Andrew Fitzherbert was convicted and sentenced to life in prison on the basis of DNA evidence alone. His case shows that it is often not the technology or the science but the supervising biologist’s subjective interpretation of the results that is the crucial factor in assessing whether a suspect sample and a crime-scene sample “match.”
by Mary Garden
On Friday, February 27, 1998, between 4 p.m. and 8 p.m., veterinarian Kathleen Marshall (subsequently dubbed the “Catwoman”) was murdered in the downstairs clinic of her home in Wilston, on Brisbane’s northside. Her decomposing body with 52 shallow stab wounds was not discovered until Sunday afternoon when two directors of the Cat Protection Society of Queensland (CPSQ), of which Marshall was president, visited her home. Sixteen cats and three dogs, unfed and distressed, were upstairs.
The police investigation initially focused on members of the CPSQ where power struggles and infighting had been a common occurrence, intensifying during the six months before 52-year-old Marshall’s death. In April, however, Ken Cox, a forensic biologist from the John Tonge Center, announced that he had found male blood in the crime-scene samples. Investigators decided to eliminate every male person involved in the deceased’s life, beginning with male members of the CPSQ and any male connected to a female member.
Palm reader and spiritualist Andrew Fitzherbert was the partner of Ruth Bennett, a recent CPSQ female member. He also practiced palmistry at the Windsor Spiritualist Church frequented by Virginia Houston, the chief suspect in the early stages of the investigation due to her tempestuous relationship with Marshall. Houston would be subsequently eliminated as a suspect.
At first 49-year-old Fitzherbert chose not to give a DNA sample because of his mistrust in “the system” and his belief that such evidence could be manipulated. This refusal made him a “person of interest” to the police.
Based solely on DNA presented at trial, Fitzherbert was sentenced to life imprisonment for the murder in August of 1999. His conviction marked the first time in Australia – and only the third in the world – where a person was convicted of a crime on the basis of DNA evidence alone.
The DNA evidence used to convict him was extracted from a few tiny drops of blood found at the crime scene.
Fitzherbert, acting as his own attorney, appealed his conviction in March 2000. He was unsuccessful. Fitzherbert had the right to retest the blood samples at the government’s expense at both his trial and appeal, but was not aware of this right. He also never received a five-page review of the DNA evidence conducted by Dr. Bentley Atchison of the Victorian Institute of Forensic Medicine in March 1999 for Legal Aid Queensland, which represented Fitzherbert. This review concludes: “The claim that the stains came from the accused and no one else cannot be justified.”
And yet Ken Cox, a forensic biologist from the John Tonge Center, told the jury there was “one chance in 14 trillion the blood found at the crime scene was from someone other than Fitzherbert.” This staggering probability is very impressive and persuasive, given that the population of the earth is less than 7 billion, but what does it mean, and how accurate is it? It does not mean that we will encounter such a “match” only once in 14 trillion people. The calculated frequency is only an estimate and can be off by an order of magnitude in either direction.
One of Fitzherbert’s later appeal lawyers, Laura-Leigh Cameron-Dow, says the DNA results do not support this probability. “I’ve been given alternate estimates from other experts of one in 100, 000,” she says. Professor Barry Boettcher, a leading DNA expert in Australia and influential in uncovering problems with evidence in the Azaria Chamberlain case, also challenges Cox’s interpretations. “There is just no possibility of determining the accuracy of the figure given to the jury. But it would have sounded highly impressive to them.”
More worryingly, Boettcher says when he examined the graphs of the DNA profiles the DNA from the male blood at the scene had deteriorated more than the DNA from the unknown female’s blood at the crime scene. “There is a chance the greater deterioration of the DNA in the male blood was due to it being deposited some days earlier than Marshall’s murder.”
Criminologist Paul Wilson has expressed concerns about the case for years and says there are strong doubts as to Cox’s DNA results. He says evidence based on small samples of genetic material, known as “low copy DNA,” is no longer used in Britain, although still routinely used in Australia. Disturbingly, Atchison expressed surprise that Cox uses a method to report population frequency statistics (i.e., a match in nine DNA tests is sufficient to say the stain came from the accused and no one else) that is simply not supported by “the majority of scientists in the world,” and quotes eminent statistician Dr. Ian Evett of the Home Office Forensic Service in England as saying this concept of individualization is “arrant garbage.”
Astonishingly, the victim’s DNA was never profiled. Cox claims her blood was too decomposed, even though earlier he had claimed that DNA “will work on degraded specimens.” But why weren’t other tissue samples, such as hair, skin or liver, examined? Even though six female profiles were obtained from the 25 blood samples collected from the crime scene, it subsequently was assumed these belonged to the victim. But what if the murderer was a woman? And crucial evidence was not tested, including a pair of bloodied shoes. It is unlikely these were Marshall’s, as she did not wear shoes of this style, and when her body was found her feet were clean, without bloodstains. And no one bothered to check whether the shoes were her size, or fit.
Public Defender Andrew Haesler says that in many cases it is not the technology or the science but the supervising biologist’s subjective interpretation of the results that is the crucial factor in assessing whether a suspect sample and a crime-scene sample “match.” (Note two individuals can coincidentally match at all loci tested). “What she/he is doing is looking at the Profiler Plus readouts and coming to a conclusion. In some cases the readouts will be clear and conclusive, in some the readings will not be so clear and in others they will be far from clear at all.” Not only did Atkinson note discrepancies between Cox’s reported results and those in his notes, but states that “Mr. Cox has assumed the type of the deceased”!
Wilson and Dianne McInnes, in their co-authored book Five Drops of Blood, point to inconsistencies in the records of the relevant crime-scene samples by the John Tonge Centre, and claim that the DNA findings from a handkerchief of Fitzherbert were “unsupported by a journal record or any computer records, and the sole record in the spreadsheets was impossibly dated. Either it was not genuine or there were serious mistakes in the report.”
At the time Cox did the testing, the John Tongue Centre had not obtained accreditation by the National Association of Testing Authorities (NATA). In late 1998, investigations discovered shortcomings including poorly documented records, unsealed evidence, unrestricted access to specimens, and potentially contaminating specimens in the refrigerator. By the time of the trial the centre had managed to get accreditation and while Cox made a point of mentioning this, he did not mention the shortcomings found by NATA the previous year. In March 2002, the centre was told to lift its game or lose its accreditation as it was found to fall short in areas of record control, methodology, handling of exhibits and reports of results.
Atkinson, in his five-page review, specifically challenges Cox’s analysis of sample 18, which had been reported as a mixture “consistent” with a mixture of the accused and the deceased. Atchison points out that this is not supported by the data, as the profile clearly does not match the accused/deceased at four loci. This mixed sample also showed a peak of 14 in the first loci, which must have come from someone else’s DNA as it does not match Fitzherbert’s profile or the profile of the other person (whether Marshall or some other female).
Aside from the problematic evidence itself, another worrying feature of the trial is what is known in the legal profession as the “prosecutor’s fallacy.” This is the fallacious reasoning that the rarity of the DNA profile is equated to the likelihood of guilt. A match probability ratio is not an expression of the probability that the accused is guilty, but the probability that the accused is responsible for leaving the crime-scene stain. It can never be enough to prove a case beyond reasonable doubt in the absence of other evidence. Yet prosecutor Paul Rutledge described Cox’s numbers as “awesome.” He said there was only one rational explanation for the presence of the male blood, “Fitzherbert has left traces of himself at the scene of Kathleen Marshall’s murder”, and hence “the man who left the blood in that surgery is guilty of the murder.” Rutledge called Fitzherbert a liar, reminding the jury they had been given the “expert evidence” of a scientist compared to the “garbage bin of the sciences of palmistry.”
Justice Ken MacKenzie committed the same “prosecutor’s fallacy” in his summing up to the jury: “You would exclude as unreasonable on the basis of the scientific evidence any suggestion that someone else may have left the blood at the scene and you would therefore be satisfied beyond reasonable doubt that the accused was the killer.”
DNA evidence if unchallenged has persuasive appeal to juries and judges. The problem is that jurors think it is absolute and infallible. On hearing of my interest in the case, one of the jurors on the Fitzherbert trial contacted me and explained that they had been swayed by the huge statistical odds provided; after all Ken Cox was touted as the expert witness, and the figures were all beyond them. The juror now believes they made a mistake, and that Andrew Fitzherbert was not given “a fair go”.
In an April 5, 2000 interview for the Southern Star newspaper, Cox boasted that “by far his most satisfying and successful case to date was providing crucial evidence in the brutal murder of Cat Protection Society president Kathleen Marshall.” This “evidence” was indeed crucial for putting a man behind bars, but it now appears that the evidence was flawed.
Fitzherbert, behind bars for over a decade, continues to protest his innocence. He claims he had never met the victim, nor been to her house. No motive has ever been established, although strangely Public Defender Andrew Haesler in his review of cases which had relied mostly on DNA evidence claims that in the Fitzherbert case “there was evidence of animosity and contact between the accused and the victim.” What animosity? There was no evidence that Fitzherbert knew Marshall, let alone knew where she lived. Fitzherbert also had a solid alibi for the Friday night that Marshall was murdered. Although the crown case rested on the assumption that the murder occurred on Thursday night (when Fitzherbert claimed he was home with his partner), this assumption relied heavily on the evidence of the entomologist Russell Luke who estimated the likely time of death from the maggots found on Marshall’s body. But at least six reliable witnesses, who knew Marshall well, including the local chemist, claimed to have seen Marshall on the Friday afternoon.
Andrew Fitzherbert’s conviction based solely on DNA results is disturbing and appears to be a blatant miscarriage of justice. Since 2003, there have been requests to several consecutive attorney generals to release crime-scene samples for testing, but these have been refused. The Innocence Project in the United States (where in most states it is mandatory that convicted people have access to samples) claims that one of the most common causes of wrongful conviction is improper forensic testimony.
Forensics techniques are sometimes improperly conducted or inaccurately conveyed in trial testimony so that juries are left with the impression that the evidence is more scientific than it is. In some cases analysts provide inaccurate statistics for the percentage of the population who share the perpetrator’s blood type.
In August 2010 the Queensland Government released guidelines for applications to the attorney general to request post-conviction DNA testing. One of these is that the attorney general has to consider whether testing would unnecessarily and adversely affect the family of a deceased victim. That is understandable but there also needs to be consideration of the awful suffering of a person who may have been charged and imprisoned for a crime he or she did not commit. Not to mention the wrongfully convicted person’s family. Another is that the biological material had not been previously tested. Although the shoes that were recovered from the scene were extensively bloodstained, Cox declared he did not test these. These should be tested, along with a range of other times retrieved including rags, a button and some bags. In a landmark case which relied on DNA evidence that has subsequently been challenged by experts, including the Victorian Institute of Forensic Medicine, such re-testing may well prove that Fitzherbert is innocent.