Skip to main content

Flowers Case

The Reliability of Forensic Evidence: The Case of Curtis Flowers

Compiled by Jiaxin Zhu, Liangcheng Yi, Wenqian Ma, Ziyue Zhu, and Guillem Esquius



The case against Curtis Flowers involves three threads of forensic evidence: the bullets found in crime scene, the bloody shoeprint next to a victim’s body, and the gunshot residue found on Flowers’s right hand. These forensic evidence together helped send Flowers to death row, but critics argue that this field is subjective and unscientific.

For decades, the forensic science disciplines have produced valuable evidence that has contributed to the successful prosecution and conviction of criminals as well as to the exoneration of innocent people. Those advances, however, also have revealed that, in some cases, substantive information and testimony based on faulty forensic science analyses may have contributed to wrongful convictions of innocent people. In this report, we will dig into the reliability of some forms of forensic evidence including ballistics, shoeprints and gunshot residues.


What Happened in the Flowers Case

The investigators in the Curtis Flowers case never found the murder weapon that was used in 1996 to kill the four employees at Tardy Furniture. From the projectiles in the crime scene, they could determine that the murder weapon was a .380 caliber handgun. The same day of the murders, a man named Doyle Simpson, who was Curtis Flower’s step-uncle, reported to the police that his .380 had been stolen from his unlocked car that morning.

Investigators wanted to match the bullets in the crime scene to Doyle Simpson’s weapon. In order to do that, they dug two bullets from a wooden post on the property of Simpson’s mother, which Doyle had used as a shooting range. They compared those two bullets with one bullet that investigators had recovered from a mattress at the crime scene. The Mississippi Crime Lab in Jackson matched the bullets.

After Flowers had been found guilty in the first trial, the Prosecutors wanted a second opinion, and the court appointed David Balash, “a retired Michigan state trooper who’s a ballistics and explosives expert with decades of experience in the field,” to review the ballistics evidence in the case. At the subsequent five trials David Balash testified that the bullets found at the crime scene and the bullets found in the wooden post came from the same weapon. That is, the murder weapon and Doyle Simpson’s gun matched. In the last trial, Balash claimed absolute certainty regarding the match: “It has to be 100 percent or I will not offer that opinion.”1

Ballistics Evidence in a Larger Context


When guns are shot, the softer metal of the bullets explodes forward impacting the stronger metal of the barrel’s ridges and groves, leaving striations. The softer metal of the cartridge case explodes backward with equal force against the stronger metal of the mechanism that absorbs the recoil, called the breech face, leaving impressions.2

These striations and impressions are known as “toolmarks”.3

According to ballistics experts, the theory of firearm identification holds that these striations and impressions in bullets and cartridge cases, have microscopic imperfections due to random variations in manufacturing and repeated use,4 and that therefore, they are unique, like “ballistic fingerprints”. Consequently, “If investigators recover bullets or cartridge cases from a crime scene, forensic examiners can test-fire a suspect’s gun to see if it produces ballistic fingerprints that match the evidence”.5

To link a bullet with a weapon, the task of the examiner is a difficult one. He needs to compare the bullets and casings of a crime scene with samples from the weapon and “identify the individual characteristics of microscopic tool marks apart from the class and sub-class characteristics and then to assess the extent of agreement in individual characteristics in the two sets of toolmarks to permit the identification of an individual tool or firearm.”6 In other words, the examiner has to be able to tell apart the unique impressions caused by a gun, from those that the same model of gun cause (class characteristics) and even from those which not only the same model of gun but also were consequently manufactured (sub-class characteristics).

-  The bullets are compared by the sight of the examiner through a comparison microscope, which consists of two microscopes joined by a comparison bridge that allows the viewing of two objects at the same time.7

-  The samples are usually obtained by test firing the weapon in a controlled environment. Typically, the gun will be fired into a tank of water. By doing so, they reduce the risk that the marks on the bullet are due to the impact and not to the barrel’s “fingerprint.”8


     Unproven assumption

     The scientific community has simply not been able to prove that a gun will produce a unique mark, which is the fundamental assumption of ballistics evidence. Alicia Carriquiry, director of the Center for Statistics and Applications in Forensic Evidence research group at Iowa State University pointed out in an interview to APM Reports that the method “hasn’t been subjected to the kind of massive peer-reviewed testing that would be required to prove such a claim.”9

     “The validity of the fundamental assumptions of uniqueness and reproducibility of firearms-related toolmarks has not yet been fully demonstrated.”10

     “A significant amount of research would be needed to scientifically determine the degree to which firearms-related toolmarks are unique or even to quantitatively characterize the probability of uniqueness.” 11

     “Sufficient studies have not been done to understand the reliability and repeatability of the methods.”12


     Even if the fundamental assumption of guns leaving a unique mark is accepted, experts point out that the analysis is subjective: “what we’re actually talking about is somebody looking at two bullets and determining if they look similar,” “today this is still a largely subjective science,” and “you have this very undesirable situation where two examiners looking at the exact same samples might reach different conclusions.”13

     Lack of consistent standard and guidelines

     There is no fixed standard regarding the point in which the extent of agreement on individual characteristics between two sets of tool marks is enough to permit the identification of an individual tool or firearm nor any defined guidelines as to how should marks look like or with how many similar marks could one affirm a match with statistical significance. The Association of Firearm and Tool Mark Examiners (AFTE)14 refers to the standard as “sufficient agreement,” but it fails to clearly define what that is.15 “A fundamental problem with toolmark and firearms analysis is the lack of a precisely defined process.”16 It therefore relies fully on what each examiner decides based on their years of experience. “The decision of the toolmark examiner remains a subjective decision based on unarticulated standards and no statistical foundation for estimation of error rates.”17

     “Because not enough is known about the variabilities among individual tools and guns, we are not able to specify how many points of similarity are necessary for a given level of confidence in the result.”18

Weakness of the Ballistics Evidence in the Flowers Case

Apart from the abovementioned general deficiencies, the ballistics evidence in Flowers case has its own problems. First, the compared bullets were not obtained by test firing the weapon in a controlled environment. Since Doyle Simpson’s gun was never found, the examiners could not obtain a bullet in an optional condition and had to rely on the bullets extracted from the wooden post. Second, there might be possible damage to the bullets, as the investigators extracted the bullets from the wooden post using a knife and therefore possibly creating additional markings on the bullet. Proof of this danger is that in similar situations some experts like Kurt Moline, with nearly 30 years of experience, advised investigators to send whole sections of the tree containing the bullet rather than extracting it themselves.19

The Future of Ballistics Evidence

As we have seen, the landmark report from the National Academy of Sciences20 pointed out the major problems causing unreliability in the currently employed methodology.

The community of firearms analysts responded with new social sciences studies. A 2016 report from the President’s Council of Advisors on Science and Technology entitled “Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods”, identified “one notable advance since 2009”: “the completion of the first appropriately designed black-box study of firearms. The work was commissioned and funded by the Defense Department’s Forensic Science Center and was conducted by an independent testing lab (the Ames Laboratory, a Department of Energy national laboratory affiliated with Iowa State University). The false-positive rate was estimated at 1 in 66, with a confidence bound indicating that the rate could be as high as 1 in 46.” This study constitutes a good step towards being able to assess statistically the reliability of the method.

More recently, a team of researchers at the National Institute of Standards and Technology (NIST) published study in Forensic Science International.21 This study is more promising not only because it continues to develop a statistical approach able to assess error rates, but also because it tries to address the problem of subjectivity by eliminating the human factor. The authors built a mathematical algorithm that “compares three-dimensional surface scans of the breech face impressions on cartridge cases. Their method, called Congruent Matching Cells, or CMC, divides one of the scanned surfaces into a grid of cells, then searches the other surface for matching cells. The greater the number of matching cells, the more similar the two surfaces, and the more likely they are to have come from the same gun”.22 The NIST itself admits that the sample size was small and not enough to “calculate realistic error rates for actual casework”. However, they are optimist about the possible results of scaling up the study “with much larger and more diverse datasets”.23



What Happened in the Flowers Case

A bloody shoeprint was found next to one of the victims’ body. Because Sam Jones, the person who discovered the crime scene, stated that he did not remember seeing the shoeprint when he entered initially, police officers checked the footwear of all personnel at the scene and found that none was consistent with the shoeprint.

The shoeprint was later determined by Andrews to be consistent with Fila Grant Hill size 10 ½ tennis shoes, which were of the same size as worn by Flowers. A Fila Grant Hill II Mid shoe box, size 10 ½, was found at the home of Flowers’s girlfriend. The shoe box was submitted to Andrews and the Mississippi Crime Lab for the examination for latent prints. As Fila manufacturer told Andrews that “all of the Grant Hill II designs have the exact same outsole pattern,” an investigator purchased a similar pair of shoes so that they could be compared to the impressions from the scene. Andrews concluded that the type of shoes purchased by the investigator could not be excluded as having made the impressions at the crime scene, and that the type of shoes which would have originally been in the shoe box found at Flower’s girlfriend’s house would have also made the same type of impressions.

Shoeprint Evidence in a Larger Context


Shoeprints are a type of two-dimensional impression evidence (a tire track in mud, for example, is a three-dimensional impression evidence).

A 1995 study sending questionnaires to forensic laboratories across the world showed that shoeprint was the most common impression evidence examined by those laboratories.24


The goal of impression evidence analysis is to identify a specific source of the impression, and the analytical process that this follows generally is an accepted sequence: (i) identifying the class (or group) characteristics of the evidence, and (ii) followed by locating and comparing individual identifying (also termed accidental or random) characteristics.25

With respect to shoeprints, class characteristics are “an intentional or unavoidable characteristic that repeats during the manufacturing process and is shared by one or more other shoes” such as design, pattern, size, and shape; and, individual identifying characteristics are caused by wear and tear, such as cuts, scratches, gouges, holes, or random inclusions that result from manufacturing.26


     Lack of consistent standard and guidelines

     There is no defined threshold of matching characteristics that must be surpassed, nor are there any studies that associate the number of matching characteristics with the probability that the impressions were made by a common source. It is generally accepted that the specific number of characteristics needed to assign a definite positive identification depends on the quality and quantity of these accidental characteristics and the criteria established by individual laboratories.

     An 1996 European survey using six identical mock cases showed that there were considerable differences in the conclusions reached by different laboratories.27


     The Scientific Working Group for Shoeprint and Tire Tread Evidence (SWGTREAD) suggests the use of terminology such as “identification,” “probably made,” “could have made,” “inconclusive,” “probably did not made,” “elimination,” and “unsuitable,” instead of binary opposition terms such as “match/no match,” “responsible for/not responsible for,” and “caused with/not caused with.”

Weakness of the Shoeprint Evidence in the Flowers Case

The shoeprint evidence in Flower’s case is rather weak. Because the investigators only found the shoe box instead of the shoes, Joe Andrews and his laboratory could only compare the class characteristics but not the individual identifying characteristics of the shoeprint, which could only lead to a very broad conclusion. It is also worth noting that the lack of common standard between forensic labs makes the reports questionable.


What Happened in the Flowers Case

Joe Andrews, a forensic scientist specializing in trace evidence, testified at trial that Flowers’s gunshot residue test revealed one particle of gunshot residue on the back of Flowers’s right hand. Flowers contended that the trial court erred in admitting this evidence, because the prejudicial effect of the evidence greatly outweighed the probative value and thus rendered this evidence inadmissible according to Mississippi Rule of Evidence 403.

Supreme Court on certiorari review found this issue to be without merit. Here, Andrews through his testimony set out three scenarios that could result in a person having gunshot residue on his hands: (i) the person had discharged a weapon; (ii) the person had been in close proximity to a discharged weapon; or (iii) the person had handled an object that has gunshot residue on it. As a result, Andrews clearly explained that the gunshot residue did not unequivocally show that Flowers had fired a gun. Because Andrews’s testimony was clearly explained, there is little risk that the jury was confused or misled by the testimony.

Gunshot Residue Evidence in a Larger Context


Gunshot residue is principally composed of burnt and unburnt particles from the explosive primer, and would be deposited when someone discharges a firearm.

The most definitive method to determine whether a particle is characteristic of or consistent with gunshot residue is by its elemental profile. An approach to the identification of particles characteristic of or consistent with gunshot residue is to compare through SEM-EDX analysis the elemental profile of the recovered particulate with that collected from case-specific known source items, such as the recovered weapons, cartridge cases or victim-related items whenever necessary.

A positive result for gunshot residue can mean many things. Mainly it indicates that the person sampled was in the vicinity of a gun when it was fired, handled a gun after it was fired, or touched something that was around the gun when it was fired.



Gunshot residue lasts nearly forever. It can land on anything and stay there indefinitely. Because many people in this country shoot guns and the things in their lives are covered with gunshot residue, an individual will be covered with gunshot residue when he or she comes in contact with these people or anything they touched.28 On the other hand, it can be easily tampered with and altered. Having a consistency similar to very fine talcum powder, its distribution is easily ruined by simple actions like dusting down clothes or washing hands.29

     Lack of compositional uniqueness

Since the test only identifies the principal elements of the residue, it can be fooled by a careful composition of the same elements obtained from different sources such as fireworks or even auto brake dust.30

To be specific, fundamentally, gunshot residue examiners are seeking microscopic particles containing lead (Pb), barium (Ba), and antimony (Sb) which, for the purposes of this article, are found uniquely in combination only in the primers of firearms cartridges. It was generally accepted that the three element particle was truly unique to gunshot residue. However, in 2002, a paper was published by Torre and others reporting the finding of rounded particles of PbBaSb in automotive brake lining and wear products. This has caused most gunshot residue examiners to downgrade the three parts particle from “unique” to “characteristic”.31

     Lack of consistent standard and guidelines

The standards for expressing an opinion in one crime laboratory may be very quantitatively different than the standards in another laboratory.32

One conducting the test must have unique particles to opine that the source of the elements is from gunshot residue and that characteristic particles support the uniqueness. The controversy now begins – how many unique and characteristic particles are needed to opine that gunshot residue is identified? The answer to this question varies widely from lab to lab across the United States and in the United Kingdom.33

     Jury confusion

Jurors, and sometimes judges, can be confused or misled when a crime laboratory reports an unqualified finding of the presence of gunshot residues. They cannot appreciate the significance of contamination if it is not mentioned in the results and without this information the case may go ahead without the gunshot residue findings being challenged.34


The fundamental problem underlying firearms and shoeprints analysis is the lack of a consistent standard and inherent subjectivity. In criminal prosecutions, forensic evidence is often offered to support conclusions about “individualization” (sometimes referred to as “matching” a specimen to a particular individual or other source) or about classification of the source of the specimen into one of several categories. However, no forensic method has been rigorously shown to have the capacity to consistently, and with a high degree of certainty, demonstrate a connection between evidence and a specific individual or source.

As discussed above, ballistics evidence is far behind what the National Academy of Science defines – a “precisely specified, and scientifically justified, series of steps that lead to results with well-characterized confidence limits,” and the same applies to shoeprint and gunshot residue evidence. As long as this is the State of the science, experts in delivering their testimonies should refrain from making any claim of absolute certainty regarding their conclusions, in order to avoid objectively inducing errors to the fact finders. And, courts in deciding whether to admit certain expert witnesses’ testimonies should follow the Daubert standard closely as the “gatekeeper” to exclude expert testimonies that are without scientific knowledge.


1 Rehman Tungekar, ‘Could they really match those bullets in the Tardy Furniture case?’ (APM Reports, May 8, 2018) accessed March 9, 2019.

2 Gayle Swenson, ‘How Good a Match is It? Putting Statistics into Forensic Firearms Identification’ (National Institute of Standards and Technology, February 8, 2018).

3 National Research Council (U.S.) et al. eds., Strengthening Forensic Science in the United States: A Path Forward (The National Academies Press 2009), 150.

4 Id, at 151.

5 Swenson (n 2).

6 National Research Council (n 3), 154.

7 Id at 152.

8 Tungekar (n 1).

9 Id.

10 National Research Council (U.S.) et al. eds., Ballistic Imagine (The National Academies Press 2008), 3.

11 Id.

12 National Research Council (n 2), 154.

13 “Theory of Identification, Range of Striae Comparison Reports and Modified Glossary Definitions – An AFTE Criteria for Identification Committee Report” (1992) 24 Journal of the Association of Firearm and Tool Mark Examiners, 336-340.

14 ‘Theory of Identification, Range of Striae Comparison Reports and Modified Glossary Definitions – An AFTE Criteria for Identification Committee Report’ (1992) 24 Journal of the Association of Firearm and Tool Mark Examiners, 336-340.

15 Sufficient agreement exists when “when it exceeds the best agreement demonstrated between tool marks known to have been produced by different tools and is consistent with the agreement demonstrated by tool marks known to have been produced by the same tool.” Id. at 336.

16 National Research Council (n 2), 155.

17 Id. at 153.

18 Id. at 154.

19 Tungekar (n 1).

20 National Research Council (n 3).

21 John Song et al., ‘Estimating error rates for firearm evidence identifications in forensic science’ (2018) 284 Forensic Science International, 15.

22 Swenson (n 2).

23 Id.

24 Liukkonen, Markku et al., “The role and duties of the shoeprint/toolmark examiner in forensic laboratories” (1996) 82.1 Forensic science international, 99-108.

25 Id.

26 National Research Council (n 2).

27 Majamaa, Heikki and Anja Ytti, “Survey of the conclusions drawn of similar footwear cases in various crime laboratories” (1996) 82.1 Forensic science international, 109-120.

28 Don Penven, ‘Gun Shot Residue Testing: Is it Still a Viable Form of Physical Evidence?’ (Crime Scene Investigator Network) accessed March 8, 2019.

29 Okorie Okorocha, ‘The Reliability Of Gunshot Residue’ (Los Angeles Forensic Toxicology Expert Witness, December 8, 2017) accessed March 8, 2019.

30 Id.

31 Dennis L. et al., ‘The Controversy Concerning Gunshot Residues Examinations’ (Forensic on the Scene and In the Lab Magazine, August 1, 2008) accessed March 8, 2019.

32 Id.

33 Id.

34 Id.