Before we had the many tools of forensic science we employ in crime investigation today, it was easier to get away with murder. At the same time, it was easier to be convicted of a crime you didn’t commit. Over the years, but by bit, scientific ways of determining what happened were developed, and each are tied with its “first use” case. Consider Spanish scientist Mathieu Orfila, who was forced to solve a problem he himself identified.
Mathieu Orfila determined that, in cases when the victim had been buried for some time, the body might become contaminated with arsenic from the ground. The accused might then be executed for the death of a "victim" who had actually died of natural causes.
His philosophy came back to bite him in the 1830s, when he was called in to argue for the prosecution in the case of a man who was accused of poisoning his own son. The body had been exhumed, and tested positive for arsenic. The defense was insisting that it came from the ground in which the body was buried. Orfila fought back, first doing tests on exactly how a body in the ground picks up arsenic, and then testing the ground around where that particular body had been buried for traces of arsenic. He proved that, although a body might absorb arsenic from the ground, this one hadn't. The man was convicted, and from then on, those who exhumed bodies collected soil samples as well.
There are nine other stories like this, and a bonus, at io9 covering the use of fingerprints, ballistics, lie detectors, and more. -via the Presurfer
Fingerprints aren't unique. DNA has many false-positives from the improper way law-enforcement tries to use it. Anybody and everybody can claim to be an expert witness with no vetting from the court. Blood spatter analysis, footwear impressions and bite-mark comparisons have zero basis in science. Coroners are political appointees who don't need any medical credentials. There's no science behind, or training for, arson investigators. etc., etc.
http://www.pbs.org/wgbh/nova/tech/forensics-on-trial.html
http://www8.nationalacademies.org/onpinews/newsitem.aspx?RecordID=12589