It has never been proven that Andrew Fitzherbert knew Kathleen Marshall or knew where she lived. The phone call Fitzherbert made a few days before her murder was to a phone number of the Cat Protection Society; he claimed he did not know that that was Marshall's (home) phone number, and the person answering the phone did not identify herself. The call was evidently a query about his membership application. His explanation is quite plausible.

I've quoted Laura-Leigh Cameron-Dow as to the problems with the DNA evidence as it was presented to the jury; according to her it should have been ruled inadmissible. Professor Boettcher, as I've also quoted, thinks the blood samples tested as male seem to have been more degraded than Marshall's which indicates they were left there earlier than the murder.

I don't agree with Andrew Fitzherbert's theory presented at his appeal (note he acted for himself) that the lab results were fraudulent or faked.

Enough question marks hang over this case, so what is the harm of re-retesting? As I said on 4BC Radio on Tuesday I don't know whether Andrew is innocent or not. But it is very dangerous to convict on DNA evidence alone. The most this shows in this case was that some man, with a similar DNA profile to Fitzherbert, was at the surgery at some time, maybe days even weeks before.

Note that in the appeal there is note that the 'alibi evidence was vigorously attacked by the Crown at the trial'. This was for Thursday night. Should the four reliable witnesses who saw Marshall on Friday afternoon be disregarded?? The alibi for Friday night was strong and the Crown spent very little time questioning this!

Was the defence weak? You'll have to read the transcript of the trial and make up your own mind.

This is my field. The discussion here is superb!

I would add two things:

1)"It has something to do with the psyche of our society, which elevates the image of the scientist to that of the uncompromising, incorruptible, all knowing, almost mystical, authority. After all, if these people in their white lab coats weren’t as pure as driven snow, our mobile phones and microwave ovens wouldn’t work and aeroplanes would fall out of the sky. Scientists have become our infallible clergy for modern times, and the authority of the clergy must be upheld at all costs." (Quote: Mick V.)

Note the above. It is true; and it must be acknowledged.
In my State this is VERY relevant. It has caused major problems. (I have no connection with any of the matters it has affected).

2) The legal 'system' is such that it will protect its decision to an extraordinary length. It becomes nothing to do with a convicted person; it has everything to do with protecting ITS verdict; not being seen to be wrong. Not to have its structure questioned.

I do not know if this man is guilty or innocent. However, it is no surprise to me that in this instance questions have been asked and deserve to be answered.

MaryG's theory that a man with "similar DNA profile" was in the surgery just highlights the unfamiliarity that many commentators have of DNA profiling. If the DNA in the surgery matched Fitzherberts, then it was Fitzherberts without a doubt. That placed him at a crime scene that he denied being anywhere near and even denied having contact. It's in the documents that Hamlet provided the link for (good work BTW). Someone with "similar profile" would not leave such a DNA profile, not even a brother or father! It is a matter of routine that DNA profiles of suspects and evidence are performed in different labs, so unless the DNA evidence was actually tampered with, that is a sample of Fitzherberts blood was added to it later, then it could be possible and hence the fraud allegation. However, Fitzherbert was the last person to give a sample, so how could that happen? The only other way was for someone to plant that blood sample before the police turned up. But Fitzherbert has never mentioned anyone siphoning off his blood and using it to frame him, so what is the most plausible explanation? I'll give you a clue: it isn't that the DNA test is wrong.

Lisa Faber (a forensic scientist who works for the N.Y.Police Department) says: "We never use the word 'match' ... the terminology is very important. We say 'similar' or 'come have come from' or 'is associated with'."

And as I've pointed out, all sorts of problems can arise in the collection of samples and mistakes in the lab. A recent Texan study found that error rate in forensic laboratories is about 1 in 10.

There are prominent critics of the way DNA evidence is used in courtrooms (e.g.Professor M Saks, Arizona University, Professor W Thompson, Uni California). Saks says there are two kinds of forensic science: 1. a dead body, let's see what chemicals are in the body? Straightforward science. 2. trace analysis: comparing a sample from the crime scene and a sample from a suspect. He says the latter is faith-based science, that there is no objective basis for making the link.

So if there is a dilemma at the heart of forensic science itself (and the 'experts' disagree), clearly DNA evidence needs to be treated with caution: it is NOT infallible.

Professor Thompson points out that DNA evidence is especially worrisome if it’s the only evidence presented. “There’s one chance in millions that any single person will win the lottery. But if you sell enough lottery tickets, somebody’s going to win,”. It is the same with a genetic lottery. “Even though there might be one chance in millions or billions that a particular person would by chance lose by matching, if you have enough people in the data base somebody’s going to lose, even if they’re innocent.”

And if Andrew Fitzherbert is such a liar, as the prosecution argued, then why didn't he simply make up a story about being at the surgery some days or weeks before the murder (he could have said he took their cat for treatment)?

Regardless of DNA evidence, if the witnesses are to be believed, Marshall was most likely murdered on the Friday night. That's where the focus should be. Fitzherbert's alibi 6 pm Friday - 1am Saturday was not challenged.

Mary Garden's excellent article "Bedazzled by DNA" asks a lot of "uncomfortable" questions which need some answers. The on-line opinion expressed by Lloyd Hamilton on 9/8/07 in forum.onlineopinion.com.au shows sound reasoning in view of the doubts surrounding this case. It is not too late to help Mr Fitzherbert in his struggle to prove his innocence at his own expense. Every Australian should have this right, after all, history shows that people have been convicted in the past for crimes they did not commit, and where there is doubt, there should also be the opportunity to redress a wrong.

MaryG, while Professor Thompson's statement may be correct, the chances that someone of "similar genetic profile" being in the same city at the same time and knowing someone by one degree of separation and also being involved with a murder are even more unlikely. You ask, if Fitzherbert was such a liar then why didn't he make up a story about taking the cat in to the surgery a few days before? Well, that is a point isn't it? Did he deny ever knowing Marshall BEFORE the DNA test was done? Or after? You see, if he denied it to police before he knew they had forensic evidence, then he cannot likely change his story can he? And the evidence has shown the story to be very likely to be false.

As for error rates, defining what an error rate is or means in a legal sense is also proving problematic. For instance, would you say that accidental contamination of a sample so that the profile matches the forensic scientist doing the work (or others who have handled the sample)is an error? It is the most likely scenario and yet legally it is meaningless because it would not result in a conviction.