They called it the hammer murder case, but so bad was the state’s case that the judge ruled that their account of the crime was “highly improbable, if not impossible”.

To see how bad the evidence was, you need look no further than the hammer itself.

More novelty item than wood- work tool, the hammer, a Christmas gift from the victim’s family, has a bottle-opener where a carpenter would expect to find a claw- grip with which to remove nails. It is 28cm long and weighs only 330g but, because it is top-heavy, it feels heavier and more lethal.

It is this that the state alleged Fred van der Vyver used to bludgeon to death his beautiful, brilliant girlfriend, Inge Lotz, in March 2005. And, lying on the evidence table in Court 4 of the Cape High Court over the months of the trial, it took on a deeply sinister aspect.

Yet, despite the best efforts of police and prosecutors to dress up their mutton as lamb, the evidence that this tool was used in this crime was spectacularly weak.

On February 15 , Sergeant Peta Davitsz of the police forensic science laboratory testified that chemical tests revealed that the hammer had on it invisible traces of a substance that might be blood.

It was only in June, however, in literally the last act of making its case, that the state handed in a statement from another police forensic scientist. It said the tests on the traces revealed that if it was blood on the hammer, it definitely wasn’t Inge’s: the DNA was male.

If these tests were a washout, the prosecution offered another line of argument. This was that the shape of the wounds on Inge’s body was exactly what you’d expect if she’d been beaten with this particular hammer.

To prove this, they flew to Cape Town Frans Maritz, a one-time police forensic scientist who now works in the crime lab servicing the needs of South Dakota’s cops in the US. Maritz produced a set of photos of Inge’s wounds and transparent overlays of the hammer which showed that it fit neatly a number of Inge’s wounds.

But there was a serious problem with this testimony, too: the pictures of the wounds had been taken at an angle, creating significant distortion. So bad was this that a photo of one of the wounds suggested that it was between 17mm and 21mm across, a size that fit beautifully with the striking surface of the hammer.

The trouble was that the pathologist who had measured the actual wound, rather than a photo, said it was 30mm across, nearly 50percent wider than the hammer.

Placing a transparency of the hammer over badly taken photos proved nothing, and Maritz’s testimony was either embarrassing incompetence or deliberately misleading. Maritz’s greatest humiliation, however, arose from a photograph he took himself.

The picture in question is of the hammer buried bottle-opener first into a sheep’s head. Such was the appalling violence of the assault on Inge, the photo is intended to show how one of her head-wounds, whose depth and shape were not consistent with the striking surface of the hammer, may have been caused by the bottle-opener.

Under cross-examination, however, Maritz was forced to concede — with a petulant, grudging reluctance — that the hammer in the picture was not, in fact, the hammer that had belonged to Van der Vyver.

He couldn’t risk using that one in these tests, you see, because, when he used it the first time to test the potential of the hammer to cause this wound, the bottle-opener bent and almost broke. How do we know this? Because the defence obtained a video of Maritz’s tests and in it one can see the bending of the hammer quite plainly and one can hear the disappointed “ooohh” of Maritz and his colleagues.

As an exhibition of forensic science, the state’s case about the hammer verged on the farcical: there was no DNA linking it to the crime; the physical similarities between the weapon and wounds was dubious and unproved; and police experts had failed to tell the court that the bottle-opener, which they said could smash through a human skulls, had a tendency to bend. Not exactly an episode of CSI: Cape Town.