Forensic Failures: Philip Scott Cannon & Bullet Analysis Blunders

Forensic Failures: Philip Scott Cannon & Bullet Analysis Blunders

In December 2009, Philip Scott Cannon of Polk County, Oregon was released from prison after his conviction was found to be based on ‘junk science’. By this time, he had served over 10 years.

The story of his wrongful imprisonment began on 23rd November 1998, when Bimla Boyd noticed that the mobile home of a neighbour appeared to be on fire. Upon investigating, she discovered the bodies of three people; Jason Kinser, his girlfriend Suzan Osborn, and Celesta Graves. All three victims had been shot. Boyd promptly called the police, and a murder investigation ensued.

Boyd stated that earlier in the day she had noticed a maroon van parked nearby, a van which happened to belong to Philip Scott Cannon. Around the same time, local men Jeremy Olsen and Larry Weaver were on their way to deliver water to the victims’ trailer. Upon arrival at the trailer, Olsen and Weaver claimed that they were met by Cannon, who was said to be “acting strangely”. Cannon informed the men that they should not go into the trailer because Jason Kinser was upset and in the midst of a heated argument with an unknown Hispanic man. Olsen and Weaver subsequently left without entering the trailer. Based on this eyewitness testimony and the fact that Cannon owned the maroon van spotted nearby, he soon became the prime suspect.

Credit: Polk County Itemizer-Observer

When questioned by police, Cannon maintained his innocence and claimed Kinser had called him over to give an estimate for fixing a plumbing problem in the trailer, after which he promptly left when Kinser began arguing with another man. However the police had a very different impression of the situation, believing that Cannon was a meth user and Kinser his dealer. Further suspicion fell on Cannon when a prisoner, Steven Brobston, informed police that he had entrusted Cannon with a box containing $16,000 to be used to support Celesta Graves, Brobston’s girlfriend and one of the victims. With the circumstantial evidence mounting, investigators searched Cannon’s home, finding the lockbox but no sign of the money. They did however stumble upon a number of weapons and ammunition. This was sufficient to arrest Cannon, who was taken into custody on 3rd December and charged with three counts of aggravated murder and the illegal possession of a firearm.

During the trail, Olsen, Weaver and Boyd were all called upon to recount their experiences of seeing Cannon near the trailer on the day of the murder. Of course this evidence was purely circumstantial, so an expert witness was called upon to study the bullets collected from the crime scene and those recovered from the suspect’s home. Michael Conrady of Oregon State University’s radiation center conducted a metallurgic analysis of the bullets known as comparative bullet lead analysis. This technique involves the application of various analytical techniques, but primarily atomic emission spectroscopy, to bullet composition determination. The method aims to establish the composition of metals in the bullet, such as copper, tin, antimony and silver, and compare profiles to ascertain whether two bullets are chemically identical. Based on this analysis, Conrady testified that the bullets from the crime scene and those from Cannon’s home were chemically identical, therefore Cannon’s ammunition was used to kill the three victims. However the weapons found in Cannon’s home were not connected to the murders, nor did police establish a reasonable motive for the triple homicide. Despite these shortcomings, on 28th February 2000, Cannon was found guilty and sentenced to three life sentences with no parole.

At the time of Cannon’s trial, the use of comparative bullet lead analysis was already under scrutiny, with some believing the reliability of the technique was unfounded. In 2005, the national Academy of Science discredited the technique and deemed it ‘junk science’, and soon after the FBI abandoned the use of this method altogether. As Cannon’s conviction was so heavily reliant on the bullet analysis, in 2009 a Polk County Circuit judge vacated Cannon’s original conviction. Incidentally it was now apparent that police involved in the original trial had hired Conrady to conduct the bullet analysis because the Oregan State crime lab had refused on the basis of the technique being scientifically unreliable. In order for a re-trial to take place, the original bullets were demanded in order to conduct further analysis. Polk County prosecutors insisted that the original trial evidence had been sent to the Department of Justice when Cannon had appealed his conviction, however Assistant Attorney General Susan Gerber, who had been assigned the case, claimed she had never received this evidence. It later came to light, when Gerber was suspended from her position on assault charges, that she had had the evidence all along, locked away in a filing cabinet.

In light of all of this, Cannon’s conviction was dismissed and he was released from prison. By this point he had spent over a decade behind bars. No other arrests have been made in relation to the murder of Kinser, Osborn and Graves.



Michigan State University National Registry of Exonerations. Philip Scott Cannon. [Available online]

Photo Credit: Polk County Itemizer Observer. Cannon retrial up to Polk DA. [Available online]


Forensic Failures: Ray Krone & Bite Mark Blunders

Forensic Failures: Ray Krone & Bite Mark Blunders

In 2002, Ray Krone became the 100th wrongfully imprisoned person to be exonerated by DNA analysis in the US, but only after spending ten years of his life detained in Arizona prisons, including a number of years on death row.

On the morning of 29th December 1991, the owner of the CBS Lounge in Phoenix, Arizona went to his bar to discover the door unlocked, the lights on, and the naked body of 36-year-old Kim Ancona on the floor of the men’s bathroom. The victim, who had worked in the bar as a waitress, had been brutally stabbed to death.

The subsequent examination of the scene was somewhat fruitless and, although saliva was recovered from Kim’s body, little other physical evidence could be found. The only piece of evidence investigators had to work with was a series of bite marks found on the victim’s breast and neck. As the investigation continued, police began interviewing those close to the victim in attempts to shed light on the events leading up to her death. It transpired that the victim had told a friend that Ray Krone, a regular customer at the bar, was to help her close up the bar the previous night. Investigators jumped at the possibility of a potential suspect.

During an interview with Krone, a detective noticed that he had a very distinctive deformity of his front teeth, no doubt causing a unique bite mark. This characteristic would later lead to the nickname of ‘the Snaggletooth Killer’. Krone was happy to oblige when asked to provide a Styrofoam impression of his teeth for comparison purposes.

However unfortunately for him, a forensic odontologist soon declared that the Styrofoam impression matched the bite marks found on the victim’s body.

Krone maintained his innocence, insisting that he was at home in bed at the time of the murder, a story corroborated by his roommate. Despite this, Ray Krone was arrested and charged with kidnapping, sexual assault and murder.

As the trial commenced, it was clear there was little evidence for the prosecution to present, so they focused their efforts on the bite mark comparison. They hired forensic odontologist Raymond Rawson to conduct the comparison between the bite marks on the victim’s body and the impression of Krone’s teeth. Using compelling video footage attempting to show the physical match between the two, Rawson informed the jury that the match was “100 per cent”, and that only Krone could have made those bite marks. The defence chose not to call upon their own court-appointed forensic odontologist.

Despite a lack of DNA analysis or eyewitness testimony linking Krone to the crime, and the bite mark comparison being the only physical evidence implicating him, Ray Krone was found guilty and sentenced to death.

Three years later Krone was awarded a re-trial due to the prosecution team concealing the persuasive video tape concerning the bite mark evidence until a day before the original trial, but once again he was found guilty. Despite the opportunity to rectify the conviction being wasted, trial judge James McDougall aired his uncertainty: “the court is left with a residual or lingering doubt about the clear identity of the killer. This is one of those cases that will haunt me for the rest of my life, wondering whether I have done the right thing”.

Forensic Odontology & Bite Mark Comparison

As DNA analysis was not carried out during this investigation, Krone’s conviction was almost entirely based on the ‘expert’ opinion that his teeth matched the bite marks on the victim’s body.

Dental identification is based on the theory that every individual’s dentition is unique, and thus bite marks made by a person will be distinguishable. In theory, this is true – we all have different combinations of jaw sizes, varying dental work and unique wear patterns to our teeth. Bite mark comparison may involve a variety of methods, including overlaying appropriately-sized photographs of teeth and bite marks and fitting together physical moulds.

At the time, there was little reason to doubt the testimony of the forensic odontologist hired by the prosecution. Raymond Rawson was a well-established expert who was certified by the American Board of Forensic Odontology, his findings in this case were supported by another expert, and the discipline of bite mark comparison had been practiced for almost 20 years. Furthermore, a 1984 study had provided “statistical evidence for the individuality of human dentition”. The expert witness testimony seemed perfectly reputable.

Well, that is until we look a little closer.

The seemingly convincing 1984 study was actually research conducted by Rawson himself, and has since been widely criticised as being a flawed study, largely because he used hand-traced dental impressions for his comparisons, a non-randomised subject selection process, and statistical tests not relevant to his type of data. Other experts had quite rightfully stated that the results of the study should absolutely not be used in a legal case.

A study conducted ten years previously comparing bite marks in wax and pig skin to the teeth of subjects stated that, although bite marks in wax were easily assessed, those made in pig skin were difficult to examine and the results unreliable. The research concluded that incorrect identification of bite marks on pig skin were made 24% of the time under laboratory conditions, and even as high as 91% of the time when based on photographs taken 24 hours after the bite marks were initially made. The study highlights the clear difficulties in subjective fields of work such as forensic odontology. Experts will often be required to examine bite marks that are hours or even days old, obscured by bruising and abrasions and typically not entirely representative of the biter’s teeth. At times it is challenging enough to merely identify an injury as a bite mark, let alone successfully compare it to a set of teeth.

Despite these apparent shortcomings, Ray Krone was to spend a decade of his life behind bars.


Fortunately for Krone, he had an undeterred family behind him maintaining his innocence and the means of hiring proficient legal help and in 2002, with the help of attorney Alan Simpson, he successfully appealed.

DNA analysis had become, by this point in time, a well-established technique frequently utilised in criminal investigations. Analysis of bodily fluids recovered from the crime scene a decade earlier soon proved not only Krone’s innocence, but also the identity of Kim Ancona’s real killer. Kenneth Phillips, a man with a long history of repeated violent sex offenses, was serving time in prison for the sexual assault of a 7-year-old girl, but at the time of Kim’s murder was living a mere 600 yards from the scene of the crime. Despite his close proximity to the bar, his deviant history and the fact that he was at the time of the murder on probation for the assault of a neighbouring woman, Phillips was never considered a suspect.

On 8th April 2002, Ray Krone left prison a free man, the 100th person to be exonerated by DNA evidence. He would certainly not be the last.

Krone now lives in Tennessee, where he has since dedicated his time to criminal justice reform and the campaign for the abolition of the death penalty.

“I would not trust the state to execute a person for committing a crime against another person. I know how the system works” – Ray Krone.



Innocence Project. Ray Krone. [online] Available:

New Scientist. Bite-mark evidence can leave false impression. [online] Available:

Rawson, R. D. et al. Statistical evidence for the individuality of the human dentision. J For Sci. 29(1984), pp245-253.

State v. Krone, 897 P.2d 621, 182 Ariz. 319 (Ariz. 06/22/1995)

Whittaker, D. Dome laboratory studies on the accuracy of bitemark identification. Int Dent J. 25(1975) pp. 166-171.

Cover Image –


Forensic Failures: Three Men, Three Hairs, Three Wrongful Convictions

Forensic Failures: Three Men, Three Hairs, Three Wrongful Convictions

In the summer of 1978, 63-year-old taxi driver John McCormick was robbed and shot on his own doorstep in his Washington, D.C home. His wife, roused from sleep by the cries of her husband pleading for his life, ran to his aid to briefly glimpse the gunman, a man with a stocking mask shielding his face. McCormick died as a result of the attack, shot by a .32 calibre handgun. A police officer and his dog later found the stocking mask nearby, and from this a number of hairs were recovered.


Santae Tribble (Source: The Innocence Project)

Suspicion soon fell on 17-year-old Santae Tribble, implicated by an informant, Bobby Jean Phillips, who claimed Tribble had recently sold a .32 calibre handgun to her roommate. A weapon that, incidentally, was never successfully linked to the gun used to kill McCormick. Tribble was pulled in for questioning. Hairs recovered from the stocking used by the killer were compared to samples collected from the suspect and, according to an FBI analyst, “matched in all microscopic characteristics”. At trial, the prosecution went on to declare that there was perhaps a “one chance in ten million” that the hair belonged to someone else. Seemingly compelling “statistics” and apparently sufficient in helping the jury reach a verdict. Despite having testified that he was in Maryland at the time of the attack, an alibi supported by half a dozen witnesses, in January 1980 Santae Tribble was found guilty of murder and sentenced to 20 years to life in prison.

Meanwhile, as Tribble was adjusting to prison life, another man’s world was about to be turned upside down thanks to a single hair. In February 1981, a 27-year-old woman was bound, raped and robbed in her apartment, briefly glimpsing her attacker in the dim light. The distraught victim helped police construct a composite sketch of her attacker; a young, cleanly shaven African American man. Unfortunately for 18-year-old Kirk Odom in the following weeks, a passing police officer decided that Odom resembled the composite sketch, passing on his suspicions to the detective leading the case. As a result of this Odom was presented to the victim as part of a somewhat dubiously organised line-up, with Odom standing on a box in order to match the height of the other men in the line-up, who were actually police officers. The victim identified Odom as her attacker.


Kirk Odom (Source: The Innocence Project)

During the trial, FBI Agent Myron Scholberg testified that the “Negroid hair” found on the victim’s nightgown was microscopically similar to a hair sample taken from Odom, “meaning the samples were indistinguishable”. He declared this to be “a very rare phenomenon”, and in the thousands of occurrences in which he had compared hair samples, on only eight or ten occasions had he failed to distinguish the hairs of two different people. Understandably, this implied a very substantial likelihood that the hair did in fact belong to Odom, and after only a few hours deliberating he was found guilty and sentenced to 20 to 66 years in prison.

In June that same year, a worryingly similar scenario ensued after 21-year-old student Catherine Schilling was found raped and murdered in Rock Creek Park, Washington. She was discovered naked and shot five times in the head. A somewhat questionable police informant, Gerald Mack Smith, relayed to police that he had been drinking in the park with a man who had admitted to killing the woman after attempting to rob her. That man was allegedly Donald Eugene Gates. It is worth noting that, not only was Smith paid by police for his information, but at the time he had two prior felony convictions and had recently been indicted for a third felony, which was conveniently dismissed after he aided the police in incriminating Gates. This information was not relayed to the defence team.

Donald Gates - Washington Post

Donald Gates (Source: The Washington Post)

So Donald Gates, aged 30 at the time, was arrested and charged with murder, not only based on the testimony of Gerald Smith, but also largely on the comparison of hair samples. FBI forensic analyst Michael Malone testified in this case, establishing that hairs recovered from the scene were “microscopically indistinguishable” from a hair sample taken from Gates. The expert witness claimed that in his many years of hair comparison work, there were perhaps two in ten thousand cases where hairs from two people could not be distinguished. Once again, seemingly convincing statistics to a lay jury. In September 1982, Gates was found guilty and sentenced to 20 years to life in prison.

Forensic Hair analysis

What do these three cases have in common? All men were wrongfully convicted for crimes they did not commit. And all convictions were largely based on flawed expert witness testimony relating to hair analysis.

The potential evidential value of human hair stems from its prevalence at crime scenes and the ease with which it can be imperceptibly transferred. It is fairly resilient and can persist at a crime scene for many years after the incident took place, all-in-all making hair a potentially vital piece of evidence well worthy of examination.

What does this examination involve? Forensic hair analysis is a type of ‘comparison analysis’ – that is an examination that is typically based on individual or class characteristics rather than incorporating numerical analyses. It generally involves the study of microscopic characteristics relating to whether the hair is human or animal, the colour, length, diameter, likely racial group, area of bodily origin, phase of growth, and so on.

Hair 2 - Wikimedia

Microscope image of hair sample

The examination may include an even closer scrutiny of patterns and intricacies along the shaft of the hair. The human hair can be described as being composed of three distinct regions: the medulla, the cortex and the cuticle. The medulla refers to the inner core of the shaft, the cortex as the surrounding material, and the cuticle the outermost protective layer. The cuticle is composed of a scale-like complexion, which can be particularly beneficial in distinguishing between animal hairs and human hairs. Cuticle characteristics will be noted in the comparison of hair samples, particularly the thickness of the cuticle, the colour, and the general shape of the “scales”. The cortex is the portion of the hair containing pigment granules which give hair its colour, obviously vital in the comparison of different samples. These pigment granules vary between racial groups, thus are pivotal in establishing whether a hair is likely Caucasian or Negroid in origin, for instance. The medulla in the centre of the hair shaft will also be scrutinised to determine whether the core is continuous, interrupted, fragmentary or absent.

This is a limited list of the features focussed on during forensic hair analysis. In reality there is a myriad of other features, both natural such as those above and artificial such as dyeing and damage, that can be used in studying the similarities and differences between hair samples. So it would seem there is an abundance of information with which to reach a conclusion. However the problem lies in that these are all class characteristics whose description is based on the individual analyst’s subjective observations.

Obviously the ideal situation would involve some form of statistical analysis to determine the reliability of comparisons, but the vast number of variable factors in hair comparison makes statistical analysis tremendously complex if not impossible. This in itself does not render such analyses worthless in criminal proceedings, provided accurate testimony is given and caveats made clear. In the aforementioned cases, and potentially others, expert witness testimony strongly implied that exact matches between hair samples had been made, and in some instances fictitious statistics were even expressed to the jury. When such evidence is presented by a seemingly distinguished scientist who may be an expert in their field, naturally members of a lay jury with little or no knowledge of the topic are unlikely to challenge this evidence. The FBI’s own guidelines on forensic hair analysis conclude that “the science of microscopic hair examination can never result in an identification” but can “provide a strong basis for an association”, but unfortunately this ethos was not always carried into the courtroom.

Collectively, Tribble, Odom and Gates served nearly 80 years in prison. All three men were eventually released and exonerated by DNA testing and the actual perpetrators identified. However by the time this was achieved, perpetrators had passed away or were able to evade conviction due to expiration of the statute of limitations. The wrongfully convicted men were awarded millions of dollars in compensation, but hardly a prize worthy of losing so many years.

As a result of these incidents, and numerous others, ongoing investigations are aiming to identify further miscarriages of justice in the United States, initially largely powered by the Innocence Project, a non-profit organisation which aims to exonerate wrongly convicted people through DNA testing. Furthermore, efforts are underway to focus on how forensic science standards can be strengthened to avoid future errors, particularly in relation to forensic ‘comparison analyses’.



Federal Bureau of Investigation. Forensic hair comparison: background information for interpretation. [online] Available:

Innocence Project. Donald Eugene Gates. [online] Available:

National Registry of Exonerations. Kirk Odom. [online] Available:

Taupin, J. M. Forensic hair morphology comparison – a dying art or junk science? Sci & Justice. 44 (2004), pp. 95-100.

The Washington Post. Santae Tribble cleared in 1978 murder based on DNA hair test. [online] Available:

Forensic Failures: Eyewitness Testimony & The Ronald Cotton Trial

Forensic Failures: Eyewitness Testimony & The Ronald Cotton Trial

Eyewitness testimony is far from being a form of scientific evidence and is in fact based on perhaps one of the most flawed tools available: human memory. Despite this, the testimony of eyewitnesses can still have an enormous bearing on the outcome of a criminal investigation but unfortunately, as we will see, this evidence can range from being mildly misleading to catastrophically incorrect. And when it is the matter of someone’s life and freedom on the line, the criminal justice system cannot afford to get it wrong.

So let’s delve into a case that shows some of the great flaws of eyewitness accounts and how they can lead to devastating consequences if not treated with care.

The Case

One night in North Carolina in 1984, 22-year-old student Jennifer Thompson lay unsuspectingly in her bed when an assailant broke in, held a knife to her throat and raped her. During the ordeal, Thompson decided that she would try to remember as much about her attacker as she could, hoping to run straight to the police and get this man identified. So she took in everything she could about his face, skin colour, hair, height, weight and voice. She was convinced that she had etched this person’s image onto her brain.

Later at the police station, an investigating officer sat down with Thompson and the two of them produced a composite sketch of the rapist. This was enough for police to round up potential suspects. Thompson was later presented with a police line-up of six men (first by photograph and later a physical line-up), from which she picked out with “100 percent certainty” her attacker – a man named Ronald Cotton.

“Looking at a series of police photos, I identified my attacker. I knew this was the man. I was completely confident. I was sure”.

The suspect fit her description perfectly and he was already known to the police following a few other legal incidents, including attempted rape when he was 16. Cotton was tried, a trial at which Thompson looked him dead in the eye and declared that this was in fact the man who had raped her. Based on the victim’s testimony, and other pieces of circumstantial evidence, he was found guilty and sentenced to life imprisonment.

Meanwhile in the prison where Cotton was being held, another inmate eventually joined the ranks, a man named Bobby Poole. Poole had been convicted of a series of violent rapes and was serving consecutive life sentences. Incidentally this man bore a striking resemblance to Cotton, so much so that prison staff were constantly getting them confused. Any alarm bells ringing? Cotton subsequently confirmed his suspicions and discovered that Poole was in fact the actual rapist of Thompson. This was enough to win Cotton a retrial, but when presented with the two men, Thompson once again stated that Cotton was the attacker and she had never before seen Poole, her actual attacker. Fortunately Cotton’s break came in 1995, when he heard about a new technique known as DNA analysis. Using a semen sample collected from Thompson at the time of the attack, DNA evidence was able to prove that Cotton was innocent and Poole was the actual rapist. Case closed.

The Problems with Eyewitness Identification

The consequences of mistaken identification can be disastrous, and the people affected by this error is larger than one might initially think. In this case, an innocent man of course spent over a decade of his life behind bars for a crime he did not commit, the implications of which do not even need to be stated. The victim must then live with the knowledge that, firstly, she is in a way responsible for the incarceration of this innocent man and, secondly, her actual attacker has been on the loose! Not to mention the effects on the family of both the victim and the wrongly accused. The range of those affected can then extend to anyone who has since been the victim of the actual criminal. In this case Poole went on to rape a series of women, crimes which perhaps may not have occurred had Cotton not been wrongfully charged, forcing police to continue their investigation.

So just how can a case of mistaken identity occur?

In this instance, the assailant was mere inches away from the victim’s face for a considerable length of time (a length of time that no doubt felt like a lifetime), and yet she still made an incorrect identification. How could this happen? There are a number of factors which are known to reduce the accuracy of eyewitness identifications, many of which can into play in Thompson’s case. Extreme stress when viewing the perpetrator and the presence of weapons at the scene can affect a person’s ability to remember and recall, as can racial differences between the perpetrator and the witness. A suspect’s lack of distinctive characteristics will render them more difficult to remember, and of course the person may be wearing some kind of disguise to hide their features. The viewing conditions can drastically affect eyewitness identification too, such as the distance between the witness and the suspect, the lighting in the environment, and the length of time the witness sees the suspect (Ellis et al, 1977). In the United Kingdom following a case in 1977 (R v Turnbull), a set of guidelines relating to cases involving eyewitness testimony were introduced. These guidelines were intended to assist members of the jury in treating eyewitness evidence with the care and attention required, taking into account the circumstances surrounding the identification by the witness.

Then there are the biases that can occur as a result of the management of the investigation. The police line-up itself can introduce prejudice from the start. Had it been somehow suggested to Thompson that the perpetrator was definitely somewhere in the line-up, she may have been far more likely to choose someone even if, as in this case, he was not actually there. The process of merely being asked to choose from a selection of faces can force people to choose incorrectly. Admittedly, Cotton and Poole did look very similar at first glance, just as many people in the world share looks to a certain extent. So in Cotton’s retrial, why would Thompson still point the finger at Cotton when her real attacker was stood right there? Because reinforcement can alter memory. After years of her positive identification of Cotton being reinforced, her memory had replaced her real rapist’s face with Cotton’s. Human memory is a funny old thing, and certainly can’t necessarily be trusted.

The eyewitness statement of a witness or victim is certainly not evidence based on science, and yet this evidence can often carry such weight that the pointing finger of a witness may as well be an airtight scientific report. The Ronald Cotton case was not a one-off incident in which eyewitness testimony failed. The list of cases of people being wrongfully convicted based on eyewitness testimony, only to be later exonerated, is far longer than it should be (in fact the Innocence Project states that eyewitness misidentification has played a role in over 70% of convictions which were later overturned by DNA evidence). Looking at all the flaws of eyewitness testimony and all the cases of wrongful conviction based on it, it makes you wonder whether eyewitness testimony has any place in the courtroom at all.


Eliis, H. D. et al. Experimental studies of face identification. Journal of Criminal Defense. 1977(3) pp. 219-234.

Hansen, M. Scoping out eyewitness ids. Am Bar Assoc J. 2001(87).

New York Times. ‘I was certain, but I was wrong’. [online] Available:

Scientific American. Why science tells us not to rely on eyewitness accounts. [online] Available:

Forensic Failures: The Shirley McKie Fingerprint Scandal

Forensic Failures: The Shirley McKie Fingerprint Scandal

In light of the recent FBI hair analysis outrage, it seemed appropriate to revisit an old classic in the history of failing forensic science. The Shirley McKie fingerprint scandal. Back in the 1990s, Shirley McKie was a police constable whose life, along with an important murder investigation, was essentially ruined due to mistakes made by forensic experts.

At the beginning of 1997, 51-year-old Marion Ross of Kilmarnock, Scotland was found murdered in her home, with suspicion quickly falling on David Asbury, a handyman who had previously carried out some work on the house. A number of fingerprints were recovered throughout the investigation, including one belonging to Asbury on a gift tag in the victim’s home. Further incriminating Asbury was a tin containing nearly £2,000 found in his house, which incidentally had the victim’s fingerprints on.

But by far the most controversial piece of evidence in this case was another fingerprint recovered from the crime scene which did not belong to neither the victim nor the suspect. A thumbprint was recovered from a doorframe at the murder scene and, according to the experts of the Scottish Criminal Record Office (SCRO), that thumbprint belonged to police constable Shirley McKie.

One might think the fingerprint of a police officer at a crime scene is nothing of great note, though perhaps some slightly sloppy police work, however that was not the case. Because in this instance, Shirley McKie adamantly denied that she had ever set foot in the victim’s house. So how did her fingerprint materialise at this crime scene? Well, it didn’t. McKie was telling the truth.

Unfortunately for McKie, her claims fell on deaf ears and she was subsequently suspended, fired and then arrested by Strathclyde Police in 1998 and charged with perjury (lying under oath), even though not one of the dozens of police staff involved could recall seeing her at the crime scene. A gruelling trial ensued, dragging both McKie’s reputation and life through the mud, along with the reliability of fingerprint evidence. Four fingerprint experts from SCRO concurred that the fingerprint belonged to McKie, the same experts who had identified the fingerprint found in Asbury’s home as belonging to the victim.

The fingerprint evidence in this case became something of a double-edged sword. If the latent print comparison conducted by the SCRO was accurate, Asbury could be reasonably named as the perpetrator but Shirley McKie would surely be lying about visiting the crime scene. Conversely, if McKie was truthful in her statement, then the fingerprint evidence was flawed and the evidence against the suspect useless. With the fingerprint identification being the only significant evidence incriminating Asbury, it naturally became a vital aspect of the case. At this point in time fingerprint evidence was perhaps viewed as an infallible gold standard in forensic science, and the jury agreed that the latent print evidence presented by the SCRO was accurate, thus Asbury was convicted and McKie assumed just as guilty.

Fingerprint comparison can be a subjective technique (

Fingerprint comparison can be a subjective technique (

Thankfully the investigation did not end there. Two fingerprint experts from the U.S. were called upon to offer their expertise, and both declared that the mysterious fingerprint found in the victim’s home did not belong to McKie. The SCRO experts had misidentified the fingerprints, although they stubbornly refused to admit to this. Furthermore, a member of the Scottish Parliament somewhat unusually invited fingerprint experts from around the world to examine the prints. 171 experts from numerous countries all reached the same conclusion – that the two latent prints did not match.

The fingerprint evidence was ultimately rejected and McKie was unanimously cleared of all charges. Perhaps too little too late for a woman who had lost her reputation and career. With the fingerprint evidence rejected and McKie’s name cleared, Asbury’s conviction was also overturned, with there being nothing more than mere circumstantial evidence linking him to the crime. So through this misinterpretation of fingerprint evidence, not only was Shirley McKie’s career ruined and the freedom of a potentially innocent man put on the line, but a murder investigation was left unsolved with little likelihood of ever finding the real killer of Marion Ross.

Over the years following this trial, a number of inquiries were conducted examining why this incident was allowed to occur in the first place. Through the public inquiry it was ultimately concluded that McKie had simply been the victim of human error and nothing more, though many argued at the time that there had been something of a conspiracy and cover-up. The inquiry called for competency training of analysts and for independent reviews to be carried out of any fingerprint evidence that is disputed, along with a prompt change in the way in which fingerprint comparisons were made in the first place.

Most importantly, it was recommended that fingerprint evidence should be viewed as opinion evidence only as oppose to the product of a scientific technique that can produce absolute answers, and that experts should not make claims with 100% certainty. This problem has been once again highlighted in the recent FBI scandal, in which hair analysis experts overstated the evidence, implying the analysis was far more reliable than it actually was. Members of the jury are unlikely to have any significant knowledge of forensic techniques utilised by experts, thus are hardly in a position to determine the reliability of the methods used. It is up to the expert to highlight just how dependable the evidence really is.

Although the Shirley McKie case offered a slight silver lining in highlighting the fallibility of forensic evidence, this is evidently a lesson that is yet to be taken onboard.


The Guardian. Fingerprint evidence ‘based on opinion rather than fact’. [online] Available:

McKie, I. A. J. ‘There’s name ever fear’d that the truth should be heard but they whom the truth would indite’ (Presentation given by Iain McKie to the Forensic Science Conference 2003. Sci Justice. 43 (2003), pp. 161-165.

University of Dundee. ‘Road map for reform’ of fingerprint practices to be developed at the Scottish Universities Insight Institute. [online] Available:

Forensic Failures: DNA Evidence and the Phantom of Heilbronn

Forensic Failures: DNA Evidence and the Phantom of Heilbronn

In 1993, the existence of a brutal murderer came to light when an elderly woman was strangled with a length of wire and left for dead in Germany, the perpetrator leaving behind nothing more than DNA on a teacup. The criminal did not stop there, and over the next decade and a half a sequence of crimes ensued, ranging from murders to car thefts to household burglaries. What did this string of ruthless attacks and petty crime have in common? Very little, with the exception of DNA from one individual recovered at each scene. A total of around 40 crime scenes all pointed to the same culprit.

Analysis of the DNA revealed the serial killer to be, somewhat surprisingly, a woman. Analysis of mitochondrial DNA indicated that the sinister suspect was most likely of Eastern European or Russian descent, narrowing down the list of suspects but not nearly enough. A bizarre picture of this woman began to form. The notorious killer became known as the Phantom of Heilbronn, a fiendish criminal with a 300,000 euro reward on her head. Also known as the Woman Without a Face, Germany’s most dangerous woman was responsible for the murder of half a dozen people… but also had a soft spot for theft?


It sounds like a baffling crime fiction novel in which the killer is never brought to justice, slipping through the net and always staying one step ahead of the perplexed police. However there is one slightly twist to this tale… the Phantom of Heilbronn was indeed never found, and a few years ago German police made a somewhat embarrassing confession… they had been pursuing a non-existent serial killer for 16 years. The Phantom of Heilbronn did not exist.

But how could this be? A series of vicious crimes and evidence of a single perpetrator at every scene, it seems like a reasonable assumption that this was the work of the same person. And this is just what investigators believed for a rather long time.

However suspicions were raised during the investigation of a person who had died in a fire, at which point the DNA of the “serial killer” was found on the body when attempting to identify the victim. The victim transpired to be a male asylum-seeker, so why had female DNA been recovered from the scene? This was bizarre enough, but even more so when subsequent tests (presumably using different swabs) failed to find the DNA again.

It soon came to light that police had been following the wrong scent, and suspicion soon fell elsewhere… cotton swabs. It eventually emerged that the DNA recovered from all of these crime scenes belonged to none other than an unsuspecting woman working in a swab factory in Bavaria, a factory which happened to have numerous Eastern European women on its staff (at least part of the serial killer profile was right). And so the mystery of the Phantom of Heilbronn was solved. Investigators had been collecting samples from numerous crime scenes using cotton swabs inadvertently contaminated with the DNA of a slightly careless factory worker. No doubt there were some red faces the day this little tidbit of information came to light!

So how could something quite so farcical occur in the first place? Let’s review the facts. We have a female serial killer, unusual in itself (women account for about 9% of serial killers, according to research carried out by Radford University), but not unheard of so no suspicions were raised there. The crimes appeared to cover all manner of sins, from murder to drugs to burglaries. Perhaps a little unusual, but theories began floating that the suspect was a drug addict or homeless person who would stop at nothing to steal some quick cash to get their next fix. This criminal’s DNA was found in all sorts of bizarre places, including on the remains of a cookie and on a toy gun. The wrongdoings spread over a decade and across three different countries, never leaving behind anything else that might point to the identity of the culprit. One would assume this took the homeless woman theory out of play – there are probably not many people living on the streets who have the funds to hop from country to country on a crime spree. In short, this phantom was one busy woman who liked to travel and mix up her wrongdoings!

Sounds somewhat ridiculous in hindsight. So why did it never occur to anyone that perhaps something else was afoot? After spending so many years tracking this villain, perhaps the concept of her not actually existing was impossible to comprehend. Investigators had spent years trying to get inside this woman’s head and figure out her motives, and the evidence was there to support their (slightly farfetched) theories. If nothing else, this highlights the dangers of being too committed to one line of evidence. It’s a very real problem for investigators and scientists. Once a hypothesis has been developed and a person genuinely believes in this theory, it is very easy to ignore subtle (or not so subtle!) indicators that the precious theory might be wrong. Suddenly, all evidence and results can be twisted to fit the theory, whether ludicrous or not. Not great scientific analysis, but it happens.

The case shed light on the fallacies of forensic evidence, even those which we tend to place great faith in. DNA evidence is perhaps viewed as one of the more reliable practices in forensic science, provided the appropriate procedures are followed (did these investigators even run control samples?). But at the time DNA analysis was something of a forensic holy grail, certainly not a technique which would draw investigators down the wrong path for over a decade. Yet from this incident, and many others since then, it is clear that even the best of techniques and practices must be accompanied with sound logic and an open mind.

As embarrassing (if slightly amusing) as this case might be, it highlighted some very real issues. To be sent back to some laboratory basics – know your equipment, always run controls and follow the results, not your hypothesis!

Putting aside the shame-faced investigators, the millions spent, the dozens of investigations affected and the police hours wasted, it at least makes for an interesting cautionary tale to tell budding forensic scientists for years to come.


BBC News. ‘DNA bungle’ haunts German police. [online][Accessed 23 Mar 2015] Available:

The Guardian. DNA analysis: far from an open-and-shut case. [online][Accessed 23 Mar 2015] Available:

Time. Germany’s Phantom Serial Killer: A DNA Blunder. [online][Accessed 23 Mar 2015] Available:,8599,1888126,00.html