Houston Crime Lab Report 2006

• Topics: Crime/Demographics, Crime Labs

Download original document:
Brief thumbnail

This text is machine-read, and may contain errors. Check the original document to verify accuracy.

Fifth Report of the
Independent Investigator
for the
Houston Police Department
Crime Laboratory and Property Room

Michael R. Bromwich
Independent Investigator
Fried, Frank, Harris, Shriver & Jacobson LLP
1001 Pennsylvania Avenue, N.W., Suite 900
Washington, D.C. 20004
202.639.7000
http://www.hpdlabinvestigation.org
May 11, 2006

HPD Crime Lab Independent Investigation

Investigative Team
Independent Investigator
Michael R. Bromwich
Fried, Frank, Harris, Shriver & Jacobson LLP
Fried Frank Team
Tommy P. Beaudreau
Piper M. Hendricks
Beth C. McClain
Jennifer M. Wollenberg
Joshua Cowdery
Michelle Hand-Arevalo
H. Coleman Hinnant
Charmain Ho-A-Lim
Scientific Advisory Board
Margaret C. Kuo
Orange County Sheriff-Coroner‘s Office (ret.)
Douglas M. Lucas
Centre of Forensic Sciences of the Province of Ontario, Canada (ret.)
Bruce W. Vander Kolk
Illinois State Forensic Sciences Command (ret.)
Forensic Scientists
Robert P. Bianchi
DEA Special Testing and Research Laboratory (ret.)
Roger J. Bolhouse
Michael Sinke
Speckin Forensic Laboratories
Michael A. Evans, Ph.D.
American Institute of Toxicology Laboratories

HPD Crime Lab Independent Investigation

Patricia P. Hamby
International Forensic Science Laboratory & Training Centre
Edward E. Hueske
Arizona Department of Public Safety (ret.)
Karen L. Irish
Forensic Services Section, Baltimore County Police Department (ret.)
Rhonda Roby
Identity Quest, LLC
Mark D. Stolorow
Jeanine Baisch, Ph.D.
Rick W. Staub, Ph.D.
Orchid Cellmark
Theresa F. Spear
California Department of Justice, Bureau of Forensic Services
California Criminalistics Institute (ret.)
Statisticians
Jessica Pollner, Ph.D.
Arthur Baines
Kerri-Ann Cullinan
PricewaterhouseCoopers LLP

HPD Crime Lab Independent Investigation

Executive Summary
This is the Fifth Report of the Independent Investigator for the Houston
Police Department (“HPD”) Crime Laboratory and Property Room. This report,
like our previous reports, is intended to advise the City of Houston (the “City”)
and the public of our progress in fulfilling the mandate to conduct a
comprehensive and independent investigation of the Crime Lab and Property
Room.1
The investigation is divided into two phases. In Phase I, which we
completed with the issuance of our Third Report on June 30, 2005, we
investigated the historical operations, practices, and management of the Crime
Lab and Property Room as well as assessed the scope of the work to be
performed during the second phase of the investigation. Phase II, which began
with the Houston City Council’s approval of our Phase II Plan on August 24,
2005, centers on the review of hundreds of cases originally analyzed by Crime
Lab forensic scientists in seven forensic science disciplines in which the Lab
historically performed work -- serology, DNA, trace evidence, controlled
substances, firearms, toxicology, and questioned documents.
We have now completed our Phase II case reviews for all of the forensic
science disciplines with the exceptions of serology and firearms. Since our last
report, we also have conducted detailed interviews of three former members of
the DNA/Serology Section about specific cases with which they were involved
as well as about technical practices and issues we have identified through the
case reviews. Unfortunately, our further attempts to secure the cooperation of
Donald Krueger, the former head of the Crime Lab; Christy Kim, a longtime
analyst in the Crime Lab’s Serology and DNA Sections; and James R. Bolding,
the former heard of the DNA/Serology Section, have not been successful. This is
particularly unfortunate because Ms. Kim and Mr. Bolding personally performed
the analytical work in many of the most problematic cases we have reviewed.
Therefore, our inability to gather information from them in connection with our
case reviews has hampered our ability to determine whether any of the most
troubling cases we have found were the product of intentional scientific fraud.
This report provides additional detail about the results of our case reviews
and the major and minor issues we have identified with respect to the work
1

Our reports are posted on our Web site at www.hpdlabinvestigation.org.

HPD Crime Lab Independent Investigation

performed during the relevant time periods in each of the sections of the Crime
Lab. We have continued to find that many of the Crime Lab’s forensic scientists
performed high quality technical work and accurately reported their results.
This is especially true for the Firearms, Toxicology, and Questioned Documents
Sections. Even in the areas in which the Crime Lab performed consistently
reliable work, however, certain fundamental practices -- for example, thorough
documentation of all analytical work, up-to-date standard operating procedures
(“SOPs”), quality assurance and quality control, technical reviews, training, and
outside inspections -- were either inconsistently applied or completely absent
from the Lab during the periods we are reviewing.
Since our last report, we have continued to uncover major issues in the
serology and DNA cases analyzed by the Crime Lab, dating back to 1980 in the
area of serology. Thus far, our investigation has identified a total of 43 DNA
cases and 50 serology cases analyzed by the Crime Lab that we have determined
to have major issues, which we have defined to mean problems that raise
significant doubt as to the reliability of the work performed, the validity of the
analytical results, or the correctness of the analysts’ conclusions. Many of the
problems we have observed in the serology cases infected the Crime Lab’s DNA
profiling operations as DNA testing gradually supplanted serology in the early
1990s. This is unfortunate, but not surprising, since many of the Crime Lab’s
serologists, such as Mr. Bolding and Ms. Kim, became DNA analysts. In
addition, although the technology for analyzing biological evidence changed
with the advent of DNA profiling, the Crime Lab’s flawed management and
practices -- as well as the inadequate level of support and attention the Lab
received from HPD -- did not.
A.

Serology

In our initial serology case reviews last fall, we identified severe problems
with the Crime Lab’s serology work performed during the period 1987 to 1990.
As a result, in December 2005, we recommended to the Stakeholders Committee
and HPD that we modify the scope of the serology case reviews as follows. First,
we recommended that the time period of our case reviews be expanded to cover
the years 1980 through 1993 in order to include the entire period in which
Mr. Bolding performed serology at HPD as well as a short period prior to his
involvement. Second, we recommended changing the basis of our case selection
in serology from random sampling, which is designed to provide a cross section
of the work performed by the Crime Lab, in order to concentrate our efforts on
systematically attempting to identify cases in which flawed serology work by the
Lab may have played a role in a criminal conviction, either by guilty plea or as a

HPD Crime Lab Independent Investigation

result of a trial verdict. These recommended changes reflect the fact that our
initial serology case reviews revealed such a high percentage of major issues and
other problems that it made sense to focus on those cases in which there was the
potential to identify specific miscarriages of justice rather than to continue using
a random sampling technique, which almost surely would have continued to
show an unacceptably high rate of similar analytical problems.
HPD and the Stakeholders Committee approved our recommended
changes in the selection criteria for our serology case reviews, and we
implemented the revised methodology in January 2006. We have faced
significant logistical challenges in identifying serology cases that might be
related to actual convictions of individual defendants, particularly cases from the
early 1980s that pre-dated the Crime Lab’s use of an electronic case tracking
system. HPD has been fully supportive of our efforts and has devoted
significant investigative resources to assisting us in identifying serology
conviction cases dating back to 1980. Nevertheless, progress in identifying these
cases has been slow.2
We have now completed reviews of 336 serology cases performed by the
Crime Lab between 1980 and 1993, which includes the 80 substantive cases
covered in our Fourth Report and 256 cases related to convictions by guilty plea
or trial verdict that we have reviewed since January 1, 2006. We have identified a
total of 50 serology cases involving major issues.
Although we continue to observe all of the same serious deficiencies that
we discussed in our Fourth Report, we focus in this report on the following
significant problems with the serology work performed in the Crime Lab during
the 1980s and early 1990s:

We recently received the approval of HPD and the Stakeholders Committee to focus our
ongoing review of serology cases on those cases analyzed by the Crime Lab during the
period 1980 through 1993 that might have played a role in the conviction of a person who
currently remains incarcerated. Because many defendants who were convicted during
this period, either by guilty plea or at trial, have already served their sentences, this will
significantly reduce the number of serology cases that we need to identify and review.
We will consult with the City Council, the Stakeholders Committee, and HPD as to
whether to expand this review to include all serology conviction cases whether or not
the defendant remains incarcerated.

HPD Crime Lab Independent Investigation

•

Failing to perform potentially probative, inculpatory or exculpatory,
ABO typing in a large number of cases, particularly sexual assault
cases.

•

Failing to report probative ABO typing results due to an apparent
reluctance to report exclusions, findings of blood types not consistent
with those of victims or known suspects, and findings indicating the
presence of relatively rare ABO type AB factors.

•

Misinterpreting and inaccurately reporting ABO typing results.

•

Reporting test results that are unsupported by documented analysis.

•

Engaging in poor forensic scientific practices, such as misapplying
substrate controls, reporting ABO typing results indicating no ABO
activity as “inconclusive,” and producing incorrect typing results due
to poor training or technique.

In this report, we discuss in detail the case of Derrick Leon Jackson, a
capital murder case, in which an initial failure to report ABO typing results was
reversed when a new suspect was learned to have the very ABO blood type
found -- and not reported -- at the time of the original analysis.3 In 1988,
Mr. Bolding obtained ABO typing results from a bloodstain sample taken from
the scene of a grisly double homicide that indicated it was foreign to both the
victims and the individual who was then suspected of the killings. At the time,
however, Mr. Bolding reported these results as “inconclusive.” The investigation
languished until 1995 when Mr. Jackson became the prime suspect.
Mr. Jackson’s ABO type was consistent with the foreign ABO factor Mr. Bolding
had detected in tests he described as “inconclusive” in 1988. Without performing
any additional testing, Mr. Bolding altered his worksheets to include previously
absent conclusive interpretations of his original typing results performed in 1988
and issued a new report stating that ABO activity consistent with Mr. Jackson’s
ABO type was found in two bloodstain samples recovered from the crime scene.
B.

DNA

In light of the seriousness and pervasiveness of the major issues we found
in our case reviews in 2005, we recommended to the Stakeholders Committee
3

Texas v. Jackson, Cause No. 748752 (Harris County, Tx.).

HPD Crime Lab Independent Investigation

and HPD that we focus the case reviews on the 69 DNA cases that were part of
the DNA re-test project previously undertaken by HPD in conjunction with the
Harris County District Attorney’s Office. The re-test project was intended to
identify all cases in which DNA analysis performed by the Crime Lab might
have played a role in a conviction, either by guilty plea or after trial. The 69 cases
were ones in which either (a) evidence had not yet been tested by outside
laboratories or (b) re-testing by outside laboratories had failed to confirm the
Crime Lab’s original DNA testing results. We suggested this approach in order
to focus resources on reviewing the cases in which there had been a conviction
and outside testing had not confirmed the Crime Lab’s original results. We
believed that these cases posed the greatest risk for potential injustice related to
flawed DNA profiling work performed by the Crime Lab. The Stakeholders
Committee and HPD approved our recommendations, and we changed the focus
of our DNA case reviews accordingly.
We have completed our review of all 69 of these DNA conviction cases in
which either the evidence has yet to be re-tested or the Crime Lab’s original
DNA results have not been confirmed through re-testing. We have identified
major issues in 25 -- or approximately 36% -- of these cases. In total, we have
reviewed 135 DNA cases analyzed by the Crime Lab from the entire period in
which the Lab performed DNA analysis and have identified major issues in 43 of
them, which is approximately 32% of the cases we have reviewed.
Many of the major issues we have identified in the cases we have
reviewed since our Fourth Report are similar to those we discussed in that
report, including the failure to report probative, even potentially exculpatory,
results; poor technical work and potential contamination; questionable
interpretation of results; and misleading reporting of the statistical significance of
DNA profiling results in mixture cases.
Similar to what we have found in many serology cases, we have found
further evidence of a pattern of reluctance on the part of the Crime Lab’s DNA
analysts to report typing results inconsistent with the known profile of either a
victim or a suspect. With the ability of DNA profiling to generate powerful
analytical results that point to, or away from, a suspect, the Crime Lab’s practice
of failing to report probative, and in some cases potentially exculpatory, DNA
typing results became all the more disturbing.
It is not clear whether this pattern and practice of avoiding the reporting
of DNA typing results that were not consistent with a victim or known suspect is
attributable to the DNA analysts’ lack of confidence in their ability to obtain

HPD Crime Lab Independent Investigation

reliable results or to a more sinister manipulation of analytical results. It is
apparent, however, that DNA analysts in many cases tended toward reporting
only those results that, from their perspective, were “safe” in the sense that they
were consistent with other evidence in the case or with the investigators’
expectations. This sometimes meant that analysts suppressed potentially
exculpatory RFLP results in favor of reporting less reliable or less discriminating
PCR-based typing results that appeared to reflect an association between the
suspect and evidence in the case. When such selective reporting was coupled
with the Crime Lab’s systematic overstatement of the statistical significance of
these weaker PCR results, a very significant potential for injustice was created.
C.

Trace Evidence

We identified 223 cases as having been opened by the Trace Evidence
Section during the period covered by our review. We have completed our
review of all 223 of these trace evidence cases, and we found that 129 involved
the type of substantive analysis by the Trace Evidence Section that was within
the scope of our review. We found that 5 cases (4% of the substantive sample)
had major issues and that 44 cases (34% of the substantive sample) had minor
issues. All of these issues were described in detail in our Fourth Report; no new
types of deficiencies were identified during the remainder of our review of trace
evidence cases.
We also selected 40 serology and DNA cases to determine (a) whether any
trace evidence was examined in connection with those cases and (b) if so,
whether there were any problems with that work. We selected these 40 cases
because they involved either death penalty or other high profile convictions.4
We identified no major issues in these 40 cases, although 5 cases (12.5%) involved
minor issues. These minor issues primarily related to documentation, which is
consistent with our findings with respect to the general sample of 223 trace
evidence cases.
D.

Controlled Substances

We reviewed three separate Controlled Substances Section case samples
during the Phase II review -- one comprised of general controlled substances
4

It is not uncommon for serology and DNA cases to contain trace evidence components.
Cases such as homicides and sexual assaults in which there may be biological evidence
(for example, blood or semen) may also involve evidence, such as hairs, that could be
subjected to trace evidence examination and comparison.

HPD Crime Lab Independent Investigation

cases,5 a second sample of cases analyzed by Vipul Patel, and a third sample of
cases analyzed by James Price. We established separate samples for Mr. Price
and Mr. Patel because they are former Controlled Substances Section analysts
associated with apparent fabrication of scientific findings. We also reviewed an
additional 50 files to evaluate how the Crime Lab handled and analyzed “bulk”
or “bulky” cases, which, as the terms suggest, are cases involving large quantities
of evidence. We have completed our review of the bulk controlled substances
cases, as well as all three Controlled Substances Section case file samples.
1.

Results of the General Controlled Substances Case
Reviews

We identified over 200 cases in the general controlled substances sample
that contained minor issues, and we consistently noted that most of those issues
should have been detected through administrative and/or technical reviews.
Major issues were identified in 116 (nearly 23%) of the 513 cases in the general
controlled substances sample. In particular, we found the following types of
deficiencies:

•

Analysts in the Controlled Substances Section, as a matter of custom
and practice, reported the quantity of a controlled substance in a liquid
or tablet as if it had been determined through analytical testing, when
in fact the analysts established quantitations based solely on
presumption, Crime Lab protocol, or visual (also referred to as
“physical” or “pharmaceutical”) identification.

•

Similarly, analysts often reported the identity of a controlled substance
sample based only on physical or visual inspections and failed to
acknowledge in those reports that chemical tests were not conducted
to identify the substance. Instead, samples were identified by merely
comparing tablet markings, imprints, etc. to those found in reference
sources, such as the Physician’s Desk Reference.

•

We identified two cases in which analysts did not report the presence
of a controlled substance and therefore failed to report a probative
finding.

The general controlled substances sample does not include cases handled by Mr. Patel or
Mr. Price.

HPD Crime Lab Independent Investigation

•

Two other cases involved types of errors in the recording of lab data
that could seriously affect the prosecution of a defendant. In each of
those cases, however, the ultimate outcomes for the defendants were
not affected by the errors.

•

Finally, one major issue case from the general controlled substances
sample involved an unexplained test result change. Due to the poor
documentation in the case file, we could not determine if the change
was made because the original analyst misinterpreted the tests or if the
substance was tested a second time and produced different results.
2.

Results of the Bulk Evidence Case Reviews

We reviewed 50 bulk evidence cases; 2 contained deficiencies
characterized as major issues, and 42 involved minor issues. The major issues
involved the same issues relating to reporting liquid quantitations without
performing quantitative analyses.
A number of the minor issues identified in the bulk evidence case review
involved documentation deficiencies and the failure to follow generally accepted
forensic science practices. Analysts regularly failed to follow SOPs that
specifically related to the handling of bulk evidence. Moreover, the Crime Lab’s
SOPs and other manuals offered little guidance regarding which cases were to be
classified as bulk cases.
3.

Results of the Patel Case Reviews

We have completed our review of the 366 cases in the Patel sample; 18
cases involved major issues, and most (14) of those cases were described in our
Fourth Report. The 4 new major issue cases that we identified involved the
following deficiencies:
•

Mr. Patel reported an undocumented finding in a case involving evidence
identified as LSD. Although preliminary tests indicated that LSD was
indeed present in the sample and we concluded that it is likely that the
substance was correctly identified, the documentation in the case file
relating to confirmatory gas chromatography/mass spectrometry
(“GC/MS”) testing did not adequately support that conclusion.

•

In three other cases in the sample, Mr. Patel failed to clearly report that
sample identifications were based on visual or physical comparisons,

HPD Crime Lab Independent Investigation

rather than on analytical testing. As discussed above, we characterize
such an identification as a major issue because it is potentially misleading.
4.

Results of the Price Case Reviews

We have reviewed all 342 cases in the Price sample. We identified major
issues in 11 of those cases; 7 were previously described in our Fourth Report.
Two of the 4 newly-identified major issue cases involved the identification of
dihydrocodeinone (Vicodin) based on a comparison of the mass spectrum of the
evidence to the mass spectrum of a standard that did not show the principal
peaks and ions necessary to identify the substance. Identification based on an
inadequate standard is not consistent with generally accepted forensic laboratory
practices.
In a third case, Mr. Price failed to perform a critical examination when
identifying evidence. Finally, Mr. Price reported the identity of a tablet or
capsule after performing a physical identification, without reporting that the
identification was based only on visual or physical testing.
E.

Toxicology

We have completed our review of the toxicology case file sample. Of the
total 396 files selected in the samples, only about half (213) involved actual
toxicology casework. Although we made adjustments in our search criteria to
isolate true toxicology cases, our reviews showed that most of the remaining 183
cases in the combined samples involved analysis associated with controlled
substances casework, rather than toxicology.
We consulted PricewaterhouseCoopers LLP (“PwC”) to determine
whether the scope of our review was sufficient under statistical standards to
establish confidence in the results of our toxicology case file review. PwC
confirmed that, in light of the very conservative parameters that were used to
establish the sample, the results maintained a high level of statistical precision
and that we did not need to conduct additional sampling of toxicology case files.
We identified only one major issue case in our review of the 213
toxicology cases, and that case was discussed in detail in our Fourth Report.6 We
6

We identified only one major issue in the toxicology sample, which involved the
identification of three drugs in a blood sample -- heroin, cocaine, and PCP -- without the
use of a confirmatory test. Our review showed that some aspects of the test results were
Footnote continued

HPD Crime Lab Independent Investigation

found that 86 of the 213 toxicology files that we reviewed (40%) had minor
deficiencies. Most involved situations in which:
(1)

drugs were identified on the basis of only one independent test,

(2)

there were indications of potential sample contamination, or

(3)

case files did not contain adequate documentation of all the work
that may have been performed.

In all of these cases, other controlled substances or drugs of abuse were properly
identified, and we therefore concluded that these minor issues would not have
had a significant impact on the cases.
F.

Firearms

We have reviewed 94 firearms cases since the publication of our Fourth
Report, for a total of 184 firearms case reviews during Phase II of our
investigation. Overall, most cases in the Firearms Section were properly
examined and reported in a timely manner. We have identified only minor
issues, primarily involving slight deficiencies in documentation, deviations from
Crime Lab policies, and deviations from generally accepted forensic science
practices. We noted that the examinations of the vast majority of the cases are
technically correct and that most issues are of an administrative nature. We
found that HPD firearms examiners operated with an impressive level of skill
and diligence in performing their examinations and correctly making
identifications.
Conclusion
This report summarizes the results of the case reviews we have conducted
thus far in Phase II of this independent investigation of the Crime Lab and
Property Room. Over the past eight months, we have reviewed nearly 2,300
cases analyzed by the Crime Lab in the forensic science disciplines of serology,
DNA profiling, trace evidence, controlled substances, firearms, toxicology, and
Footnote continued from previous page
pharmacologically unlikely and indicated possible sample contamination. Because of
these questionable results, the possibility of contamination, and the absence of a second
confirmatory test, we concluded that the work performed by Crime Lab analysts in this
case was inconsistent with generally accepted forensic science practices.

HPD Crime Lab Independent Investigation

questioned documents. Our case reviews now are complete in all of these areas
except for firearms and serology, which --- in light of the serious and pervasive
problems we have identified with the Crime Lab’s serology work -- has been
expanded to include cases dating back to 1980. We are continuing to work with
HPD to identify convictions in the 1980s and early 1990s in which serology
performed by the Crime Lab might have played a role and then to review those
cases to evaluate the reliability of the serologists’ analysis. Unfortunately, the
process is not one that can be completed quickly because of the inherent
difficulties in identifying the relevant cases.
The case reviews we have completed since our last report continue to
reveal widespread problems with the Crime Lab’s analysis of biological
evidence -- beginning with serology and continuing after the advent of DNA
profiling -- during the entire period of our review from 1980 through 2002. In the
cases we have reviewed since the publication of our Fourth Report in January
2006, we have found additional examples of serologists and DNA analysts failing
to report probative results -- results that might have helped identify and convict
the guilty as well as results that might have exonerated the innocent. We have
found a clear and troubling pattern of reluctance in the Serology and DNA
Sections to report typing results that were not consistent with the blood types or
DNA profiles of either the victim or a known suspect; in many such cases, the
serology or DNA results were reported as inconclusive. We have also seen
persistent problems with analysts’ techniques, the interpretation of results, and
the failure to address potential sample contamination. The Crime Lab never
issued written reports containing the statistical significance of its serology typing
results, and the frequency estimates presented by DNA analysts in cases
involving mixtures of body fluids were often overstated by orders of magnitude.
We have found no semblance of an effective technical review program or quality
assurance regime to detect and correct these problems. As a result, they
continued unabated.
We have identified 50 serology cases and 43 DNA cases in which work
performed by the Crime Lab was unreliable or the reported results were
misleading. The names of each of the suspects or defendants involved with these
cases are listed in Appendices B and C to this report. We will continue to
provide information about each of these cases to the Innocence Project and
affiliated organizations so that these cases can be reviewed and evaluated to
identify prisoners who may have been wrongfully convicted.
We still have significant work ahead of us in completing this
investigation. In the coming months, we will review the current operations of

HPD Crime Lab Independent Investigation

both the Crime Lab and the Property Room and formulate recommendations
regarding the forensic science work performed by HPD. The shared goal of the
City of Houston, HPD, the Stakeholders Committee, and our investigative team
is not only to produce a comprehensive accounting of the Crime Lab’s historical
problems but also to help develop a blueprint for ensuring that the Lab generates
sound, well-documented, and reliable forensic scientific results for use in the
criminal justice system.

HPD Crime Lab Independent Investigation

TABLE OF CONTENTS
Page
E ecutive Summary...........................................................................................................1
Introduction .......................................................................................................................1
The Investigative Team ....................................................................................................2
Status of the Investigation ...............................................................................................5
I.

The Case Reviews .................................................................................................5

II.

Interviews...............................................................................................................8

III.

The Innocence Project...........................................................................................8

Results of Phase II Case Reviews....................................................................................9
I.

Process for Reviewing Cases ...............................................................................9
A.

B.
II.

Definitions of Major and Minor Issues ................................................10
1.

Major Issues .................................................................................10

Minor Issues.................................................................................10

Quality Assurance Review ....................................................................11

Serology ................................................................................................................11
A.

Status of the Serology Case Reviews....................................................14

Serious Problems Identified in Serology Cases ..................................20
1.

Systemic Failure to Perform Potentially
Probative ABO Typing in Cases Involving
Known Suspects ..........................................................................21

Failure to Report Probative Findings .......................................23
a.

Serology in the Derrick L. Jackson Case ......................24

HPD Crime Lab Independent Investigation

Page
b.

III.

IV.

Additional Examples of the Crime Lab’s
Failure to Report Probative Findings ...........................28

Misinterpretation and Inaccurate Reporting of
Results...........................................................................................31

Reported Findings Unsupported by
Documented Analysis ................................................................33

Other Examples of Poor Forensic Science
Practices........................................................................................34

DNA ......................................................................................................................36
A.

Status of the DNA Case Reviews..........................................................38

Serious Problems Identified in DNA Cases ........................................40
1.

Failure to Report Potentially Exculpatory
Results...........................................................................................41

Failure to Mathematically Confirm RFLP
Results...........................................................................................43

Poor Quality PCR Analysis .......................................................47
a.

Problems with D1S80 Analysis .....................................48

Questionable Interpretation of PCR
Results...............................................................................51

Problems with Interpretation of STR Results .........................53

Routine Misreporting of the Statistical
Significance of DNA Profiling Results.....................................55

Trace Evidence.....................................................................................................58
A.

Status of the Trace Evidence Review ...................................................58

Findings ....................................................................................................59
1.

The Trace Evidence Sample.......................................................59

HPD Crime Lab Independent Investigation

iii

Page
2.
V.

Review of Selected Serology and DNA Cases ........................60

Controlled Substances ........................................................................................61
A.

Results of the General Controlled Substances Case
Reviews.....................................................................................................62
1.

3.
B.

Major Issues .................................................................................62
a.

Reports Based on Visual or Physical
Identifications Without Analytical
Testing ..............................................................................62

Presumptive Quantitation of Liquids
and Tablets .......................................................................63

Failure to Report Probative Findings ...........................64

Mistakes and Inaccuracies in Reports ..........................64

Unexplained Test Result Change .................................65

Minor Issues Identified in the General
Controlled Substances Sample..................................................65
a.

Inadequate Documentation ...........................................66

Failure to Follow Generally Accepted
Forensic Science Practices ..............................................66

Failure to Follow Crime Lab Standard
Operating Procedures.....................................................68

Handling and Labeling of Evidence.........................................68

Results of the Patel Case Reviews ........................................................69
1.

Reporting an Undocumented Finding.....................................69

Reports Based on Visual or Physical
Identifications Without Analytical Testing .............................70

Inadequate Documentation .......................................................70

HPD Crime Lab Independent Investigation

Page

VI.

VII.

Failure to Follow Generally Accepted Forensic
Science Practices ..........................................................................70

Failure to Follow Standard Operating
Procedures....................................................................................71

Results of the Price Case Reviews ........................................................71
1.

Major Issues .................................................................................71

Minor Issues.................................................................................72
a.

Failure to Follow Generally Accepted
Forensic Science Practices ..............................................72

Failure to Follow Standard Operating
Procedures........................................................................72

Insufficient Documentation and
Administrative Errors ....................................................73

Results of Bulk Case Reviews ...............................................................73

Lack of Consistent Technical Reviews .................................................74

Firearms................................................................................................................74
A.

Method Related to Firearms Examinations.........................................75

Documentation of Examinations ..........................................................76

Inefficient or Inappropriate Deviations from
Generally Accepted Laboratory Practices ...........................................77

Other Issues .............................................................................................79

Toxicology............................................................................................................79
A.

The Toxicology Case Sample.................................................................80
1.

Confirming Initial Results With A Second
Independent Test ........................................................................81

Indications of Sample Contamination .....................................82

HPD Crime Lab Independent Investigation

Page
3.

Documentation Issues ................................................................82

Conclusion .......................................................................................................................83

Appendix A

Stakeholders Committee Members

Appendix B

Serology Major Issue Cases

Appendix C

DNA Major Issue Cases

Appendix D

Acronyms

HPD Crime Lab Independent Investigation

Introduction
This is the Fifth Report of the Independent Investigator for the Houston
Police Department (“HPD”) Crime Laboratory and Property Room. This report,
like our previous reports, is intended to advise the City of Houston (the “City”)
and the public of our progress in fulfilling the mandate to conduct a
comprehensive and independent investigation of the Crime Lab and Property
Room.1
The investigation is divided into two phases. In Phase I, which we
completed with the issuance of our Third Report on June 30, 2005, we
investigated the historical operations, practices, and management of the Crime
Lab and Property Room as well as assessed the scope of the work to be
performed during the second phase of the investigation. Phase II, which began
with the Houston City Council’s approval of our Phase II Plan on August 24,
2005, centers on the review of hundreds of cases originally analyzed by Crime
Lab forensic scientists in seven forensic science disciplines in which the Lab
historically performed work -- serology, DNA, trace evidence, controlled
substances, firearms, toxicology, and questioned documents.
We have now completed our Phase II case reviews for all of the forensic
science disciplines with the exceptions of serology and firearms. Since our last
report, we also have conducted detailed interviews of three former members of
the DNA/Serology Section about specific cases with which they were involved
as well as about technical practices and issues we identified through the case
reviews. Unfortunately, our further attempts to secure the cooperation of
Donald Krueger, the former head of the Crime Lab; Christy Kim, a longtime
analyst in the Crime Lab’s Serology and DNA Sections; and James R. Bolding,
the former heard of the DNA/Serology Section, have not been successful. This is
particularly unfortunate because Ms. Kim and Mr. Bolding personally performed
the analytical work in many of the most problematic cases we have reviewed.
Therefore, our inability to gather information from them in connection with our
case reviews has hampered our ability to determine whether any of the most
troubling cases we have found were the product of intentional scientific fraud.
This report provides additional detail about the results of our case reviews
and the major and minor issues we have identified with respect to the work
1

Our reports are posted on our Web site at www.hpdlabinvestigation.org.

HPD Crime Lab Independent Investigation

performed during the relevant time periods in each of the sections of the Crime
Lab. We continue to find that many of the Crime Lab’s forensic scientists
performed high quality technical work and accurately reported their results.
This is especially true for the Firearms, Toxicology, and Questioned Documents
Sections. Even in the areas in which the Crime Lab performed consistently
reliable work, however, certain fundamental practices -- for example, thorough
documentation of all analytical work, up-to-date standard operating procedures
(“SOPs”), quality assurance and quality control, technical reviews, training, and
outside inspections -- were either inconsistently applied or completely absent
from the Lab during the periods we are reviewing.
Since our last report, we have continued to uncover major issues in the
serology and DNA cases analyzed by the Crime Lab, dating back to 1980 in the
area of serology. Thus far, our investigation has identified a total of 43 DNA
cases and 50 serology cases analyzed by the Crime Lab that have major issues,
which we have defined to mean problems that raise significant doubt as to the
reliability of the work performed, the validity of the analytical results, or the
correctness of the analysts’ conclusions. Many of the problems we have
observed in the serology cases infected the Crime Lab’s DNA profiling
operations as DNA testing gradually supplanted serology in the early 1990s.
This is unfortunate, but not surprising, because many of the Crime Lab’s
serologists, such as Mr. Bolding and Ms. Kim, became DNA analysts. In
addition, although the technology for analyzing biological evidence changed
with the advent of DNA profiling, the Crime Lab’s flawed management and
practices -- as well as the inadequate level of support and attention the Lab
received from HPD -- did not.

The Investigative Team
We have assembled a highly experienced team of lawyers and forensic
scientists for the Crime Lab investigation. Our team is led by Michael R.
Bromwich, who is a partner in the Washington, D.C. and New York offices of
Fried, Frank, Harris, Shriver and Jacobson LLP (“Fried Frank”) and heads the
Firm’s internal investigations, compliance, and monitoring practice group.
Mr. Bromwich is a former federal prosecutor and, from 1994 to 1999, served as
Inspector General of United States Department of Justice. Mr. Bromwich is
supported by a team of Fried Frank lawyers and legal assistants.
Our Scientific Advisory Board, comprised of three renowned forensic
scientists who are experienced crime laboratory managers, has worked closely on

HPD Crime Lab Independent Investigation

the investigation throughout Phases I and II.2 Each member of the Scientific
Advisory Board has visited the Crime Lab and Property Room, participated in
interviews, and conducted quality control and quality assurance reviews of the
work performed by our team. In addition, throughout Phases I and II we have
consulted -- and will continue to consult -- with the Scientific Advisory Board in
order to discuss the status of the investigation and to receive the members’ input
and guidance. The members of the Scientific Advisory Board are:
Margaret Kuo retired as Deputy Director of Forensic Science Services after
27 years with the Orange County Sheriff-Coroner’s Office. Among other things,
Ms. Kuo has participated in or led approximately 30 crime laboratory inspections
or audits.
Douglas M. Lucas is the retired Director of the Centre of Forensic Sciences
of the Province of Ontario, Canada. Among his many leadership positions in the
forensic science community, Dr. Lucas is a past president of the American
Society of Crime Laboratory Directors (“ASCLD”) and has led approximately 13
accreditation inspections performed by the American Society of Crime
Laboratory Directors/Laboratory Accreditation Board (“ASCLD/LAB”), as well
as audits of 12 other crime laboratories.
Bruce W. Vander Kolk retired in 2001 as the Commander of the Illinois
State Forensic Sciences Command, where he oversaw the operations of eight
regional forensic science laboratories and a research and development
laboratory. During his career, Mr. Vander Kolk has, among other things, served
on the strategic planning committee as well as the Board of Directors of ASCLD.
Our team includes a Scientific Coordinator, Roger J. Bolhouse, who is
responsible for overseeing and coordinating the case reviews performed by our
forensic scientists. Mr. Bolhouse also is our primary expert in trace evidence
examination. He was an officer with the Michigan State Police (“MSP”) for 26
years, including 22 years in the MSP’s crime laboratory system. He retired in
2000 as Director of the MSP’s Grand Rapids Laboratory and currently is a
forensic scientist with Speckin Forensic Laboratories in Okemos, Michigan.
The following forensic scientists involved with the investigation have
been drawn from across North America and are experts in their respective fields.
2

The curricula vitae for each of the members of the Scientific Advisory Board and the
forensic scientist members of our investigative team are posted on our Web site.

HPD Crime Lab Independent Investigation

These scientists are responsible for conducting the case reviews during Phase II
of the investigation.
Jeanine M. Baisch, Ph.D., is the Director of the Research and
Development Laboratory, Orchid Identity Genomics in Dallas, Texas.
Robert P. Bianchi is the former Director of the Drug Enforcement
Administration Special Testing and Research Laboratory in McLean, Virginia.
Michael A. Evans, Ph.D., is the President and Chief Executive Officer of
the American Institute of Toxicology Laboratories located in Indianapolis,
Indiana.
Patricia P. Hamby has over 30 years of experience in forensic serology
and has been a criminalist in several law enforcement crime laboratories.
Edward E. Hueske is a firearms and toolmark expert who retired as the
Supervising Criminalist for the Arizona Department of Public Safety in 1996.
Karen L. Irish retired in 2003 as the Director of the Forensic Services
Section of the Baltimore County Police Department.
Rhonda Roby is the founder and Chief Executive Officer of Identity
Quest, LLC. Among other things, she is a former Technical Leader of the
Mitochondrial DNA Section with the Department of Defense DNA Registry,
Armed Forces DNA Identification Laboratory.
Michael Sinke spent 20 years as a forensic scientist with the Michigan
State Police Crime Laboratory and is a questioned documents examiner with
Speckin Forensic Laboratories
Theresa F. Spear has over 25 years of experience as a criminalist and
recently retired from the California Department of Justice, Bureau of Forensic
Services, where, among other things, she was a supervisor in the California
Criminalistics Institute’s Biology Program.
Rick W. Staub, Ph.D., has a doctorate in genetics and is a Senior Manager
for Forensics and Laboratory Director at Orchid Cellmark.
Mark D. Stolorow is the Executive Director for Forensic Science at Orchid
Cellmark and has been a forensic serologist for over 30 years.

HPD Crime Lab Independent Investigation

Status of the Investigation
I.

The Case Reviews

Pursuant to our agreement with the City and HPD, our investigation into
the management, operations, and performance of the Crime Lab and Property
Room is divided into two phases.
During Phase I, we investigated the current and historical operations and
practices of the Crime Lab and Property Room. Among other things, this phase
of the investigation was designed to lead, in consultation with HPD, to the
development of a detailed plan for the second phase of the investigation. We
provided our Phase II Plan to HPD on July 6, 2005.3
The second phase of our investigation involves reviewing samples of cases
analyzed by the Crime Lab during defined time periods. The samples have been
drawn from the seven forensic science disciplines applied in the Crime Lab -serology, DNA, controlled substances, toxicology, trace evidence, questioned
documents, and firearms. These cases are being reviewed by our team of
forensic scientists and evaluated with reference to the Crime Lab’s SOPs in place
at the time, as well as applicable standards and practices generally accepted
within the forensic science community during the time the Lab conducted its
examination of the cases.
During Phase I of the investigation, we reviewed the methodology by
which HPD arrived at its suggested sample size of 1,966 cases. We determined
that it would be prudent to consult with expert statisticians to develop our
sample populations. After advising HPD and gaining the approval of the
Stakeholders Committee, which oversees our investigation, we retained and
consulted with statisticians from PricewaterhouseCoopers LLP (“PwC”),
including PwC partners Dr. Jessica Pollner and Arthur Baines. With PwC’s
guidance, we developed appropriate sample sizes for the case reviews to be
performed by our forensic scientists in each of the following forensic science
disciplines:
•
•
•
3

Serology
DNA
Trace Evidence

•
•
•

The Phase II Plan is posted on our Web site.

Controlled Substances
Firearms
Toxicology

HPD Crime Lab Independent Investigation

For the last discipline, Questioned Documents, because of the relatively
small number of cases examined, we decided to review all of HPD’s questioned
documents cases since 1998. We also selected separate statistically-based sample
populations of the controlled substances cases analyzed by former HPD
Criminalists Vipul Patel and James Price, both of whom were involved in
instances of drylabbing in the Controlled Substances Section.
In light of the pervasive and serious problems in the serology and DNA
cases we reviewed last fall, in December 2005 we recommended to the
Stakeholders Committee and to HPD that we modify the scope and focus of our
case reviews in both of these areas.
With respect to serology, we suggested that (a) the relevant time period
for the reviews be extended earlier in time to January 1980, prior to when
Mr. Bolding began analyzing biological evidence in the Crime Lab, and (b) the
statistical sampling be abandoned in favor of reviewing every serology case
related to a conviction, either by guilty plea or trial verdict, from 1980 through
the early 1990s.
With respect to DNA, we also recommended against completing the
review of all the cases in the statistical sample as originally drawn in favor of
focusing on the cases identified by HPD and the Harris County District
Attorney’s Office for re-testing that either (a) had not yet been re-tested by an
outside laboratory or (b) had been re-tested and the Crime Lab’s original DNA
findings could not be confirmed. As we discussed in our previous reports, the
re-test project was designed to identify all cases in which DNA analysis was
performed and a conviction resulted, either by guilty plea or after trial.
These recommended changes reflect the fact that our initial serology and
DNA case reviews revealed such a high percentage of major issues and problems
that it made sense to focus on those cases that may have resulted in miscarriages
of justice rather than to continue using a random sampling technique, which
almost surely would have continued to show an unacceptably high rate of
analytical errors. The Stakeholders Committee and HPD approved both of these
recommendations, and, beginning in January 2006, we adjusted our selection of
serology and DNA case for review accordingly.
We have now completed all of the case reviews in the nine categories of
cases, with the exceptions of firearms and serology. The following chart
summarizes the information relating to our case reviews.

HPD Crime Lab Independent Investigation

Category of Cases

Completed
Reviews

DNA4

135

100%

Serology5

336

n/a

513

100%

James Price

342

100%

Vipul Patel

366

100%

Toxicology

396

100%

223

100%

Firearms

222

364

61%

Questioned Documents8

100%

2,288

2,430

94%

Controlled

Trace

Substances6

Evidence7

Total Cases (excl. serology)

Total Number of
Cases

Percentage of Reviews
Completed

Finally, we have begun the process of formulating recommendations for
HPD and the Crime Lab. In order to provide the most relevant and helpful
recommendations tailored specifically to the Crime Lab as it functions now, we
4

The DNA case reviews include all 18 death penalty cases involving DNA testing by the
Crime Lab.

As discussed in detail in the serology section of this report, we have been working with
HPD since January 2006 to identify all serology cases related to a conviction, by guilty
plea or trial verdict, during the years 1980 through 1993. To date, we have identified a
total of 899 serology conviction cases and have reviewed 256 of them. However, we have
only completed the process of identifying serology convictions for 1980 and 1987.
Accordingly, we cannot report a total number of serology cases to be reviewed. As more
fully discussed below, our review of serology cases now will focus on those cases related
to convictions involving defendants who currently remain incarcerated.

This figure includes 150 cases from our original controlled substances sample and 363
cases from a reconfigured sample that we prepared, in consultation with PwC, to target
substances analyzed in the Crime Lab other than marijuana and cocaine. In addition to
the samples, we also reviewed 50 “bulk” controlled substances cases, not reflected in the
above chart, in order to evaluate how the Crime Lab and Property Room handle bulk
seizures of controlled substances.

In addition to these 223 trace evidence cases, our trace evidence expert reviewed the trace
evidence component of 40 cases identified as serology or DNA cases.

In our Phase II Plan, we estimated that the total number of Questioned Documents
Section cases that we would review was approximately 200. This estimate was based on
the total number of cases identified on the Questioned Documents Section case log. Once
we began our case review, it became clear that only 91 of the cases on the Questioned
Documents Section case log involved substantive work that we could review.

HPD Crime Lab Independent Investigation

need to review the Lab’s current operations. The Crime Lab is now accredited in
most areas, with the significant exceptions of DNA analysis and trace evidence
examination. In addition, HPD and the Crime Lab have taken a number of other
steps since December 2002 to improve the quality of the forensic scientific work
performed in the Lab. Indeed, HPD expects that the Crime Lab will receive
accreditation and be approved to resume DNA testing in the near future. Our
review of current operations and our recommendations will be designed to help
the Crime Lab make the most of its fresh start and to solidify and build upon the
improvements it already has made.
II.

Interviews

As the case reviews have progressed, we have learned a great deal about
the operations of the Crime Lab and about the various problems experienced in
each of the sections of the Lab. We have periodically met with and interviewed
current Crime Lab personnel who worked in the Lab during the relevant time
periods to get answers to our questions and to obtain perspective on issues
relating to the Lab’s operations and work product. Current Crime Lab personnel
have been extremely cooperative with and helpful to this investigation.
However, we continue to experience difficulty obtaining the cooperation
of certain critical former Crime Lab personnel, including Mr. Krueger and
Ms. Kim. Mr. Bolding agreed to be interviewed twice during the first phase of
the investigation, but declined our most recent requests to speak with him. In
January 2006, Representative Kevin Bailey, the Chairman of the General
Investigating and Ethics Committee of the Texas House of Representatives,
generously offered us the use of the Committee’s subpoena power to compel
recalcitrant former Crime Lab employees to cooperate with this investigation.
This appeared to be the most promising vehicle for us to obtain information from
witnesses whose testimony has been unavailable. A number of discussions took
place over the course of the past several months; however, it now appears that
the Committee is not willing to provide us with its subpoena authority to assist
in our investigation of the Crime Lab. We very much appreciate Chairman
Bailey’s efforts to assist us in our investigation and regret that obstacles of
various kinds, the details of which have not been communicated to us, are likely
to prevent that from happening.
III.

The Innocence Project

From the outset of this investigation, we have been concerned about
whether any action will be taken to explore legal remedies for individuals who
may have been victims of inadequate or flawed work by the Crime Lab’s

HPD Crime Lab Independent Investigation

serologists and DNA analysts, including, in particular, persons currently serving
jail terms or on death row. That concern is shared by the City Council, the
Stakeholders Committee, and many members of the public. In February 2006,
we received a request from the Innocence Project that we provide information
about each of the major issue serology and DNA cases identified by our
investigation to the Innocence Project, the Texas Innocence Network, and the
Texas Center for Actual Innocence. HPD and the Stakeholders Committee
approved our doing so, and, in April 2006, we forwarded information to these
organizations about each of the 25 major issue DNA cases and 18 major issue
serology cases referred to in our Fourth Report. We will update the case
information we have provided to the Innocence Project and affiliated
organizations to include the additional major issue DNA and serology cases we
have identified since the issuance of our Fourth Report.9

Results of Phase II Case Reviews
I.

Process for Reviewing Cases

We have not altered the basic process for reviewing Crime Lab cases from
what we described in our Fourth Report. The investigative team performs many
tasks on site at the Crime Lab in HPD headquarters. We have had access to
Crime Lab staff, documents, and raw data, including autoradiographs and
photographs of DNA test strips. When necessary, we have reviewed available
underlying evidence10 to assess the reasonableness of the Crime Lab’s original
forensic science work. The Fried Frank legal team continues to work extensively

Not all of the major issue DNA and serology cases we have identified during the course
of our investigation may be appropriate for review by the Innocence Project and its
affiliates. For example, in some of the cases in which we identified unreliable forensic
science work, there was no known suspect or the suspect was not charged. In other
cases, the defendant has been released from prison or has passed away.

Consistent with the scope of our mandate, we have not re-tested any evidence. We
reviewed underlying evidence only in cases where information and documentation in the
Crime Lab file -- such as photographs, narrative descriptions of the evidence and the
analyses conducted, and laboratory notes -- are inadequate to permit us to assess the
reasonableness of the original forensic science work.

HPD Crime Lab Independent Investigation

with the forensic scientists and has managed the legal, technical, and
administrative aspects of the Phase II review.11
A.

Definitions of Major and Minor Issues

Our primary method of distinguishing between deficiencies identified
during our case reviews is to characterize them as “major issues” or “minor
issues.” These are defined terms that were developed at the beginning of the
case review process in consultation with our team of forensic experts. 12
1.

Major Issues

Major issues are matters that raise significant doubt as to the reliability of
the work performed, the validity of the analytical results, or the correctness of
the analyst’s conclusions. They can include significant errors in the testing,
evaluation, or handling of evidence or in the reporting of results; serious
omissions where an analyst failed to perform a critical examination or analysis;
or analytical work that was undocumented or insufficiently documented to
permit an outside reviewer to assess the basis of the analyst’s conclusions.13
2.

Minor Issues

Minor issues are matters that involve deviations from generally accepted
forensic science practices or from the Crime Lab’s SOPs, but do not appear to
raise significant doubt as to the reliability of the work performed, the validity of
the analytical results, or the correctness of the analyst’s conclusions. Some
deficiencies, though significant, are categorized as minor issues if we conclude
that they would not have had a material impact on the results of the forensic
scientific work in the case.
11

For a more detailed discussion of the Phase II case review process, see Fourth Report at
pages 8-11.

Examples of deficiencies that would be characterized as major and minor issues can be
found in our Fourth Report at pages 10-11.

In general, with respect to the prosecutions of any individual discussed in this report, our
investigation is limited to reviews of the original forensic scientific work performed by
the Crime Lab and the presentation of analysts’ findings in any related criminal
proceedings. We have not reviewed or considered other evidence, such as eyewitness
testimony or confessions, that might be available in such cases. We also make no
assessment as to the likely guilt or innocence of any of the suspects or defendants, or the
appropriateness of any punishment, discussed in this report.

HPD Crime Lab Independent Investigation

Quality Assurance Review

Distinctions between major and minor issues are sometimes subtle and
involve case-specific judgments. We have developed a quality assurance (“QA”)
process that enables the forensic scientists on the investigative team to confer
about such matters and reach consensus before any final determinations are
made regarding the appropriate classification for a particular deficiency. QA
reviews are performed in each forensic science area by members of the Scientific
Advisory Board. We perform a QA review for every case that has been
preliminarily identified as involving a major or minor issue. We also review
randomly-selected files from the case sample to confirm that cases are being
assessed consistently and in accordance with our review standards.
In the following sections, we present the findings of our case reviews
conducted subsequent to those described in our Fourth Report. We discuss the
case reviews in the following order: (1) serology, (2) DNA, (3) trace evidence,
(4) controlled substances, (5) firearms, and (6) toxicology.14
II.

Serology

The term serology refers to the study of blood and other body fluids,
particularly blood group interactions. The forensic serology practiced in the
Crime Lab during the 1980s and early 1990s primarily involved genetic marker
typing relating to the four nominal blood types -- A, B, AB, and O.15 For
example, if a bloodstain on an item of evidence is determined through ABO
typing to contain ABO type A factors, and a suspect is determined to be ABO
type A, then the suspect is included in the population of potential contributors to
the evidence. On the other hand, if the suspect is determined to be ABO type B,
then he or she is excluded as a potential donor of the evidence. Until it was
replaced by the growing use of DNA profiling technology in the early 1990s, the
typing tests used in forensic serology were the only techniques available to
14

Because our case reviews related to the Questioned Documents Section were completed
last fall and are discussed in detail in our Fourth Report, no additional discussion
regarding the Crime Lab’s examination of questioned documents is included in this
report.

As discussed in our Fourth Report, from the 1960s on, forensic serology also included the
identification of other biochemical genetic markers present in body fluids such as certain
polymorphic enzymes and proteins. Our case reviews suggest that the Crime Lab only
rarely used such enzyme testing results to associate or disassociate evidence stains with
respect to a victim or suspect. See Fourth Report at 14.

HPD Crime Lab Independent Investigation

forensic scientists to develop information as to whether specific individuals
might be associated with biological evidence -- such as blood, semen, or saliva -related to crimes, particularly homicides and sexual assaults.16
ABO typing is not limited to blood samples. In many people, ABO factors
also are present in other body fluids, such as semen, saliva, and vaginal
secretions. The population is divided into two groups with respect to the
presence or absence of ABO factors in body fluids other than blood.
Approximately 80% of the population has detectable levels of their ABO type in
their other body fluids and are known as ABO “secretors.” The remaining 20%
of the population lacks normally detectable levels of their ABO factors in their
secretions; these people are known as ABO “non-secretors.” Although useful in
the investigation of homicides and other crimes, ABO testing of body fluid
secretions was particularly valuable in the analysis of biological evidence related
to sexual assaults.
The first step in forensic serology is to determine through presumptive
testing whether biological material is present on evidence items collected from a
crime scene, a victim, or a suspect. This examination is conducted using various
presumptive tests that can indicate whether a stain is likely to contain blood or
semen and therefore be susceptible to genetic marker typing. The Crime Lab
typically would screen suspected bloodstains by applying a color test using the
chemical phenolphthalein, which reacts to the protein hemoglobin found in
blood.17 The Crime Lab also commonly used a confirmatory test for blood
known as the Takayama test, which is a micro-crystalline test also directed at
detecting the presence of hemoglobin. If these tests resulted in positive readings,
the stain would be confirmed to contain blood.
The Crime Lab generally used three types of tests to detect the presence of
semen in evidence related to suspected sexual assaults. First, a serologist would
use a microscope to attempt to visually confirm the presence of sperm cells
16

For ease of reference and to provide helpful background for the discussion of significant
issues we have identified in the serology work performed by the Crime Lab, we include
in this section a brief general description of serology and certain common testing
methods. For a more detailed discussion of technical aspects of serology, please refer to
Appendix B of our Fourth Report, “Discussion of Serology Techniques Used by the
Crime Lab.”

Hemoglobin is a protein found in red blood cells that is responsible for transporting
oxygen and carbon dioxide in the bloodstream. Hemoglobin gives blood its red coloring.

HPD Crime Lab Independent Investigation

extracted from an evidence sample and applied to a microscope slide. Second,
the serologist might use a presumptive color test on a cutting from an evidence
stain to detect the presence of acid phosphatase (“AP”), which is an enzyme
secreted by the prostate gland into seminal fluid. Finally, Crime Lab serologists
also used testing techniques to determine the presence or absence of a prostatic
protein called p30, which is unique to seminal plasma.
If these tests indicated that an evidence specimen in fact contained blood
or semen, then the serologist could attempt ABO genetic marker testing on the
evidence. Absorption elution (“AE”), a form of direct ABO typing, is the
generally accepted forensic serology testing method for determining the ABO
factors present in bloodstain evidence.18 Serologists sometimes used AE testing
of bloodstains in conjunction with a reverse blood typing technique called the
“Lattes Crust” test.19 Serologists often used Lattes testing to obtain ABO typing
results from scrapings of dried blood crust collected from hard surfaces, such as
glass or a weapon (hence the term “Lattes Crust” test).20 Absorption inhibition
(“AI”) was the generally accepted forensic serology method for determining
ABO factors expressed in stains related to body fluids other than blood -- such as
semen, saliva, vaginal secretions, perspiration, nasal mucous, or mixtures of
these fluids. AI, unlike AE, is an indirect method for the detection of ABO
antigens.21

AE is known as a “direct” ABO test because the agglutination observed as a result of the
antigen-antibody interaction in AE testing directly indicates which ABO antigenic
factors, if any, are present in the sample. In other words, in AE testing, the presence of a
specific ABO factor is indicated by observation of agglutination in the test well for that
ABO factor.

The Lattes Crust test is known as a reverse typing test because it detects the presence of
naturally occurring ABO antibodies in the plasma or serum portion of a bloodstain. The
ABO antibodies in a person’s serum are complementary to his or her ABO antigens
present on the person’s red blood cells. Thus, the detection of ABO antibodies through a
Lattes Crust test can assist the forensic serologist to infer the ABO type of the donor of a
bloodstain.

The Lattes Crust test, however, is less sensitive than AE. Consequently, more bloodstain
material must be consumed to conduct a Lattes Crust test than the amount needed for
AE.

AI is referred to as an indirect test because the presence of an ABO factor in an evidence
stain is determined by observation of a diminished level or absence of agglutination in
the test solution related to that particular ABO factor. AI also is used to test known
reference saliva standards obtained from a victim or suspect to determine whether he or
Footnote continued

HPD Crime Lab Independent Investigation

If an evidence stain tested positive for the presence of blood, semen, or
other body fluid, it was generally possible to determine the genetic
characteristics (e.g., ABO type) reflected in the evidence sample. The serologist
then could compare the genetic types of the evidence sample with those of
known reference standards. A known reference standard is a sample of blood or
saliva collected from a victim or a suspect who is potentially associated with an
evidence stain. Depending on the nature of the crime under investigation and
whether the evidence includes secretion stains, a crime laboratory might subject
known reference standards both to ABO testing, in order to determine the
victim’s or suspect’s ABO blood type, and to Lewis blood group testing, which is
helpful to predict or confirm whether the donor can be expected to be a secretor
whose ABO type is expressed in body fluids other than blood.22
During the 1980s and early 1990s, the Crime Lab compared the ABO types
of victims and suspects with ABO activity detected in evidence samples to
establish associations -- or exclusions -- between individuals and biological
evidence in only a small proportion of the cases sent to the Lab for analysis. As
discussed in detail below, we have found that the Crime Lab performed ABO
typing and comparisons in only a small percentage of “serology” cases. This was
true even where presumptive screening indicated the presence of blood or semen
in evidence stains and where there was a known suspect against whom ABO
typing results related to such evidence could and should have been compared.
The Crime Lab’s chronic failure to perform ABO typing and comparisons in
cases where the serology may have produced probative or even exculpatory
results is extremely disturbing.
A.

Status of the Serology Case Reviews

Our original sampling of serology cases was drawn from cases assigned to
analysts working in the Serology Section of the Crime Lab from 1987 through
1990. Based on our initial serology case reviews, we found that the Crime Lab
continued to perform ABO typing well into 1993, even after the Lab had
established its DNA analysis capability. We also found that our original sample
of serology cases, which was derived from the Crime Lab’s ledger of cases
Footnote continued from previous page
she is a “secretor” -- i.e., a person whose ABO type is expressed in body fluid other than
blood.
22

Lewis genes are related to an individual’s ABO secretor status. Lewis testing of a
person’s known reference blood sample may be used to infer one’s ABO secretor status.

HPD Crime Lab Independent Investigation

assigned to analysts in the DNA/Serology Section during the relevant time
period, included a large proportion of cases that did not involve any substantive
forensic science work by the Lab and, therefore, would not provide a basis to
assess the quality of the Lab’s serology work. For example, we found that, upon
receiving a sexual assault kit, the Crime Lab would assign a Lab number and
generate a Lab file. However, if no suspect was identified or no samples were
provided for comparison with the evidence in the sexual assault kit, the Crime
Lab typically would only inventory the contents of the sexual assault kit and
would not analyze the evidence.
In late 2005, in order to identify cases involving substantive analytical
work, such as ABO typing, performed by the Crime Lab’s serologists, we
developed a database of cases derived from raw data records that the Lab
preserved and we reconfigured our sample based on that database.23 As of our
Fourth Report, we had completed reviews of 80 of these substantive serology
cases. We identified major issues in 18 -- or approximately 22.5% -- of these
cases.
The case reviews we performed last fall revealed a number of serious
problems with the serology work performed by the Crime Lab. The five most
significant issues we found were:

•

The absence in the serologists’ reports of genetic profile frequency
statistics or any discussion of the significance of the statement that a
suspect could not be excluded as a potential donor of evidence samples;

•

The failure of serologists to use substrate controls and positive and
negative controls in their ABO typing;

•

The routine and common failure to report the results of testing and
probative findings;

•

The lack of any documentation of administrative or technical reviews of
the serologists’ work; and

The Crime Lab has not been able to locate and provide us with raw data related to the
work performed by all of the serologists employed by the Lab during the relevant period.
Most of the serology raw data records that have been preserved are those originally
maintained by Ms. Kim dating back only to 1989.

HPD Crime Lab Independent Investigation

•

The absence of generally accepted documentation and evidence control
procedures -- such as assignment of unique identification numbers to
items of evidence, descriptions of evidence, and preparation of complete
tables of testing results -- and errors by analysts in transferring their test
results to worksheets.

We also identified two very troubling cases -- related to defendants Dwight H.
Riser and Charles E. Hodge -- in which the Crime Lab reported incorrect
conclusions that were inconsistent with the actual ABO testing performed by the
analysts.24
In our initial serology case reviews last fall, we identified severe problems
with the Crime Lab’s serology work performed during the period 1987 to 1990.
As a result, in December 2005, we recommended to the Stakeholders Committee
and HPD that we modify the scope of the serology case reviews as follows. First,
because of the number of major issue serology cases we identified in which
Mr. Bolding was the responsible serologist, we recommended that the time
period of our case reviews be expanded to cover the years 1980 through 1993 in
order to include the entire period in which Mr. Bolding performed serology at
HPD as well as a short period prior to his involvement.25 Second, we suggested
that our reviews be limited to cases in which the Crime Lab’s serology work
related to a suspect who was convicted of the crime under investigation, either
by guilty plea or as a result of a trial verdict. We suggested changing the basis of
our case selection in serology from random sampling, which is designed to
provide a cross section of the work performed by the Crime Lab to permit
assessments of its overall quality, in order to concentrate our efforts on
systematically attempting to identify cases in which flawed serology work by the
Lab may have played a role in a criminal conviction. In short, because the
original serology work we reviewed appeared to be so deeply and pervasively
flawed, we believed that continuing with the random sample would have simply
further confirmed the conclusions that we had already reached.
24

Detailed discussions of the major issues we identified last fall, including the Riser and
Hodge cases, appear in our Fourth Report at pages 15-29.

Based on interviews with Mr. Bolding and his employment file with HPD, we estimated
that he probably began performing serology casework in the Crime Lab in approximately
the spring of 1981. See Third Report at 16. Our review of serology cases from the early
1980s has shown that Mr. Bolding actually began issuing serology reports in October
1980. Mr. Bolding did not attend a formal outside training program in bloodstain
analysis until the summer of 1982.

HPD Crime Lab Independent Investigation

HPD and the Stakeholders Committee approved our recommended
changes in the selection criteria for our serology case reviews, and we
implemented the revised approach in January 2006. We have faced significant
logistical challenges involved in identifying serology cases that might be related
to actual convictions of individual defendants, particularly in cases from the
early 1980s that pre-dated the Crime Lab’s use of an electronic case tracking
system. HPD has been fully supportive of our efforts and has devoted
significant investigative resources to assist us in identifying serology conviction
cases. Nevertheless, progress in identifying these cases has been slow.26
As of April 10, 2006, the only years for which we have completed the
identification of all serology conviction cases are 1980 and 1987.27 In 1980, there
were 408 convictions that can be tied to a Crime Lab serology case. For 1987, we
have identified 383 such cases. The status of our serology conviction case
reviews for each year during the period 1980 through 1991 is reflected in the
chart below.

We selected 1980 as one of the first years for which we would identify serology
conviction cases in order to make an assessment as to whether to recommend that the
review of serology cases be extended even further back in time. We selected 1987 as a
second year with which to start this process because it is the earliest year for which we
already had processed electronic case tracking data maintained by the Crime Lab in
order to create a database of potential serology cases.

HPD Crime Lab Independent Investigation

Status of Serology Convictions Case Reviews 1980 - 199128
Year

Serology Conviction
Cases Identified

1980

408

1981 – 1985

1986

1987

383

1988

1989

1990

1991

899

256

Totals

Reviews
Complete

Cases With
Major Issues

Our review of serology cases related to convicted suspects is limited by
the lack of documentation contained in the Crime Lab’s files. In most of the cases
from the 1980s, it is not possible to evaluate the serologists’ interpretation of the
ABO testing results they obtained because the laboratory notes and worksheets
contained in the file record only the analysts’ conclusions about the ABO factors
detected and provide no information about the raw test data that would indicate,
for example, the relative intensity of the agglutination upon which the analysts’
conclusions were based. Also, with the exception of several notebooks kept by
Ms. Kim beginning in 1989 and the 1990 p30 test log maintained by serologist
Grace Daz, there are no logs or other records reflecting the raw data related to
testing performed by the Crime Lab’s serologists, apart from the raw data record
occasionally present in the case file.29

To date, 1980 and 1987 are the only years for which all convictions where there is a
related Crime Lab serology case have been identified.

Because the raw data notebooks that we have obtained were prepared primarily by
Ms. Kim, we are able to evaluate the serology results she reported much more
thoroughly and effectively during the period for which we have such raw data than we
are able to with other analysts. Also, we have found that Ms. Kim was an extremely
prolific analyst and handled more cases than any other serologist in the Crime Lab,
including Mr. Bolding. For these reasons -- the volume of her casework and our ability to
more effectively review her results --many of the major issue cases we have identified
were analyzed by Ms. Kim. However, the problems we have observed with the Crime
Lab’s serology work, including the major issues, are endemic to the Serology Section as a
Footnote continued

HPD Crime Lab Independent Investigation

Finally, for cases analyzed during the early 1980s, Crime Lab serologists
did not report conclusions as to whether, based on ABO typing of evidence and
comparison to reference samples, a suspect was included (or excluded) as a
potential contributor to the evidence sample. Typically during this period, the
Crime Lab report only presented the results of ABO typing -- i.e., which ABO
factors, if any, were detected -- and included no interpretation with respect to
inclusions or exclusions. Therefore, in such cases it is not possible to evaluate
whether the serologist correctly expressed the significance of his or her ABO
typing results. Essentially, for most of the serology cases performed in the 1980s,
even in the relatively few cases where ABO typing actually was performed and
the results were reported, no conclusions were reported as to inclusion or
exclusion; consequently, we are able to evaluate only whether the serologist
accurately reported the test results recorded in the analysts’ worksheets.
This combination of factors -- the relatively few cases in which genetic
marker testing was performed, a paucity of documentation in the analysts’
worksheets, the absence of raw data records, and the lack of reported statements
regarding inclusions or exclusions based on ABO testing results -- likely explains
the relatively small number of cases involving major issues (only 4) that we have
identified among the 92 serology conviction cases from 1980 that we have
reviewed thus far. 30 In other words, the data in the files is so sparse that there is
insufficient information on which to base a thorough analysis. However, the
small number of major issues we have identified among the 1980 cases should
not be understood to mean that the work performed was more proficient than in
later years; it is simply a reflection of the fact that the analytical paper trail is so
thin that it precludes reaching such judgments.

Footnote continued from previous page
whole, and there is no reason to conclude that Ms. Kim’s performance as a serologist was
uniquely deficient.
30

In adopting our recommendation to expand the relevant time period of our serology case
reviews to 1980, HPD and the Stakeholders Committee left open the possibility of
expanding our serology case reviews even further to pre-1980 cases. Based on the lack of
documentation contained in the case files from the early 1980s, and our expectation that
documentation would only deteriorate the further back in time we search, we do not
believe it would be fruitful to review pre-1980 serology cases. Accordingly, we have
recommended to HPD that we not embark on a review of serology conviction cases prior
to 1980, despite our profound reservations about the overall quality of the forensic
serology work that might have been performed in those cases.

HPD Crime Lab Independent Investigation

We have now completed reviews of 336 serology cases performed by the
Crime Lab between 1980 and 1993. These include the 80 substantive serology
cases on which we reported in our Fourth Report and the 256 serology cases
related to investigations that resulted in a conviction by guilty plea or trial
verdict that we have reviewed since January 1, 2006.31 We have identified a total
of 50 serology cases involving major issues. 32
B.

Serious Problems Identified in Serology Cases

Our case reviews since January 2006 have reinforced our previous finding
of pervasive and serious problems with the quality of work performed by
serologists in the Crime Lab as well as with the presentation of the ABO
grouping results obtained by Lab analysts using various serology testing
methods. These problems are present in virtually every serology case we have
reviewed, even in those cases that we determined did not contain major issues.
Moreover, these very significant deficiencies are not the result of mistakes or
interpretive errors made by individual serologists. Rather, they are the product
of defective procedures employed in the Serology Section throughout the
relevant time period -- from 1980 through the early 1990s -- as well as the Crime
Lab’s systematic failure to adequately train and supervise its serologists.
We continue to observe the same serious deficiencies that we discussed in
our Fourth Report. In order to provide additional detail and illustrative
examples of the most significant problems with the serology work performed in
the Crime Lab during the 1980s and early 1990s, in this report we focus on the
following issues: (1) failing to perform potentially probative, or even
exculpatory, ABO typing in a large number of cases, particularly sexual assaults;
(2) failing to report probative ABO typing results due to an apparent reluctance
to report exclusions, findings of blood types not consistent with those of victims
or known suspects, and findings indicating the presence of the relatively rare
31

Because we are no longer basing our review of serology cases on a statistically-based
sampling methodology, it is not appropriate to report a major issue rate in terms of a
percentage of the cases we have reviewed. Moreover, a significant majority of the 256
serology conviction cases we have reviewed since January 1, 2006 did not involve any
ABO typing work performed by the Crime Lab. Because most of these cases involved
either no analysis by the Crime Lab or only presumptive screening for blood or semen,
they were very unlikely to contain any major issues.

In Appendix B to this report, “Serology Major Issue Cases,” we include the names of the
suspects and defendants related to each of the 50 cases we have identified thus far as
involving major issues.

HPD Crime Lab Independent Investigation

ABO type AB factors; (3) misinterpreting and inaccurately reporting ABO typing
results; (4) reporting test results that are unsupported by documented analysis;
and (5) generally poor forensic scientific practices, such as misapplying substrate
controls, reporting ABO typing results indicating no ABO activity as
“inconclusive,” and poor training or technique producing incorrect typing
results.
1.

Systemic Failure to Perform Potentially Probative ABO
Typing in Cases Involving Known Suspects

Since changing the methodology for our serology case reviews to focus on
serology cases handled by the Crime Lab in which there was a known suspect
who was convicted of a crime, we have completed, in conjunction with HPD, the
process of identifying such serology conviction cases for the years 1980 and 1987.
In 1980, there were 408 serology conviction cases, of which we have reviewed 92.
We have reviewed 77 of the 383 serology conviction cases identified for 1987.
Early in our review of the 1980 and 1987 serology conviction cases, we
found that the Crime Lab performed genetic marker analysis and comparison of
ABO blood type factors detected in evidence, on the one hand, with the ABO
types of victims and known suspects, on the other, in an alarmingly small
proportion of the cases. This is troubling because each of these cases, by
definition, involved a known suspect (whose name is reflected in the Crime Lab
report) who eventually was convicted of an offense related to the crime for which
evidence was sent to the Lab for analysis. We have categorized each of the
serology conviction cases from 1980 and 1987 that we have reviewed based on
the type of analysis applied.

HPD Crime Lab Independent Investigation

Type of Analysis Performed in Serology Conviction Cases
1980 and 1987
Category

Type of Analysis Performed

1980

1987

Inventory of Evidence Only
Screening for Blood or Semen
Results Positive
Screening for Blood or Semen
Results Negative
ABO Typing of Evidence Only
Comparison of Results of ABO
Typing of Evidence with Known
Reference Samples
Evidence Sent Out for DNA
Analysis

9 (9.8%)

10 (13%)

36 (39.1%)

26 (33.8%)

30 (32.6%)

13 (16.9%)

15 (16.3%)

5 (6.5%)

2 (2.2%)

23 (29.9%)

2A
2B
3
4
5

As reflected in the above chart, it appears that the Crime Lab performed
the full typing and comparison serology analysis necessary to develop probative
information as to whether a known suspect could be included -- or excluded -- as
a potential contributor to biological evidence in only a small fraction of the cases
where such analysis was possible. The Crime Lab’s failure to perform genetic
marker testing of evidence, such as bloodstains or swabs from a sexual assault
victim, is particularly disturbing in Category 2A above where presumptive
screening of evidence showed that body fluids, such as blood or semen, were
present in the evidence. In 30% to 40% of both the 1980 and 1987 serology
conviction cases we have reviewed, presumptive tests for blood or semen were
positive and there was a known suspect for comparison, and yet no genetic
marker analysis was performed. In a significant number of cases, ABO testing
was performed on evidence, and yet no testing to determine the victim’s or
suspect’s ABO type was performed so that a comparison to the evidence could
be made.
The Crime Lab’s failure to generate potentially probative ABO testing
results in cases where it was possible to conduct such testing and comparisons to
known reference samples were possible is very troubling. This failure has
implications both for ensuring that the guilty are convicted and that the innocent
are exonerated. From the perspective of making sure the guilty are convicted,
this data indicates that the Crime Lab routinely failed to develop information
that potentially could have guided investigators and strengthened the ability of
the investigators and prosecutors to associate suspects with evidence in the case.
From the perspective of making sure the innocent are exonerated, the Crime Lab
failed to perform genetic marker analyses that, in some cases, might have
excluded an individual suspect as a potential donor of evidence, such as semen

HPD Crime Lab Independent Investigation

stains related to a sexual assault. Particularly in light of a 1995 Federal Bureau of
Investigation (“FBI”) study which found that, between 1989 and 1995, suspects
were excluded by DNA testing in approximately 23% of cases,33 it is entirely
possible that properly performed ABO testing would have established a scientific
basis for excluding individual suspects in some percentage of these untested
cases.34
The Crime Lab’s chronic failure to conduct potentially probative genetic
marker analysis on available evidence in these major crimes against persons
reflects a troubling failure to fully and properly use the available tools of forensic
serology in the criminal justice system during the 1980s and early 1990s, thus
undermining the ability of the system to properly serve victims, suspects, and
the public.
2.

Failure to Report Probative Findings

Twenty of the major issue serology cases we have identified involve the
serologist’s failure to report the potentially probative ABO typing results that the
analyst in fact obtained, as reflected in raw data or in the analyst’s laboratory
notes or worksheets. We found in these cases an apparent reluctance on the part
of Crime Lab serologists to report typing results obtained from evidence that
were not consistent with the known ABO type of either a victim or a suspect.
Ethical standards in the practice of forensic science require full disclosure of all
relevant analytical test results without regard for the potential impact on an
investigation or prosecution. Our serology case reviews reflect many instances
of failure to report analytical results that would have weakened the prosecution’s
case or strengthened the case for exonerating the defendant.
Without speaking to the Crime Lab’s two central serologists active during
the 1980s and early 1990s -- Mr. Bolding and Ms. Kim -- about these cases, it is
33

See Convicted by Juries, Exonerated by Science: Case Studies in the Use of DNA Evidence to
Establish Innocence After Trial, United States Department of Justice, Office of Justice
Programs, National Institute of Justice (June 1996), at 20.

Obviously, the discriminatory power of ABO genetic marker testing is much weaker than
that of DNA analysis. Nevertheless, ABO typing can result in the exclusion of an
individual suspect where, for example, based on ABO blood typing of evidence and
comparison to the victim’s and suspect’s known blood types, the suspect can be
eliminated as a potential contributor because an ABO factor present in evidence is
foreign to both the victim and suspect, thereby indicating that a third person contributed
to the evidence sample.

HPD Crime Lab Independent Investigation

difficult to determine whether this pervasive pattern of avoiding the reporting of
results inconsistent with the victim or a known suspect is attributable to the
analyst’s lack of confidence in his or her ability to obtain reliable ABO typing
results or to scientific fraud of the most pernicious kind with the motive to secure
convictions rather than do justice. Regardless of the individual serologist’s state
of mind, however, such practices are intolerable in a forensic science laboratory
and undermine confidence in the integrity of the criminal justice system.
Below, we discuss examples of cases in which the Crime Lab obtained
potentially probative ABO typing results but did not report them.
a.

Serology in the Derrick L. Jackson Case35

Forrest Henderson was a singer with the Houston Grand Opera. His
friend, Richard Wrotenbury, was an elementary school music teacher and also
participated in the Houston Grand Opera. After Mr. Wrotenbury failed to
appear for work at the school on Monday, September 12, 1988, the manager of
the building in which Mr. Henderson and Mr. Wrotenbury shared an apartment
entered the apartment and discovered the men had been brutally murdered.
Both men had been beaten and stabbed, and there were bloodstains throughout
the apartment. Mr. Bolding accompanied HPD investigators to the crime scene
and personally collected blood samples from the apartment in order to perform
serological tests on the evidence.
According to entries in his worksheets, between September 15 and
September 23, 1988, Mr. Bolding performed ABO genetic marker tests on over 30
blood samples taken from various spots at the crime scene. He also typed
known reference samples from the victims and Calvin Dorne, HPD’s initial
suspect in the killings. Mr. Bolding found that both of the victims were ABO
type A and that Mr. Dorne was ABO type O. Mr. Bolding’s worksheets from
September 1988 indicate that he failed to observe any agglutination as a result of
AE testing of 13 of the bloodstains.36 The worksheets also reflect, however, that
35

Texas v. Jackson, Cause No. 748752 (Harris County, Tx.).

Mr. Bolding’s reported failure to observe agglutination in 13 of the bloodstain samples he
tested is suspect. In light of the quantity of bloodstain evidence in the apartment and the
relative freshness of the stains when Mr. Bolding tested them, it seems highly unlikely
that he would have failed to obtain interpretable results with respect to so many samples.
Indeed, bloodstains on three of these items of evidence actually yielded results when
subjected to RFLP DNA testing eight years later in 1996. Since RFLP testing requires a
much larger and higher-quality sample to yield results than does ABO testing, we
Footnote continued

HPD Crime Lab Independent Investigation

he was in fact able to obtain results indicating ABO type A activity, consistent
with the ABO blood type of both Mr. Henderson and Mr. Wrotenbury, in certain
samples taken from the bedrooms and bathroom in the apartment. In his Crime
Lab report dated March 15, 1989, Mr. Bolding reported that he detected “type ‘A’
human blood” in these samples.
Mr. Bolding’s September 20, 1988 worksheet also indicates that he
obtained strong Lattes results indicating ABO type O in a sample taken from the
“North Bedroom Door.”37 A separate AE test indicated ABO type B activity in
this sample. In his March 15, 1989 Crime Lab Report, Mr. Bolding correctly
reported the combined Lattes and AE results for this sample as “inconclusive.”
Mr. Bolding’s September 23, 1988 worksheet records that he also obtained an AE
result indicating ABO type B activity on a “Swab of N. Bedroom Door.” In the
March 15, 1989 Crime Lab report, however, Mr. Bolding failed to report that he
had found ABO type B activity in this swab sample from the north bedroom door
in the apartment. Instead, Mr. Bolding reported that “human blood having
inconclusive grouping activity was detected on samples from . . . bedroom door.”
The unreported ABO type B activity results that Mr. Bolding obtained from the
swab sample taken from the apartment’s north bedroom door were clearly
probative because ABO type B was foreign to both of the victims and to HPD’s
initial suspect, Mr. Dorne.
HPD’s investigation of the opera singers’ killings went cold. A break in
the case came in 1995 when the Harris County Sheriff’s Department acquired the
capability to use the Automated Fingerprint Identification System (“AFIS”),
which enables law enforcement agencies to compare unknown latent fingerprints
with a database of known prints. In April 1995, HPD submitted a latent
fingerprint taken over six years earlier from Mr. Henderson and
Mr. Wrotenbury’s apartment to the Sheriff’s Office to run through AFIS. Using
AFIS, the Sheriff’s Office obtained a match for the unknown fingerprint, and
Footnote continued from previous page
question whether the absence of agglutination recorded for these samples is either (a) a
product of poor analytical technique on the part of Mr. Bolding or (b) an instance of
drylabbing where Mr. Bolding may have recorded “no agg.” without having actually
performed ABO testing on the evidence.
37

Mr. Bolding’s original worksheets from September 1988 do not assign identifying
numbers to individual items of evidence. Rather, the worksheets identify individual
items of evidence only by Mr. Bolding’s descriptions of where the sample was taken from
the apartment.

HPD Crime Lab Independent Investigation

Derrick Leon Jackson was identified as a suspect in the killings of Mr. Henderson
and Mr. Wrotenbury.
On April 26, 1995, HPD investigators submitted a blood sample taken
from Mr. Jackson to the Crime Lab for analysis. Mr. Bolding analyzed the
reference sample and determined that Mr. Jackson was ABO type B. At some
point, Mr. Bolding revisited the ABO typing work he had performed over six
years earlier. Without performing any additional ABO testing on the evidence
samples and without dating his changes to the raw data worksheets, Mr. Bolding
wrote the finding of “’B’ act” on the September 20, 1988 and September 23, 1988
worksheets under the columns related to the two samples taken from the north
bedroom door. 38 Mr. Bolding previously reported his results with respect to
both of these samples as “inconclusive” in the March 15, 1989 Crime Lab report.
On June 28, 1995, Mr. Bolding issued another Crime Lab report in which he
stated: “Derrick Jackson was determined to have type ‘B’ blood” and “human
blood having type ‘B’ activity was present on two samples taken from the north
bedroom door.”
In March 1998, Mr. Jackson was tried for capital murder in connection
with the slayings of Mr. Henderson and Mr. Wrotenbury. On March 11, 1998,
Mr. Bolding testified about his serology work. On cross examination, there was
the following exchange between Mr. Bolding and defense counsel regarding
Mr. Bolding’s ABO typing results related to samples from the north bedroom
door:

But you did specifically say that in direct examination
that you found Type B blood on the door, is that
correct, on the north bedroom door?

Yes, sir.

On your original report generated in 1988 –

We have located copies of both the original worksheets prepared in September 1988 and
the worksheets that were altered to include new conclusions about Mr. Bolding’s original
ABO testing results. The original worksheets were obtained from microfilmed archives
of the Crime Lab file related to the serology work performed in 1988. The amended
worksheets were found in the paper file related to the DNA analysis that was performed
by the Crime Lab in 1996 after Mr. Jackson was identified as a suspect. The DNA
analysis in Mr. Jackson’s case is discussed in the DNA section of this report.

HPD Crime Lab Independent Investigation

Yes, sir.

-- do you show that anywhere?

No, sir, I don’t.

So you took those samples – and this report was
generated at the time you evaluated those samples; is
that correct?

Yes, sir.

And, so, at the time you generated those samples, you
were inconclusive about that; is that correct?

Yes, sir.

So, now nine years later, you’re saying Type B; is that
correct?

The Type B activity is basically what I’m saying.39

Although Mr. Bolding testified about changing his conclusions with
respect to his ABO typing results of the evidence samples taken from the north
bedroom door, defense counsel did not question Mr. Bolding as to why he
reported his original findings as “inconclusive” or the basis for his amended
findings. Although we cannot draw any firm conclusions about Mr. Bolding’s
motivations based on the information currently available to us, the most obvious
explanation is that Mr. Bolding decided not to report his original ABO typing
result finding ABO type B activity in the swab sample from the north bedroom
door because that result was not consistent with either the ABO types of the
victims or Mr. Dorne, HPD’s initial suspect. Rather than report his serology
work suggesting that someone other than the victims or Mr. Dorne bled in the
room, Mr. Bolding reported these findings as inconclusive. Over six years later,
after Mr. Jackson was identified as the suspect in the killings and his blood ABO
blood type was determined to be consistent with Mr. Bolding’s earlier
unreported findings, Mr. Bolding amended his worksheets and issued a
supplemental report to reflect a consistency between the evidence and HPD’s
current suspect. This case is a troubling example of the head of the Crime Lab’s
39

Jackson Tr., Vol. 23, at 181:15-182:11.

HPD Crime Lab Independent Investigation

Serology Section seemingly tailoring his reported results to fit with investigators’
pre-existing expectations.40
b.

Additional Examples of the Crime Lab’s Failure to
Report Probative Findings

As discussed above, the failure of Crime Lab serologists to report
potentially probative findings is not isolated to a few cases or to a single rogue
analyst. We have identified 18 cases in which it appears that serologists failed to
report seemingly reliable results obtained from ABO typing or other genetic
marker analysis. These cases indicate a disturbing pattern of selective reporting
and disregard for scientific integrity in the Crime Lab’s Serology Section.
In 1990, Ms. Kim performed serological analysis, involving ABO typing as
well as enzyme analysis,41 in connection with the investigation of the sexual
assault and murder of Norma Torres. Ms. Kim received reference samples from
Ms. Torres, Ms. Torres’s boyfriend, and the suspect, Roland Salazar. Ms. Kim
determined that the victim, boyfriend, and suspect were all ABO type A with
EsD and PGM 1 enzyme markers.42 Ms. Kim also determined through Lewis
testing that the boyfriend and Mr. Salazar were both secretors whose ABO type
could be expected to be expressed in body fluids such as semen.
In her November 15, 1990 report, the only typing results that Ms. Kim
reported with respect to the evidence samples she tested was that “human blood
having ‘A’ activity was detected on the bedspread, and yellow/purple shirt.”
The ABO type A activity detected in these bloodstains was consistent with the
ABO type of the victim and both the boyfriend and Mr. Salazar.
40

On March 12, 1998, Mr. Jackson was convicted of capital murder, and he is currently on
death row. Our review does not question the reliability of Mr. Bolding’s original
serology tests that determined ABO type B activity to be present on the north bedroom
door of the crime scene. Rather, the issues illustrated by this case are selective reporting
by Mr. Bolding and his failure to disclose results that potentially exonerated HPD’s initial
suspect, Mr. Dorne, when Mr. Bolding originally obtained them.

As discussed in our Fourth Report, although the Crime Lab has retained electrophoresis
logs reflecting the results of enzyme testing, serologists in the Lab rarely reported the
results of this testing. This case is an exception.

PGM (phosphoglucomatase) and EsD (esterase D1) are enzyme proteins found in blood.
Because these and other enzymes and proteins are polymorphic, meaning they exist in
different forms, serologists found the analysis of these enzymes useful in distinguishing
between the blood of different individuals.

HPD Crime Lab Independent Investigation

Ms. Kim also reported that “semen was detected on the vaginal and rectal
swabs, fitted bed sheet, top bed sheet, peach colored rag and blue striped towel.”
She did not report any typing results related to the semen stains she found on
these items of evidence. However, her raw data logbook and worksheets
indicate that Ms. Kim detected the presence of ABO types A and B activity on the
stains from the fitted bed sheet, peach colored wash rag, and blue striped towel.
The unreported ABO type B activity that Ms. Kim detected in these items of
evidence was probative, and potentially exculpatory, because it was foreign to
the victim and to both the boyfriend and Mr. Salazar. In other words, Ms. Kim
failed to report serology findings indicating that, if there were a sole semen
donor, the semen stains on these three items of evidence could not have
originated from either of the two men.43
After Ms. Kim completed her serology work, HPD sent eleven samples
related to this case to the Kleberg laboratory at the Baylor College of Medicine
for DNA testing. On October 17, 1991, the Kleberg laboratory reported that the
DNA pattern obtained from the semen stain on the peach colored wash rag “is
not consistent with the DNA pattern from either the suspect Roland Salazar or
from” the boyfriend.44 These DNA typing results were consistent with Ms. Kim’s
unreported ABO typing results, indicating that someone other than Mr. Salazar or
the boyfriend contributed to the semen stain on the wash rag.45
In a 1986 sexual assault case in which the original serology work was
performed by Crime Lab analyst David Coffman, we found another example of
the Crime Lab obtaining a potentially probative ABO typing result but failing to
report it. Mr. Coffman’s original serology work in this case was performed
before any suspect had been identified. In his October 20, 1986 report,
Mr. Coffman reported only the results of his presumptive screening for blood
43

Ms. Kim’s unreported finding of ABO type B activity could not exclude Mr. Salazar or
the boyfriend because ABO type A activity, common to both men (as well as the victim, if
she is determined to be a secretor), also was detected. At a minimum, however,
Ms. Kim’s unreported results indicate that there was a contributor to the stains on the
fitted sheet, wash rag, and towel other than the victim, the boyfriend, or Mr. Salazar.

The Kleberg laboratory was not able to perform DNA testing on samples from the fitted
bed sheet or blue striped towel because either no DNA could be extracted from the
samples or the DNA was degraded.

Although the homicide happened over fifteen years ago, the Harris County criminal case
tracking system indicates that Mr. Salazar was arrested on March 30, 2005 in connection
with the murder of Ms. Torres and that he is currently awaiting trial.

HPD Crime Lab Independent Investigation

and semen on various items of evidence, including the detection of human blood
on a pair of white tennis shoes belonging to the victim. Mr. Coffman did not
report that, according to his worksheet dated July 21, 1986, he determined
through an AE test that the bloodstain on the victim’s tennis shoes contained
ABO types A and O activity, which indicated an ABO type A donor.
Mr. Coffman also failed to report that he had detected ABO type O activity on
the vaginal swab contained in the sexual assault kit.
On March 11, 1987, a worksheet prepared by Ms. Kim indicated that she
performed ABO testing on reference samples obtained from the two suspects
who by then had been identified as Leonard N. Francis and Benjamin Silma Mao,
a reference sample from the victim, and the vaginal swab from the rape kit.
Ms. Kim found that the victim was an ABO type O secretor, Mr. Francis was
ABO type B, and Mr. Mao was ABO type O. Ms. Kim also detected ABO type O
activity on the vaginal swab, which was consistent with the results of
Mr. Coffman’s analysis.
In her February 9, 1988 report, however, Ms. Kim reported only that
Mr. Francis was ABO type B, that Mr. Mao was ABO type O, and that ”type ‘O’
secretor activity was detected on the vaginal swab previously submitted.”
Ms. Kim did not report that the victim was an ABO type O secretor and that,
therefore, no ABO factors foreign to her were detected on the vaginal swab.
Accordingly, because no ABO factors foreign to the victim were detected on the
swab, no semen donor in the male population could be excluded as a contributor
of the semen on the swab -- a fact Ms. Kim did not report. Even more
significantly, neither Mr. Coffman nor Ms. Kim reported the type A activity
Mr. Coffman detected in the bloodstain on the tennis shoes, which was foreign to
the victim and to both of the suspects, Mr. Francis and Mr. Mao.46 The
unreported ABO type found on the tennis shoe certainly was probative in that it
suggests that an individual other than Mr. Francis or Mr. Mao could have been
involved in the assault.

On March 23, 1987, Mr. Mao pleaded guilty to aggravated robbery and was sentenced to
45 years in prison. The kidnapping and sexual assault charges against him were
dismissed. Similarly, Mr. Francis pleaded guilty to aggravated robbery on April 9, 1987,
and the rape, kidnapping, and sexual assault charges against him were dismissed.
Mr. Francis was sentenced to life in prison. Curiously, Ms. Kim’s Crime Lab report is
dated February 9, 1988, nearly a year after these guilty pleas.

HPD Crime Lab Independent Investigation

Misinterpretation and Inaccurate Reporting of Results

We have reviewed 17 cases that contain major issues related to the
misinterpretation of serology test results and the inaccurate recording and
reporting of results. These errors include mistakes in transferring data from test
run logbooks to worksheets and from worksheets to Crime Lab reports. These
are the types of errors in interpretation, documentation, and presentation that
would have been detected, and presumably corrected, if there had been an
effective supervisory control and quality assurance system in the Serology
Section. As discussed in our earlier reports, however, it is clear that neither
Mr. Bolding nor anyone else routinely reviewed the work performed by the
Serology Section in order to identify technical issues related to testing and
interpretation of results or, for administrative purposes, to ensure that work
performed was adequately and accurately documented. The cases discussed
below are illustrative examples of the errors that went unchecked in the Crime
Lab as a result of the absence of such supervisory and quality controls.
In the 1990 sexual assault case involving a suspect named Jose Luna, the
victim reported that a man she later identified in a lineup as Mr. Luna broke into
her motel room and raped her. Ms. Kim determined that the victim was an ABO
type O secretor and that Mr. Luna was an ABO type O non-secretor, meaning
that his ABO type is not expressed in his body fluids such as semen. Ms. Kim
performed ABO typing on various items of evidence, including a vaginal swab
from the victim’s sexual assault kit, two stains on the motel bed sheets, and a
white towel. The chart below compares the results Ms. Kim obtained through
ABO testing on these items of evidence, as reflected in her raw data logbook,
with the results she recorded in her worksheets and with the results she reported
in her December 3, 1990 Crime Lab report.

HPD Crime Lab Independent Investigation

Evidence Sample

Kim’s Logbook47

Case Worksheet

December 3, 1990
Report

Vaginal Swab

A and O activity

Weak A activity

No activity

Sheet Stain #2

A and B activity

A activity

Sheet Stain #3

A and B activity

A activity

Towel

A activity

A and O activity

Ms. Kim’s reporting of the vaginal swab as “no activity” is particularly
troubling because the actual results recorded by Ms. Kim in her logbook are
particularly probative with respect to the investigation of Mr. Luna. The ABO
type A activity that the raw data shows Ms. Kim detected on the vaginal swab is
foreign to the victim. Therefore, if the case involved only one possible semen
donor (i.e., only one assailant and no recent consensual partners), then Mr. Luna
would be excluded as a potential contributor to the sample on the vaginal swab
since the assailant would be ABO type A.48 Although this is clearly a case of
misreporting the results of ABO testing, we cannot determine on the basis of
currently available information whether the errors are the result of unintentional
mistakes in recording the results reflected in raw data or, as appears more likely,
selective reporting of serology results. At a minimum, however, HPD lacked an
effective quality control regime to detect and correct the inaccurate reporting of
results illustrated by the Luna case.49
In the 1989 sexual assault case involving a suspect named Roy Anthony
Qualls, we observed another instance in which the reported ABO typing results
were not consistent with the results reflected in the serologist’s raw data notes.
In his July 15, 1989 report, Mr. Bolding reported both the victim and
Mr. Qualls as being ABO type A secretors. However, Mr. Bolding’s July 6, 1989
raw data worksheet reflects that Mr. Qualls was an ABO type B secretor.
Mr. Bolding also reported that “semen was detected on the vaginal swab and
47

The results reflected in this column of the chart are our interpretation of the ABO
agglutination intensities recorded by Ms. Kim in her raw data logbook.

If there were the possibility of more than one assailant or a recent consensual sexual
partner, then Mr. Luna could not be eliminated by the ABO typing results related to the
vaginal swab simply on the basis of his non-secretor status; however, he could not be the
sole contributor of the semen detected on the evidence sample..

Mr. Luna pleaded guilty to burglary on May 1, 1991, and the original sexual assault
charge against him was dismissed. He was sentenced to twelve years in prison.

HPD Crime Lab Independent Investigation

smear” from the sexual assault kit and that “the vaginal swab examined
contained type ‘A’ grouping activity.” This statement in the report also is
contradicted by the typing results reflected on Mr. Bolding’s raw data
worksheets. The raw data shows that Mr. Bolding’s AI test results relating to the
vaginal swab were negative for ABO type A and type H [O]50 activity and that
the results for type B activity were uninterpretable due to lysis (rupture of the
cell wall) of the type B indicator cells. The bottom line is that Mr. Bolding’s
reported ABO typing results, which suggested a potential association between
Mr. Qualls and genetic material on the vaginal swab, were unsupported and
contradicted by his recorded ABO test results. In fact, the raw data demonstrates
that the results of Mr. Bolding’s ABO tests were inconclusive as to whether
Mr. Qualls was a potential contributor to the semen sample on the vaginal
swab.51
4.

Reported Findings Unsupported by Documented Analysis

We also have reviewed 5 serology cases where the findings reported by
the Crime Lab analyst were not supported by documented analytical work
reflected in the case file. In such cases, the absence of documented work
supporting conclusions reported by Crime Lab serologists is a gross deviation
from generally accepted forensic science practices. If the reported conclusions in
fact were not supported by analytical work, then those findings are the product
of scientific fraud.
In the 1987 sexual assault case in which Patrick Dewayne Tyler was the
suspect, the Crime Lab reported results that were not supported by documented
analysis. In a report dated March 17, 1987, Ms. Kim stated that “semen was
detected on the vaginal smear” contained in the sexual assault kit. According to
a worksheet in the case file, this conclusion was based on a microscopic
examination of the vaginal smear that found sperm cells. The worksheet does

There is no common human antibody against ABO type O blood cells. Therefore, in AI
testing, serologists use an extract from gorse seeds, Ulex Europeus, to cause type O cells to
agglutinate. The seed extract, called lectin, agglutinates the H antigen found on all ABO
cells, but the agglutination occurs in much higher concentration in the presence of type O
cells. Therefore, the reaction to lectin observed in AI testing indicates type H antigenic
activity, from which ABO type O activity is inferred.

Mr. Qualls pleaded guilty to the lesser offense of assault and was sentenced to five
months in prison.

HPD Crime Lab Independent Investigation

not reflect that an AP or p30 screening test for semen was performed on the
vaginal swab submitted to the Crime Lab.52
On July 13, 1987, upon a motion made by Mr. Tyler, the court ordered that
genetic marker analysis be performed on the evidence and compared to a known
reference sample provided by Mr. Tyler. The Crime Lab, however, never
performed ABO testing on the vaginal swab.53 On September 24, 1987, Ms. Kim
issued a second report stating that “[n]o semen was detected on the vaginal
swab.” There is no documentation in the case file indicating that Ms. Kim
performed any work to arrive at the conclusion that no semen was present on the
vaginal swab. Nor does the report or documentation in the file provide any
explanation as to why an AP or p30 test for semen was not attempted on the
vaginal swab. It is not clear whether this is a case where work was performed,
but not documented, or whether it is an example of drylabbing54 that effectively
denied a defendant the genetic marker testing for which he had petitioned and
that the court had ordered.55
5.

Other Examples of Poor Forensic Science Practices

Our serology case reviews also have revealed cases involving specific
flawed scientific practices that -- while consistent with an overall culture
characterized by unreliable reporting of serology results and the absence of any
effective quality control regime -- we have not observed to be present across
multiple cases. Examples of such poor scientific practices are misapplication of
substrate controls, misunderstanding the significance of a “no activity” finding
in serological analysis, and incorrect results that can be directly attributed to
poor training in serology analytical techniques.

The worksheet does indicate, however, that an AP test for semen was performed on a
pair of undergarments submitted to the Crime Lab and that the test was negative.

The vaginal smear slide Ms. Kim observed microscopically to contain sperm cells likely
was made from the same vaginal swab she later reported as negative for semen. The
inconsistency between these reported results is difficult to reconcile.

“Drylabbing” is a colloquial term for a form of egregious scientific fraud involving the
fabrication and reporting of scientific results for tests that actually never were conducted.

Mr. Tyler had been charged with indecency with a child. On October 23, 1987, he
pleaded guilty to burglary and was sentenced to eight years in prison.

HPD Crime Lab Independent Investigation

In our Fourth Report we discussed our finding that Crime Lab serologists
generally failed to run substrate56 controls in connection with AE and AI tests for
ABO activity, which is a significant departure from generally accepted forensic
science practices.57 A forensic serologist must use substrate controls to determine
whether the ABO factors detected in the questioned stain were part of the
background material -- i.e., were contained in the substrate before the questioned
stain was deposited on the substrate material -- rather than present in the body
fluid evidence being tested. If background ABO factors are detected in the
substrate control, the significance of the presence of those same factors in the
questioned stain must be taken into consideration in the interpretation of the
ABO factors detected. The appropriate response to a failure of the substrate
control usually is to report the failure, avoid interpreting the results of the initial
test, and repeat the test whenever possible.
In a 1989 sexual assault case involving a suspect named Porfino
Ayarzagoitia, we found a rare instance in which the Crime Lab serologist, in this
case Ms. Kim, actually used a substrate control. However, Ms. Kim failed to
properly apply the control. The substrate control run by Ms. Kim exhibited ABO
type A activity. In light of this failure of the control, the results of the AI test
should not have been reported. Nevertheless, Ms. Kim did not report the failed
substrate control from the A, B, and H[O] activity detected in the stain and chose
not to repeat the AI test. Instead, she simply did a subtraction and reported the
ABO types B and O activity on the stain without disclosing the detection of the
type A activity in the substrate control and the evidence stain. This represents an
egregious violation of principles of AI interpretation and alone would raise
significant doubt about Ms. Kim’s competence as a serologist.58
We have seen two cases in which serology was attempted by a DNA
analyst who was not trained in the interpretation of serology results. One is a
1992 sexual assault case analyzed by Joseph Chu. Mr. Chu determined that the
56

Substrate material is the fabric or surface upon which the questioned stain was
deposited.

See Fourth Report at 20.

Another major issue in this case is that Ms. Kim erroneously reported finding “no
activity” on the vaginal swab. This finding is contradicted by Ms. Kim’s raw data, which
indicates that she detected ABO type A activity on the vaginal swab, which is foreign to
both the victim and Mr. Ayarzagoitia who were both ABO type O secretors. These
results would have eliminated Mr. Ayarzagoitia as a sole contributor to the semen
detected on the vaginal swab.

HPD Crime Lab Independent Investigation

victim was an ABO type O non-secretor and that the suspect was an ABO type A
secretor. Mr. Chu’s ABO testing found no antigenic activity on the vaginal swab
contained in the sexual assault kit. Mr. Chu, who was trained as a DNA analyst
and not a serologist, appears not to have understood that a “no activity” result is
significant in ABO typing. For example, in this case, where the victim was a
non-secretor and the suspect was a secretor, a “no activity” result would be
consistent with the absence of any foreign ABO activity on the swab and would
either eliminate the suspect as a potential contributor of the semen or, if the
semen in fact originated from a secretor (such as the suspect in this case), indicate
that semen was present in an amount below the detectible level. Mr. Chu,
however, incorrectly interpreted the absence of antigenic activity on the vaginal
swab as an “inconclusive” test result.59
We also found a 1993 case in which Baldev Sharma, a DNA analyst,
attempted to perform ABO typing on a bloodstain located on a shirt worn by a
man suspected of robbing and beating the victim with a tire iron. Although no
serology report was issued in this case, Dr. Sharma’s worksheet reflects that his
ABO typing results would have excluded the victim as the source of the
bloodstain on the suspect’s shirt and, therefore, would have failed to establish a
potential association between the shirt worn by the suspect and the victim.
Subsequent DNA testing, however, concluded that the blood on the suspect’s
shirt matched the victim’s DNA profile. It appears that Dr. Sharma, who was not
formally trained as a serologist, failed to obtain reliable ABO typing results.
III.

DNA

Similar to the serology techniques that preceded it, forensic DNA profiling
of evidence samples involves the analysis of genetic markers to ascertain
associations among suspects, victims, and evidence. The uniqueness and
durability of DNA make it ideal for use by forensic scientists, and DNA profiling
has many advantages over earlier conventional serology procedures. In addition
to the immensely improved discriminatory power of DNA profiling, the DNA
molecule itself is a particularly robust test target compared to the less stable
genetic markers involved with serology. Another significant advantage of DNA
testing is the ability to use a technique called differential extraction by which the
sperm (male) components of a mixture can be separated from the epithelial
(female) components. Differential extraction is, therefore, extremely useful in
59

Testing by an outside laboratory in 1993 found no DNA foreign to the victim on the
vaginal swab.

HPD Crime Lab Independent Investigation

typing DNA evidence in sexual assault cases because it is frequently capable of
producing two separate DNA extracts that can be used to produce profiles that
represent the DNA types of the female and male contributors to a mixed
sample.60
The first step in DNA analysis is to determine whether a body fluid stain,
potentially containing DNA, is present on the evidence items. Forensic scientists
perform preliminary screening tests to determine what type of body fluids are
present. After a sample is identified as blood or semen, several techniques may
be used to extract DNA from the evidentiary sample. With mixed specimens
such as those typically found in sexual assault cases, a differential extraction
procedure is used to separate the “male” from the “female” components of the
mixture, which are then purified and analyzed separately.
Restriction fragment length polymorphisms (“RFLP”) analysis was used
in crime laboratories until the mid-1990s. The RFLP analysis process, while very
discriminating, is time consuming and requires a relatively large amount of
non-degraded, high molecular weight DNA.
DNA profiling technology made a major advance in the late 1980s with
the development of a technique known as polymerase chain reaction (“PCR”),
which is an amplification process designed to copy or multiply specific segments
of DNA. Development of the PCR process gave forensic scientists the ability to
analyze much smaller quantities of DNA and made DNA profiling possible in
some cases involving sample amounts too small or too degraded for effective
RFLP analysis. The early PCR-based methods used in the Crime Lab were
known as DQ Alpha, Polymarker, and D1S80.
The most common form of DNA typing used today is a form of
PCR-based typing based on markers known as STRs (“short tandem repeats”).
STRs are regions of human DNA that contain a series of short repeated units.
The forensic science community in the United States has standardized DNA
typing using a set of 13 core STR loci. This set of 13 core STR loci is used for

For ease of reference and to provide helpful background for the discussion of significant
issues we have identified in the DNA work performed by the Crime Lab, we include in
this section a brief, general description of certain forms of DNA testing. For a more
detailed discussion of the technical aspects of DNA profiling, please refer to Appendix C
of our Fourth Report, “Discussion of DNA Profiling Technology and Techniques Used by
the Crime Lab.”

HPD Crime Lab Independent Investigation

entry into the national DNA profiling database known as the Combined DNA
Index System (“CODIS”),61 which is managed by the FBI.
DNA profiles obtained from biological evidence samples can inculpate a
suspect associated with that evidence with a high degree of scientific certainty.
The statistical meaning of comparisons between DNA profiles developed from
known reference samples and the DNA profiles developed from evidence items
must be properly calculated and routinely reported in the laboratory reports
prepared by DNA analysts. The true significance of a DNA “match” cannot be
properly conveyed without an appropriate estimate of how rare that profile is in
the human population. Because the frequency of occurrence of the DNA profile
conveys the weight of this evidence, it must be presented accurately and clearly
by the DNA analyst.
A.

Status of the DNA Case Reviews

We originally drew our sample of DNA cases to be reviewed during
Phase II from cases analyzed in the Crime Lab from 1991 through the closure of
the DNA Section in 2002. Similar to our experience with serology cases, we
found that our original sample of DNA cases included a large number of cases
that did not involve substantive forensic scientific work by the Crime Lab and,
therefore, would not provide a basis to assess the quality of DNA analysis
performed in the Lab. In order to identify cases involving substantive analytical
work, we developed a database of cases derived from raw data records
maintained by the Crime Lab; then, with the assistance of PwC, we modified our
sample based on that database. Through this process, we identified a total of
1,288 “substantive” DNA cases, from which PwC developed a sample of 296
DNA cases.62 Last fall, we also began reviewing all 18 death penalty cases that

CODIS is a system that “enables federal, state, and local crime labs to exchange and
compare DNA profiles electronically, thereby linking crimes to each other and to
convicted offenders.” CODIS is a hierarchical database with three tiers -- the National
DNA Index System (NDIS) is the highest tier, with state (SDIS) and local (LDIS)
databases flowing into it. See www.fbi.gov/hq/lab/codis/brochure.pdf.

As described in the Phase II Plan, the original sample size PwC developed for DNA cases
(which was derived from a population including both substantive and administrative
cases) totaled 358 DNA cases.

HPD Crime Lab Independent Investigation

involved DNA analysis by the Crime Lab as well as cases that have not yet been
confirmed through the post-conviction re-testing process.63
At the time of our last report, we had completed reviews of 67 DNA cases,
including all 18 of the DNA death penalty cases except for the case of Derrick L.
Jackson.64 We identified major issues in 27 of these cases, or approximately 40%
of the DNA cases we had reviewed prior to the publication of our Fourth Report.
Three death penalty cases were identified as having major issues.65 The most
significant and pervasive issues we identified in the DNA cases we summarized
in our Fourth Report were:
•

Failure to report typing results, including potentially exculpatory results.

•

Prevalence of low quality analytical results, particularly with respect to
PCR-based DQ Alpha, Polymarker, and D1S80 testing, likely attributable
to some combination of poor technique on the part of the Crime Lab’s
DNA analysts and contamination.

•

Misleading reporting of the statistical significance of the Crime Lab’s
DNA profiling results, particularly in cases involving mixture evidence.

•

Failure to use and show proper regard for scientific controls, especially
negative controls in PCR testing, and failure to compare typing results at

In early 2003, the District Attorney’s Office and HPD began a process designed to re-test
all cases that resulted in a conviction -- whether at trial or through a guilty plea -- in
which DNA evidence analyzed by the Crime Lab may have played a role. The central
purpose of the re-testing program has been to identify any cases in which the results of
DNA analysis performed by the Crime Lab cannot be confirmed. As of December 22,
2005, re-testing had been ordered for 416 cases.

We were unable to complete our review of the Derrick Jackson death penalty case last fall
because the original RFLP autorads related to that case were missing from the Crime
Lab’s raw data records. On February 3, 2006, the Crime Lab produced a recentlydiscovered box containing, among other things, case notes, raw data materials such as
RFLP autorads and DQ Alpha test strips, and tubes of DNA extracts. The missing
autorads related to the Derrick Jackson case were in the box.

The three death penalty cases involving major issues that we discussed in our Fourth
Report relate to death row inmates Franklin Dewayne Alix, Juan Carlos Alvarez, and
Gilmar Alex Guevara. In this report, we discuss the Crime Lab’s work in the case of
Derrick Jackson, a fourth death penalty case in which we identified major issues related
to the DNA analysis (as well as the serological testing) performed by the Lab.

HPD Crime Lab Independent Investigation

the redundant loci when two STR reagent kits were used to type the
same evidence samples.
•

Failure to perform and document meaningful technical and
administrative reviews of the work performed by DNA analysts.

•

Absence of a system assigning a unique identifier to track evidence
samples from submission through analysis to reporting.

In light of the seriousness and pervasiveness of the major issues we found
in our case reviews last fall, we recommended to the Stakeholders Committee
and HPD that we focus the case reviews on the 69 DNA re-test cases which either
(a) had not yet been tested by outside laboratories or (b) re-testing by outside
laboratories had failed to confirm the Crime Lab’s original DNA testing results.
We suggested this approach in order to focus resources on reviewing the cases in
which there had been a conviction and outside testing had not confirmed the
Crime Lab’s original results. These cases pose the greatest risk for potential
injustice related to flawed DNA profiling work performed by the Crime Lab.
The Stakeholders Committee and HPD approved our recommendations, and we
changed the focus of our DNA case reviews accordingly.
We have completed our review of all 69 of these DNA conviction cases in
which either the evidence has yet to be re-tested or the Crime Lab’s original
DNA results have not been confirmed through re-testing. We identified major
issues in 25 -- or approximately 36% -- of these cases. In total, we have reviewed
135 DNA cases analyzed by the Crime Lab from across the entire period in which
the Lab performed DNA analysis and have identified major issues in 43 -- or
approximately 32% -- of these cases.66 However, it is important to note that,
because we are no longer basing our review of DNA cases exclusively on a
statistically-based sampling methodology, this rate of occurrence of major issues
is not easily extrapolated to the set of all DNA cases analyzed by the Crime Lab.
B.

Serious Problems Identified in DNA Cases

Many of the major issues we have identified in the cases we have
reviewed since our Fourth Report are similar to those we have discussed
previously, including the failure to report probative, even potentially
66

There are 45 suspects or defendants involved in the 43 major issue DNA cases we have
identified. Each of these individuals is identified in Appendix C, “DNA Major Issue
Cases.”

HPD Crime Lab Independent Investigation

exculpatory, results; poor technical work and potential contamination;
questionable interpretation of results; and misleading reporting of the statistical
significance of DNA profiling results in mixture cases. In this report, we expand
on these and other major issues, as well as discuss additional troubling cases
involving seriously flawed or unreliable DNA work performed in the Crime Lab.
1.

Failure to Report Potentially Exculpatory Results

As discussed in the previous section, we found a disturbing number of
instances where the Crime Lab’s serologists appeared to be unwilling to report
typing results that were inconsistent with the known ABO type of either a victim
or a suspect. This indefensible practice continued into the DNA era as many of
the same criminalists who had been serologists, including Mr. Bolding and
Ms. Kim, became the Crime Lab’s senior DNA analysts. With the exponentiallyimproved discriminatory power of DNA profiling over serology, the Crime Lab’s
practice of failing to report probative, and in some cases potentially exculpatory,
DNA typing results became even more egregious.
As with serology, it is not clear whether this pattern and practice of
avoiding the reporting of DNA typing results that were not consistent with a
victim or known suspect is attributable to the DNA analysts’ lack of confidence
in their ability to obtain reliable results or to fraudulent reporting designed to
maximize the chances of convictions and avoid the consequences of potentially
exculpatory results. It is apparent, however, that DNA analysts in many cases
tended toward reporting only those results that, from their perspective, were
“safe” in the sense that they were consistent with other evidence in the case or
with the investigators’ expectations. As in the cases of Franklin Dewayne Alix
and Garland Davis discussed in our Fourth Report and Michael Mingo discussed
below, this sometimes meant that analysts failed to report potentially
exculpatory RFLP results in favor of reporting less definitive or less powerful
PCR-based typing results that appeared to reflect an association between the
suspect and evidence in the case. When such selective reporting was coupled
with the Crime Lab’s systematic exaggeration of the statistical significance of
these weaker PCR results, a very significant risk of injustice was created.
Mr. Mingo’s case related to the investigation of a suspected sexual assault
on a 12-year-old girl that took place on December 30, 1996. A sexual assault
examination was performed on the girl the following day. She told the
examiners that she had had vaginal intercourse for the first time with a
25-year-old man named Mingo and that the intercourse occurred with the use of
a condom. She also told the examiners that she had showered prior to the

HPD Crime Lab Independent Investigation

examination. In December 1997, a reference sample taken from Mr. Mingo was
submitted to the Crime Lab for DNA analysis and comparison to genetic material
on the vaginal swab included in the victim’s sexual assault kit.
In January 1998, DNA analyst Raynard Cockrell performed RFLP testing
on extracts taken from the vaginal swab. He initially ran two RFLP probes. He
exposed the first probe for two hours and then extended the exposure overnight,
and exposed the second probe for one and a half hours. Neither of these RFLP
tests on the vaginal swab included a profile consistent with Mr. Mingo. A
telephone log in the Crime Lab file reflects that, on January 21, 1998, Mr. Cockrell
told a prosecutor in the District Attorney’s Office that the “suspect could not be
ruled out” and that “PCR analysis might be able to give a conclusive result.” On
January 23, 1998, Mr. Cockrell informed a second prosecutor that “her suspect
could not be ruled out” and that “PCR analysis [is] needed.” He also noted that
“SUSPECT is in jail” and that the trial currently was scheduled for the following
Monday.
On February 5, 1998, Mr. Cockrell ran a third RFLP probe on the extract
from the vaginal swab. After relatively short exposure times of only two hours
and then four hours, faint bands consistent with the victim’s DNA profile were
apparent on the RFLP autoradiograph (“autorad”). Also visible on the autorad
were two very faint bands suggesting the presence of a DNA profile that was
consistent with neither the victim nor Mr. Mingo. In other words, this third
autorad indicated a potential unknown contributor to the sample from the
vaginal swab. In light of these faint results, the appropriate action would have
been for Mr. Cockrell to subject the third probe to a longer exposure time in
order to further develop this DNA profile related to an unknown suspect
potentially present on the vaginal swab. Instead, Mr. Cockrell prematurely
terminated the third probe, thus never permitting these faint bands to develop.
On February 19, 1998, Mr. Cockrell and DNA analyst Joseph Chu advised
the prosecutor that “RFLP results are inconclusive” and that Mr. Chu would
perform PCR analysis. The results of Mr. Cockrell’s RFLP testing, including the
potentially exculpatory bands detected by the third RFLP probe suggesting the
presence of the victim’s and an unknown person’s DNA profiles, were never
reported.
Mr. Chu then performed PCR-based tests -- specifically, D1S80,
Polymarker, and DQ Alpha tests -- on the evidence sample. In his February 26,
1998 report, Mr. Chu reported that “the DNA type detected on the vaginal swab
matches the DNA type of Michael Mingo.” Consistent with the DNA Section’s

HPD Crime Lab Independent Investigation

flawed and misleading practice of reporting statistical calculations of the
suspect’s reference sample rather than the statistical calculations of the profile
detected in the mixture sample, Mr. Chu reported that “the DNA type of Michael
Mingo can be expected to occur in 1 out of 300,000 people among the American
Black population.”67 We calculated the relevant frequency estimate, based on
Mr. Chu’s PCR-based results, to be 1 in 1,022 for the African American
population, 1 in 157 for the Caucasian population, and 1 in 135 for the Hispanic
population.68
The case against Mr. Mingo went to trial in June 1998. On June 18, 1998,
after the trial began but before a jury verdict was rendered, Mr. Mingo pleaded
guilty to aggravated sexual assault of a child. He was sentenced to ten years in
prison.69
2.

Failure to Mathematically Confirm RFLP Results

The results of RFLP testing are visualized as x-ray film images known as
autorads and appear as translucent films with dark bands on them. These bands
reflect the presence of RFLP alleles detected in the evidence and known reference
samples. The DNA analyst determines the size of each band through
comparison with known sizing standards. The size of the evidence bands
reflected on the autorad is then compared with the sizes of the known reference
sample bands. If the size of the evidence bands falls within an appropriate
“match window,” the evidence and reference samples are said to “match.”
67

This statement by Mr. Chu is completely irrelevant to the significance of the association -reported as a “match” -- between Mr. Mingo and the evidence based on Mr. Chu’s PCR
testing. The Crime Lab’s consistent practice of including misleading statistical
calculations in mixture cases is the most pervasive of the major issues we have identified
in the Lab’s DNA cases. This issue is discussed further below.

As discussed further below, forensic DNA analysts express the strength of the association
of an individual with a specific sample of biological evidence through the calculation of a
frequency estimate called a “random match probability.” That estimate quantifies the
likelihood that a person randomly drawn from the population could be the source of the
genetic profile detected in the evidence sample. A proper presentation of frequency
estimates includes calculations for all significant populations because it is inappropriate
to assume that a contributor to the evidence was from the same racial population as the
suspect against whom the DNA profile comparison was made.

The outside laboratory that re-tested the vaginal swab in Mr. Mingo’s case was able to
develop a partial DNA profile. The outside laboratory calculated a frequency estimate of
1 in 59 unrelated individuals in the African American population.

HPD Crime Lab Independent Investigation

Initially, this comparison between the bands related to evidence samples
with the bands related to the victim’s and suspect’s known reference samples is
performed visually by the DNA analyst. Allelic bands in evidence samples that
appear to line up at approximately the same positions as bands in victim or
suspect reference samples can be said to be visual matches, indicating common
alleles may be present in the evidence sample and the matching reference
sample.
The Crime Lab’s SOP regarding “Procedures for the Detection of
Restriction Fragment Length Polymorphisms in Human DNA” (“RFLP SOP”),
dated 1991, required that “[v]isual matches must be confirmed or rejected
through application of the appropriate mathematical procedures.”70 The RFLP
SOP prescribed a mathematical confirmation procedure under which the DNA
analyst was required to calculate “match windows” with a range of 2.5% of the
base pair size for the DNA fragments related to the known specimen and
evidence sample that had been determined to be visual matches. If the sizes of
the known and evidence specimen bands fall within 2.5% of one another, then
the visual match was confirmed. The SOP stated that, “[i]f the ranges do not
overlap, the presumptive equality of the fragment sizes is either inconclusive or
exclusionary” -- in other words, there would be no valid DNA profile match
between the reference and evidence samples being compared when the 2.5%
match window is exceeded.
The death penalty case of Derrick Jackson is unusual because the passage
of time between the murders in 1988 and the identification of Mr. Jackson as a
suspect in 1995 spanned the technological transition in the Crime Lab from
serology to DNA profiling. In the Jackson case, the Crime Lab performed ABO
typing, RFLP DNA analysis, and PCR-based DNA analysis.71 The reported
results of the RFLP testing in Mr. Jackson’s case, which was performed by DNA
analyst Mary Childs-Henry, were flawed because Ms. Childs-Henry failed to
follow the Crime Lab’s RFLP SOP guidelines regarding mathematical
confirmation of presumed matches.
From July to September 1996, Ms. Childs-Henry performed RFLP testing
on various evidence items recovered from the apartment in which the two
victims, Mr. Henderson and Mr. Wrotenbury, had been murdered eight years
70

RFLP SOP at 19.

The major issues we identified with respect to Mr. Bolding’s serological analysis in the
Derrick Jackson case are discussed in the previous section.

HPD Crime Lab Independent Investigation

earlier. Ms. Childs-Henry obtained RFLP results related to samples extracted
from a beige towel and a red towel found in at the crime scene. Based on these
RFLP results, a report issued by the Crime Lab, dated March 14, 1997, stated
“[t]he DNA type detected on the beige towel (item 26) and the red towel
(item 27) matches the DNA type of Derrick Jackson.” Regarding the significance
of the match between Mr. Jackson’s DNA profile and the DNA profiles from
bloodstains on the towels, the Crime Lab reported that “[t]he DNA type of
Derrick Jackson can be expected to occur in 1 out of 5.8 million among the
American Black population (based on RFLP results).”72
On March 10, 1998, the second day of Mr. Jackson’s capital murder trial,
defense counsel requested a hearing to challenge the admissibility of evidence
related to the DNA analysis performed by the Crime Lab on the grounds that the
Lab’s DNA testing was unreliable.73 Specifically, the defense sought to
challenge, through the testimony of its expert Dr. Elizabeth Johnson, whether the
“match windows” related to the Crime Lab’s RFLP testing were properly
calculated.74 After hearing briefly from Dr. Johnson, the court declined to hold a
pre-admissibility hearing regarding the validity of the Crime Lab’s DNA analysis
techniques and permitted the State to introduce the DNA testing results.75
72

This statement is not a proper presentation of a frequency estimate relating to the DNA
profile detected on the towels, which more appropriately should have been expressed as
random match probabilities with respect to at least the three most significant North
American racial populations --African American, Caucasian, and Hispanic. However,
because the bloodstains involved were not mixtures, these statistics are less problematic
and misleading than the statistics the Crime Lab routinely calculated based on suspects’
known reference samples in cases involving mixed body fluid samples from more than
one donor.

Under the principles established in United States Supreme Court case Daubert v.
Merrill-Dow Pharmaceuticals, 509 U.S. 579 (1993), and its Texas state court analog, Kelly v.
State, 824 S.W.2d 568 (Tex. Crim. App. 1992), the trial judge must perform a
“gatekeeping” function with respect to the admission of scientific testimony to ensure
that such evidence is premised on valid scientific theory and is reliable.

Jackson Tr., Vol. 22, at 9:11-9:23; 31:8-22.

On appeal, the Texas Court of Criminal Appeals agreed with Mr. Jackson that the trial
court “abandoned its ‘gatekeeping’ function” in refusing to hold a Daubert/Kelly hearing,
but held that “in this case the error was harmless because the State’s DNA evidence was
in fact reliable.” Jackson v. Texas, No. 73,081 (Tex. Ct. Crim. App. May 17, 2000). This
Court of Criminal Appeals ruling is surprising and difficult to square with the fact that,
as discussed below, at trial the State and Ms. Childs-Henry admitted that she had in fact
miscalculated the RFLP match windows.

HPD Crime Lab Independent Investigation

That evening, the prosecution advised Mr. Jackson’s counsel that the
statistics related to the Crime Lab’s RFLP results had been recalculated and that,
rather than a frequency estimate of 1 in 5.8 million in the African American
population, the Lab’s revised statistical calculation was 1 in 224.76 The following
day, on March 11, 1998, Ms. Childs-Henry acknowledged on the witness stand
that the RFLP results of two of the four autorads related to DNA tests on the
towels fell outside the appropriate match windows and, therefore, could not be
used in the statistical calculation. Accordingly, the results relating to those two
autorads were disregarded and the frequency estimate recalculated, with the
effect of reducing the Crime Lab’s reported 1 in 5.8 million African Americans
frequency estimate to 1 in 224.77 The next day, Mr. Jackson was convicted of
capital murder.78
Standing alone, the Crime Lab’s failure in Mr. Jackson’s case to follow its
own established procedures for mathematically confirming DNA profile
“matches” between evidence samples and the suspect’s reference sample is
troubling. However, this case also illustrates the broad failure of the Crime Lab
to have any semblance of a quality assurance program in the DNA Section to
detect and remedy errors. We have found no evidence that routine and
competent technical reviews of DNA analysts’ work were ever performed in the
Crime Lab, even in capital murder cases such as this case. It is precisely errors
such as those reflected in the Derrick Jackson case that should be detected and
remedied through an effective quality assurance program.79

Jackson Tr., Vol. 23, at 137:4-138:9.

Id. 252:11-254:8.

On August 13, 2003, upon a motion filed by Mr. Jackson’s counsel, the court ordered that
the stains on the two towels be subjected to DNA re-tests performed by LabCorp
Laboratory, an outside laboratory retained by the defense. After attempting to contact
Mr. Jackson’s current counsel, Steven Rosen, by telephone, on January 31, 2006, we sent
Mr. Rosen a letter requesting raw data relating to testing in Mr. Jackson’s case, including
the results of LabCorp.’s re-testing of evidence in this case. We have received no
response from Mr. Rosen.

The DNA analysis in the Derrick Jackson case was performed after Dr. Baldev Sharma
had been removed as the Criminalist III supervisor in the DNA Section and placed in a
Quality Assurance/Quality Control position for the entire Crime Lab in August 1996
following the Lynn Jones scandal. See Third Report at 21. Neither Dr. Sharma nor
Mr. Bolding, the head of the DNA Section, appeared to have appreciated the important
role of a quality assurance program in producing reliable DNA profiling results.

HPD Crime Lab Independent Investigation

Poor Quality PCR Analysis

In our Fourth Report, we discussed the alarming frequency with which
PCR-based testing -- D1S80, Polymarker, and DQ Alpha -- performed by the
Crime Lab generated multiple DNA profiles that were matched to a suspect and
the victim, plus one or more unknown donors. The analysts’ failure to recognize
that this abundance of alleles was unreasonable is an indicator that they were not
adequately trained to be aware of and alert for the potential for contamination in
any PCR-based analysis.
Indeed, it appears that at some point the Crime Lab became concerned
about PCR contamination. In Mr. Bolding’s files, we found a memorandum from
Mr. Bolding dated March 2, 2001 and addressed to Mr. Chu, a prolific PCR
analyst. In the memorandum Mr. Bolding wrote:
ON [sic] February 21, 2001, we spoke about contamination in the
PCR Process, I requested a [sic] documentation of the
contamination and what steps you took to alleviate the problem.
To date I have not received that document.
The corrections for contamination problems, via formal document
are due in my hands on Monday Mar. 5, 2001 by 10 a.m. This
document should be in a form that can be presented to defense
experts.
Mr. Chu could not recall whether a specific incident or event gave rise to
Mr. Bolding’s directive that he document and address contamination problems
related to the PCR process. Mr. Chu told us that he believed the Mr. Bolding’s
memorandum related to concerns about contamination at the DNA extraction
stage, before evidence was transferred to the DNA analysts for testing. He also
could not recall what action, if any, he took in response to this memorandum.,
and we have found no evidence of any response to Mr. Bolding’s directive that
contamination in the PCR process be investigated.
Below, we describe the problems we have observed in the PCR-based
typing performed by the Crime Lab in several cases. We use the Reginald
Jackson case as an illustrative example to show the results of a combination of
poor analytical technique and possible contamination in the Crime Lab’s use of
D1S80, a PCR-based test.

HPD Crime Lab Independent Investigation

Problems with D1S80 Analysis

D1S80 is an early form of PCR-based testing related to the D1S80 locus,
which is found on human chromosome number 1. The D1S80 locus was
attractive to forensic DNA analysts because it exhibits a very high degree of
polymorphism, or variability, between individuals. D1S80 typing tests involve
electrophoresis of D1S80 products through a gel to determine the number of
tandem repeats present in evidence or reference samples. After the D1S80
product is loaded and run on the gel, the D1S80 allelic bands are visualized to
produce gel image similar in appearance to an RFLP autorad. Sizing ladders are
run in the gel along with the evidence and known reference samples in order to
permit the DNA analyst to interpret the allelic bands produced through the
electrophoretic process. Each individual has a maximum of two alleles at the
D1S80 locus (homozygous persons have only one type of allele). An individual’s
D1S80 type is expressed as the combination of these alleles. For example, one
person might be typed for D1S80 as a type “18, 24” and another person as D1S80
type “22, 31.”
In the Reginald Jackson case, the Crime Lab was asked by investigators to
analyze evidence relating to a stabbing that occurred in 1997. Investigators
submitted four items of evidence to the Crime Lab: (1) blood from a steak knife,
(2) blood taken from a walkway in a parking lot, (3) blood from a foyer floor, and
(4) a bloodstain from an article of clothing alternately described as “jeans” or
“pants” in difference Lab documents. The Crime Lab also received reference
samples from the victim and the suspect, Mr. Jackson. Ms. Kim performed
DQ Alpha, Polymarker, and D1S80 testing on the evidence and reference
samples.80
Below is an image of the original gel reflecting a set of results obtained by
Ms. Kim through D1S80 testing in the Reginald Jackson case. Each lane is
labeled to reflect the sample, sizing ladder, or control placed in that lane.

Ms. Kim’s February 6, 1998 report indicates that “DNA extracted from the . . . evidence
except for 4 microliters was transferred from Criminalist C. Kim to R. Cockrell for DNA
(RFLP) analysis.” Mr. Cockrell never reported any results related to the evidence in this
case. A “Post-it” note attached to one of Mr. Cockrell’s RFLP data sheets indicated: “No
results on unknowns.”

HPD Crime Lab Independent Investigation

Positive Control

Victim

Allelic Ladder

Victim

Walkway

Foyer Floor

Allelic Ladder

Pants/Jeans

Negative Control

Allelic Ladder

Negative Control

Reginald Jackson

Allelic Ladder

Knife

Negative Control

Allelic Ladder

D1S80 Gel in the Reginald Jackson Case

The above gel image reflects that Ms. Kim obtained a strong homozygous
D1S80 type “34” allele in the “knife” sample. This 34 allele also is present in the
foyer floor, walkway, and victim’s reference samples. The walkway and victim’s
reference samples, however, also reflect the presence of numerous other alleles.
The profile indicated for Mr. Jackson appears to be D1S80 type “24,28,” although
several extraneous bands are apparent in his reference sample as well.81 The
81

A “24,28” profile is consistent with the profile developed for the jeans/pants evidence
sample.

HPD Crime Lab Independent Investigation

unexplained presence of these extraneous alleles in both the victim’s and
suspect’s reference samples illustrates some of the problems we have observed
generally with the Crime Lab’s PCR-based DNA testing.
First, in addition to the 34 allele, the victim’s reference sample contains
five other D1S80 alleles: 18, 22, 24, 27, and 31. At most, an individual has only 2
alleles at a particular locus such as D1S80. The fact that Ms. Kim detected a total
of six alleles in the victim’s reference sample, which should be a pristine,
single-source sample, is extremely troubling. This result indicates that the
victim’s reference sample was contaminated at some point in the handling of this
sample. It is theoretically possible for a victim’s reference sample to have
become contaminated at the medical examiner’s or coroner’s office. In this case,
however, subsequent to the D1S80 testing, Mr. Cockrell performed RFLP typing
on the victim’s reference sample and obtained a single donor profile. Moreover,
the outside laboratory that later performed a DNA re-test was able to obtain a
single donor profile for the victim sample using the original DNA extract
prepared in the Crime Lab. This suggests that, if the result reflected on the
D1S80 gel for the victim reference sample is the product of contamination, the
contamination occurred after extraction, most likely at the PCR amplification
stage.
It is possible that Ms. Kim added too much DNA from the evidence and
reference samples during the PCR amplification phase of the testing process,
which could have led to over-amplification and the presence of multiple
extraneous bands in the reference and evidence samples. When we asked
Mr. Chu to discuss what he believed accounted for the extraneous alleles in
several of the evidence and reference samples reflected on the Reginald Jackson
gel, he mentioned overloading and over-amplification. In light of the band
pattern observed for the victim’s reference sample, overloading or overamplification seem like unlikely explanations for the multiple bands seen in the
victim’s reference sample.
Extraneous alleles also are visible in Reginald Jackson’s reference sample
and the walkway sample and could involve crossover or contamination from the
allelic ladder. If the sizing ladder were overloaded in the gel, it is possible that it
could carry over into an adjacent lane. In this case, however, while the Reginald
Jackson reference sample is adjacent to an allelic ladder lane, the walkway
sample is not. The walkway sample appears to be an example of mixing the
allelic ladder with an evidence sample prior to loading it into the gel.
Overloading and over-amplification is another theoretical possibility to explain

HPD Crime Lab Independent Investigation

this “laddering” phenomenon when a specimen seems to exhibit all the bands
just as an allelic ladder does.
Regardless of the cause of the serious problems with the D1S80 testing in
this case, Ms. Kim was unable to resolve the issues. Ultimately, Ms. Kim only
reported that “the DNA type detected from the knife is not consistent with that
of Reginald Jackson.” Although the victim reference sample shared a strong 34
allele with the knife sample, it appears Ms. Kim was unwilling to interpret the
mixed profile she obtained in the victim’s reference sample. Ironically, in several
cases, including the Carlos Segura case discussed below, where the Crime Lab
has obtained questionable multiple profiles in evidence samples, analysts either
selectively reported profiles consistent with a suspect’s or victim’s known profile
or reported out multiple profiles, including matches with known reference
samples from the victim and suspect(s) as well as one or more “unknown”
individuals.
b.

Questionable Interpretation of PCR Results

In a 1998 homicide case involving three suspects -- Carlos Segura, Mark
Zavala, and Francisco Zapata -- we observed Ms. Kim’s flawed interpretation of
PCR-based results in a mixture case. This case involved a stabbing in which
blood was identified on a knife and on several samples from the crime scene,
including a blood trail leading away from the victim’s body. Ms. Kim performed
PCR-based testing on the blood trail samples, a sample from the knife, reference
samples from the three suspects, and a sample described as “white tissue/blue
diaper” that appears to contain the victim’s DNA profile.82
In her July 7, 1999 report, Ms. Kim stated that “[t]he DNA pattern from
the blood trail samples . . . is consistent with that of Francisco Zapata.” She also
reported finding “[a] mixture of DNA type [sic] consistent with DNA type [sic]
of Francisco Zapata, Carlos Segura, and Mark Zavala and the donor of the blood
trail sample and white tissue/blue diaper was detected on the back handle, the
82

Ms. Kim’s analysis in this case is further confused by the fact that there is no clear
profiling data related to a reference sample from the victim. It appears that Ms. Kim was
not able to obtain DNA typing results on the blood obtained from the victim’s autopsy.
However, the Crime Lab obtained results from the “white tissue/blue diaper” sample,
which was labeled as “DNA extract from white tissue/blue diaper -- Moises Ayala [the
victim]” when the sample was sent to an outside laboratory for DNA testing in 1999.
This indicates that the “white tissue/blue diaper” sample contained the victim’s DNA
profile.

HPD Crime Lab Independent Investigation

bolster and handle of the knife.” Ms. Kim reported finding four DNA profiles on
the knife -- all three suspects and a fourth person whose profile is consistent with
the DNA profile obtained from “white tissue/blue diaper” and who actually is
the victim.
The raw data reflecting the results of Ms. Kim’s PCR-based testing, and
the results of her DQ Alpha testing in particular, strongly indicate that her
interpretation of the DNA testing results to include all three suspects and a
fourth profile (the victim) is flawed and misleading. Unfortunately, the suspects
and victim share several common DQ Alpha and D1S80 alleles, which made
elimination of a suspect as a possible contributor to the evidence on the knife
sample difficult. However, a reasonable interpretation of the raw DQ Alpha
results would be consistent with a mixture of the victim and Mr. Zapata only.
Such an interpretation certainly would be more reasonable than Ms. Kim’s
finding that all three suspects and the victim bled on the knife. At a minimum,
Ms. Kim should have tabulated the allelic findings in her report and explained
that she could not eliminate Mr. Zavala or Mr. Segura as potential contributors
due to the common alleles they shared with either the victim or Mr. Zapata.
Instead, she issued a report that was extremely muddled and confusing to the
point of being nearly incomprehensible.
In 1999, the evidence in this case was sent to an outside laboratory for STR
testing. In a report dated August 25, 1999, the outside laboratory concluded that
“both [the victim] and Francisco Zapata are included as potential contributors to
the stains on the knife and sheath. Both Carlos Segura and Mark Zavala are
excluded as contributors to the stains on the knife and sheath.”83 These STR
results are consistent with the raw data related to Ms. Kim’s PCR-based testing.
However, Ms. Kim interpreted and reported her data to include all three
suspects as well as a fourth contributor, whom she failed to identify as the
victim. Her interpretive error was exacerbated when she presented misleading
frequency estimates, calculated based on the suspects’ reference profiles, that
suggested strong associations between Mr. Segura (1 in 11,300) and Mr. Zavala (1
in 758,000) and the DNA evidence from bloodstains on the knife.84

HPD has acknowledged that re-testing by an outside laboratory in this case has reversed
the Crime Lab’s original DNA typing results as to Mr. Segura and Mr. Zavala.

We calculated the frequency estimate, based on Ms. Kim’s typing results for the
bloodstains on the knife, to be 1 in 48 for the Hispanic population.

HPD Crime Lab Independent Investigation

Problems with Interpretation of STR Results

Errors by Crime Lab DNA analysts in the interpretation of raw data
related to DNA tests were not limited to early PCR-based testing. Similar
interpretative errors continued into the STR testing era. As discussed in our
Fourth Report, a critical control in the STR system used by the Crime Lab is
confirmation that the alleles detected at the redundant D3 and D7 loci85 in the
COfiler and Profiler reagent kits are in concordance.86 We have identified several
cases in which DNA analysts reported STR results and developed profiles
despite discordance between COfiler and Profiler typing results at these
redundant loci.
The presence of these redundant loci in the COfiler and Profiler reagent
kits used by the Crime Lab for STR testing is a built-in quality control feature
designed to detect possible sample mix-ups. This feature also is a tool to ensure
that both kits are working properly. If the alleles detected with the COfiler
reagent kit for D3, D7 and amelogenin markers are not in concordance with those
detected using the Profiler reagent kit, it is a warning that there is a problem,
including the possibility that the sample is of poor quality. If the allele
identifications at the redundant loci are not the same for the same samples, it is
imperative that the problem be resolved by re-analyzing the original samples.
The Crime Lab’s SOPs specifically required that the D3 and D7 loci for the
COfiler and Profiler systems must agree in each sample run through the STR
process.87
The case of Ronald Cantrell presents an example of questionable
interpretation and reporting of STR results in the face of COfiler and Profiler
discordance at the redundant D3 and D7 loci. This case involved a reported
sexual assault on an 8-year-old girl. According to the sexual assault examination
form prepared on December 12, 2001, the victim reported that a suspect named
“Ronny” forced the victim to perform oral sex, resulting in a “semen” stain on
her shirt. On December 17, 2001, the Crime Lab received several articles of
clothing from the victim, including a blouse.
85

“D3” is shorthand for the D3S1358 locus and “D7” is shorthand for the D7S820 locus.

Fourth Report at 46-48.

The Crime Lab’s SOPs, however, provide no guidance as to what procedures the DNA
analyst should follow with respect to sample tests where the D3 and D7 loci are not in
concordance.

HPD Crime Lab Independent Investigation

On February 8, 2002, Crime Lab analyst Audrey Tims reported that
“semen was detected on the blouse.” Ms. Tims’s identification of semen on the
blouse is questionable. It does not appear that she performed a microscopic
examination to detect the presence of sperm cells on the blouse. Ms. Tims’s
worksheet, dated December 17, 2001, reflects that the AP screening tests for
semen were negative when she tested stains on the blouse. She also ran a p30
Abacard test for semen on the blouse, the results of which she recorded as “POS
weak.” Based on the limitations of the p30 Abacard test system and the negative
AP test for semen, this weak positive result probably was not a sufficient basis to
support a finding that semen was present. Nevertheless, a differential extraction
from the blouse’s stain was performed, and the extracts were forwarded to
Ms. Kim and Mr. Cockrell for DNA analysis.
Mr. Cockrell performed the STR analysis of the DNA extract from the
blouse. On August 30, 2002, Mr. Cockrell reported that “[a] mixture of DNA
types was detected on the blouse.” He also reported that “[the victim], Ronald
Cantrell, and at least two other donors are included in this mixture.”
Mr. Cockrell reported these results despite a clear discordance between COfiler
and Profiler typing results at the D3 and D7 loci. The following chart reflects
alleles Mr. Cockrell obtained at the D3 and D7 loci as shown by his original STR
electropherogram. The alleles in bold are those that appear on the
electropherogram below the 150 rfu threshold necessary under the Crime Lab’s
SOPs in order for an allele to be interpreted as present in the sample.
Locus

Locus and Sample

COfiler

Profiler

D3
D3

Epithelial Fraction
Sperm Fraction

14, 15, 16, 17, 18
15, 18

14, 15, 16, 17, 18
15, 16, 18

D7
D7

Epithelial Fraction
Sperm Fraction

8, 9, 10, 11, 12
10, 12

10, 12
9, 10

As shown, there is general concordance of the typing results between
COfiler and Profiler at the D3 locus. Profiler, despite being less sensitive than
COfiler, detected a weak 16 allele in the sperm fraction, but it was below the
150 rfu threshold. There is, however, significant discordance of the typing
results at the D7 locus. With respect to the epithelial fraction, COfiler results
show the 9 and 11 alleles above the 150 rfu threshold and an 8 allele below the
threshold that was not detected by Profiler. Although this is a significant
discordance, the slightly greater sensitivity of CoFiler might explain the extra
alleles detected by that system and does not necessarily indicate that the control
failed and the test must be disregarded. However, the discordance in the sperm

HPD Crime Lab Independent Investigation

fraction at the D7 locus cannot be explained by the greater sensitivity of COfiler.
With respect to that sample, Profiler detected weak 9 and 10 alleles, and COfiler
(the more sensitive system) failed to detect the 9 allele at all. This discordance at
the D7 locus cannot be reconciled and should have invalidated the results and
caused Mr. Cockrell to re-perform the analysis.88
Re-testing by outside laboratories has failed to confirm the original results
reported by Mr. Cockrell. HPD reported that the “raw evidence” stain on the
victim’s blouse had been consumed.89 Therefore the re-testing laboratory had to
use the DNA extracts prepared by the Crime Lab. A report issued by an outside
laboratory on May 2, 2003 stated that “[a] mixture of male and female DNA
profiles was obtained from the epithelial fraction of the extracted DNA from the
blouse” and that “Ronald Cantrell is excluded as being a potential donor to the
mixture.” The outside laboratory reported that “[n]o DNA was obtained from
the sperm fraction of the extracted DNA from the blouse.” A review by the same
outside laboratory of the Crime Lab’s original testing in this case questioned
whether any semen was present in the stain on the victim’s blouse. The outside
laboratory concluded that “the evidentiary value of an inclusion [of Mr. Cantrell
is] extremely limited, and makes it impossible to accurately assess the statistical
significance of the conclusion that the suspect could not be excluded from the
mixture.”90
5.

Routine Misreporting of the Statistical Significance of
DNA Profiling Results

The most pervasive major issue we have identified in our review of the
Crime Lab’s serology cases is the routine and systemic misreporting of the
statistical significances of “matches” reported by the Crime Lab in cases
88

Also, there is no evidence that the crime Lab ran a substrate control with respect to the
stain on the blouse. A substrate control would have been helpful in determining whether
the alleles detected by the STR tests pertained to the suspected semen stain or to
substrate material on the blouse.

It is unclear what the size of the purported semen stain might have been because the
Crime Lab file did not contain a description of the stain. Good laboratory practice calls
for the routine retention of a portion of an evidence specimen to be properly preserved
for possible additional testing in the future. Since the STR system is a sensitive test that
requires only a small amount of DNA, it would be unnecessary and problematic to
consume an entire specimen.

On August 19, 2002, Mr. Cantrell pleaded guilty to aggravated sexual assault and was
sentenced to six years in prison.

HPD Crime Lab Independent Investigation

involving evidence samples comprised of body fluid mixtures containing more
than one DNA profile. Indeed, 23 of the 43 major issue DNA cases we have
identified -- or approximately 53.5% of the major issue DNA cases -- involve
reported statistics that are misleading because the Crime Lab calculated them
based on the profile of the suspect’s known reference sample, rather than the
profiles identified in the evidence sample.91
We have prepared the chart below to demonstrate the sometimes
exponential difference between the statistics reported by the Crime Lab and our
calculation of the correct frequency estimates for the interpretable DNA typing
results originally obtained by the Crime Lab.92 In many cases, the disparities are
staggering.

In our Fourth Report, we provided a detailed discussion regarding the meaning of
random match probabilities in describing the significance of associations between an
individual and a specific sample of biological evidence. It is clear that Crime Lab
analysts, including Mr. Bolding, the head of the DNA Section, failed to grasp the
scientific basis of calculating frequency estimates related to DNA profiles obtained from
evidence samples. See Fourth Report at 43-46.

The Crime Lab’s convention, which was inappropriate, was to report only the statistics
related to the racial population with which the suspect is identified. Properly reported
frequency estimates include calculations for the three most significant ethnic populations
in North America -- African American, Caucasian, and Hispanic. It is not appropriate in
the calculation and reporting of random match probabilities to assume that a contributor
to the evidence sample is from the same racial demographic as the suspect.

HPD Crime Lab Independent Investigation

Comparison of Statistics Reported by the Crime Lab
with Properly Calculated Frequency Estimates
Suspect’s Name
Alix, Franklin
Boudreaux, Raymon
Carter, Harold93
Emory, Gregory
Guevara, Luis/Fernandez, Sixto94
Harris, Erskin
House, Dillard
Johnson, Arthur
Lawson, David
Lopez, Segundo
Meza, Alfredo
Pineda, Johnny
Napper, Laurence95
Parra, Carlos
Rayson, Carl Lee
Segura, Carlos/Zavala, Mark96
Southern, Ronnie
Sutton, Josiah
Valdez, Richard
Vanzandt, Lonnie
Vaughn, Artice
Ware, Marshall
Washington, Dedrick

HPD Reported Stats
(1 in ___)
81,000
11,200
9%
13,000
663 million/61 trillion
158,000
2,773
11 million
1.8 million
1.7 million
2.6 million
110,000
statistical match
146,00
1.8 million
11,300/758,000
6.3 million
694,000
15,000
15
988
2.9 million
1,800

Recalculated Stats
(1 in __)
11
37
75%
23
5,900/9,100
8 and 6
83
113
55
400
9
110
232,000
119
145
48
30
14
50
2
42% (~1 in 2)
22% (~1 in 5)
428

In this case, HPD presented its statistics in terms of the percentage of the relevant
population that that could be expected to have a DNA profile in common with the
suspect’s reference sample. For the sake of comparison with HPD’s presentation in this
case, we have re-calculated the frequency estimate related to the evidence in terms of a
percentage of the relevant population that could provide a random match.

The re-calculated frequency estimate for Sixto Fernandez is based on the results of a
re-test performed by an outside laboratory rather than the original DNA profile
developed by the Crime Lab.

The re-test performed by an outside laboratory of the evidence in the Napper case
developed a partial suspect profile. The outside laboratory calculated a frequency
estimate of 1 in 255 in the African American population based on that partial profile.

There is only one re-calculated frequency estimate in this case because the properly
re-calculated frequency estimate is based on the DNA profiles developed in the mixed
evidence sample, as opposed to HPD’s method of calculating statistics not based on the
DNA profile developed from the individual suspect’s known reference sample.

HPD Crime Lab Independent Investigation

This failure to properly calculate frequency estimates exacerbated the poor
quality of the Crime Lab’s technical work in developing DNA profiles from
evidence samples. As discussed above, Crime Lab analysts often developed and
reported DNA profiles reflecting multiple donors, which frequently were
reported as including the suspect and one or more “unknown donors.” As
reflected in the above chart, often the Crime Lab then went on to grossly
exaggerate the significance of finding the suspect’s DNA profile among the other
DNA profiles from the evidence sample by calculating and reporting frequency
estimates based on a suspect’s known reference sample.
IV.

Trace Evidence

Trace evidence can consist of many different types of material found at a
crime scene, on the victim of a crime, on a suspect, or in places visited by a victim
or suspect. Fibers, hairs, paint, glass, and fire debris are some of the most
common types of trace evidence examined in forensic labs. Forensic scientists
examine trace evidence to determine the physical and chemical properties of the
material, to compare these properties with those of known samples so that
possible common origins can be identified, and to provide investigative leads.
The Crime Lab currently is not performing trace evidence examinations,
although it did handle this type of evidence until October 2003. When the Trace
Evidence Section was operating, it was normally staffed with two analysts and a
supervisor. The Crime Lab stopped performing trace evidence examinations
after the section supervisor was appointed the Quality Assurance/Quality
Control Leader for the Crime Lab in late 2003. Since then, trace evidence
collected by HPD has been examined by the Texas Department of Public Safety
Crime Lab.
A.

Status of the Trace Evidence Review

We identified 223 cases as having been logged in by the Trace Evidence
Section during the period covered by our review. We initially selected a sample
of 141 trace evidence cases for our review. However, for reasons explained in
more detail in our Fourth Report,97 many of those cases did not involve any
substantive trace evidence examinations. As a result, we expanded the scope of
our trace evidence review to include all 223 trace evidence cases, as well as
certain DNA and serology cases that may have included a trace evidence
97

See Fourth Report at 52.

HPD Crime Lab Independent Investigation

component. We have completed our review of all 223 of these trace evidence
cases, as well as 40 DNA and serology cases, some of which involved trace
evidence casework.
B.

Findings
1.

The Trace Evidence Sample

From the total universe of 223 cases that were logged in by the Trace
Evidence Section during the period of our review, 129 involved the type of
substantive analysis by the Trace Evidence Section that was the focus of our
review.98 We identified 5 cases (4% of the substantive sample) with “major
issues” and 44 cases (34% of the substantive sample) with “minor issues.”
We found the following four primary types of issues in the trace evidence
sample:
•

Lengthy delays and lack of follow-up that occurred at several points in the
Crime Lab’s trace evidence examination process;

•

Minimal or non-existent attempts to examine evidence that could have
generated useful investigative leads;

•

Failure to follow generally accepted laboratory procedures involving the
documentation and use of controls, including reagent blanks that are used
to ensure that test results are not influenced by variations in the reagents;
and

•

Sparse documentation in many trace evidence case files.

These issues were discussed in additional detail in our Fourth Report.99
Additionally, a few files lacked documentation of any technical or
administrative review in cases in which the notes indicate that no trace evidence

Cases involving the examination of fire debris and latent fingerprints were not within the
scope of our trace evidence review.

See Fourth Report at 53-54.

HPD Crime Lab Independent Investigation

examinations were performed. The Crime Lab’s SOPs require such reviews in all
cases, whether or not an examination was actually performed.100
2.

Review of Selected Serology and DNA Cases

Some cases handled by the Crime Lab involve examinations in more than
one area of forensic science; this is especially true in cases involving violent
crime. We selected 40 serology and DNA case files to determine (a) whether any
trace evidence was examined in connection with those cases and (b) if so,
whether there were any issues with that work. These cases were selected
because they involved either death penalty or other high profile convictions. 101
We did not identify any major issues with respect to trace evidence
examinations in the selected 40 serology and DNA cases. We identified minor
issues in 5 (12.5%) of these cases. These minor issues primarily involved
documentation problems that are consistent with those previously identified and
reported in the general trace evidence sample.
Two death penalty case files had trace evidence deficiencies that we
determined were minor. In the first case, the work notes contained insufficient
documentation of the hair comparison that was performed. In the second case,
hair samples were collected, but the case file contained no Trace Evidence
Section report.
The trace evidence examinations performed in a third, non-death penalty
case were classified as involving minor issues because of discrepancies between
the thoroughness of the work notes relating to the known and questioned
samples. 102 The work notes in this case, which involved hair comparisons,
contained appropriately detailed observations of the microscopic characteristics
100

It might be appropriate for the Crime Lab to review the application of this policy to cases
in which no examinations are performed.

101

It is not uncommon for there to be a trace evidence component to serology and DNA
cases. Cases such as homicides and sexual assaults in which there may be biological
evidence (for example, blood or semen) may also involve evidence, such as hairs, that
could be subjected to trace evidence examination and comparison.

102

As was noted in our Fourth Report, examinations performed by the Trace Evidence
Section typically involve the comparison of “known” and “questioned” samples.
“Questioned” samples may involve evidence from an unknown source that is collected at
the crime scene. The “known” sample is collected from an identified source, often the
suspect or the victim.

HPD Crime Lab Independent Investigation

of the known hairs. However, there was no detailed description of the
questioned hairs other than that they were “microscopically consistent” with the
known hairs. We reviewed the transcript of the examiner’s trial testimony in this
case, and it appropriately reflected the limits of the observations that were made
and their interpretation.103
Similar issues were identified in a fourth case, in which numerous hair
associations were made linking questioned hairs to the victim. The victim’s hair
characteristics were well described, but the questioned hairs were not described
in sufficient detail to document the association.
In the fifth minor issue case, we found that the file contained no
information regarding the source of some items of evidence and that some
descriptions in the notes and report are too vague to be informative. As a result,
it is not possible to interpret the significance of the findings that are documented
in this file. Finally, the work notes indicate that several items were examined,
but those items are not referred to in the report.
V.

Controlled Substances

The Controlled Substances Section analyzed the majority of the cases in
the Crime Lab, handled a variety of controlled and non-controlled substances,
and employed more analysts than any other section in the Lab.104 Our review
was designed to take into consideration the large numbers of cases, substances,
and analysts involved. HPD and the City also requested a review specifically
focusing on cases analyzed by Vipul Patel and James Price, two former
Controlled Substances Section analysts associated with drylabbing incidents.105
At the outset of Phase II, we therefore established three separate samples of
controlled substances cases -- one comprised of general controlled substances
cases,106 a second sample comprised of cases analyzed by Mr. Patel, and a third
sample comprised of cases analyzed by Mr. Price.
103

This was the highly publicized case involving George Rodriguez.

104

See Fourth Report at 56-58 for a description of the techniques and instruments used to
identify controlled substances.

105

The Price and Patel drylabbing incidents are described in detail in our Second and Third
Reports.

106

The general controlled substances sample does not include cases handled by Messrs.
Patel and Price.

HPD Crime Lab Independent Investigation

Our initial review of the general controlled substances sample revealed
that most cases handled by this section involved basic marijuana and cocaine
identifications. We adjusted the general controlled substances sample to include
cases involving more complex and challenging analyses, and we ultimately
reviewed 513 general controlled substances case files.107 Our review of
Mr. Patel’s sample included 366 cases, and our review of Mr. Price’s sample
included 342 cases. Finally, we also reviewed an additional 50 files to evaluate
how the Crime Lab handled and analyzed “bulk” or “bulky” cases, which, as the
terms suggest, are cases involving large quantities of evidence. We have
completed our review of all three controlled substances case samples, as well as
the “bulky” controlled substances cases.
A.

Results of the General Controlled Substances Case Reviews
1.

Major Issues

We identified major issues in 116 (nearly 23%) of the 513 cases in the
general controlled substances sample after the sample was reconfigured to focus
on more challenging cases. In our Fourth Report, we discussed four major issue
cases. In one of those cases, the analyst reported a finding without having made
a definitive identification of the substance. Three of those cases involved the
reporting of quantitative results for a liquid substance, even though quantitative
analyses were not performed. We identified numerous cases with deficiencies
similar to these two categories of major issues in the remainder of the general
controlled substances sample.
a.

Reports Based on Visual or Physical Identifications
Without Analytical Testing

During this stage of our review, we identified 72 major issue cases related
to scientifically unsound policies and procedures.108 These practices were first
noted in our Fourth Report discussion of issues identified in the Price and Patel
samples. The Crime Lab allowed analysts to identify unknown tablets by using a
reference source, such as the Physician’s Desk Reference (“PDR”) or the Drug
Enforcement Administration’s Logo Index (“Logo Index”). These results were
107

There were 383 cases in the original general controlled substances sample.

108

In a few of these cases, we identified more than one major issue. Therefore, the total
number of times that our review identified a particular type of major issue is greater than
the total number of cases containing one or more major issues.

HPD Crime Lab Independent Investigation

sometimes reported as if the identity of the tablets was established or confirmed
through actual analytical testing. While it is conditionally acceptable to identify
tablets by comparing their physical characteristics to the description provided by
a reference source,109 it is unacceptable to report such results without
acknowledging that analytical procedures were not used to identify the
substance. If the reports had stated, for example, that the items were “physically
identified as” a particular substance, they would have been marginally
acceptable.110 However, in these 72 cases the report did not provide any
indication that no analytical testing had been performed on the tablet.
b.

Presumptive Quantitation of Liquids and Tablets

It also appears to have been customary practice in the Controlled
Substances Section to presume that liquid codeine cough syrup would not have a
concentration greater than 200 mg of codeine per 100 mL of liquid. Although
this might generally be the case, it might not always be true because it is possible
for higher concentrations of codeine to be present in a solution. Since the
issuance of our Fourth Report, we have identified an additional 36 cases that
involved this type of “presumptive” quantitation of liquid samples.
We identified similar issues in cases in which identifications and
quantitations were based merely on the visual inspection of tablets. Two cases
involved the reporting of quantitative results based on the quantitation given in
the Logo Index. In both of those cases, the analyst compared the physical
109

This process is also referred to as “visual” or “pharmaceutical” identification.

110

In contrast, some cases contained clear disclosures that no analytical testing was
performed, such as the following: “Pharmaceutical identification only. No chemical
analysis performed. If analysis is required for prosecution purposes, please contact this
laboratory at least one week prior to trial date.” Because this is a marginally acceptable
practice, we classified such cases as involving minor issues. Although these cases contain
better disclosure of the limited work performed by the analysts, they remain deficient
because they involve the identification of substances without analytical testing. Tablets
are easily counterfeited, and an analyst can only be certain of the identity of a substance
by delving deeper than its surface physical appearance. In fact, the Crime Lab changed
its policy to require analytical confirmation of visual identifications after members of the
District Attorney’s Office and the Lab expressed concerns about the risk of incorrect
identifications of counterfeit drugs. Controlled Substances Section meeting notes reflect
that, starting in April 2004, analysts were instructed to perform a full analysis on any
tablet that is reported and not retained. Once again, we note that it is important for the
Crime Lab to consolidate its policies and procedures so that analysts have clear guidance
regarding the Lab’s current policies.

HPD Crime Lab Independent Investigation

appearances of a tablet -- such as its color, shape, and markings -- to those of
known tablets listed in the Logo Index. After physically identifying the tablet as
dihydrocodeinone (Vicodin) in this manner, the analyst reported the quantitation
associated with that tablet based on information in the Logo Index.
By reporting the liquid and tablet quantitations in this manner, analysts
implied that quantitative analyses were actually performed. At the very least,
the reports should have clearly stated that quantitative results were based on
Crime Lab protocol permitting presumptive quantitation or pharmaceutical
identification, and not on the performance of actual analyses.111 However,
standard forensic laboratory practice is to perform analytical testing to determine
the quantity of the controlled substance in a sample before reporting quantitative
results, especially when the sample is a liquid found in containers such as baby
and soda bottles, rather than sealed pharmaceutical containers.
c.

Failure to Report Probative Findings

We identified two cases in which analysts did not report the presence of a
controlled substance and, therefore, failed to report a probative finding. The
GC/MS testing performed in both cases showed the presence of heroin and two
non-controlled substances, papaverine and noscapine. In addition, the physical
appearance of the samples suggested that the substances were either black tar
heroin or opium. The analysts reported that there were no controlled substances
present, perhaps because the heroin peak on the GC/MS printout was not as
strong as those of the two non-controlled substances. However, the heroin peak
was strong enough in both cases to support reporting its presence in the
substance.
d.

Mistakes and Inaccuracies in Reports

In two cases, we discovered documentation or transcription errors that
had the potential to have seriously affected the prosecution of a defendant. In
111

We understand that the Crime Lab might still be following this policy when reporting
quantitative results for certain substances. The Crime Lab’s current SOPs, which were
updated in 2004, rarely require analysts to report quantitative results. For example, HPD
analysts no longer routinely report quantitative results for cocaine, heroin, and
methamphetamine. However, analysts should report quantitative results for certain
controlled substances (such as codeine, dihydrocodienone, dihydrocodeine, and
morphine) in order to establish the applicable penalty group under Texas law. See Texas
Controlled Substances Act §§ 481.102-481.105.

HPD Crime Lab Independent Investigation

one, the weight of cocaine was reported incorrectly in terms of grams, rather than
in milligrams, thus overstating the amount. This mistake did not ultimately
affect the defendant’s prosecution because he pleaded guilty to aggravated
robbery.
The second case involved two marijuana samples and two defendants.
The HPD analyst made an error when transferring information from a worksheet
to a supplemental report and misreported the weights of both marijuana
samples. As a result, the marijuana sample that weighed less than one ounce
was reported as weighing more than fifteen ounces. This could have affected
either defendant’s sentence (if one sample of marijuana was attributed to each
defendant) because possession of the smaller amount of marijuana is a
misdemeanor under Texas law, while possession of fifteen ounces of marijuana is
a felony.112 Both defendants were ultimately found guilty of charges relating to
their possession and delivery of cocaine, not marijuana.
e.

Unexplained Test Result Change

The final major issue case identified in the general controlled substances
sample involved unexplained changes to test results. The analyst recorded that
43 Ruybal color tests exhibited a negative finding for a controlled substance.
Based on our handwriting comparison, it appears that someone other than the
original analyst changed the results recorded on the worksheet from “negative”
to “pos blue,” thus indicating the presence of methaqalone, a controlled
substance. There are no initials or other markings indicating who made the
change, when it was made, or why the change was made. We could not even
determine if the change was made because the original analyst misinterpreted
the color tests or if the substance was tested a second time and produced
different results.
2.

Minor Issues Identified in the General Controlled
Substances Sample

In addition to the major issues described above, we identified over 200
cases that contained minor issues in the general controlled substances sample.
We consistently found that most minor issues should have been detected
through administrative and/or technical reviews.

112

Texas Controlled Substances Act § 481.121.

HPD Crime Lab Independent Investigation

Inadequate Documentation

In a large number of cases, we found that case documentation lacks detail.
This problem alone made it sometimes difficult to review controlled substances
case files because we were unable to determine what processes the analysts
employed when analyzing the cases and reporting results.
The documentation issues that we discovered included reference spectra
that were mislabeled. For example, we found cases in which the spectrum for
codeine was labeled incorrectly with the promethazine label, “stdprometh.Sms.”
In some cases, documentation was lacking regarding standards and blank runs.
Additionally, incorrect case numbers and dates were recorded on instrument
printouts and in other sections of the reports. In other cases, analysts failed to
record complete quantitation calculations.
Documentation issues were also apparent in the many discrepancies we
noted in instrument printouts, worksheets, and reports regarding the number of
pieces of evidence being evaluated. Case files often lacked complete descriptions
of the physical appearance of evidence; descriptions of the extraction and
preparation processes; initials, dates, and notations to indicate who, when, and
why changes were made to a report. Some amended reports carried the same
dates as the original reports, and analysts failed to maintain both the amended
and original reports in the files.
b.

Failure to Follow Generally Accepted Forensic
Science Practices

We found pervasive evidence in the general controlled substances sample
that section analysts did not follow generally accepted forensic science practices.
For example, analysts sometimes used sample sizes that were too small when
conducting quantitative analyses. Section analysts sometimes reported results
based on inadequate instrument runs, poor instrument standards, and poor
matches between the standard and the sample. Other examples of deviations
from generally accepted forensic science practices include the following:
•

Analysts used samples extracted for UV testing for GC/MS and FTIR
testing rather than preparing separate samples.

•

Supplemental reports did not contain enough detail when analysts
reported that a controlled substance was found but did not specify which
particular item of evidence contained the substance.

HPD Crime Lab Independent Investigation

•

Analysts routinely modified the submitting officer’s description of
evidence on the Property Record by adding to, deleting from, and
completely altering the description.

•

Analysts held evidence in their custody for excessively long periods of
time.

•

An analyst conducted the technical review of his or her own report.113

Our case reviews showed that, in a number of circumstances, analysts
failed to conduct additional testing that was warranted. For example, further
testing should have been performed when minor peaks indicated the possible
presence of another controlled substance. Additional testing also should have
been performed to confirm the identity of a substance in cases for which only
non-confirmatory testing was performed. Finally, we noted cases in which
further testing should have been done to determine whether a cigar or cigarette
contained a controlled substance other than marijuana.
In cases handled by several different analysts involving various drugs, we
found the use of inadequate instrument-generated reference standards.114 Some
laboratory instruments conduct a library search and provide a list of results for
the standard that most closely matches that of the unknown substance. In the
Crime Lab, some instruments consistently provided reference standards that
were inadequate because the standards lacked sensitivity, provided incomplete
ion patterns, and even, at times, provided standards that lacked major ions.
Because the Crime Lab’s libraries of instrument-generated reference
standards were often inadequate, analysts should have compared sample results
to standards provided in forensic literature. Alternatively, if the analysts were
going to rely on the instruments’ libraries, Controlled Substances Section
supervisors should have ensured that the library standards were accurate by
providing technical review of the standards, comparing them to standards in
forensic literature, and purging and adding standards to the library as needed.
113

A “technical review” is review by another qualified person of an examiner’s notes, data,
and other documents that formed the basis for the examiner’s conclusions. While an
analyst may administratively review his or her own work for mistakes in spelling and
transposing information, an analyst should not conduct his or her own technical review.

114

“Standards” are the analytical results of known substances. An analyst can compare the
instrument run readout of an evidence sample to that of a standard to determine if the
sample is the same as the known substance.

HPD Crime Lab Independent Investigation

The Crime Lab’s use of instrument-generated standards seems to be improving,
but supervisors and analysts should be aware of this issue and be alert for any
standard that appears to be inadequate.
c.

Failure to Follow Crime Lab Standard Operating
Procedures

We classified a number of cases as involving minor issues because the
analysts failed to follow the Crime Lab’s SOPs that were in place at the time that
the analyses were performed. For example, some case files were missing
information required under the SOPs, such as the analysts’ initials and evidence
that an administrative review was performed. In others, the analysts did not
include all three of the following required items in the case file: the notice of
modification, original supplement, and modified supplement. We also found
that examination sheets were sometimes not numbered correctly.
Some case files were deficient because analysts performed testing
adequate to identify the substance but did not perform all of the SOP-required
testing. In others, results of microcrystalline testing were not peer-reviewed.
3.

Handling and Labeling of Evidence

It has been extremely difficult to track evidence in controlled substances
cases involving more than one item of evidence because analysts often identified
items differently at each stage of the process -- on the officer’s report, on the
Crime Lab’s worksheet, and on the Lab’s supplemental report. Because we had
difficulty tracking items of evidence, we suspect that officers, attorneys, and
defense experts might well have experienced similar problems in attempting to
match test results to particular pieces of evidence. This is especially important
when certain items of evidence are attributed to a suspect, but others are not.
Accepted forensic science practice is to assign a unique identifier to each item of
evidence that is used throughout the investigation.
Our review of the general controlled substances sample raised additional
concerns regarding HPD’s general procedures for submitting evidence to the
Crime Lab. In particular, we will look further at the following: (1) the process of
submitting evidence via after-hours drop boxes; (2) the training given to officers
regarding the proper packaging of evidence; and (3) the security precautions
used to ensure the integrity of evidence. These and other issues relating to
Central Evidence Receiving, the Property Room, and evidence handling will be
addressed in our final report.

HPD Crime Lab Independent Investigation

Results of the Patel Case Reviews

We have completed our review of the 366 cases in the Patel sample, and 18
of these cases involved major issues. We reported on 14 of those cases in our
Fourth Report, which included discussion of another potential drylabbing
incident.115
1.

Reporting an Undocumented Finding

In the Patel sample, we observed the reporting of an undocumented
finding, which we classified as a major issue. Mr. Patel reported that the
evidence, two paper squares, contained lysergic acid diethylamide (LSD). His
preliminary testing of the evidence, which was done through color, UV, and
thin-layer chromatography (“TLC”) testing, indicated the presence of LSD.
However, Mr. Patel failed to properly document the results of the GC/MS test he
ran as a confirmatory test. The case file contains a reference standard for lysergic
acid methylpropylamide (LAMPA), a compound that is closely related to LSD.
The case file documentation indicates that the library on the GC/MS instrument
automatically matched the mass spectrum for the evidence with that of LAMPA,
and not LSD. While the LAMPA reference standard is of poor quality and we
believe that the mass spectrum test results indeed indicated the presence of LSD,
the only documentation for the mass spectrum portion of the GC/MS testing is
the LAMPA reference standard found in the file.
Moreover, although the GC peak in this case matched that of LSD in the
gas chromatography portion of the GC/MS run, there is no GC reference
standard for LSD in the case file. Despite the lack of reference standards
showing that the GC/MS spectrum of the evidence matched known samples of
LSD, Mr. Patel recorded “pos LSD” in the GC/MS block on the worksheet for
this case. While there is little doubt that the evidence is indeed LSD, because the
preliminary tests indicated LSD and both portions of the GC/MS run match
standards for LSD, it is unclear -- based on the documentation in the case file -how Mr. Patel came to and could ultimately support that conclusion if
challenged.

115

Discussion of our previous reviews of Mr. Patel’s casework can be found in our Fourth
Report at pages 61-64.

HPD Crime Lab Independent Investigation

Reports Based on Visual or Physical Identifications
Without Analytical Testing

In three of Mr. Patel’s cases, he reported the identity of a tablet or capsule
after performing a physical identification, but he did not clearly report that the
identification was based on a physical/pharmaceutical comparison, and not on
analytical testing. As with our general controlled substances sample, we
classified such an identification as a major issue because it is potentially
misleading.
3.

Inadequate Documentation

As described in our last report, Mr. Patel’s work contained numerous
documentation deficiencies. We identified a number of cases with minor
documentation issues similar to those noted in the general controlled substances
sample. In one case, Mr. Patel failed to document and explain a discrepancy
between the weight he recorded for the evidence and the weight recorded by the
officer. A second case with deficient documentation concerned when and why
second tests were run or reanalysis was conducted on evidence. In a third case,
Mr. Patel failed to report the net weight of the evidence. In a few multiple
evidence item cases, Mr. Patel failed to give any indication, either by item
number or description of item, as to which item of evidence was being analyzed
on a particular GC run.
4.

Failure to Follow Generally Accepted Forensic Science
Practices

In addition to documentation deficiencies, we also noted a number of
other minor issues with Mr. Patel’s cases. For example, we have reviewed cases
in which Mr. Patel departed from generally accepted forensic science practices
when he modified officers’ descriptions of evidence and entered new
descriptions on property records.116 He consistently failed to initial corrections
and additional notations on worksheets, and he also held evidence for months at
a time.
Mr. Patel failed to conduct further testing, such as GC/MS testing, to
confirm that evidence was indeed negative for controlled substances after
receiving negative results on UV tests. When performing GC/MS testing,
116

The Controlled Substances Section’s SOPs instructed analysts to modify the property
records in this manner. This SOP deviates from generally accepted forensic science
practice.

HPD Crime Lab Independent Investigation

Mr. Patel reported only cocaine on the worksheet, while the GC/MS identified
both cocaine and procaine. He used poor quality reference standards, which had
incomplete ion patterns, lacked sensitivity, and/or were missing major ions.117
Finally, he failed to identify minor peaks on GC and GC/MS results, which may
have led to the identification of other substances found in the evidence.
5.

Failure to Follow Standard Operating Procedures

In a number of cases, Mr. Patel’s work did not comply with the SOPs for
the Controlled Substances Section. For example, he
•

Failed to document the performance of macroscopic analyses of
marijuana.

•

Failed to perform quantitative analysis on pipe residue.

•

Failed to perform all of the required tests.

•

Failed to have microcrystalline testing verified by another analyst.

•

Failed to record weights and dates as required.
C.

Results of the Price Case Reviews

We have reviewed all 342 cases in the Price sample. We have identified
major issues in 11 of those cases, 7 of which were previously described in our
Fourth Report.
1.

Major Issues

Two of the newly-identified major issue cases involved the identification
of dihydrocodeinone (Vicodin) based on a comparison of the mass spectrum of
the evidence to a mass spectrum of a standard. The standard was of poor quality
and did not show the principal peaks and ions necessary to identify the
substance. Mr. Price’s identification was based on an inadequate standard that
was missing necessary major ions and was not consistent with generally
accepted forensic laboratory practices. However, we concluded that the
substance likely was dihydrocodeinone because the sample spectra in the cases
were of good quality and matched the spectrum for dihydrocodeinone found in
forensic literature.
117

As discussed in greater detail above, the use of poor standards was a section-wide
problem.

HPD Crime Lab Independent Investigation

In a third case, Mr. Price failed to perform a critical examination when
analyzing evidence. Mr. Price identified evidence as benzocaine, a
non-controlled substance, based only on presumptive testing. Mr. Price should
have conducted definitive testing on the chunk substance in order to determine
that controlled substances were not present.
Finally, Mr. Price reported the identity of a tablet or capsule after
performing a physical identification, and he neglected to report that his
identification was based only on physical/pharmaceutical testing. As described
above within our discussion of the general controlled substances sample, we
determined that such identifications constitute major issues.
2.

Minor Issues
a.

Failure to Follow Generally Accepted Forensic
Science Practices

Numerous cases analyzed by Mr. Price contained minor issues, most of
which were also found in the general controlled substances sample. For
example, Mr. Price did not follow generally accepted forensic science practices
when he kept evidence in his custody for months at a time, failed to initial his
corrections, and modified the submitting officers’ descriptions of evidence. We
also found that, after screening evidence for marijuana, he frequently did not
screen it for other possible substances.
Mr. Price incorrectly used reference samples by using poor reference or
standard samples,118 comparing a mass spectrum of the sample to that from
another case rather than a laboratory standard, and using a poorly resolved
spectrum to make an identification. Moreover, he failed to identify extra peaks
on GC results after identifying one controlled substance.
b.

Failure to Follow Standard Operating Procedures

Mr. Price also failed to follow the Controlled Substances Section’s SOPs
that were in place at the time of his analyses. In some cases, for example, he
failed to document that he had conducted a macroscopic analysis on marijuana
evidence. In others, Mr. Price did not record the reagents used for
microcrystalline testing or the amount of sample used for quantitation, and he
did not perform a UV analysis on residue evidence. Additionally, many of
Mr. Price’s cases failed to comply with the required SOPs because they did not
118

This issue is discussed above with regard to the general controlled substances sample.

HPD Crime Lab Independent Investigation

receive technical review. Finally, in a bulk case, a supervisor had not initialed
the weights on the worksheets or the backs of the photographs, and the
measuring device used in photographs was unreadable.
c.

Insufficient Documentation and Administrative
Errors

We found insufficient documentation and administrative errors in several
cases in the Price sample. For example, in some of Mr. Price’s cases, we observed
that he entered incorrect dates on instrument printouts.119 We also found that
Mr. Price recorded an incorrect incident number in one case; in others, he
inaccurately numbered worksheet pages and recorded results in the wrong boxes
on the worksheets. We also observed that neither Mr. Price nor a supervisor
initialed entries or corrections on the worksheets. Finally, Mr. Price failed to
include the original and revised supplemental reports in the case files.
D.

Results of Bulk Case Reviews

We reviewed 50 bulk evidence cases; 2 contained deficiencies
characterized as major issues, and 42 involved minor issues. The major issues
involved the reporting of liquid quantitations without performing quantitative
analyses, an issue that is described in greater detail above.
We found a number of minor issues that involved documentation
deficiencies and the failure to follow generally accepted forensic science
practices. Analysts regularly failed to follow SOPs that specifically targeted the
handling of bulk evidence. Moreover, the Crime Lab’s SOPs and other manuals
offered little guidance regarding which cases were to be classified as bulk
cases.120
Even when analysts did attempt to follow SOPs for bulk cases, we noted a
number of deficiencies. For example, the supervisor’s initials, analyst’s initials,
and Crime Lab number are not present on photographs of the evidence. A
supervisor neither observed the weighing of evidence nor initialed the recorded
weight. Photographs of bulk evidence were not taken at close enough range to
119

As was noted in our Fourth Report, the Controlled Substances Section supervisors
allowed one GC/MS machine in the Crime Lab to consistently print incorrect dates,
rather than have the machine repaired.

120

The definition found in three different SOPs, including the January 2006 SOP, gives only
the following guidance: “A bulky case is defined as any controlled substance case for
which a representative sample must be taken and preserved.”

HPD Crime Lab Independent Investigation

show markings on evidence or the officer’s identifying marks, and the ruler or
marker used for establishing size and scale in photographs was either missing or
unreadable. In some cases, items that were unrelated to the case were not
hidden from view and were shown in the photographs with the bulk evidence.
E.

Lack of Consistent Technical Reviews

The nature of the pervasive problems that we discuss above strongly
indicates that many of the issues we have identified should have been detected
by Controlled Substances Section supervisors during administrative or technical
reviews. The technical review process is what enabled section supervisors to
identify the four drylabbing episodes involving Price and Patel that were
described in our Second and Third Reports. However, the deficiencies identified
during Phase II lead us to question the effectiveness and completeness of the
administrative and technical reviews completed in the Controlled Substances
Section during the period of our review.121
In particular, it appears that in many cases section supervisors merely
went through the motions of conducting technical reviews. Even more
disturbing, it appears that supervisors sometimes conducted this lackluster
review after the analyst issued the report. These poor practices sent a message to
the entire Controlled Substances Section that sloppy work would be tolerated
and accounts for many of the issues identified in our review.
Despite all of the issues we have found as a result of our review of
hundreds of cases, we believe that the Controlled Substances Section generally
performed good work and that the quality of its work has improved over time.
The section has taken steps recently to improve some of its policies. For
example, we reviewed the current Controlled Substances Section SOPs that
describe procedures for analyzing controlled substances and dangerous drugs,
and those materials are clearly written and contain the essential information.
VI.

Firearms

The forensic examination of firearms evidence involves, among other
things, microscopic bullet, cartridge casing, and shot shell comparisons. The
Crime Lab’s Firearms Section also currently performs test firing, trigger pull
determinations, serial number restorations, and muzzle-to-target distance
121

Unlike the DNA Section for much of its history, the Controlled Substances Section had a
full complement of three Criminalist III supervisors to perform such technical reviews.

HPD Crime Lab Independent Investigation

determinations. These examinations can be used to match ammunition
components to the weapon that fired them, link different crimes committed with
the same weapon, and provide investigative leads.
We have reviewed 94 firearms cases since the issuance of our Fourth
Report, for a total of 184 firearms case reviews completed during Phase II of our
investigation. With the cooperation of the Crime Lab and the Harris County
District Attorney’s Office, we are working to obtain evidence related to several
firearms cases in which the evidence was previously released to courts for use at
trial. We are progressing steadily through the variety of firearms cases selected
for our sample.
Overall, we have found that in the vast majority of cases processed in the
Firearms Section during the relevant time period -- 1998 through 2004 --analysts
conducted proper examinations and reported their findings in a timely manner.
We have identified only minor issues, generally involving slight deficiencies in
documentation, deviations from Crime Lab policies, and deviations from
generally accepted forensic science practices. We found that the examinations of
the vast majority of the cases are technically correct and that most issues are of an
administrative nature. Several of the 94 cases reviewed since our Fourth Report
included an extensive number of components and required careful analysis. We
found that HPD firearms examiners operated with an impressive level of skill
and diligence in performing their examinations and correctly making
identifications.
In our Fourth Report, we noted a tendency in the Firearms Section to
avoid reporting results as inconclusive, even when this would have been the
most appropriate conclusion. Though none of the cases in which examiners
failed to report their results as inconclusive involved identifications or
eliminations, we were troubled by examiners’ apparent reluctance to report less
than definitive results. This issue was not present in any of the cases reviewed
since our Fourth Report.
A.

Method Related to Firearms Examinations

The examinations required in firearms cases vary greatly depending upon
the types of evidence involved. Fired bullets, cartridge casings, and shot shells
are examined with a comparison microscope, which enables the examiner to

HPD Crime Lab Independent Investigation

view side-by-side images of the ammunition components. Proper use of the
comparison microscope requires a great deal of time, patience, and experience.122
Where documentation in the Crime Lab’s case files was not sufficient to
permit us to evaluate the reported conclusions, we reviewed evidence to assess
the original work performed by the Lab’s Firearms Section examiners. This has
been particularly necessary in comparison and identification cases because, prior
to 2004, the Crime Lab’s SOPs did not require Firearms Section examiners to take
photographs, make drawings, or otherwise document their observations that
form the basis for their conclusions in such cases. Some cases we reviewed
contained multiple bullets or cartridge casings, and thus a single case involving
fired ammunition components may take considerable time to review. Other
cases -- such as trigger pull examinations, serial number restorations, and test
fires -- generally can be reviewed based on the documentation contained in the
Crime Lab file and, therefore, are less time consuming.
To date, our review of firearms cases has progressed smoothly, and we
still have identified no major issues through our case reviews. The minor issues
we have identified include insufficient documentation of examinations and
inefficient or inappropriate deviations from generally accepted forensic
laboratory practices.
B.

Documentation of Examinations

In our Fourth Report, we noted that that Firearms Section examiners did
not consistently document all of the necessary information required by generally
accepted forensic laboratory practices. We have identified fewer documentation
issues in the cases that have been reviewed since then, but we continue to
observe some deficiencies in this area.
Crime Lab examiners occasionally perform serial number restorations on
firearms when the number, which can track previous ownership of the firearm,
has been altered or obliterated. Most members of the forensic science
community document serial number restorations photographically in order to
create a record of what may be a transitory restoration. It is currently not the
policy of the Firearms Section to make a photographic record of serial number
restorations. Since our Fourth Report, we have found five additional cases in
which serial number restorations were not documented. Of the five types of
122

See Fourth Report at 67-68 for a more detailed summary of the methods used in firearms
examinations.

HPD Crime Lab Independent Investigation

documentation deficiencies discussed in our Fourth Report, this was the only
deficiency of which we found further examples.
In four other cases, we found that Firearms Section examiners did not
thoroughly document the results of the examinations they performed. For
example, an examiner in one case provided inadequate notes regarding a shot
shell comparison. In another case, the examiner was able to predict a weapon
based on general rifling characteristics (“GRCs”), but failed to include notes in
the case file regarding the lands and grooves studied during the examination.
We confirmed that the identifications in both of these cases were correct, but we
were concerned not to find more detail in the examiners’ notes. As discussed in
our Fourth Report, the Firearms Section was not accredited before 2004. The
Firearms Section’s current SOPs require more thorough documentation.
We found two cases in which information that was documented in an
examiner’s notes was not included in the final reports of those cases. In other
words, the examiner performed and documented the necessary analysis but did
not include the results in the report that reached the investigators. In one of
these instances, the examination involved cross-comparisons to evidence from
other cases. While the examiner’s notes show an elimination (meaning that the
evidence from a second case did not match that of the first), the elimination was
not reflected in the final report of the first case. This failure to include
examination results in final reports may have caused investigators to use their
time ineffectively.
Lastly, in one case, we found that the final report contained different
results than the notes related to the examination would suggest. Even though the
notes state that the firearm’s “feeding is difficult,” the report states that the
firearm “functioned as designed.” While it is possible that initial difficulties with
the firearm were resolved, examiners should document all steps of their
examinations, including changes in the functioning of the firearm.
C.

Inefficient or Inappropriate Deviations from Generally Accepted
Laboratory Practices

We have identified several departures from generally accepted laboratory
procedures and documentation practices. These departures included:
•

Using correction fluid or tape on worksheets;

HPD Crime Lab Independent Investigation

•

Failing to examine cartridge casings and shot shells found inside the
chambers of submitted weapons;123 and

•

Performing a trigger pull examination on every firearm submitted to the
Crime Lab for examination.124

In the 94 cases reviewed since the Fourth Report, we found 7 additional
instances in which an examiner used correction fluid on a worksheet. The
generally accepted laboratory practice is to mark through errors in work notes
with the single stroke of a pen, write the correct information to the side, and
initial the correction. In the cases examined since our Fourth Report, we found
five instances in which an examiner wrote on top of the error, instead of to the
side, and five instances in which an examiner made a correction without signing
his or her initials. Without initials, one cannot determine whether the original
examiner or a subsequent reviewer made the correction.
A similar concern arose when we found 9 cases that lacked reviewer
information. These nine cases did not contain “date reviewed” or “reviewed by”
entries, meaning that examiners may have reviewed their own examinations.
While there is no evidence to suggest that the review was not performed by a
123

The Crime Lab has had a long-standing practice of not examining ammunition
components that are contained in the chambers of submitted firearms. The practice
apparently is based on the assumption that the cartridge casing or shot shell found in a
submitted weapon must have been fired from that weapon. While typically this
assumption is correct, failing to examine ammunition components in such instances
prevents the Crime Lab from detecting staged events in which, for example, ammunition
from another weapon is placed in a weapon at a crime scene. This possibility may
appear remote, but most laboratories would examine these cartridge casings and shot
shells for markings that might help identify bullets fired from the same firearm.
Although examining more cartridge casings and shot shells obviously requires the
examiner to spend additional time on a case, doing so may lead to useful evidence and
would be consistent with standard laboratory practice.

124

The Firearms Section performed trigger pull determinations on every firearm submitted
to the Crime Lab. The time spent on these determinations would, in our view, have been
better used comparing ammunition components received within the chambers of
submitted firearms and, where appropriate, performing muzzle-to-target distance
determinations. Where there is no issue regarding unintentional firing, an examiner does
not gain useful data by conducting a trigger pull examination. Furthermore, in the
course of test firing, an examiner can readily recognize weapons that appear to have
extremely light trigger pulls. We believe that examiners’ time would be better spent if
trigger pull examinations are performed only where there is a question regarding
unintentional firing or where the test firing identifies a light trigger pull.

HPD Crime Lab Independent Investigation

second examiner, the lack of “date reviewed” or “reviewed by” documentation is
a departure from generally accepted laboratory practices.
D.

Other Issues

As noted in our Fourth Report, we identified cases involving delays in the
reporting results of firearms examinations. While the SOPs required that
requests for analysis be performed in a “timely fashion,” two case files revealed
delays in reporting. In one case, fourteen months passed from the completion of
the examination until a report was issued. In the other case, eight months passed
from the time the case was received to the issuance of the report. Neither of the
case files documented a reason for the delay.
We also found several instances of minor typographical errors and
oversights. In addition to two cases in which an examiner did not sign a
supplemental report, we noted two instances in which an examiner did not
include the case number or date of submission on a supplemental report. We
also found five examples of minor typographical errors in reports. In two such
cases, the examiner also failed to include a printout of weapon possibilities (also
known as a CLIS file search) with the report. However, the results of the
searches were included in the supplemental report.
Lastly, as noted in our Fourth Report, Firearms Section examiners did not
perform muzzle-to-target distance determinations in any of the cases that we
reviewed. None of the cases reviewed since our Fourth Report included this
type of examination.
VII.

Toxicology

Forensic toxicology involves the analysis of body fluids and other tissues
for the presence of potential toxins, including alcohol and other drugs. At the
Crime Lab, most toxicology cases historically involved the analysis of urine and,
less frequently, blood to identify controlled substances and drugs of abuse.
However, after the Section supervisor’s performance on a competency test in
October 2003 was found to be unsatisfactory, the Crime Lab suspended all work
in the Toxicology Section.
In May 2005, the Crime Lab was accredited by ASCLD/LAB to perform
blood alcohol analysis only. It now has three analysts performing blood alcohol
testing, but still does not perform analyses for other drugs. Analysts from the
Toxicology Section also maintain and calibrate HPD’s breath alcohol analysis
equipment.

HPD Crime Lab Independent Investigation

The Toxicology Case Sample

In our Fourth Report, we discussed the selection of the toxicology case
sample from a universe of 1,555 potential toxicology cases handled by the Crime
Lab between 1998 and 2004. However, based on an initial review of 101 cases,
we determined that a significant number of the files that were identified as
potential toxicology cases actually involved other forms of analysis, particularly
controlled substances analysis.125 We developed new search criteria based on the
Crime Lab’s system for coding toxicology and controlled substances cases, and
PwC used these criteria to generate a new sample of 308 cases.
We have now completed our review of the toxicology case sample. Of the
396 cases126 comprising the first and second toxicology samples, only about half
(213) involved actual toxicology casework. Despite the adjustments that were
made in the search criteria, most of the remaining 183 cases in the combined
samples involved analysis associated with controlled substances casework,
rather than toxicology.
Our review of the 213 toxicology cases that were in the sample yielded
consistent results, and the deficiencies that we identified tended to fall into the
three primary categories described below. PwC evaluated both the scope and
the findings of our toxicology case reviews and determined that, in light of the
conservative parameters used to define the samples and the consistency of our
findings, the sample maintained a reliable level of statistical precision.
Therefore, no additional sampling of toxicology cases was necessary.
B.

Results of the Toxicology Case Review

We identified only one major issue case in our review of the 213
toxicology cases, which was discussed in detail in our Fourth Report.127
125

126

127

Our initial sample was selected based in part on the identity of the analyst performing
the casework. We discovered that some of these analysts worked, at various points in
time, in both the Controlled Substances and Toxicology Sections of the Crime Lab.
Our first sample was comprised of 101 cases and the second contained 308 cases, but the
total was reduced to 396 after duplicate cases (cases that appeared in both the first and
second samples) were eliminated from the total figure.
The only “major issue” identified in the toxicology sample involved the identification of
three drugs in a blood sample -- heroin, cocaine, and PCP -- without the use of a
confirmatory test. Moreover, some aspects of the test results were pharmacologically
unlikely and indicated possible sample contamination. These questionable results, the
possibility of contamination, and the absence of a second confirmatory test led us to
Footnote continued

HPD Crime Lab Independent Investigation

Eighty-six of the 213 toxicology cases reviewed (40%) had deficiencies that we
identified as minor issues. Most involved situations in which:
(1)

drugs were identified on the basis of only one independent test;

(2)

we found indications of potential sample contamination; or

(3)

case files did not contain adequate documentation of all the work
that may have been performed.

In all of these minor issue cases, other controlled substances or drugs of abuse
were properly identified, and we therefore concluded that these minor issues
would not have had a significant impact on the case.
1.

Confirming Initial Results With A Second Independent
Test

As we discussed in greater detail in our Fourth Report, generally accepted
forensic toxicology practices require two levels of testing, with initial screening
tests followed by independent confirmatory analyses. Screening tests commonly
used by the Toxicology Section included fluorescence polarization immunoassay
(“FPIA”) and TLC. At the Crime Lab, confirmatory tests were normally
performed using gas chromatography and mass spectrometry.
In the remainder of the toxicology file sample, we continued to find cases
in which a drug (or metabolite) was reported as present in the sample, but the
identification was based on a single GC/MS analysis -- i.e., there was no
documentation in the file of a second independent test. In other cases, the
GC/MS matches for some drugs were weak. Finally, one case involving blood
alcohol analysis was deficient because of the method of reporting the results. In
this case, the HPD analyst performed the analysis properly and in triplicate. The
reported result, however, was the average rather than the lowest of the three
results. The more generally accepted practice in forensic science is to report the
lowest result obtained. The difference was not significant in this case.

Footnote continued from previous page
conclude that the work performed by Crime Lab analysts in this case was inconsistent
with generally accepted forensic science practices.

HPD Crime Lab Independent Investigation

Indications of Sample Contamination

We identified potential sample contamination in 5 cases analyzed by the
Toxicology Section in 1998 and 1999. In each of these cases, the sample tested
positive for cocaine, but no cocaine metabolites were detected. Cocaine
metabolites are chemical substances that are rapidly produced by the human
body when it metabolizes, or processes, cocaine. When test results are positive
for cocaine but none of the several possible metabolites are detected, this is a
strong indication that the presence of cocaine is due to sample contamination
rather than ingestion of the drug.
It is important to note that, in each of these cases, the Crime Lab analyst
appropriately did not report the presence of cocaine in the samples. However, in
the face of such results, the proper practice would be to seek the cause of the
contamination and correct it. Because these problems appear to have occurred
during a limited time frame, it is possible that some corrective action was taken
by the Crime Lab, but there was no documentation in these files indicating what,
if any, remedial actions were taken. We have been told that there were times
when the Toxicology and Controlled Substances Sections shared equipment. If
so, this is one very likely potential source of contamination because of the
frequency with which cocaine is analyzed in the Controlled Substances Section.
3.

Documentation Issues

We also identified documentation deficiencies in some of the cases
reviewed. A number of toxicology case files did not contain adequate
documentation of work that may have been performed by the Crime Lab. In
some cases, the analyst’s summary sheet noted positive TLC tests for
cannabinoids or cocaine metabolites, but there was no other documentation (e.g.,
a copy of the chromatogram) in the case file. In others, the analyst’s summary
sheet indicated that more than one type of analysis had been run, but there was
no documentation of the additional results.
During our review, we also found instances of the obvious mislabeling of
chromatograms in a case file. The mislabeling was apparently not identified by
the individual who performed the original analysis or by the person who
performed the technical review of the file, raising some concern about the
thoroughness of that review.
Despite the above observations, the work that was documented in the files
we reviewed was generally satisfactory. Except for the issues that are noted
above, the files were well organized, the reviews were properly documented,
and the range of analytical procedures used was appropriate.

HPD Crime Lab Independent Investigation

Conclusion
This report summarizes the results of the case reviews we have conducted
thus far in Phase II of this independent investigation of the Crime Lab and
Property Room. Over the past eight months, we have reviewed nearly 2,300
cases analyzed by the Crime Lab in the forensic science disciplines of serology,
DNA profiling, trace evidence, controlled substances, firearms, toxicology, and
questioned documents. Our case reviews now are complete in all of these areas
except for firearms and serology, which -- in light of the serious and pervasive
problems we have identified with the Crime Lab’s serology work -- has been
expanded to include cases dating back to 1980. We are continuing to work with
HPD to identify convictions in the 1980s and early 1990s in which serology
performed by the Crime Lab might have played a role and then to review those
cases to evaluate the reliability of the serologists’ analyses. Unfortunately, the
process is not one that can be completed quickly because of the inherent
difficulties in identifying the relevant cases.
The case reviews we have completed since our last report continue to
reveal widespread problems with the Crime Lab’s analysis of biological
evidence -- beginning with serology and continuing after the advent of DNA
profiling -- during the entire period of our review from 1980 through 2002. In the
cases we have reviewed since the publication of our Fourth Report in January
2006, we have found additional examples of serologists and DNA analysts failing
to report probative results -- results that might have helped identify and convict
the guilty as well as results that might have exonerated the innocent. We have
found a clear and troubling pattern of reluctance in the Serology and DNA
Sections to report typing results that were not consistent with the blood types or
DNA profiles of either the victim or a known suspect; in many such cases, the
serology or DNA results were reported as inconclusive. We have also seen
persistent problems with analysts’ techniques, the interpretation of results, and
the failure to address potential sample contamination. The Crime Lab never
issued written reports containing the statistical significance of its serology typing
results, and the frequency estimates presented by DNA analysts in cases
involving mixtures of body fluids were often overstated by orders of magnitude.
We have found no semblance of an effective technical review program or quality
assurance regime to detect and correct these problems. As a result, they
continued unabated.
We have identified 50 serology cases and 43 DNA cases in which work
performed by the Crime Lab was unreliable or the reported results were

HPD Crime Lab Independent Investigation

Appendix A

Stakeholders Committee Members
Adrian Garcia
Houston City Council
Fran Gentry
National Association for the Advancement of Colored People
Sylvia Gonzalez
League of United Latin American Citizens
Rusty Hardin, Esq.
Rusty Hardin & Associates
Richard Li, Ph.D.
Sam Houston State University
Ashraf Mozayani, PharmD., Ph.D., D-ABFT
Harris County Medical Examiner Office
Frank Parish, Esq.
Justice for All and Parents of Murdered Children
Annise Parker
Houston City Controller
Wayne Riley, M.D.
Baylor College of Medicine
Benjamin Roa, Ph.D.
Baylor College of Medicine
Kent W. “Rocky” Robinson, Esq.
Houston Bar Association
Andrews Kurth LLP
Richard Ward, Ph.D.
Sam Houston State University
Anthony Woods, Ph.D.
Texas Southern University

HPD Crime Lab Independent Investigation

Appendix B

Serology Major Issue Cases
Aparicio, Jesus, L87-06849

Martinez, Guellermo, L90-08288

Authorlee, Carlos Miguel, L90-06462

Mays, Sammie, L89-07713

Ayarzagoitia, Porfiro, L89-04353

Miller, Roy Michael, L80-05462

Barton, Willie, L87-02871

Pacholsky, Charles, L89-04795

Block, Gerald, L86-01480

Phillips, Michael/Walker, Robert, L88-02292

Brown, Robert/Richard, Tony, L87-02402

Qualls, Roy Anthony, L89-07064

Bruno, Howard Joseph, L90-09806

Randall, Paul, L80-04956

Burkett, Harold, L89-01678

Riser, Dwight, L87-07602

Cosmi, Traian Daniel, L88-05490

Rodriguez, George, L87-02120

Garcia, Clemente, L80-05005

Salazar, Roland, L90-08015

Garcia, Eleazar, L80-04725

Salinas, Miguel, L90-07636

Gomez, Fermin Diaz, L89-09747

Sanchez, Joe, L87-03227

Gonzales, Frank, L90-09804

Stokes, Forrest, L90-05132

Guerra, Armand, L87-06353

Strom, Arling Michael, L88-08249

Hairrell, Gordon, L86-09235

Thomas, Earl, L93-04349

Hodge, Charles, L86-10546

Thomas, James, L92-00499

Ireland, Charles Franklin, L88-06846

Tyler, Patrick D., L87-00752

Jackson, Derrick, L88-08130

Unknown, L88-01139

Jackson, Norman, L89-06057

Unknown, L88-04309

Jackson, Willie Gerard, L89-05545

Unknown, L89-01224

Leblanc, Perry, L89-10423

Unknown, L89-06435

Luna, Jose, L90-02094

Unknown, L89-09783

Mangis, Troy Wayne, L87-02602

Unknown, L90-09703

Mao, Benjamin/Francis, Leonard, L86-06801

Williams, Joseph, L80-01135

Martin, Steve, L87-07876

Wyatt, Clarence/Chatman, Ray, L88-00099

HPD Crime Lab Independent Investigation

Appendix C

DNA Major Issue Cases
Alix, Franklin, L97-12163

Meza, Alfredo, L95-10460

Alvarez, Juan Carlos, L98-07378

Mingo, Michael, L97-13990

Boudreaux, Raymon, L97-00568

Napper, Laurence, L01-02205

Cantrell, Ronald, L01-17322

Nugent, Hermann, L93-12224

Carter, Harold, L94-02461

Parra, Carlos, L97-05353

Davis, Garland, L93-10985

Pineda, Johnny, L94-06976

Emory, Gregory, L96-05918

Preston, Terrance, L00-01952

Garcia, Luis, L98-09736

Rayson, Carl Lee, L96-13604

Gonzales, Jose James, L93-12489

Samuels, Michael, L95-13955

Guevara, Gilmar, L00-08053

Segura, Carlos/Zavala, Mark, L98-11877

Guevara, Luis/Fernandez, Sixto, L00-13216

Shields, Robert, L94-10028

Harris, Erskin, L95-08229

Smith, Charles Leon, L84-08387

Hayden, Robert, L94-01695

Southern, Ronnie, L95-03891

House, Dillard, L00-02780

Sutton, Josiah, L98-13476

Jackson, Derrick, L88-08130

Valdez, Richard, L96-05919

Jackson, Reginald, L97-12636

Vanzandt, Lonnie, L94-12745

James, Leon, L94-05670

Vaughn, Artice, L94-11539

Johnson, Arthur, L97-00479

Ware, Cory, L02-04949

Lawson, David, L96-03366

Ware, Marshall, L95-05151

Lewis, Leroy, L91-00532

Washington, Dedrick, L95-00745

Lewis, Roger, L93-02191

Zelaya, Alberto, L95-08103

Lopez, Segundo, L97-12346

HPD Crime Lab Independent Investigation

Appendix D

Acronyms
AE
AFIS
AI
AP
ASCLD
ASCLD/LAB
City
CODIS
DNA
FBI
FPIA
FTIR
GC
GRC
HPD
LAMPA
LSD
MS
PCR
PDR
PGM
PwC
QA
RFLP
SOP
STR
TLC
UV

absorption elution
Automated Fingerprint Identification System
absorption inhibition
acid phosphatase
American Society of Crime Laboratory Directors
American Society of Crime Laboratory Directors/Laboratory
Accreditation Board
The City of Houston, Texas
Combined DNA Index System
deoxyribonucleic acid
Federal Bureau of Investigation
fluorescence polarization immunoassay
Fourier Transform Infrared
gas chromatography
general rifling characteristic
Houston Police Department
lysergic acid methylpropylamide
lysergic acid diethylamide
mass spectrometery
polymerase chain reaction
Physician’s Desk Reference
phosphoglucomatase
PricewaterhouseCoopers LLP
quality assurance
restriction fragment length polymorphisms
standard operating procedure
short tandem repeats
thin-layer chromatography
ultraviolet