IN THE COURT OF CRIMINAL APPEALS
OF TEXAS
v.
THE STATE OF TEXAS
FROM THE THIRTEENTH COURT OF APPEALS
VICTORIA COUNTY
In his second ground for review, this pro se appellant asserts that the court of appeals erred in holding that the trial court correctly admitted expert testimony by an FBI analyst concerning comparative bullet lead analysis (CBLA). During the Daubert hearing at appellant's trial in 2003, appellant's attorney cited relevant critical studies that attacked the theoretical and methodological validity of the FBI analyst's proposed testimony. The trial judge, however, relied upon two cases from other jurisdictions that had admitted this type of expert testimony years before its flaws became apparent to the scientific community. Only a few months after this 2003 trial, the scientific basis for this expert testimony was largely discredited by National Research Council (NRC) in a study commissioned by the FBI. (1) The FBI ended all such comparative bullet lead analysis work and testimony in 2005. (2) Several recent published cases from other jurisdictions have reversed convictions that had allowed an FBI analyst to testify to this flawed scientific theory. The FBI is currently trying to identify all of the cases in which convictions were obtained in reliance on CBLA testimony so that they can be reviewed. (3) This is one of them.
Appellant was charged with the murder of a rival gang member. At trial, the State offered expert testimony from Charles Peters, a long-time FBI analyst, on comparative bullet lead analysis. The trial court held a Daubert/Kelly (4) hearing in which Mr. Peters was the sole witness. He stated that he has been doing comparative bullet lead analysis for the FBI since 1975. The FBI is the only entity that has ever done CBLA. By 2003, only two FBI analysts, Mr. Peters and Mr. Riley, did this work. No one outside the FBI had ever seen their process or how they arrived at their scientific conclusions. (5) According to Mr. Peters, "We're basically the only game in town, because of the expense of the analysis and local jurisdictions only need it on a very limited basis."
Mr. Peters's opinion, in a nutshell, was that the chemical composition of the bullets found at the murder scene matched the composition of bullets taken from appellant's apartment that had been distributed by the American Ammunition Company. He testified that they were analytically indistinguishable. Therefore, he concluded, "these bullets likely came from the same source or metal or lead at that manufacture, during the manufacturing process." In other words, the crime scene bullet came from the defendant's box of ammunition or another box that was produced from that mill at the same time. In his opinion, the bullets from the scene "likely" came from defendant's box of ammunition, "but I always put on the caveat- the many other boxes, it's as likely they were produced at the same time." Both Mr. Peters and the prosecutor repeatedly implied that the crime-scene bullet probably came from appellant's box of ammunition because they "matched," and because American Ammunition was a small manufacturer and distributor of bullets.
Mr. Peters testified that appellant's box of bullets was "finished" by American Ammunition, but he thought that the lead bullets themselves were manufactured by CCI-Speer which, he said, is one of the four largest lead bullet manufacturers in the U.S. They sell and distribute their lead bullets to innumerable smaller manufacturers for finishing (e.g., putting on the copper coat, milling, stamping, boxing, and distributing). "They sell this stuff all over the United States." (6)
Appellant's counsel told the trial judge that he had two specific issues to raise with Mr. Peters: (1) Is the scientific testing and analysis sufficient to establish that two bullets do, in fact, come from the same "pot of lead"; (8) and (2) If that comparison is possible, then what does Mr. Peters mean by "same pot of lead," how big of a sample is this, and what does he know about the manufacturing and distribution process of these particular lead bullets. During the Daubert hearing, appellant's counsel confronted Mr. Peters with several then-recent studies that contradicted his theory, methodology, and opinions. One of them, "A Metallurgical Review of the Interpretation of Bullet Lead Compositional Analysis," was written by Mr. Peters's former mentor at the FBI, Bill Tobin, and published by Forensic Science International. (12) Mr. Peters rejected his former mentor's criticisms, although they had been widely cited in the scientific and legal community. (13) Mr. Peters also downplayed the 1999 FBI-funded study by the Department of Statistics at Iowa State University-a study that led one of its principal investigators to conclude that "[t]he leap from a match to equal origin is enormous and not justified given the available information about bullet lead evidence." (14) Similarly, at the time of the 2003 Daubert hearing, appellant's counsel referred Mr. Peters to the then-pending NRC study. Mr. Peters suggested that the study would validate his methodology, (15) but, in fact, it did the reverse. (16) He also rejected the criticisms leveled at CBLA that were being raised by Professor Imwinkelried, a prominent legal scholar on forensic sciences, evidence, and the criminal justice system, (17) but Mr. Peters did admit that "the science is evolving."
Mr. Peters explained that if two bullets are "analytically indistinguishable" they must have come from the same lead "melt" at the same factory at the same time. (18) When defense counsel pointed out that the Iowa State study had expressed concern about the possibility that bullets from different "melts" might coincidentally match, (19) Mr. Peters said that he had nothing to do with that study.
Next, defense counsel asked about the "Keto" Study in the Journal of Forensic Science. (20) When defense counsel asked Mr. Peters if that study concluded that "general bullet lead analysis does not generate individualizing information," Mr. Peters responded, "It's been some time since I read it. It's a very limited study. . . . And he certainly isn't an expert in the bullet analysis world." The purport of Mr. Keto's criticism in 1999 was that even if one could say there was a perfect lead match, and even if one could say that the two bullets must have been manufactured from the same lead "melt," that tells the factfinder nothing about the number of bullets made from that "melt" or their distribution throughout the United States. (21)
After hearing Mr. Peters's testimony at the Daubert/Kelly hearing, the trial court ruled that he could testify before the jury to his expert opinions concerning his comparison between the crime scene bullets and those found in appellant's apartment. In making that ruling, the trial judge explicitly relied on two earlier cases, State v. Noel, 723 A.2d 602 (N.J. 1999), and United States v. Davis, 103 F.3d 660 (8th Cir. 1996), which had both admitted CBLA evidence as scientifically reliable. During the late 1990's, the major criticisms concerning CBLA had just begun, so those courts were not yet aware of its potential for scientific unreliability. (23) In the wake of the NRC Report, published opinions by several state courts have held that CBLA evidence is scientifically unreliable and, therefore, inadmissible. (24)
The problem in this case is not that appellant's counsel failed to raise, litigate, or preserve the right issue-the scientific reliability of Mr. Peters's proposed expert opinion testimony. He did. Admirably. But the trial record in this direct appeal does not contain all of the materials necessary to fairly resolve the issue that he preserved. Specifically, the trial record does not contain (and could not contain because it did not exist at that time) the NRC Report that defense counsel referred to, (25) as well as the more recent scientific studies and evidentiary materials that have followed in its wake.
Although we may take judicial notice of the recent cases that have reversed convictions based upon this "flawed science," we generally do not take judicial notice, on direct appeal, of other scientific or research materials that are not contained within the trial record. (26) Such "newly available" materials may, of course, be considered in post-conviction proceedings if they are placed before the trial court.
Especially when courts must deal with evolving science in the field of forensic evidence, judges will find treatises, studies, peer-reviewed articles, and the like helpful only if they are physically introduced into evidence so that both the trial and appellate courts can carefully consider the full spectrum of pertinent materials. Further, if those scientific materials are introduced into evidence, both parties have an opportunity to call witnesses to discuss whether those materials are generally accepted and relied upon in the pertinent scientific community. They may also question the source and validity of the data relied upon, the methodology used, and the conclusions reached by the authors. In sum, they may call witnesses in support or opposition to the theories, methodologies, and conclusions at issue. To ensure that only "good" science is admitted and "bad" science is excluded in our criminal trials, the parties must shoulder the responsibility of providing and explaining the appropriate educational materials for judges to make that determination. That was not entirely possible at the time of this 2003 trial as the science itself was "evolving," but a trial-court review of the testimony admitted in this case, based upon more recent scientific developments, may now be both possible and appropriate.
Thus, because the trial record is incomplete, I concur in the Court's decision to refuse appellant's petition for discretionary review on direct appeal.
Filed: February 24, 2010
Do Not Publish
1. National Research Council, Committee on Scientific Assessment of Bullet
Lead Elemental Composition Comparison, Forensic Analysis: Weighing Bullet Lead
Evidence (2004) (NRC Report) (available at:
http//www.nap.edu/catalog.php?record_id=10924).
2. The NRC Report was issued in February, 2004; appellant's trial took place in
September, 2003. Nonetheless, appellant's lawyer was aware of the pending report and its
criticism of Mr. Peters's methodology and testimony. The FBI issued a press release
discontinuing its lead bullet analysis on September 1, 2005, slightly more than two months after
the court of appeals issued its opinion in this case.
3. See Smith v. State, ___ So.3d ___, 2010 WL 21178 (Fla. Dist. Ct. App. Jan. 6, 2010)
(citing "60 Minutes" Report); Paul C. Giannelli, Comparative Bullet Lead Analysis-An Update,
23 Crim. Just. 24, 25 (2008) (noting that The Innocence Network and the NACDL had formed a
task force to work with the FBI to identify and contact defense attorneys and convicts to re-examine cases in which CBLA testimony had been admitted); Jessica D. Gabel & Margaret D.
Wilkinson, Symposium: "Good" Science Gone Bad: How the Criminal Justice System Can
Redress the Impact of Flawed Forensics, 59 Hastings L.J. 1001, 1006 (2008) (stating that FBI's
corrective action "apparently will include a nationwide review of all CBLA-related testimony and
notification to prosecutors so that the courts and defendants can be alerted"); see also "FBI
Reviews Cases Where Flawed Evidence Used," Houston Chronicle (Jan. 18, 2010) (available at
http://www.chron.com/disp/story.mpl/ap/nation/6821348.html) (last visited Jan. 26, 2010)
(noting three convictions, which had been based on CBLA evidence, that had been reversed).
4. See Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 593-94 (1993); Kelly v. State,
824 S.W.2d 568, 573 (Tex. Crim. App. 1992).
5. Mr. Peters testified in front of the jury that some scientists from the National Academy
of Scientists had recently come to his laboratory to observe his technique as they were compiling
their Report for the National Research Council.
6. The NRC Report concluded that too little is known about patterns of the distribution of
ammunition to establish whether bullets from the same compositionally indistinguishable volume
of lead (CIVL) are more or less likely to come from a given box of ammunition: 7. Id. at 6; See generally, D.H. Kaye, The Current State of Bullet-Lead Evidence, 47
Jurimetrics J. 99 (2006) (reviewing NRC Report, collecting pre- and post-Report cases and
concluding that "[u]ntil there is reason to believe that CIVLity is strongly indicative of a
defendant's association with both bullet-lead samples, the wiser course may be to exclude even
the NRC's sanitized version of CABL testimony"). -
- -
" '
- '
"
8. Mr. Peters testified that he "matches" lead samples by comparing the concentration of
seven different elements (arsenic, antimony, tin, copper, bismuth, silver, and cadmium) in the
bullet lead alloy of both the crime scene and suspect's bullets. (9)
9. Paul Giannelli, Comparative Bullet Lead Analysis, American Bar Association, 23 Crim.
Just. 24, 24 (Summer 2008) (outlining the FBI testing method and noting that "[e]xactly what the
phrase 'analytically indistinguishable' meant was the source of the problem").
""
' '
" "
"" - (10)
10. Edward J. Imwinkelried & William Tobin, Forensics Symposium: The Use and Misuse
of Forensic Evidence; Comparative Bullet Lead Analysis: Valid Inference or Ipse Dixit?, 28
Okla. City U.L. Rev. 43, 49 (2003).
"" "" (11)
11. Id.
"
' ' "
12. 127 Forensic Sci. Intl. 174, 182 (2002).
13. The article written by William A. Tobin and Wayne Duerfeldt, How Probative Is
Comparative Bullet Lead Analysis?, concluding that "the current practice of CBLA is
scientifically flawed, and that no scientifically or statistically adequate data exist to support its
foundation," appeared in the American Bar Association's publication, Criminal Justice, in the
Fall of 2002. 17 Crim. Just. 26, 33. At trial, Mr. Peters dismissed Mr. Tobin's work as that of a
turncoat after Mr. Tobin had retired from the FBI. Mr. Tobin's criticisms were ultimately cited
and relied upon in the NRC Report.
14. See Tobin & Duerfeldt at 32 (quoting Professor Alicia L. Carriquiry at Iowa State
University); see generally, Alicia L. Carriquiry, Michael Daniels, Hal S. Stern, Statistical
Treatment of Class Evidence: Trace Element Concentrations in Bullet Lead, Dept. of
Statistics, Iowa State University and Ames Laboratory (May 4, 2000).
15. Mr. Peters testified that "[i]t's my understanding the study has been completed and is
being written up and being printed at this moment and is supposed to be released in October."
He stated that he did "not believe it is a possibility" that the NRC study would deem his analysis
and testimony unreliable.
16. See note 19 infra.
17. Edward J. Imwinkelried & William Tobin, Forensics Symposium: The Use and Misuse
of Forensic Evidence; Comparative Bullet Lead Analysis: Valid Inference or Ipse Dixit?, 28
Okla. City U.L. Rev. 43, 49 (2003).
18. The NRC was highly skeptical of this inference. It concluded that the best that could be
said is that they "could have" come from the same lead melt, but there had been insufficient
comparative studies of lead "melts" to see if they are always, sometimes, or never
indistinguishable in their results. NRC Report at 7. In 2003, Professor Imwinkelried and
William Tobin concluded, "There is insufficient data to support any conclusions about the
uniqueness or individuality of each molten source." Imwinkelried, supra, at 51-52 (noting that
"[s]everal factors conspire to make it plausible that by chance, the composition of different
molten sources will be analytically indistinguishable.").
19. See Alicia L. Carriquiry, et al., supra note 10; see also NRC Report at 99 ("The
frequency of coincidentally identical CIVLs is unknown.").
20. See 21. The NRC Report stated that a single "melt" "can range from . . . as few as 12,000 to as
many as 35 million . . . compared with a total of 9 billion bullets produced each year." (22)
22. NRC Report at 6-7. The NRC Report states, "A conclusion that two bullets came from
the same melt does not justify an expert in further testifying that this fact increases the odds that
the crime bullet came from the defendant." Id. at 102.
- "
"
23. But see Andre Moenssens, et al., Scientific Evidence in Criminal Cases § 9.07,
at 541 (3d ed. 1986) (noting that elemental lead analysis of bullets to connect a crime scene
bullet with another one is problematic; "[a] flaw, however, in this logic, which may be fatal,
exists in the possibility that the number of bullets manufactured with similar elemental
characteristics by the same manufacturer may be so numerous as to make the likelihood of the
accused's link to the evidence bullet statistically tenuous.").
24. See State v. Behn 25. National Research Council, Committee on Scientific Assessment of Bullet
Lead Elemental Composition Comparison, Forensic Analysis: Weighing Bullet Lead
Evidence (2004). The NRC Report used the acronym CABL (compositional analysis of bullet
lead) instead of CBLA and includes these findings:
CABL is sufficiently reliable to support testimony that bullets from the same compositionally indistinguishable volume of lead (CIVL) are more likely to be analytically indistinguishable than bullets from different CIVLs. An examiner may also testify that having CABL evidence that two bullets are analytically indistinguishable increases the probability that two bullets came from the same CIVL, versus no evidence of match status.
Although it has been demonstrated that there are a large number of different compositionally indistinguishable volumes of lead (CIVLs), there is evidence that bullets from different CIVLs can sometimes coincidentally be analytically indistinguishable.
The available data do not support any statement that a crime bullet came from, or is likely to have come from, a particular box of ammunition, and references to "boxes" of ammunition in any form is seriously misleading under Federal Rule of Evidence 403. Testimony that a crime bullet came from the defendant's box or from a box manufactured at the same time is also objectionable because it may be understood as implying a substantial probability that the bullet came from defendant's box.
Id. at 107.
26. See Hernandez v. State, 116 S.W.3d 26, 31 (Tex. Crim. App. 2003) (although appellate
courts may take judicial notice of other appellate opinions concerning a specific scientific theory
or methodology in evaluating a trial judge's