FOR PUBLICATION
UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
SHIRLEY REE SMITH, No. 04-55831
Petitioner-Appellant,
v. D.C. No.
CV-01-04484-ABC
GWENDOLYN MITCHELL, Warden,
OPINION
Respondent-Appellee.
Appeal from the United States District Court
for the Central District of California
Audrey B. Collins, District Judge, Presiding
Argued and Submitted
February 8, 2005—Pasadena, California
Filed February 9, 2006
Before: Harry Pregerson and William C. Canby, Jr.,
Circuit Judges, and Edward C. Reed, Jr.,* District Judge.
Opinion by Judge Canby
*The Honorable Edward C. Reed, Jr., Senior United States District
Judge for the District of Nevada, sitting by designation.
1535
SMITH v. MITCHELL 1537
COUNSEL
Michael J. Brennan, Manhattan Beach, California, for the
petitioner-appellant.
Richard T. Breen, Deputy Attorney General, Los Angeles,
California, for the respondent-appellee.
OPINION
CANBY, Circuit Judge:
Shirley Ree Smith appeals the district court’s denial of her
habeas corpus petition. The State of California convicted
Smith of assault on a child resulting in death. The state courts
affirmed her conviction. Smith then filed this federal habeas
petition claiming that her conviction violated due process
1538 SMITH v. MITCHELL
because the evidence was constitutionally insufficient. On
appeal, Smith focuses her argument almost exclusively on the
absence of constitutionally sufficient evidence of one element
of the crime—the cause of the child’s death. We agree with
Smith that no rational trier of fact could have found beyond
a reasonable doubt that Smith caused the child’s death. We
further conclude that the state court’s affirmance of Smith’s
conviction constituted an unreasonable application of Jackson
v. Virginia, 443 U.S. 307, 319 (1979), which established the
standard for constitutional sufficiency of the evidence. See 28
U.S.C. § 2254(d)(1). We accordingly reverse and remand with
instructions to grant the writ.
I
Smith is the grandmother of the deceased infant, Etzel.
Smith’s daughter Tomeka Smith is the mother of the child.
Tomeka also had two older children, Yondale and Yolanda.
Prior to Etzel’s birth, the entire family lived in Illinois. In July
1996, Smith moved to California and brought her two grand-
children, Yondale and Yolanda, with her. On October 8, 1996,
Tomeka moved to California, where she joined Smith, Yon-
dale, and Yolanda. Two days later, Tomeka gave birth to
Etzel.
Tomeka delivered him two weeks early, and he weighed
five pounds, four ounces. He was born with jaundice and a
slight heart murmur. The murmur disappeared, however, three
days after it was diagnosed. The jaundice gradually disap-
peared, too. Etzel became a healthy baby and showed no signs
of child abuse.
Smith, Tomeka, Yondale, Yolanda and Etzel usually stayed
with Stephen Keys, Smith’s brother. Occasionally, however,
they stayed with Renee Townsend, Smith’s sister, at Town-
send’s apartment. Townsend’s two children, Marcus and Mar-
cellus lived there as well. When Tomeka and her children
stayed at Townsend’s apartment, Smith would accompany
SMITH v. MITCHELL 1539
them to assist Tomeka in taking care of the children. Smith
was described as always giving loving care to her grandchil-
dren, and no one had ever seen her act harshly or abusively
toward them.
On the night of Etzel’s death, the group stayed at Town-
send’s. Tomeka testified at length about the events of that
evening, from her own observations and from what Smith told
her as the emergency unfolded and thereafter.1 The facts as
Tomeka related them were not complicated. Etzel, who was
approximately seven weeks old at the time, appeared perfectly
healthy during the day and at the beginning of the evening.
Both Tomeka and Smith fed him.
That night, Tomeka fed, changed, and washed Etzel before
rocking him to sleep and laying him on the couch in the living
room, placing him on his stomach, with his face to the side.
Yondale also slept on the couch. Yolanda slept on the love seat.2
Marcus and Marcellus slept in their bedroom.3 Smith slept on
the floor, next to the love seat on which Yolanda slept.
Townsend left the apartment at roughly the same time that
Tomeka placed Etzel on the couch. Tomeka remained in the
living room for another hour or so. During this time, she
checked Etzel’s diaper while he slept, and she saw that he
moved his body at that time. She also moved Etzel back on
the couch pillow because his feet started to fall off of it. She
then went into Townsend’s room to listen to music.
Though Tomeka usually slept in the living room with
1
Smith herself testified very briefly, denying that she had shaken Etzel
on the night of his death. She was subjected to almost no cross-
examination, and her entire testimony extends for less than three pages of
transcript.
2
Yondale was approximately fourteen-months old, and Yolanda was
approximately four-years old.
3
They were seven- and ten-years old, respectively.
1540 SMITH v. MITCHELL
Smith and the children, that night she fell asleep in Town-
send’s room while listening to music. She left the door to
Townsend’s room open, at least partially. Tomeka testified
that she fell asleep around midnight. She also stated that it
was the first night that she slept in a different room from
Etzel. At approximately 1:30 a.m., Smith awoke and found
Etzel on the floor. She picked him up, rocked him back to
sleep, and placed him on the couch in the same position
(stomach down, head to the side). She did not notice anything
unusual about him.
At 3:20 a.m., Smith awoke again because she had to go to
the bathroom. After she returned from the bathroom, she
looked at Etzel and saw that he had thrown up and had blood
on his right nostril. He did not respond to her touch. She
picked him up and his head “flopped back.” She moved him
back and forth, but he did not respond.4 She then went into
Townsend’s room carrying Etzel. She woke Tomeka and told
her what had happened. Tomeka dialed 911. Over the phone,
Tomeka and Smith were instructed to give Etzel CPR, which
they did.
When firefighters and paramedics arrived, Smith was “ap-
prehensive” and stated that she thought Etzel had fallen off
the couch. Etzel was clothed and warm, but he was not
breathing and had no heartbeat. The paramedics began CPR.
Three of the rescue squad noticed blood in one of Etzel’s nos-
trils, and one consequently thought Etzel had suffered an
injury. When an ambulance arrived, two more technicians
administered CPR on the way to the hospital. Etzel appeared
“chalky.” They arrived at the hospital at 3:50 a.m. Etzel was
4
Smith demonstrated this “jostle” to a social worker, who testified at
trial. The social worker stated that Smith demonstrated picking up the
baby under his arms and moving him quickly forward and back in a
smooth motion. Smith later told police that she had given him a little
shake, but then she corrected herself, saying she had “twisted” him slowly
from side to side.
SMITH v. MITCHELL 1541
in full cardiac arrest. The attending physician pronounced him
dead and suspected he died of Sudden Infant Death Syndrome
(“SIDS”), a death with no known cause.
It was the theory of the prosecution, however, that this reci-
tation of facts left out one crucial occurrence: Smith must
have shaken Etzel so violently that it caused his death. The
evidence offered in support of this theory was the autopsy and
the controverted expert testimony, based on that autopsy, that
Etzel died of Shaken Baby Syndrome.5 The physical evidence,
however, was not typical of that usually associated with
Shaken Baby Syndrome.
Associate Deputy Medical Examiner Dr. Stephanie Erlich
performed the autopsy on Etzel.6 She testified, as did her
supervisor, Dr. Eugene Carpenter, who participated in parts of
the autopsy.7 They found recent subdural and subarachnoid
hemorrhages (i.e., bleeding on the brain). There was also evi-
dence of old subdural bleeding, and both old and new bleed-
ing around the optic nerves. There was in addition a recent
small abrasion, approximately 1/16 by 3/16 of an inch, on the
lower skull, upper neck region, and a recent bruise beneath
this abrasion. Etzel’s heart was normal.
All of the expert testimony offered by both sides agreed
that the amount of recent bleeding (approximately one or two
tablespoons) was not sufficient to have caused death, nor was
the small abrasion sufficient for that purpose. To the prosecu-
tion experts, however, the presence of blood supported the
5
Shaken Baby Syndrome is also known as Shaken Infant Syndrome or
SIS. We use the Shaken Baby terminology to avoid confusion with Sud-
den Infant Death Syndrome or SIDS.
6
This was Dr. Erlich’s first autopsy in a child abuse case. She was
assigned the autopsy because Etzel was originally believed to be the vic-
tim of Sudden Infant Death Syndrome.
7
Dr. Carpenter was a Medical Examiner and forensic pathologist for the
Los Angeles County Coroner. He had performed 3,000 to 4,000 autopsies.
1542 SMITH v. MITCHELL
diagnosis of Shaken Baby Syndrome.8 There was no dispute,
however, that the usual Shaken Baby Syndrome death occurs
from massive bleeding or swelling of brain tissue that creates
such crushing pressure against the brain stem that vital pro-
cesses are interrupted and the baby dies. It was also agreed
that in 80% or more of the cases of Shaken Baby Syndrome,
there is bleeding in the retinas of the eyes. There are also fre-
quently fractures in the arms or similar evidences of violence.
There was no swelling, and only a small, non-fatal amount
of bleeding, in Etzel’s case. Etzel had no retinal bleeding, and
no fractures or large bodily bruises common in cases of shak-
ing. The scalp abrasion was minimal, and was not even dis-
covered until well into the autopsy.
The prosecution experts testified, however, that shaking
caused the death even though the physical examination of the
brain during and after autopsy could not demonstrate that fact.
The experts testified that the shaking must have been so vio-
lent and severe that it directly tore or sheared parts of the
brain stem, causing immediate cessation of vital activity such
as breathing. This tear in the brain stem would not have been
apparent in autopsy, according to the prosecution experts,
because instantaneous death would have prevented any bleed-
ing or swelling. No microscopic examination of the brainstem
was performed following the autopsy because, as Dr. Erlich
testified, “[W]e wouldn’t have seen anything anyway.” The
fatal tear or shearing would not have been detectable. Dr. Ehr-
lich could not identify any source in the literature for her
hypothesis of undetectable brain stem shearing, but said she
had learned it from lectures and consultations.
With regard to this undetectable cause of death, defense
8
In addition to Drs. Ehrlich and Carpenter, Dr. David Chadwick, a pedi-
atrician specializing in abuse, also testified for the prosecution. His testi-
mony was briefer than, but generally paralleled, that of Drs. Ehrlich and
Carpenter.
SMITH v. MITCHELL 1543
expert Dr. Richard Siegler, said that the hypothesis was “fan-
tasy.” When pressed by the prosecutor to elaborate, he said:
“[W]hat you have said is possible, but it is also not possible
and that’s what we call fantasy. . . . There is no way to con-
firm it or deny it.” Dr. Siegler stated that he had not previ-
ously heard or read of the hypothesis of undetectable brain
stem injury in a Shaken Baby death.9 Dr. Seigler opined that
Etzel died from the lingering effects of earlier brain trauma of
unknown but quite possibly innocent cause, and that his death
was inconsistent with Shaken Baby Syndrome.
The defense also presented Dr. William Goldie, who exam-
ined the records and testified that Etzel likely died of SIDS.
Dr. Goldie, like Dr. Siegler, did not believe that Shaken Baby
Syndrome could occur without massive brain swelling or
bleeding, at least not when the brain stem appeared undam-
aged, as Etzel’s did. He noted that prematurely born infants
sometimes bleed into the head without cause. Dr. Goldie testi-
fied that SIDS was the leading cause of deaths in infants from
one to five months of age, while Shaken Baby Syndrome
deaths occur predominantly in ages four to nine months.
Dr. Goldie described some of the characteristics that led
him to conclude that Etzel died of Sudden Infant Death Syn-
drome. With SIDS, the infant usually would appear normal,
but then he or she suddenly would die. SIDS occurred more
frequently in babies who, like Etzel, were small for their age,
who had mothers who had multiple children already or
smoked or used drugs, and, most importantly, who had been
placed face-down on their stomachs. Males were more likely
victims than females. He also concluded that Etzel’s problems
—jaundice, heart murmur, and low birth weight—made him
more likely to die from SIDS, as did a background of poverty.
9
Dr. Siegler was a pathologist of many years experience associated with
the University of Southern California, and formerly with Harvard and
U.C.L.A. He had done 5,000 to 8,000 autopsies, about 50 of which were
infants.
1544 SMITH v. MITCHELL
II
On this evidence, a jury of the Superior Court of Los Ange-
les County convicted Smith of assault on a child causing
death, in violation of section 273ab of the California Penal
Code.10 The trial judge sentenced her to fifteen years to life in
prison. Smith appealed, challenging the sufficiency of the evi-
dence. The California Court of Appeal affirmed in an unpub-
lished opinion, and the Supreme Court of California denied
review.
Smith filed a petition for habeas corpus in the Central Dis-
trict of California. A magistrate judge recommended dis-
missal, and the district court dismissed the petition. The court,
however, granted a certificate of appealability on the question
“[w]hether there was sufficient evidence to convict [Smith] of
committing assault on a child in her care or custody causing
death.”
III
We review de novo the district court’s denial of a habeas
petition. See Clark v. Murphy, 331 F.3d 1062, 1067 (9th Cir.
2003). We approach this case with a firm awareness of the
very strict limits that the Antiterrorism and Effective Death
Penalty Act (“AEDPA”) places on our collateral review of
state criminal convictions. See Juan H. v. Allen, 408 F.3d
1262, 1269-70 (9th Cir. 2005), cert. denied, 2006 WL 88993,
88994 (U.S. Jan. 17, 2006). We cannot grant relief unless the
decision of the California Court of Appeal was “contrary to,
10
Section 273ab provides the following:
Any person who, having the care or custody of a child who is
under eight years of age, assaults the child by means of force that
to a reasonable person would be likely to produce great bodily
injury, resulting in the child’s death, shall be punished by impris-
onment in the state prison for [fifteen] years to life.
CAL. PENAL CODE § 273ab.
SMITH v. MITCHELL 1545
or involved an unreasonable application of, clearly established
Federal law, as determined by the Supreme Court of the
United States” or “was based on an unreasonable determina-
tion of the facts in light of the evidence presented in the State
court proceeding.” 28 U.S.C. § 2254(d). The constitutional
standard for sufficiency of the evidence established by the
Supreme Court is “whether, after viewing the evidence in the
light most favorable to the prosecution, any rational trier of
fact could have found the essential elements of the crime
beyond a reasonable doubt.” Jackson v. Virginia, 443 U.S. at
319. Our task under AEDPA, then, is to determine whether
the decision of the California Court of Appeal, holding that
the evidence was sufficient to convict Smith, was an unrea-
sonable application of Jackson. See 28 U.S.C. § 2254(d); see
also Juan H. v. Allen, 408 F.3d at 1274-75 & nn. 12-13
(establishing that AEDPA requires a layer of deference to a
state court decision in addition to that specified by Jackson).
In this most unusual case, we conclude that the Court of
Appeal unreasonably applied Jackson.
IV
[1] The troubling state of the evidence in this case was per-
haps best described by Magistrate Judge Patrick J. Walsh.
Even though, contrary to our view, he believed the strict
requirements of AEDPA for habeas relief had not been met,
Judge Walsh stated:
This is not the typical shaken baby case. Grandmoth-
ers, especially those not serving as the primary care-
takers, are not the typical perpetrators. Further,
Petitioner was helping her daughter raise her other
children (a 2-year-old and a 14-month-old) and there
was no hint of Petitioner abusing or neglecting these
other children, who were in the room with Etzel
when he died. Still further, there was no evidence of
any precipitating event that might have caused Peti-
tioner to snap and assault her grandson. She was not
1546 SMITH v. MITCHELL
trapped in a hopeless situation with a child she did
not want or love. Nor was she forced to single-
handedly care for a baby that had been crying all day
and all night. In fact, there is no evidence that Etzel
was doing anything other than sleeping the night that
he died. The medical evidence was not typical either,
in that some of the telltale signs usually found in
shaken baby cases did not exist in this case.
Nothing significant in the background suggests guilt, there-
fore, and many factors suggest innocence.11 Indeed, not only
was there no evidence of any “precipitating event that might
have caused [Smith] to snap,” but it is extremely unlikely that
even a very troublesome act by seven-week-old Etzel would
cause Smith to shake Etzel to death when his mother lay but
a few feet away and easily available. A constitutionally per-
missible finding of guilt in this case therefore depends on the
expert evidence of the cause of death.
[2] There is no question that the prosecution experts testi-
fied that a shaking had caused the death, but they conceded
the absence of the usual indicators of violent shaking such as
bruises on the body, fractured arms or ribs, or retinal bleed-
ing. There was bleeding on the brain, both old and new, but
not enough to cause death. All of the prosecution witnesses
based their opinion of Shaken Baby Syndrome on their
hypothesis that violent shaking had torn or sheared the brain
stem in an undetectable way. Their testimony was not that the
brain demonstrated death in the usual manner of Shaken Baby
Syndrome, caused by excessive bleeding or swelling that
crushes the brain stem. Instead, their testimony was that death
was caused by shearing or tearing of the brain stem and they
11
A social worker, who visited Smith and Tomeka and expressed the
view that Etzel had died of shaking, testified that Smith said “Oh, my God.
Did I do it?” A rational trier of fact, however, could not convict on this
distraught and equivocal statement in the absence of credible expert testi-
mony as to the cause of death.
SMITH v. MITCHELL 1547
reached this conclusion because there was no evidence in the
brain itself of the cause of death. Thus, as the defense expert
Dr. Siegler stated, the tearing might have occurred or it might
not have occurred; there simply was no evidence to permit an
expert conclusion one way or the other on the point. This is
simply not the stuff from which guilt beyond a reasonable
doubt can be established, especially in the face of all the other
circumstances, many of which were recited by the magistrate
judge, making the crime unlikely. An expert’s testimony as to
a theoretical conclusion or inference does not rescue a case
that suffers from an underlying insufficiency of evidence to
convict beyond a reasonable doubt. See United States v. Bois-
soneault, 926 F.2d 230, 234 (2d Cir. 1991).
V
[3] The prosecution’s expert testimony, absolutely critical
to its case, concluded that the cause of death was tearing or
shearing of the brain stem when there was no physical evi-
dence of such tearing or shearing, and no other evidence sup-
porting death by violent shaking. Absence of evidence cannot
constitute proof beyond a reasonable doubt.
[4] The California Court of Appeal, in affirming Smith’s
conviction, described the expert testimony but did not address
the problem presented by the lack of evidence, from the
autopsy or ensuing tests, that the brain stem had been sheared.
It affirmed with a statement that the experts conflicted and
resolution of the conflict was for the jury. With all due respect
to the California Court of Appeal, and even with the addi-
tional layer of deference mandated by AEDPA, we conclude
that the Court of Appeal unreasonably applied Jackson when
it held the evidence to be sufficient to convict Smith of caus-
ing Etzel’s death. There was simply no demonstrable support
for shaking as the cause of death. As a result of the unreason-
able application of Jackson, there has very likely been a mis-
carriage of justice in this case.
1548 SMITH v. MITCHELL
We reverse the judgment of the district court and remand
with instructions to grant the writ.
REVERSED; REMANDED with instructions.