Landry v. State

2016 UT App 164

THE UTAH COURT OF APPEALS

HERBERT LANDRY,
Appellant,
v.
STATE OF UTAH,
Appellee.

Opinion
No. 20140638-CA
Filed July 29, 2016

Fourth District Court, Provo Department
The Honorable Steven L. Hansen
No. 100402172

Cory A. Talbot and M. Benjamin Machlis, Attorneys
for Appellant
Sean D. Reyes and Erin Riley, Attorneys for Appellee

JUDGE GREGORY K. ORME authored this Opinion, in which JUDGE
MICHELE M. CHRISTIANSEN and SENIOR JUDGE PAMELA T.
GREENWOOD concurred. 1

ORME, Judge:

¶1 Herbert Landry appeals the district court’s order
dismissing his petition for postconviction relief from his
conviction for one count of aggravated arson. Landry’s theory is
one of layered ineffective assistance of counsel. He alleges that
he received ineffective assistance from his appellate counsel in
his direct appeal when appellate counsel failed to raise the issue
of ineffective assistance rendered by his trial counsel. Landry

1. Senior Judge Pamela T. Greenwood sat by special assignment
as authorized by law. See generally Utah R. Jud. Admin. 11-
201(6).
Landry v. State

asserts that trial counsel rendered ineffective assistance by
failing to object to the State’s use of testimony about an
accelerant-detection canine’s alerts at the scene of the fire and on
items of his clothing and by failing to consult with an arson
expert or to call an expert witness to refute the State’s arson
experts. We conclude that such failings, taken together,
constitute prejudicial ineffective assistance. Thus, we reverse the
district court’s contrary disposition and remand for a new trial
or the fashioning of such relief as is appropriate under the
circumstances. See Utah Code Ann. § 78B-9-108(1)(b) (LexisNexis
2012).

BACKGROUND

¶2 The salient facts were developed in an evidentiary
hearing on Landry’s postconviction petition, and we take the
district court’s findings of fact as our starting point. We also rely
on our prior decisions in Landry’s unsuccessful direct appeal,
State v. Landry, 2008 UT App 461U, and his successful appeal
from the district court’s denial of an evidentiary hearing on his
postconviction petition, Landry v. State, 2012 UT App 350, 293
P.3d 1092, as well as the record from Landry’s original trial.

¶3 Landry moved into an apartment complex in Provo, Utah,
in September 2005. From September 2005 until he was evicted
from his apartment in February 2006, he split his time between
Provo and Texas, where his family lived. Landry arranged to
vacate his apartment on February 26, 2006, and to conduct a final
walk-through inspection with the apartment complex manager
at 9:00 p.m. that evening.

¶4 On the day of the scheduled inspection, Landry and his
girlfriend were in his bedroom, packing. While there, his
girlfriend sat on the edge of the bed and smoked a cigarette. In
the afternoon, the two left, apparently intending to return. Some
of Landry’s belongings remained behind, including clothes,

20140638-CA 2 2016 UT App 164
Landry v. State

some small furniture items, a mattress, floor coverings, and a
suitcase that contained clothes, prescription medications, and
financial documents. Neighbors saw the pair leave in a hurry.
Some five to ten minutes later, around 4:30 p.m., the same
neighbors observed smoke coming from Landry’s apartment and
called 911 to report a fire. Around this same time, another
neighbor observed an unidentified man outside Landry’s
apartment. Fire investigators were unable to locate that
individual, but they did find a discarded lighter nearby.

¶5 During their investigation of the fire, police used a
certified accelerant-detection canine named Oscar. Oscar alerted
his handler to possible ignitable substances on the floorboards in
three places in Landry’s bedroom, and investigators determined
the bedroom to be the point of the fire’s origin. Investigators
then sent samples from each of those three locations to the crime
lab. The crime lab concluded that heptane was the only ignitable
substance that could be present on the floorboard samples.
According to the district court’s findings, heptane is an
ingredient in “a common construction adhesive that is often
used to glue carpet down and to glue the components of a sub-
floor,” and at trial the State conceded that heptane was not the
cause of the fire. Oscar was also taken to the motel room where
Landry had moved many of his belongings. Oscar alerted to one
of Landry’s socks and one of his shoes. The crime lab
subsequently tested Landry’s clothing, including the shoe and
sock, and found no ignitable substances.

¶6 In addition to Oscar’s alerts, investigators identified a
distinct v-pattern on the floor of Landry’s bedroom, which
indicated, as one investigator testified at trial, that this was “an
intentionally set fire from the ignition of ignitable poured
liquid.” Police interviewed Landry, who did not attempt to
explain the cause of the fire or express concern about the
possessions left in his bedroom. Landry agreed during the
interview that the fire was likely not due to accidental electrical
or chemical causes. He denied smelling smoke or other odors

20140638-CA 3 2016 UT App 164
Landry v. State

consistent with fire at the time he left the apartment with his
girlfriend. He also said there were no tobacco products or other
smoking materials in his bedroom. Following its investigation,
the State charged Landry with aggravated arson, a first degree
felony. See Utah Code Ann. § 76-6-103(2) (LexisNexis 2012).

¶7 During the ensuing two-day jury trial, the State’s case
included testimony regarding Oscar’s training and the meaning
of his alerts as well as expert testimony regarding the cause and
origin of the fire. First, Oscar’s handler, himself a fire
investigator, testified that Oscar was trained to “find[] ignitable
liquids, the typical type of liquid that you’d find in . . . a
hardware store, for instance, kerosene, Coleman fuel, lighter
fluid, charcoal lighter gasoline is the most common type of
accelerant[] that he’s trained on, kerosene.” But Oscar was not
trained to alert to alcohol. On direct examination, when asked if
“the fact that Oscar [alerted] on an odor necessarily means that
there’s an ignitable liquid there,” the handler replied that
“[t]here are times when he’s alerted and it comes back not
identifiable hydrocarbon.” 2 But on cross-examination the
handler testified that when the lab determines a sample contains
no identifiable hydrocarbon, “[i]t doesn’t necessarily mean there
is nothing there, there may . . . have been but they couldn’t

2. “In the field of arson and fire investigations, combustible
vapors or vapors resulting from accelerants and ignitable liquid
residues are often referred to as ‘Hydrocarbons.’ These
hydrocarbons are a by-product of the original petroleum based
material that may have been used as the original ignitable liquid
residue used to start the incendiary fire (e.g., gasoline,
kerosene).” Matthew D. Baldwin, Practical Applications of
Hydrocarbon and Photoionization Detection Units in Arson
Investigations, Crime Scene Investigator Network (Feb. 25, 2015),
http://www.crime-scene-investigator.net/practical-applications-
of-hydrocarbon-and-photoionization-detection-units-in-arson-
investigations.html [https://perma.cc/3RWQ-LW7K].

20140638-CA 4 2016 UT App 164
Landry v. State

identify what it was.” And according to the district court on
remand, Landry’s trial counsel “elicited testimony” from the
handler “that a general rule of thumb is that the error rate is
probably in the range of 10%.” During his trial testimony, the
Provo City fire marshal stated, “[T]he dog is much more
sensitive than some of the scientific equipment that we use.” The
district court later found that “[f]ar from accepting the Crime
Lab’s findings, the State used them instead to bolster the
evidence regarding Oscar’s alerts,” pointing to the Provo fire
marshal’s testimony in particular.

¶8 At trial, the State also put on expert testimony from the
Orem City and Provo fire marshals regarding the origin and
cause of the fire. The Orem fire marshal testified that the fire
“beg[a]n at the floor level” and that the burn marks created a v-
pattern, which he explained occurs when combustion materials
separate from the air around them and which directed the
investigators to what they believed was the point of origin. He
testified that “the patterns on the floor were indicative of a
poured or flowing liquid that had been on the floor and
ignited[,] causing increased damage in the specific areas on the
floor consistent with ignitable poured liquid.” Despite conceding
that he found no accelerants or ignitable liquids in Landry’s
bedroom, the Orem fire marshal persisted in testifying that “this
was an intentionally set fire that occurred from the ignition of
ignitable poured liquid.” He also refuted a defense theory
advanced by trial counsel—that a lit cigarette that inadvertently
came into contact with spilled alcohol could have caused the
fire—by pointing to Landry’s denial that smoking materials
were in his bedroom and his failure to mention spilled alcohol
during his initial interview with investigators and by saying that
he believed there would have been an explosion that was visible
to witnesses if that had been the cause. The Provo fire marshal
also testified that the patterns on a piece of subfloor introduced
into evidence resulted from “an ignitable liquid that was poured
or otherwise distributed” on Landry’s bedroom floor.

20140638-CA 5 2016 UT App 164
Landry v. State

¶9 Landry’s trial counsel called no witnesses to refute the
State’s experts’ testimony regarding the investigation methods
and the cause and origin of the fire. Trial counsel instead
defended Landry by suggesting that Landry did not set the
fire—at least not intentionally. In support of the theory that the
fire was an accident, Landry testified that he hosted a party in
his apartment the night before the fire and that alcohol was
spilled in the general area where the experts testified that the fire
had started. This was the first time that Landry mentioned
anything about a party or spilled alcohol. He further testified,
contradicting his pretrial statements to investigators, that his
girlfriend was smoking on the bed shortly before the fire started.
Again, this was the first time Landry had mentioned this fact.
And when asked on cross-examination for specific details about
the party and his girlfriend, Landry could provide none—not
even his girlfriend’s last name. But apparently hoping to plant
the seed of reasonable doubt, trial counsel suggested in her
closing argument that the cigarette could have fallen into the
alcohol or onto the bed and started the fire.

¶10 In support of the defense theory that whatever the cause
of the fire, Landry was not it, trial counsel also called a neighbor
who testified that around four o’clock on the afternoon of the fire
he heard a knock, apparently on Landry’s door, and then saw a
man who was “[k]ind of tall, narrow, had glasses, goatee, shaggy
beard” and wearing “dark clothes, dark hat,” and whom the
neighbor previously had seen “around.” Ten minutes later the
neighbor saw that cardboard used as Landry’s temporary front
window had been pushed in. And at 4:30, he saw the fire. Police
never located the unidentified man. Hoping to create a
reasonable doubt in the mind of the jury, trial counsel suggested
in closing argument that the unidentified man may well have
started the fire.

¶11 And at least some doubt apparently existed in the jurors’
minds. After closing arguments, the jury deliberated for several
hours and was at one point deadlocked. State v. Landry, 2008 UT

20140638-CA 6 2016 UT App 164
Landry v. State

App 461U, para. 3 n.1. After the jury informed the trial court of
the deadlock, the court explained that the jury would have to
return and continue deliberations the following day. When
faced, however, with the prospect of having to return the next
day to continue their deliberations—potentially extending the
trial, which was initially scheduled for two days, to an
unexpected third day and causing at least one juror a hardship—
the jurors asked for the opportunity to reconvene briefly that
night. Only then did they reach a consensus and return a guilty
verdict.

¶12 Landry appealed his conviction, and on direct appeal,
appellate counsel argued “that there was insufficient evidence to
convict [Landry] of the crime charged.” Id. para. 1. This court
affirmed. Id. Landry then filed the instant petition for relief
under the Post-Conviction Remedies Act (PCRA), Utah Code
Ann. §§ 78B-9-101 to -405 (LexisNexis 2012), raising a number of
theories, including ineffective assistance of appellate counsel. See
Landry v. State, 2012 UT App 350, ¶¶ 1, 4–5, 293 P.3d 1092.
Landry’s postconviction petition asserted that his appellate
counsel was ineffective for her failure to raise trial counsel’s
ineffectiveness as an issue on direct appeal. Id. ¶ 1. The State
moved to dismiss the postconviction petition, and the district
court granted the State’s motion. Id. ¶ 5. On appeal, “[w]e
reverse[d] and remand[ed] for an evidentiary hearing . . . on the
limited issue of whether appellate counsel was ineffective when
[she] did not raise on direct appeal the ineffective assistance of
trial counsel.” Id. ¶ 1. We affirmed as to Landry’s other
postconviction issues. Id.

¶13 The district court held the required evidentiary hearing.
Landry presented testimony from trial and appellate counsel,
and both Landry and the State presented expert testimony
assessing the expert testimony from his trial as well as
addressing the likely cause of the fire.

20140638-CA 7 2016 UT App 164
Landry v. State

¶14 Trial counsel testified in a deposition, the recording of
which was played during the evidentiary hearing, that although
she had “been to CLEs that touched on arson,” she did not recall
participating in any other arson cases and did not “have any
personal training in arson.” Despite her inexperience, trial
counsel consulted only the Provo fire marshal, who was the
State’s expert and who would later testify on behalf of the State.
Moreover, she admitted that she “didn’t rebut their expert
testimony at trial” or present any evidence supporting a
nonarson theory. She acknowledged that her defense strategy
rested on the possibility that a lit cigarette could have fallen into
spilled alcohol and started the fire, or alternatively, that an
unidentified man may have started the fire. Trial counsel also
stated that it did not occur to her to challenge the testimony
regarding Oscar’s alerts as improper expert testimony, or to its
“coming in at all.”

¶15 Landry elicited testimony that appellate counsel, like trial
counsel, did not consider it problematic that trial counsel relied
exclusively on consultation with the State’s expert witness and
did not consult an independent arson expert. Furthermore,
appellate counsel testified that she did not review State v.
Schultz, 2002 UT App 366, ¶¶ 26–29, 39, 58 P.3d 879 (limiting the
admissibility of accelerant detection canine alerts), as a part of
her preparation, even though the State cited the decision in its
brief on appeal. She did, however, testify that if she had
reviewed Schultz, she would have argued plain error, ineffective
assistance of trial counsel, and that the State’s introduction of
Oscar’s alerts as substantive evidence was improper under
Schultz. 3

3. In its conclusions of law, the district court disregarded
counsel’s testimony that she overlooked State v. Schultz, 2002 UT
App 366, 58 P.3d 879, because the court concluded that Oscar’s
alerts were not substantive evidence of the presence of
accelerants. But when Oscar’s handler testified to the presence of
(continued…)

20140638-CA 8 2016 UT App 164
Landry v. State

¶16 Also during the evidentiary hearing on remand, Landry
and the State relied on arson experts, who prepared reports and
testified. The State’s expert said the fire was properly classified
as arson, while Landry’s expert concluded that the fire could not
definitely be classified as arson and instead should have been
classified as of “undetermined origin.” The experts focused
much of their testimony on whether the fire investigators
properly conducted the investigation. Based on their testimony,
the district court found that it was unclear whether the
investigators used the fire-investigation guide NFPA 921 4 in
investigating this particular fire, which guide both experts relied

(…continued)
accelerant on Landry’s sock and shoe even though the laboratory
had not corroborated those alerts, he was citing Oscar as the
source of the evidence without satisfying the test for expert
testimony—a conclusion that the State does not refute. We
conclude that the handler’s testimony regarding Oscar’s alerts
was substantive evidence—all the more so given the Provo fire
marshal’s contention that Oscar was more “sensitive” than some
scientific equipment and the district court’s characterization of
the crime lab results as being used “to bolster the evidence
regarding Oscar’s alerts.” See supra ¶ 7.

4. The NFPA 921 is a guide for “scientific-based investigation
and analysis of fire and explosion incidents . . . [and] the
foremost guide for rendering accurate opinions as to incident
origin, cause, responsibility, and prevention.” NFPA 921: Guide
for Fire and Explosion Investigations, National Fire Protection
Association, http://www.nfpa.org/codes-and-standards/all-codes-
and-standards/list-of-codes-and-standards?mode=code&code=921
[https://perma.cc/FNR5-DZWQ]. It is used in the field and by
courts. Id. The experts who testified in the evidentiary hearing
on remand both agreed NFPA 921 is a leading guide for arson
investigators.

20140638-CA 9 2016 UT App 164
Landry v. State

upon as authoritative in their own reports. The experts also
agreed, the court found, that “the [S]tate’s original fire
investigators made a mistake when they concluded that fire
patterns on the floor definitely indicated a pour pattern for
ignitable liquids.”

¶17 The State’s expert believed the v-pattern could have been
from poured ignitable fluid, but he indicated that conclusion
could not be substantiated because the fire was characterized by
“full room involvement.” The district court found that Landry’s
expert was similarly unwilling to rule out the possibility that the
fire started because of ignitable liquids, but he believed any such
liquids were on the mattress, not the floor. Landry’s expert
testified that the mattress was the “first fuel ignited in this fire”
and pointed to the common role cigarettes play in starting home
fires as a potential cause.

¶18 The district court further found that the experts disagreed
about whether the original investigators properly conducted
their investigation. According to the court’s findings, Landry’s
expert first said the investigators failed to identify the
arrangement of the bedroom, did not consider “target and
secondary fuels,” did not “identif[y] the volume or fire load,”
and “failed to identify [the] type or heat release rate” of each of
the fuels present in the room. The State’s expert disagreed with
each of these conclusions. Regarding the heat release rate, the
State’s expert acknowledged that there was no documented
proof that the State’s trial experts considered the heat release
rates, but he opined that they still could have considered them.
Landry’s expert also said the investigation was flawed because
the investigators failed to conduct a full electrical arc survey,
while the State’s expert was not troubled by that decision
because such a survey is not the only way to identify a fire’s
point of origin. Landry’s expert further criticized the
investigation because the investigators did not conduct a
complete reconstruction of the bedroom, but the State’s expert
said this was unnecessary as the original investigators left the

20140638-CA 10 2016 UT App 164
Landry v. State

items in the room and relied on photographs and Landry’s
characterizations to visualize an accurate layout. Finally,
Landry’s expert faulted the original investigators for not
excluding other possible ignition sources, while the State’s
expert said that they listed all such sources and evaluated the
likelihood that each could cause a fire in the relevant time frame.

¶19 The postconviction experts agreed that Cognac, which
Landry testified to having had at the previous night’s party, is
flammable and that “alcohol is miscible in water.”5 But they both
also testified that they would have said “no” if they had been
called at trial and asked if alcohol that had been spilled the
previous day could have been the fuel for the fire. And both
stated that it was unlikely that a cigarette on a bare mattress
could have started the fire.

¶20 The court found that the experts disagreed about what
arson investigators should consider in classifying a fire. Landry’s
expert did not read the trial transcript or Landry’s statements to
the police. He also asserted that even if a person stated the intent
to burn a house, that statement would not be one of the factors
that fire investigators should consider in deciding whether to
classify a fire as arson. The State’s expert, on the other hand, said
arson investigators should consider circumstantial, nonfire
related details of an investigation, such as witness statements, in
classifying a fire. And the court found that the State’s
postconviction expert did consider such evidence.

¶21 Ultimately, the district court on remand dismissed
Landry’s claim with prejudice. Landry appeals that disposition.

5. The district court’s findings state that miscible means that “[i]f
alcohol were used to ignite a fire, and then firemen shot a lot of
water on the scene, they would not expect lab testing to be able
to determine that alcohol was used.”

20140638-CA 11 2016 UT App 164
Landry v. State

ISSUE AND STANDARD OF REVIEW

¶22 This case presents one issue: whether Landry’s appellate
counsel was ineffective for failing to bring a claim of ineffective
assistance of trial counsel. Specifically, Landry contends that trial
counsel was ineffective for failing to object to the admission of
an accelerant detection canine’s alerts and for failing to put on a
nonarson defense through an independent arson expert who
could testify on Landry’s behalf or, at minimum, for failing to
consult with an arson expert so as to be better prepared to
highlight the weaknesses in the State’s case. In reviewing a
ruling on a petition for postconviction relief, we review the
district court’s findings of fact for clear error and its conclusions
of law for correctness. Tillman v. State, 2005 UT 56, ¶ 14, 128 P.3d
1123.

ANALYSIS

¶23 Parties seeking postconviction relief under the PCRA will
prevail if they show that they received ineffective assistance of
counsel during the conviction, sentencing, or appellate phase of
their criminal case. See Utah Code Ann. § 78B-9-104(1)(d)
(LexisNexis 2012). Postconviction relief is warranted if
“counsel’s performance was deficient” and “the deficient
performance prejudiced the defense.” See Strickland v.
Washington, 466 U.S. 668, 687 (1984). See also Kell v. State, 2008 UT
62, ¶¶ 27–28, 194 P.3d 913 (employing the Strickland standard
when reviewing a petition for postconviction relief that raised an
ineffective assistance of counsel claim). 6 And we use the same

6. The Strickland prejudice requirement is the same standard a
petitioner must demonstrate to obtain postconviction relief. See
Utah Code Ann. § 78B-9-104(2) (LexisNexis 2012) (“The court
may not grant relief from a conviction or sentence unless the
petitioner establishes that there would be a reasonable likelihood of
(continued…)

20140638-CA 12 2016 UT App 164
Landry v. State

standard when a petitioner bases a claim of ineffective assistance
of appellate counsel on appellate counsel’s failure to raise a
claim of ineffective assistance of trial counsel. Kell, 2008 UT 62,
¶ 42. The petitioner “has the burden of meeting both parts of”
the ineffective assistance of counsel test. State v. Templin, 805
P.2d 182, 186 (Utah 1990).

¶24 Because Landry’s claim is that appellate counsel was
ineffective only for failing to raise trial counsel’s ineffectiveness,
whether appellate counsel provided ineffective assistance
depends on whether trial counsel was ineffective. We therefore
first assess trial counsel’s performance according to Strickland,
after which we address that of appellate counsel.

I. Trial Counsel

A. Trial Counsel’s Performance Was Deficient.

¶25 To prove that counsel’s performance was deficient, a
claimant “must show that counsel’s representation fell below an
objective standard of reasonableness.” Strickland, 466 U.S. at 687–
88. Reasonableness is evaluated “under prevailing professional
norms.” Id. at 688. And while we “indulge a strong presumption
that counsel’s conduct falls within the wide range of reasonable
professional assistance,” a claimant who shows that under the
circumstances there is no way that counsel’s actions “might be
considered sound trial strategy” will overcome that

(…continued)
a more favorable outcome in light of the facts proved in the post-
conviction proceeding, viewed with the evidence and facts
introduced at trial or during sentencing.”) (emphasis added).
Because we conclude that both trial counsel’s and appellate
counsel’s errors prejudiced Landry under Strickland, see infra
¶¶ 39–43, 45, we conclude that the PCRA standard is also
satisfied.

20140638-CA 13 2016 UT App 164
Landry v. State

presumption. Id. at 689 (citation and internal quotation marks
omitted).

¶26 Landry argues that trial counsel’s performance was
deficient for two reasons: (1) she “fail[ed] to object to the
admissibility of an accelerant sniffing dog’s alerts as substantive
evidence of the presence of an ignitable liquid at the crime scene
and on Landry’s clothing when the alerts were uncorroborated
by laboratory testing” and (2) she “fail[ed] to consult with an
arson expert or call an arson expert to testify at trial” even
though she “lacked experience or training regarding arson and
the State’s case relied primarily on expert testimony regarding
arson.”

¶27 First, we agree that trial counsel’s failure to object to the
admission of testimony about Oscar’s alerts at the motel fell
below the objective standard of reasonableness. 7 It is objectively
unreasonable for counsel to forgo a valid objection to the
admissibility of incriminating evidence, when that evidence
provides no benefit to the defendant. State v. Doutre, 2014 UT
App 192, ¶ 21, 335 P.3d 366. See id. ¶ 24 (“If clearly inadmissible
evidence has no conceivable benefit to a defendant, the failure to
object to it on nonfrivolous grounds cannot ordinarily be
considered a reasonable trial strategy.”).

7. Landry also argues that counsel should have objected to
testimony regarding Oscar’s alerts in the apartment. Because
those alerts were corroborated by laboratory analysis, at least
with respect to heptane, see State v. Schultz, 2002 UT App 366,
¶ 39, 58 P.3d 879 (permitting “testimony about canine detection
. . . offered as substantive proof of the presence of an accelerant”
if it is confirmed by a laboratory), and because we do not have to
reach the issue given our disposition, we decline to address trial
counsel’s effectiveness with regard to Oscar’s alerts in the
apartment.

20140638-CA 14 2016 UT App 164
Landry v. State

¶28 Here there was an obvious, nonfrivolous ground upon
which trial counsel should have objected to the evidence of
Oscar’s alerts to the shoe and sock—evidence that not only had
no benefit to Landry but which was also rather incriminating. In
State v. Schultz, 2002 UT App 366, 58 P.3d 879, we concluded
that, to be admissible, testimony about canine alerts on
accelerants must either be corroborated by laboratory analysis or
satisfy the test for admissibility of expert testimony. Id. ¶ 39. In
Schultz, the prosecution’s witnesses testified about a dog’s alerts
at the crime scene even though laboratory results for those
samples were negative for accelerants. Id. ¶¶ 15–16. Although
the Schultz court ultimately held that the trial court’s admission
of the dog’s alerts was harmless because of the “significant
amount of other evidence to support [the defendant’s]
conviction,” id. ¶ 40, it first concluded that “if testimony about
canine detection is offered as substantive proof of the presence of
an accelerant, without laboratory confirmation, [a] three-part
analysis . . . is required because such evidence is based on novel
scientific principles and methods,” id. ¶ 39. Specifically, where
the State seeks to prove the presence of accelerants by testimony
that an accelerant detection canine alerted without also
providing corroborating laboratory results, the State must prove
that (1) the alert is “inherently reliable,” (2) the techniques used
were “properly applied to the facts of the particular case by
sufficiently qualified experts,” and (3) the evidence of the alert is
“more probative than prejudicial.” Id. ¶¶ 27–29 (citations and
internal quotation marks omitted). If the State does not satisfy
these three elements, defense counsel should object to the
admission of testimony about the alert on the authority of
Schultz, unless there is some reason the evidence would benefit
the defendant. See Doutre, 2014 UT App 192, ¶ 24.

¶29 During Landry’s trial, the State called witnesses who
testified to Oscar’s alerts and who asserted that Oscar was more

20140638-CA 15 2016 UT App 164
Landry v. State

sensitive than laboratory testing. 8 The district court found that
the State used the alerts as evidence of arson. Yet the State
provided no evidence that accelerant detection canine alerts
generally, or Oscar’s alerts on Landry’s sock and shoe
specifically, met the admissibility test for expert testimony
outlined in Schultz. See 2002 UT App 366, ¶¶ 27–29. And Oscar’s
handler’s testimony was, as noted above, quite incriminating.
Not only did Oscar’s alerts on the sock and shoe suggest to the
jury that an accelerant was used, but it also tied the accelerant to
Landry’s person, undermining the defense’s theory that perhaps
someone other than Landry was responsible for causing the fire.

¶30 Yet trial counsel did not object. Instead, as spun by the
State in its brief, she “dealt with the dog-alert evidence in cross-
examination, showing its limitations both generally and in this
case.” On remand, the district court found that through cross-
examination, trial counsel “elicited testimony that a general rule
of thumb is that the error rate”—i.e., false alerts—of accelerant
detection canines “is probably in the range of 10%.” And in the
evidentiary hearing on remand, the district court found that the
reason Landry’s counsel did nothing more to defend against the
dog-alert testimony was that “she really didn’t think about” it.
Nonetheless, the State suggests her cross-examination was
sufficient to constitute objectively reasonable representation. But
while cross-examination can be an effective trial strategy, “we
cannot conceive of a sound basis for failing to object” under
Schultz “when it should have been obvious” that the State could
not support the assertion that Oscar was a better source than
laboratory testing for reliably discerning the presence of

8. The olfactory abilities of canines might in some respects be
more sensitive than scientific equipment, as the Provo fire
marshal testified, but whatever the sensitivity of a dog’s nose,
evidence of an accelerant detection canine’s alerts has not yet
been accepted as admissible proof of arson without independent
corroborating laboratory testing. Id. ¶ 26.

20140638-CA 16 2016 UT App 164
Landry v. State

accelerants in an evidentiarily significant way, see Doutre, 2014
UT App 192, ¶ 21, considering that Schultz expressly states that
this is not the settled view of the scientific community, see 2002
UT App 366, ¶ 26 (“[I]f testimony concerning canine detection is
used as substantive proof that an accelerant was used in a fire,
without laboratory confirmation, such evidence is based on
novel scientific principles or techniques.”).

¶31 No reason has been suggested for why trial counsel
would not object. And she could have objected without
contradicting or interfering with her alternative defense that
whatever the cause, Landry did not start the fire, because the
two defenses “offered alternative grounds for reasonable
doubt.” See Dugas v. Coplan, 428 F.3d 317, 331 (1st Cir. 2005) (“‘A
tactical decision to pursue one defense does not excuse failure to
present another defense that would bolster rather than detract
from the primary defense.’”) (brackets omitted) (quoting Foster v.
Lockhart, 9 F.3d 722, 726 (8th Cir. 1993)). Indeed, the only
explanation we can fathom for trial counsel’s failure to object to
this harmful evidence is that “perhaps because of the lack of
adequate preparation to meet [the State’s] expert testimony, trial
counsel failed to grasp the problems with the testimony and to
object to it.” See Doutre, 2014 UT App 192, ¶ 24. Her testimony
during the evidentiary hearing on remand is not inconsistent
with our surmise. Thus, we conclude that it was objectively
unreasonable for trial counsel not to object under Schultz to the
testimony of Oscar’s uncorroborated alerts to the sock and shoe
found in Landry’s motel room.9

9. The State argues that under State v. Clark, 2014 UT App 56, 322
P.3d 761, trial counsel has no obligation to raise futile objections
so it was not unreasonable for trial counsel not to object on the
authority of Schultz. See id. ¶ 31. We do not agree that a Schultz
objection to the handler’s testimony about Oscar’s alerts to the
sock and shoe and to the Provo fire marshal’s statement that “the
dog is much more sensitive than some of the scientific
(continued…)

20140638-CA 17 2016 UT App 164
Landry v. State

¶32 Second, we agree with Landry that trial counsel’s failure
to engage an expert of her own, at least to advise and possibly to
testify, was also objectively unreasonable. Although we are
generally reluctant to question trial strategy, including whether
to call an expert witness, where “there is no reasonable basis for
that decision,” we will conclude there was deficient performance
by trial counsel. See State v. Tyler, 850 P.2d 1250, 1256 (Utah
1993). See also Houskeeper v. State, 2008 UT 78, ¶¶ 39–40, 197 P.3d
636 (determining that counsel’s performance was deficient
because counsel failed to investigate or provide any expert or

(…continued)
equipment that we use”—implying that the alerts were credible
despite the lack of corroborating laboratory analysis—would
have been futile. Had trial counsel objected under the authority
of Schultz, she likely would have prevailed in excluding the
evidence because it was the State’s burden, as the proponent of
the evidence, to prove that the evidence of Oscar’s alerts was
admissible as substantive evidence. See Schultz, 2002 UT App
366, ¶ 27 (stating that to satisfy the first element “a proponent
must” provide evidence of the alerts’ foundation and reliability).
Further, Landry has shown that the State likely could not have
proven Oscar’s alerts were inherently reliable because Schultz
itself stands for the principle that accelerant detection canine
alerts are not generally accepted in the scientific community. See
id. ¶ 26. See also Yell v. Commonwealth, 242 S.W.3d 331, 345 (Ky.
2007) (Scott, J., concurring and dissenting) (cataloguing cases
from across the United States, including Schultz, that determined
“uncorroborated canine accelerant alerts . . . to be novel scientific
evidence, not generally accepted in the scientific community of
arson investigators”). And the State has not shown how it might
have “‘proffer[ed] a sufficient foundation to demonstrate the
inherent reliability’” of accelerant-detection dog alerts generally
or Oscar’s alerts in particular. See Schultz, 2002 UT App 366, ¶ 27
(quoting State v. Brown, 948 P.2d 337, 340 (Utah 1997)).

20140638-CA 18 2016 UT App 164
Landry v. State

character witnesses while the State presented at least four
medical experts). The specific facts of a case may require trial
counsel to investigate potential witnesses to determine whether
such testimony would be appropriate. See State v. Templin, 805
P.2d 182, 187–88 (Utah 1990) (concluding that trial counsel’s
performance was deficient when counsel did not adequately
investigate potential witnesses identified by the defendant
because “a decision not to investigate cannot be considered a
tactical decision. It is only after an adequate inquiry has been
made that counsel can make a reasonable decision to call or not
to call particular witnesses for tactical reasons.”). See also Dees v.
Caspiri, 904 F.2d 452, 454–55 (8th Cir. 1990) (per curiam) (stating
that when expert evidence is critical to the case, “counsel ha[s] a
duty to make a diligent investigation of the forensic evidence
and its potential weaknesses” and “garner the expertise
necessary to cross examine” the expert).

¶33 Dugas v. Coplan, 428 F.3d 317 (1st Cir. 2005)—a case we
referenced in the appeal from the initial dismissal of Landry’s
petition, see Landry v. State, 2012 UT App 350, ¶ 8, 293 P.3d 1092,
and which the State makes no attempt to distinguish—
demonstrates why an expert is needed in cases like Landry’s. In
our previous decision, we noted that Dugas supported at least
further proceedings in this case. In Dugas,

the defendant was convicted of arson for setting
fire to [a] grocery store . . . . At trial, the
prosecution called fire experts, who testified that
the fire was intentionally set based on the lack of
flame or charring damage expected from an
electrical fire and the burn patterns on a pile of
papers that indicated the pile had been tightly
packed. Additionally, a canine trained to detect fire
accelerants “alerted” to the pile of papers, some of
which subsequently tested positive for ignitable
liquids. The defendant’s conviction was upheld
on appeal, and he . . . [subsequently claimed]

20140638-CA 19 2016 UT App 164
Landry v. State

ineffective assistance of counsel for inadequately
pursuing a non-arson defense, particularly for not
consulting an expert. . . . On review, the First
Circuit . . . determined that counsel’s failure to
fully investigate a non-arson defense was deficient
and prejudicial in this case for several reasons. For
one, “challenging the state’s arson case was critical
to [the] defense” because the only other available
defense—another suspect—was “difficult to mount
and fraught with evidentiary problems.” In
addition, the state’s case depended on arson
evidence whose weaknesses would have been
revealed by a thorough, expert-assisted defense
investigation. The First Circuit also found it
significant that although this was defense counsel’s
first arson case and he lacked knowledge of fire
investigation principles, he decided to accept the
state’s characterization of the scene rather than
seek guidance from an outside expert. Counsel’s
challenge to the state’s experts therefore was
limited to cross-examination . . . . But because of
counsel’s inexperience and unfamiliarity with
arson, his cross-examination was unfocused and
failed to challenge critical conclusions asserted by
the state’s experts. The First Circuit therefore
concluded that counsel’s failure to mount a non-
arson defense, particularly his failure to consult an
arson expert, “cannot be classified as a conscious,
reasonably informed tactical decision” . . . .

Id. (final alteration in original) (internal citations omitted)
(quoting Dugas, 428 F.3d at 320, 328–32).

¶34 With the benefit of the evidentiary hearing on remand
and the district court’s findings, we are now convinced that, as
in Dugas, more was reasonably to be expected of trial counsel.
First, as in Dugas where counsel was working his first arson case,

20140638-CA 20 2016 UT App 164
Landry v. State

here trial counsel testified to having never before worked on an
arson case and to relying on the limited knowledge that can be
conveyed during continuing legal education courses (CLEs). On
remand, the district court found that counsel had attended
multiple CLEs that “touched on arson,” but the court made no
findings about, and the State does not point us to any place in
the record that shows, how many CLEs she attended, how recent
these CLEs were, or how extensively these CLEs dealt with
effective strategies for defending arson cases. Thus, the findings
and record in this case suggest counsel’s prior experience did not
independently provide her with the knowledge necessary to cast
doubt on the State’s case through effective cross-examination.

¶35 Second, despite trial counsel’s inexperience, she made
only minimal efforts to educate herself on fire investigation
principles. The district court on remand found that the only
arson expert with whom trial counsel met was the Provo fire
marshal—the State’s lead arson investigator. In effect, she
“accept[ed] the state’s characterization of the scene” just like
counsel in Dugas had done. See Landry, 2012 UT App 350, ¶ 8
(discussing Dugas, 428 F.3d at 329–30). Even assuming the Provo
fire marshal spoke objectively and candidly in sharing his views
with Landry’s trial counsel, she should have anticipated that the
information he conveyed would have foretold his later
testimony on behalf of the prosecution, given his obligation to
testify truthfully. Obviously, he could not have helped her see
the flaws in the State’s experts’ opinions when he would later
testify to the validity of those opinions.

¶36 Third, we now know from the two postconviction experts
that there were inadequacies in the State’s investigation, of
which trial counsel would have learned had she more
thoroughly investigated the State’s experts with the assistance of
her own expert—an expert who could then have testified on
behalf of Landry at trial. See Templin, 805 P.2d at 188.
Specifically, had she secured the services of the expert who
testified for Landry at the evidentiary hearing on remand, she

20140638-CA 21 2016 UT App 164
Landry v. State

would have discovered that it was unclear whether the State’s
investigators employed the fire-investigation principles set out
in NFPA 921; that the original investigators who testified for the
State improperly concluded that the fire’s burn pattern definitely
proved ignitable liquids were poured on the floor; that if alcohol
was spilled at a party it was not a problem that lab results did
not find it because alcohol is miscible in water and would have
disappeared when the firefighters used water to extinguish the
fire, see supra note 5; that it was possible the original
investigators improperly evaluated the potential fuel sources in
the room, including the mattress; and that perhaps the State’s
experts improperly considered circumstantial evidence in
classifying the fire as arson and that its origin should more
properly have been classified as “undetermined.”

¶37 Fourth, as a result of trial counsel’s failure to consult her
own expert, she missed the opportunity to understand the
several problems with the State’s case and to highlight them for
the jury. According to the district court’s findings, the only error
in the State’s experts’ views that trial counsel was able to elicit
through cross-examination was that “just because an
investigator is not able to identify an accidental cause of the fire,
it doesn’t mean that it doesn’t exist.” Likewise, an independent
investigation and consultation with her own expert might have
pointed her to the scientific community’s skepticism about
accelerant detection canines, see supra ¶ 30 & note 9, which
would have helped her exclude that evidence instead of merely
eliciting testimony on cross-examination that such alerts have an
error rate of about 10%. Thus, as in Dugas, discussion with and
use of an independent expert would have allowed trial counsel
to more effectively challenge the State’s experts’ conclusions.

¶38 Fifth, as in Dugas, where it was critical to challenge the
State’s theories because of the difficulty in the developing the
theory that another suspect committed the arson, here it was
critical to Landry’s case that trial counsel effectively challenge
the State’s theory of the case and supporting expert testimony

20140638-CA 22 2016 UT App 164
Landry v. State

because the other theories available to Landry’s defense were
“difficult to mount and fraught with evidentiary problems.” See
Dugas, 428 F.3d at 329. Trial counsel’s alternative theories—that
the fire was either caused by a cigarette dropped on a day-old
alcohol spill or by a stranger—were only somewhat supported
by the evidence. Both postconviction experts agreed that the
cigarette-in-alcohol theory was weak, which suggests that if
counsel had employed an expert, he or she would have advised
against staking the defense case on this theory, which was only
speculative in any event. Landry testified that he had a party the
night before, which he never mentioned before trial, including
during his interviews with investigators on the night of the fire.
He testified that Cognac had been spilled and that the next day
his girlfriend sat on his bed and smoked a cigarette. But Landry
did not testify that his girlfriend actually dropped the cigarette
into the spilled alcohol. Instead trial counsel relied on testimony,
as stated in the district court’s findings, from a State witness that
“it is possible to start a fire with a discarded cigarette” to connect
Landry’s girlfriend’s potentially-dropped-and-still-lit cigarette to
the alcohol spill. Similarly, while the unidentified-arsonist
theory was on firmer ground than such stories often are—after
all, it was a disinterested neighbor who testified to seeing the
unidentified man, not Landry—the man could not be found.
And the neighbor said nothing about the man carrying gasoline,
flicking a lighter, smoking a cigarette, or the like, although he
did testify that, after he saw the man, he noticed that cardboard
used as a temporary window replacement in a window frame
next to Landry’s front door had been pushed in. Thus, by not
consulting an expert of her own, trial counsel forwent the
opportunity to pursue a more effective trial strategy, instead
pursuing theories that were much weaker. See id. at 329 (noting
that the “someone else did it” defense is both weak and difficult
to establish). Consequently, her representation of Landry was
objectively deficient.

20140638-CA 23 2016 UT App 164
Landry v. State

B. Trial Counsel’s Errors Prejudiced Landry.

¶39 “When a defendant challenges a conviction, the question
is whether there is a reasonable probability that, absent the
errors, the factfinder would have had a reasonable doubt
respecting guilt.” Strickland v. Washington, 466 U.S. 668, 695
(1984). “In making this determination, a court hearing an
ineffectiveness claim must consider the totality of the evidence
before the judge or jury.” Id. “Some errors will have had a
pervasive effect on the inferences to be drawn from the evidence,
altering the entire evidentiary picture, and some will have had
an isolated, trivial effect.” Id. at 695–96. And when the other
evidence in the case is “‘less compelling’ we ‘will more closely
scrutinize the conduct.’” State v. Thompson, 2014 UT App 14,
¶ 83, 318 P.3d 1221 (quoting State v. Troy, 688 P.2d 483, 486 (Utah
1984)).

¶40 Considering the totality of the evidence in this case, trial
counsel’s errors “had a pervasive effect” and “alter[ed] the entire
evidentiary picture.” See Strickland, 466 U.S. at 695–96. We
conclude that a reasonable jury would have had a reasonable
doubt but for trial counsel’s deficient performance. We reach this
conclusion for three reasons: (1) postconviction expert testimony
revealed substantial errors in the State’s arson case, which
should have come out at trial; (2) under Schultz, trial counsel
likely could have kept out Oscar’s alerts to the sock and shoe;
and (3) the jury’s struggle to reach a decision at trial indicates
that this was a close case—even with trial counsel’s deficient
performance. We now explain each reason more fully.

¶41 First, trial counsel’s acceptance of the State’s expert’s
characterization, instead of consulting independent experts,
prejudiced Landry because it meant that she was unprepared to
effectively cross-examine the State’s witnesses, see supra ¶¶ 35–
37, and that she did not introduce opposing expert opinion
testimony regarding the investigation methods, origin, and
cause of the fire. The reliability of the State’s investigation and

20140638-CA 24 2016 UT App 164
Landry v. State

the conclusions its experts drew as to the cause and point of
origin of the fire were the cornerstones of the State’s case. And,
as expert testimony in the postconviction evidentiary hearing
revealed, the investigation may not have been reliable. It was, for
example, unclear whether the investigation was based on
established fire-investigation principles, and it may not have
taken full account of all of the fire’s fuel sources. Landry’s
postconviction expert also testified about numerous other
possible errors. Significantly, the State’s experts’ conclusions
regarding the origin of the fire were erroneous, as they were
based on a misunderstanding of the v-pattern on the floor and
the effect of full-room involvement on such patterns. And
trained experts disagreed as to whether the fire originated from
poured ignitable fluid—making this fire the result of arson—or
was more likely an accident. Further, other than the State’s
expert witnesses and Oscar’s alerts, the State’s only evidence
implicating Landry was circumstantial, namely that he was
being evicted, that he had left very few items in the apartment
when he left, 10 that a lighter was found near the crime scene
(although never connected to Landry), and that he was seen
leaving the apartment five to ten minutes before the fire was
reported. Where counsel’s errors result in a failure to undermine
evidence that “the State’s case relied almost entirely upon” and
where that evidence could easily have been undermined had
counsel used appropriate judgment, we will conclude that the
defendant was prejudiced. See Thompson, 2014 UT App 14,
¶¶ 84–86. See also Dugas, 428 F.3d at 331–32, 341 (concluding that
“the crucial role of the arson evidence to the state’s case” meant

10. We are not convinced that the lack of items in the apartment
supports the State’s case circumstantially because, although
Landry acknowledged he was almost done moving, at least
some of the items he left—prescription medications and financial
documents—are not the sort of items a person would
intentionally leave behind to burn.

20140638-CA 25 2016 UT App 164
Landry v. State

counsel’s failure to effectively challenge that evidence was
prejudicial). Thus, if the jury had heard an arson expert testify
that the cause of the fire was not necessarily arson and that, even
if it was arson, the State’s experts were wrong as to the origin
and spread of the fire, the jury very likely would have
entertained a reasonable doubt that Landry set fire to his
apartment.

¶42 Second, had counsel objected, on the authority of Schultz,
to the dog handler’s testimony regarding Oscar’s alerts to
Landry’s shoe and sock, the State would have been required to
prove the reliability of Oscar’s alerts—and it would have been
difficult to do so. See supra ¶¶ 30–31 & note 9. Trial counsel likely
would have been able to keep out the evidence of the alerts, see
State v. Schultz, 2002 UT App 366, ¶ 27, 58 P.3d 879; Yell v.
Commonwealth, 242 S.W.3d 331, 345 (Ky. 2007) (Scott, J.,
concurring and dissenting), as the State has not pointed us to
any evidence of “‘a sufficient foundation to demonstrate the
inherent reliability of the underlying principles and techniques’”
as an alternative way to prove the inherent reliability of Oscar’s
alerts, Schultz, 2002 UT App 366, ¶ 27 (quoting State v. Brown,
948 P.2d 337, 340 (Utah 1997)). And the testimony that Oscar was
more reliable than lab tests, if allowed at all, would have
appeared very questionable indeed. Thus, the State could likely
not have satisfied the three-part test outlined in Schultz, and
Oscar’s alerts would not have come in as substantive evidence of
ignitable fluid on Landry’s sock and shoe. If Oscar’s alerts had
not come in as substantive evidence, there would have been no
solid evidence directly connecting Landry to the cause of the fire
alleged by the State: poured ignitable fluids.

¶43 Third, we know that even with evidence of Oscar’s alerts
and without the benefit of a contradicting expert, the jury still
struggled in reaching the decision to convict. As we noted in our
initial opinion affirming Landry’s convictions, “at one point the
jury was deadlocked.” State v. Landry, 2008 UT App 461U, para.
3 n.1. And the trial record reflects that the jury only reached a

20140638-CA 26 2016 UT App 164
Landry v. State

consensus after the court explained that it would have to come
back to deliberate for an extra, unplanned day, causing a
hardship for at least one juror, if it did not decide the case at that
time. This demonstrates that the jury struggled to reach its
decision, making it much more likely that it would have reached
a different conclusion but for trial counsel’s ineffectiveness.

II. Appellate Counsel

¶44 Finally, “[t]o show that appellate counsel was ineffective
in failing to raise a claim, [Landry] must show that the issue
[was] obvious from the trial record and . . . probably would have
resulted in reversal on appeal.” Kell v. State, 2008 UT 62, ¶ 42, 194
P.3d 913 (third alteration and omission in original) (citation and
internal quotation marks omitted).

¶45 In the previous appeal, we determined that “[t]he
prejudice resulting from appellate counsel’s failure to raise a
claim of ineffective assistance of trial counsel . . . is . . .
automatic” when, “but for trial counsel’s ineffectiveness, it was
reasonably likely that [Landry] would not have been convicted.”
Landry v. State, 2012 UT App 350, ¶ 11, 293 P.3d 1092.

If trial counsel’s deficiencies were prejudicial,
appellate counsel’s failure to raise those
deficiencies is necessarily prejudicial in the same
way and to the same extent. The prejudice from
Landry’s claim that appellate counsel was deficient
in failing to assert on appeal trial counsel’s
ineffective representation is therefore implicit in
his argument that he was prejudiced by trial
counsel’s deficient performance.

Id. Because we conclude that trial counsel’s deficient
performance prejudiced Landry, we conclude that appellate
counsel’s failure to raise a claim of ineffective assistance of trial

20140638-CA 27 2016 UT App 164
Landry v. State

counsel on appeal was likewise objectively deficient and
prejudicial.

CONCLUSION

¶46 We conclude that trial counsel’s performance was
deficient and that there is a reasonable probability that, without
this deficiency, the jury would have reached a more favorable
outcome for Landry; therefore, we also conclude that Landry has
established ineffective assistance of appellate counsel. Thus,
Landry is entitled to appropriate relief under the PCRA. This
relief would ordinarily be a new trial with the assistance of new
counsel, who would presumably engage a qualified expert and
object under the authority of Schultz to much of the State’s case.
But because Landry is no longer imprisoned and has relocated to
Texas, or so we were advised at oral argument, 11 it is far from
clear that this would be a reasonable use of judicial resources or
even something Landry would welcome. We therefore remand
to the district court to implement an appropriate remedy under
the circumstances.

11. According to the sentencing report, which is not a part of the
record in this case but of which we take judicial notice, Landry
was sentenced on August 16, 2006, to an indeterminate sentence
of five years to life. We were advised at oral argument that he
was paroled from the Utah State Prison in February 2014.

20140638-CA 28 2016 UT App 164