas of homicide and suicide.

                   5
          Now, one thing here is a loss has been found to be associ-
          ated with or precipitate or lead to suicide over and over and
          over. The more loss in someone's life, the more likely they
          are to become self-destructive. And it seems that killing a
          family member or killing a close friend is an act of self-
          destruction. They are, after all, killing something that is part
          of them, very close to them, very important to their self.
          They are destroying them. So in the act of killing another
          person they are in fact destroying part of [themselves, com-
          mitting] a self-destructive act.

J.A. 684-85. Dr. Humphrey then described the types of losses to
which he was referring--for example the loss of a parent or sibling.
Further, Dr. Humphrey testified that he had interviewed Boyd and
learned of the losses Boyd had experienced. Dr. Humphrey testified,
"And what struck me [was] the consistency of Mr. Boyd's life with
what we found to be true of homicide offenders in general." J.A. 687.
Dr. Humphrey continued:

          It seems that people who are threatened with loss, and
          mainly these are losses of someone very close to them, wife,
          girlfriend, some close relationship, at that point that they are
          threatened with this loss they become depressed, very com-
          monly depressed, and depression is in a sense anger turned
          toward yourself. Now, at that point people react either
          toward themselves totally or they will react outwardly and
          inwardly at the same time. Those people who destroy some-
          one or something at that point will not destroy a stranger,
          will not indiscriminately kill. They don't constitute a threat
          to the public. They constitute a threat to that which they fear
          losing the most, the person closest to them. And it is that
          person that is unfortunately in harm's way. And having
          extended that aggression toward other people they are in
          fact aggressing toward themselves. They are destroying that
          which they fear losing the most.

J.A. 688. Following voir dire, the State argued that Dr. Humphrey's
testimony should not be admitted, asserting that the study was not
"scientific" and that the testimony told the jury "[n]othing." J.A. 715.
The trial court sustained the objection. The jury sentenced Boyd to

                     6
death, finding two aggravating factors--that the murder was espe-
cially heinous, atrocious, or cruel and that Boyd previously had been
convicted of a felony of violence.

The North Carolina Supreme Court affirmed Boyd's conviction
and sentence, holding that the exclusion of Dr. Humphrey's testimony
was not error because the testimony was not mitigating. See State v.
Boyd, 

, 197-99 (N.C. 1984). The United States
Supreme Court denied certiorari on April 15, 1985. See Boyd v. North
Carolina, 

(1985).

Thereafter, Boyd sought postconviction relief from his convictions
and sentences in state court by filing a motion for appropriate relief
(MAR). See N.C. Gen. Stat. § 15A-1415 (1997). The state court con-
ducted an evidentiary hearing and denied relief. The North Carolina
Supreme Court subsequently denied certiorari.

In February 1989, Boyd filed a § 2254 petition in the district court.
This petition was held in abeyance pending a decision by the Supreme
Court in McKoy v. North Carolina, 

(1990), and during
Boyd's unsuccessful attempt to obtain postconviction relief under
McKoy in state court. In October 1996, a magistrate judge recom-
mended granting the State's motion for summary judgment as to all
claims. The district court adopted the magistrate judge's recommen-
dation and denied Boyd's application for a certificate of probable
cause to appeal.

Boyd now seeks review in this court of the decision of the district
court denying his petition for habeas corpus relief. 4 He raises five
_________________________________________________________________
4 Boyd's request for a certificate of probable cause to appeal is granted
because at least one judge on the panel concludes that Boyd "has made
a substantial showing of the denial of a constitutional right." 4th Cir. R.
22(a). There has been no argument concerning whether Boyd should be
granted a certificate of probable cause to appeal, as he requested, or a
certificate of appealability. And, we need not address that question here
because the certificate would be granted based on the conclusion that
Boyd made a substantial showing of the denial of a constitutional right
irrespective of which type of certificate technically should be issued
under these circumstances. Compare Lozada v. Deeds, 

,

                    7
allegations of error: (1) that the sentencing court deprived him of his
Eighth and Fourteenth Amendment rights to present mitigating evi-
dence by refusing to permit Dr. Humphrey to testify; (2) that a rea-
sonable likelihood exists that the instructions to the sentencing jurors
concerning their consideration of mitigating evidence led the jurors to
conclude that they could not consider a factor as mitigating unless the
jurors unanimously concluded that the factor was mitigating in viola-
tion of the Eighth and Fourteenth Amendments; (3) that the prosecu-
tor's closing argument in the sentencing phase was so inherently
flawed that it deprived Boyd of a fair trial in violation of the Due Pro-
cess Clause of the Fourteenth Amendment; (4) that the prosecution's
knowing use of perjured testimony violated his Fourteenth Amend-
ment right to due process; and (5) that the district court erred in con-
cluding that Boyd's argument relating to the use of his state nolo
contendere plea as a basis for a prior conviction was procedurally
defaulted. We address these arguments in turn.

II.

Boyd first asserts that the state trial court deprived him of his rights
under the Eighth and Fourteenth Amendments by refusing to permit
his expert witness, Dr. Humphrey, to present mitigating evidence dur-
ing sentencing. "`[T]he Eighth and Fourteenth Amendments require
that the sentencer ... not be precluded from considering, as a mitigat-
ing factor, any aspect of a defendant's character or record and any of
the circumstances of the offense that the defendant proffers as a basis
for a sentence less than death.'" Eddings v. Oklahoma, 

,
_________________________________________________________________
431-32 (1991) (per curiam) (explaining that to warrant the grant of a cer-
tificate of probable cause to appeal, a habeas petitioner must "make a
substantial showing of the denial of [a] federal right" and that to satisfy
this showing, the petitioner "must demonstrate that the issues are debat-
able among jurists of reason; that a court could resolve the issues [in a
different manner]; or that the questions are adequate to deserve encour-
agement to proceed further" (alterations in original) (internal quotation
marks omitted)), with Murphy v. Netherland, 

, 101 (4th Cir.)
(denying certificate of appealability under 28 U.S.C.A. § 2253 (West
Supp. 1998) in habeas corpus action seeking relief from death sentence
when petitioner failed to make a substantial showing of the denial of a
constitutional right), cert. denied, 

(1997).

                     8
110 (1982) (second alteration in original) (quoting Lockett v. Ohio,

, 604 (1978) (plurality opinion)). Such evidence includes
evidence of a defendant's troubled upbringing, see 

as well
as evidence bearing on whether the defendant will pose a danger in
the future, see Skipper v. South Carolina, 

, 5 (1986). See
also 

(noting "that the sentencer may not refuse to consider or
be precluded from considering any relevant mitigating evidence"
(internal quotation marks omitted)). The Due Process Clause of the
Fourteenth Amendment may require the admission of mitigating evi-
dence even if state-law rules of evidence (e.g. , hearsay) would
exclude it. See Green v. Georgia, 

, 97 (1979) (per
curiam). Similarly, this court has observed that"the Supreme Court
has been very sensitive to any impediment to the consideration of any
type of mitigating evidence in a death sentencing hearing" and that
"subject only to the loose evidentiary requirement of relevance, capi-
tal defendants have a right to offer any evidence they choose on char-
acter or record or circumstances of the offense." Hutchins v.
Garrison, 

, 1437 (4th Cir. 1983) (internal quotation
marks omitted); see Howard v. Moore, 

, 418 (4th Cir.
1997) (en banc) (recognizing that the Eighth Amendment requires
that all proffered relevant mitigating circumstances be presented to
the sentencer for consideration in determining whether to impose a
death sentence), petition for cert. filed, 66 U.S.L.W. ___ (U.S. May
22, 1998) (No. 97-9263); see also 

(explain-
ing that "[r]elevant mitigating evidence is evidence which tends logi-
cally to prove or disprove some fact or circumstance which a fact-
finder could reasonably deem to have mitigating value" (internal quo-
tation marks omitted)). The question of whether an evidentiary ruling
excluding testimony prevented the jury from considering mitigating
evidence is a mixed question of law and fact that this court reviews
de novo. See 

.

As discussed by the district court, Dr. Humphrey's proffered testi-
mony addressed two distinct potentially mitigating factors. First, Dr.
Humphrey explained that, based on his research, individuals in North
Carolina who had committed a homicide of someone close to them
had been subjected to more stressful life events in the form of losses
and that based on his interview of Boyd, Boyd fit the profile of these
individuals. Second, Dr. Humphrey opined that individuals who have
suffered significant losses become depressed to the point that they act

                    9
in a self-destructive manner, which may include the destruction of
that which they most fear losing.

The district court concluded that a portion of Dr. Humphrey's testi-
mony was not mitigating, reasoning:

          Dr. Humphrey's opinion testimony that Petitioner Boyd, as
          a result of losses in his life, fit the profile of a man more
          likely to kill a friend than to kill a stranger is simply not mit-
          igating. Standing alone, it is neutral on the question of future
          dangerousness, and it is also entirely without implication or
          inference that could affect a jury in forming a reasoned
          moral response to the question of whether Boyd should be
          given the death penalty.

J.A. 299 (internal quotation marks omitted). From this portion of Dr.
Humphrey's testimony, Boyd argues, a reasonable juror could con-
clude that he would not pose a future danger because unlike some
other first-degree murderers, he was not likely to kill at random and
those circumstances under which he might be dangerous would be
unlikely to reoccur in prison. Cf. 

(explaining
that "evidence that the defendant would not pose a danger if spared
(but incarcerated) must be considered potentially mitigating"). Addi-
tionally, he asserts that this portion of Dr. Humphrey's testimony pro-
vided the basis for a conclusion that Boyd fit within the category of
offenders who act in a self-destructive manner in taking the life of
someone close to them. Although we have serious questions concern-
ing whether this portion of Dr. Humphrey's proffered testimony accu-
rately may be characterized as mitigating,5 we agree with the
_________________________________________________________________
5 The State contends that this evidence could not be mitigating evi-
dence of Boyd's lack of future dangerousness because Dr. Humphrey
never testified that Boyd was not homicidal or that he would not kill
again. Rather, the State maintains that this portion of Dr. Humphrey's
testimony could at most support a conclusion that Boyd was dangerous
only to those who "established an intimate or family-type relationship
with him." Brief of Appellees at 24. The evidence, the State asserts, is
not mitigating and, on the contrary, is aggravating because it demon-
strates that Boyd is precisely the dangerous killer of those close to him
the State attempted to portray him as.

                     10
conclusion of the district court that the portion of Dr. Humphrey's tes-
timony concerning the self-destructive motivation of those who have
suffered great loss was mitigating because Boyd could have argued
that he acted out of a self-destructive impulse rather than the selfish
impulse advanced by the State.

While we conclude that the trial court committed constitutional
error in excluding relevant mitigating evidence, the question remains
whether that error was harmless. It is now well established that not
all errors of constitutional dimension warrant a federal court to over-
turn a state conviction or sentence. See Chapman v. California, 

, 23-24 (1967); Sherman v. Smith, 

, 1137 (4th
Cir. 1996) (en banc), cert. denied, 

(1997); Smith v.
Dixon, 

, 974-75 (4th Cir. 1994) (en banc). Although fed-
eral habeas courts play an important role in protecting the constitu-
tional rights of state criminal defendants, that role is circumscribed
and secondary to that of state courts. See Brecht v. Abrahamson, 

, 633 (1993). Once the principal avenue for review of a state
criminal conviction and sentence--direct review--has been com-
pleted, "`a presumption of finality and legality attaches to the convic-
tion and sentence.'" 

v. Estelle, 

,
887 (1983)). Respect for the finality of a presumptively valid state-
court conviction and sentence dictates that a federal court may not
grant habeas corpus relief on the basis of trial error of constitutional
dimension unless the court is convinced that "the error `had substan-
tial and injurious effect or influence in determining the jury's ver-
dict,'" 

(quoting Kotteakos v. United States, 

,
776 (1946)), or at a minimum entertains grave doubt that it had such
_________________________________________________________________
We need not address this argument because even if we were to agree
with the State that this portion of Dr. Humphrey's testimony--that
Boyd's history of personal loss typifies the profile of a killer who mur-
ders those who are emotionally closest to him when he fears losing them
--is not mitigating with respect to Boyd's future dangerousness, the tes-
timony nevertheless would have been admissible to provide the founda-
tion for Dr. Humphrey's opinion that killers who have experienced this
type of repeated personal loss may kill as a self-destructive act, which we
determine is mitigating. Thus, for purposes of this opinion, we assume
that both of these portions of Dr. Humphrey's testimony are mitigating.

                    11
an effect, see O'Neal v. McAninch, 

, 437 (1995) (holding
that when "the record is so evenly balanced that a conscientious judge
is in grave doubt as to the harmlessness of an error," the judge must
resolve that doubt in favor of the habeas petitioner).6

In applying this standard, a federal habeas court does not ask
whether the evidence of guilt was sufficient, whether the jury would
have reached the same conclusion if the error had not occurred, or
whether the jury reached the correct result based on the evidence pre-
sented. See Satcher v. Pruett, 

, 567-68 (4th Cir.), cert.
denied, 

(1997). Rather, the court reviews the record de
novo to determine whether the error "substantially sway[ed] or sub-
stantially influence[d] the response" of the jury to the question put to
it--i.e., in the guilt context, whether the defendant is guilty or not
guilty and in the penalty context, whether the defendant should
receive the death penalty. Cooper v. Taylor, 

, 370 (4th
Cir. 1996) (en banc), cert. denied, 

(1997); see 

(explaining that in making the harmlessness determi-
nation, a federal habeas judge must review the record to assess
whether "the judge[ ] think[s] that the error substantially influenced
the jury's decision" (internal quotation marks omitted)); 

(holding that an error does not have a substantial and inju-
rious effect on a jury verdict unless "it resulted in `actual prejudice'"
to the habeas petitioner (quoting United States v. Lane, 

,
449 (1986)).

Boyd contends that because neither his guilt nor the circumstances
surrounding the murder were subject to serious dispute, his trial strat-
egy was to show that the murder of his victim was the product of two
factors--his significant impairment as a result of drug and alcohol
consumption and his background of repeated losses of those closest
to him. The State, in turn, attempted to portray Boyd as a cold, selfish
_________________________________________________________________
6 The Brecht Court left open the possibility that under unusual circum-
stances "a deliberate and especially egregious error of the trial type, or
one that is combined with a pattern of prosecutorial misconduct, might
so infect the integrity of the proceeding as to warrant the grant of habeas
relief, even if it did not substantially influence the jury's verdict." 

n.9. This does not appear to be such a case.

                    12
man who, faced with imprisonment on an unrelated criminal charge,
killed Hartman to prevent her from seeing other men.

We agree with the district court that the refusal of the state trial
court to permit Dr. Humphrey to testify did not have a substantial or
injurious effect on the determination of the jury that Boyd should be
sentenced to death. Boyd's actions were indisputably premeditated.
He had threatened Hartman in the weeks prior to the murder and had
purchased a lock-blade knife just prior to his attack. Immediately
before the murder, Boyd talked to Hartman calmly, assuring her that
he would not hurt her. His sudden attack on Hartman was a brutal and
hideous one in which he inflicted a total of 37 wounds while her
family--including her young daughter--and friends watched in horri-
fied helplessness. Hartman suffered a physically agonizing death. Fur-
thermore, the underlying circumstances upon which Boyd was to
argue that this was an act of self-destruction were before the jury.
Boyd testified about the loss of his father and grandfather and about
his love for Hartman.7 We cannot conclude that, viewed in this con-
text, any juror's sentencing decision would have been substantially
influenced by hearing an expert criminologist opine that murderers
who have experienced great personal losses are more likely to kill a
family member or someone close to them than a stranger and are
more likely to murder as an act of self-destruction and that Boyd's
history of loss fit the pattern of someone in this category.
Dr. Humphrey's testimony simply was not adequate to have had such
an effect in the circumstances of this case. Thus, we hold that any
error in refusing to permit Dr. Humphrey to testify does not provide
a basis for federal habeas corpus relief.

III.

Boyd next challenges the instructions provided to the jury concern-
ing the use of mitigating evidence. Jury instructions that require jurors
to find unanimously the existence of a mitigating factor before that
_________________________________________________________________
7 We recognize that the testimony of Boyd and his mother may not sub-
stitute for Dr. Humphrey's expert testimony concerning these issues.
However, we note that the state trial court did not compound the error
by excluding the underlying mitigating factual evidence or prohibiting
Boyd from arguing that the underlying facts were mitigating.

                    13
factor may be weighed in determining whether mitigating evidence
outweighs aggravating factors are unconstitutional under the Eighth
and Fourteenth Amendments. See 

; Mills
v. Maryland, 

, 374-75 (1988). Although Boyd concedes
that the trial judge did not give the jurors an explicit direction that
they could not consider mitigating evidence unless they found its
existence unanimously, he contends that read as a whole, there is a
reasonable likelihood that the jury understood the instructions to have
required such unanimity.

The instructions that Boyd challenges are identical to those
recently held not to be violative of the Constitution in Noland v.
French, 

, 213-14 (4th Cir. 1998). We view our decision
in Noland as dictating a conclusion that Boyd's challenge to the jury
instructions lacks merit.

IV.

Boyd also maintains that the prosecutor's closing argument during
the sentencing phase of his trial deprived him of due process. He con-
tends that during the closing argument in the sentencing phase of the
trial, the prosecutor made repeated references to his personal opinions
concerning various matters, including Boyd's credibility; the credibil-
ity of Boyd's witnesses; the weight to be given various mitigating fac-
tors; certain biblical quotations and references; and the
appropriateness of the death penalty for Boyd, including a reading of
a North Carolina Supreme Court case suggesting that mercy was not
appropriate in death cases, and referring to a later-repudiated system
of mandatory capital punishment.

In determining whether a closing argument by a prosecutor violates
due process, this court must look to "whether the proceeding at issue
was rendered fundamentally unfair by the improper argument."
Bennett v. Angelone, 

, 1345 (4th Cir.) (internal quotation
marks omitted), cert. denied, 

(1996). This determina-
tion requires the court to look to "the nature of the comments, the
nature and quantum of the evidence before the jury, the arguments of
opposing counsel, the judge's charge, and whether the errors were
isolated or repeated." 

(internal quotation marks omit-
ted).

                    14
Undoubtedly, all of the arguments of which Boyd complains were
improper. A prosecutor should refrain from stating his personal opin-
ions during argument and misleading the jury about the law. See
Drake v. Kemp, 

, 1459-60 (11th Cir. 1985) (en banc).
Furthermore, religiously based arguments are "universally con-
demned." 

. The remaining factors, however,
weigh in favor of a conclusion that the prosecutor's argument did not
deprive Boyd of a fair trial. The evidence that Boyd committed the
offense was overwhelming. Further, the murder unquestionably was
heinous, atrocious, or cruel, and Boyd had entered a stipulation that
he had committed a prior felony of violence. In addition, although the
improper remarks occurred intermittently throughout the prosecu-
tion's argument, some of the biblical references were invited by
Boyd's testimony concerning his salvation experience while in prison
awaiting trial and Boyd's explanation of the murder as having
resulted from his being beguiled by Satan. Cf. United States v. Young,

, 12-13 (1985) (explaining that in determining whether
prosecutor's improper argument was prejudicial to defendant, review-
ing court must consider whether prosecutor's comments were invited
response to defense and "did no more than respond substantially in
order to right the scale" (internal quotation marks omitted)). Addition-
ally, the state trial judge instructed the jurors that they were to decide
the facts based on the evidence presented. Cf. 

-47 (concluding prosecutor's improper argument did not deny
due process in part because trial court instructed jury: "What the law-
yers say is not evidence. You heard the evidence. You decide what
the evidence is." (internal quotation marks omitted)). Our review
leads us to determine that the prosecutor's closing argument did not
deprive Boyd of a fair trial.

V.

Boyd further asserts that his conviction resulted from the prosecu-
tion's knowing use of perjured testimony. A conviction acquired
through the knowing use of perjured testimony by the prosecution
violates due process. See Napue v. Illinois, 

, 269 (1959).
This is true regardless of whether the prosecution solicited testimony
it knew to be false or simply allowed such testimony to pass uncor-
rected. See Giglio v. United States, 

, 153 (1972); 

. And, knowingly false or misleading testimony by a

                     15
law enforcement officer is imputed to the prosecution. See Wedra v.
Thomas, 

, 717 n.1 (2d Cir. 1982); Curran v. Delaware,

, 712-13 (3d Cir. 1958) (citing Pyle v. Kansas, 

(1942)); cf. Boone v. Paderick, 

, 450-51 (4th Cir.
1976) (recognizing that withholding of exculpatory evidence by
police is imputed to the prosecution). But see Koch v. Puckett, 

, 530-31 (5th Cir. 1990) (rejecting habeas petitioner's claim
that sheriff and investigators testified falsely at trial on the basis that
petitioner had failed to show that the prosecutor knew the testimony
was perjurious). As this court has explained:

          The police are also part of the prosecution, and the taint on
          the trial is no less if they, rather than the State's Attorney,
          were guilty of the nondisclosure. If police allow the State's
          Attorney to produce evidence pointing to guilt without
          informing him of other evidence in their possession which
          contradicts this inference, state officers are practicing decep-
          tion not only on the State's Attorney but on the court and
          the defendant.

Barbee v. Warden, Md. Penitentiary, 

, 846 (4th Cir.
1964) (footnote omitted). The knowing use of perjured testimony con-
stitutes a due process violation when "`there is any reasonable likeli-
hood that the false testimony could have affected the judgment of the
jury.'" Kyles v. Whitley, 

, 433 n.7 (1995) (quoting United
States v. Agurs, 

, 103 (1976)); see United States v. Ellis,

, 915 n.5 (4th Cir. 1997), cert. denied, 

(1998); United States v. Kelly, 

, 933 (4th Cir. 1994).

During Boyd's trial, each of the State's witnesses who testified
regarding Boyd's condition either immediately before or just after the
murder indicated that Boyd was not intoxicated. For example, the taxi
driver who drove Boyd to the mall stated that Boyd did not appear to
be intoxicated. The two salesmen in the store where Boyd purchased
the knife immediately before the murder testified that Boyd did not
appear to have been drinking or under the influence of anything. Hart-
man's father and a family friend both testified that they saw Boyd just
prior to the murder and that he did not appear to be intoxicated. Fol-
lowing this evidence, the State presented the testimony of officers
who observed Boyd just after the murder. Officer Sumner gave an

                     16
opinion that Boyd was not under the influence. Agent Perry stated his
opinion that Boyd did not appear to be drunk or intoxicated. Detective
Armstrong, the chief investigating officer, was asked by defense
counsel, "[B]ased upon your observation of the defendant out there on
the occasion you have described, in your opinion was he drunk or
intoxicated?" J.A. 410. Detective Armstrong responded, "He didn't
appear to me to be, no, sir." 

state MAR hearing, Boyd's attorney had the following col-
loquy with Detective Armstrong:

         Q. ... [T]hinking back on the day that you saw Mr. Boyd,
         do you have an opinion, yourself, on that date based on the
         limited time you had to observe him as to whether or not he
         was subject to some impairing substance at that time?

         A. I felt like he was at the time, yes.

         Q. What [made you think so?]

         A. Well, I had seen him sober. And I had seen him drunk
         on many occasions over the years.

         Q. What observation about him did you make that day that
         caused you to think he was subject to some impairing sub-
         stance?

         A. I just felt like he was effected [sic] to some degree, that
         he was under the influence. I was a right good distance from
         him. But I've been at this same distance from him in the
         past when he was drinking. And just the way, the way he
         called my name and said that he was, what had we arrested
         him for, being under the influence.

         Q. To what extent do you think that he was impaired? Do
         you have a word that you can describe the extent of his
         impairment with?

         A. It would be appreciably.

                    17
          Q. Appreciably means to you noticeably or clearly?

          A. To me. It may have not been noticeable to someone else
          that didn't know him. But to me he had been, he was under
          the influence of something.

J.A. 883-84. When confronted during cross-examination by the State
with the inconsistency of his testimony, Detective Armstrong admit-
ted that he had not remembered his prior testimony. When asked by
the State whether he told the truth when he testified at trial that Boyd
was not intoxicated, Detective Armstrong answered affirmatively.
And, Detective Armstrong agreed with the State that his trial testi-
mony was closer to the murder, and he was a law enforcement officer
at the time.

At the state MAR proceeding, Officer Perry was asked, "From the
observations that you were able to make of [Boyd] that afternoon do
you have an opinion as to whether he was under the influence of or
intoxicated from any kind of drug or alcohol?" J.A. 914. He
responded:

          In my opinion he was not under the influence. He did
          appear to have either been drinking or maybe have taken
          some drugs. He was somewhat glassy-eyed. But he was
          walking, he was not swaying or staggering. He, in my opin-
          ion, was not under the influence of--to an appreciable
          degree.

MAR court determined that the State did not withhold
exculpatory evidence from Boyd and that even if the evidence from
the habeas hearing had been presented at trial it would not have
affected the outcome of the proceedings. However, the state habeas
court failed to render an express factual finding with respect to
whether the officers knowingly presented misleading testimony.
Because the state MAR court failed to make a factual finding on the
question of whether the law enforcement officers knowingly pre-
sented false or misleading testimony, a federal evidentiary hearing to

                     18
resolve this issue normally would be required. See Townsend v. Sain,

, 312-13 (1963). Such a hearing is not necessary in this
instance, however, because we conclude that there is no reasonable
likelihood that the officers' testimony, if false, could have affected the
judgment of the jury. The jury heard a wealth of testimony concerning
the amount of alcohol and drugs that Boyd had ingested in the hours
prior to the murder; undoubtedly the jury recognized that Boyd must
have been impaired to some degree. However, the testimony of the
lay witnesses and police officers established that despite the alcohol
and drugs, Boyd's demeanor prior to and immediately after the mur-
der was calm and controlled. As such, testimony by the officers that
Boyd was under the influence would not have affected the verdict of
the jury.

VI.

Finally, Boyd contends that the district court erred in holding fed-
eral habeas review of his argument concerning the use of his nolo
contendere plea to establish a prior violent felony was barred because
Boyd procedurally defaulted the claim. Absent cause and prejudice or
a miscarriage of justice, a federal habeas court may not review consti-
tutional claims when a state court has declined to consider their merits
on the basis of an adequate and independent state procedural rule. See
Harris v. Reed, 

, 262 (1989). Such a rule is adequate if
it is regularly or consistently applied by the state court, see Johnson
v. Mississippi, 

, 587 (1988), and is independent if it does
not "depend[ ] on a federal constitutional ruling," Ake v. Oklahoma,

, 75 (1985).

Boyd seeks to argue that his prior nolo contendere plea to a 1963
violent felony--assault with intent to commit rape--did not constitute
a prior conviction for a felony involving the use of violence against
the person of another within the meaning of N.C. Gen. Stat.
§ 15A-2000(e)(3) (1997). He contends that a conviction qualifies
under § 15A-2000(e)(3) only if it was treated as a conviction under
state law at the time and that North Carolina law prior to 1981 did not
permit such treatment for pleas of nolo contendere.

Boyd recognizes that he did not raise this claim at trial--indeed,
counsel stipulated that Boyd had a prior conviction within the mean-

                     19
ing of § 15A-2000(e)(3)--or on direct appeal. Further, the attorney
who represented Boyd in his first state MAR proceeding failed to
review personally trial counsel's files, in which the information con-
cerning Boyd's prior conviction was contained, or to investigate the
basis for the prior conviction. Because counsel did not discover the
claim, it was not raised in Boyd's first MAR. However, attorneys for
Boyd ultimately discovered this information and returned to state
court to exhaust the claim. The state court that heard Boyd's second
MAR found that "Boyd presented no evidence to show he was pre-
vented by some objective factor external to the defense from raising
the claim." J.A. 1036. Therefore, the state habeas court ruled that the
claim was procedurally defaulted under N.C. Gen. Stat.
§ 15A-1419(a)(1) (1997). The North Carolina Supreme Court summa-
rily denied Boyd's petition for certiorari. See Ylst v. Nunnemaker, 

, 805-06 (1991) (holding that in applying procedural default
provisions an unexplained appellate state-court decision is presumed
to be based on the last reasoned decision).

Boyd acknowledges that the state court expressly relied on an inde-
pendent state procedural ground to refuse to consider the merits of
this claim, but argues that the state procedural rule is not "adequate"
because it is not regularly or consistently applied. 8 This court has con-
sistently held, however, that § 15A-1419 is an adequate and indepen-
dent state-law ground for decision foreclosing federal habeas review.
See Williams v. French, No. 97-19, 

, at *3 (4th Cir.
May 18, 1998); Ashe v. Styles, 

, 87-88 (4th Cir. 1994)
(explaining that a federal habeas petition should have been denied on
the basis of procedural default because the state court denied relief
pursuant to § 15A-1419(a) which is "an adequate and independent
state law ground of decision"); see also O'Dell v. Netherland, 

, 1241 (4th Cir. 1996) (en banc) (holding that unambiguous pro-
cedural rules derived from state statutes or court rules are necessarily
"firmly established" (internal quotation marks omitted)), aff'd, 

(1997); 

& n.10 (concluding that
_________________________________________________________________
8 Boyd also asserts that "cause" exists to excuse the default because the
attorney who represented him during his first MAR was constitutionally
ineffective in failing to raise this issue. This argument lacks merit. See
Mackall v. Angelone, 

, 446-49 (4th Cir. 1997) (en banc),
cert. denied, 

(1998).

                    20
§ 15A-1419 is an adequate and independent state-law ground for deci-
sion).

VII.

In sum, we conclude that the refusal of the state trial court to per-
mit Boyd to present the mitigating testimony of his expert witness,
Dr. Humphrey, did not have a substantial or injurious effect on the
verdict. Similarly, we are convinced that the verdict of the jury would
not have changed if it had heard the police officers testify that Boyd
was impaired on the day of the murder. And, Boyd's remaining argu-
ments lack merit. Accordingly, we affirm the judgment of the district
court.

AFFIRMED

MURNAGHAN, Circuit Judge, concurring:

I concur in the result the majority has reached but, with respect to
Part II, I find necessary and sufficient only the initial paragraph and
the conclusion that Dr. Humphrey's testimony was not erroneously
excluded since it was not mitigating:

          [W]e have serious questions concerning whether this portion
          of Dr. Humphrey's proffered testimony accurately may be
          characterized as mitigating . . . we conclude that the refusal
          of the state trial court to permit Boyd to present the mitigat-
          ing testimony of his expert witness, Dr. Humphrey, did not
          have a substantial or injurious effect on the verdict.

The proffered testimony of Dr. Humphrey was that"prisoners con-
victed of homicide had suffered over the course of their lives more
stressful events than nonviolent offenders" and that "individuals
whose victims were close to them tended to have experienced more
loss in their lives than those who had killed strangers." Even if that
proffered testimony was mitigating, it was harmless error to exclude
it.

                     21