Human volunteers pioneered modern safety tests

It's not a pleasant thing to contemplate while driving, but the design of almost every aspect of a modern car interior is based on the tolerance of different parts of the human body to violent impacts.

Safety engineers have tuned the padding on the dashboard, the strength of the seat belt webbing, the stiffness of the steering column, the thickness of the windscreen glass, and even the rear view mirror mounting, according to how much of an impact the body can withstand without being seriously injured.

If you want to build a dashboard which doesn't fracture the skull when a head slams into it in a crash, you need to know how much of a blow the head can take without being fractured. The padding must be as stiff as it can be - make it too soft and in a high speed crash, the head will plough right through, into the metal behind. Make it too hard, and even at low speeds it may cause fatal brain injuries.

Well into the 1950s, car manufacturers tended to claim that accidents could never be made survivable - the violence was simply too much for the body to bear. But they were wrong.

Experiments on monkeys caused uproar in the US

Thanks to a gruesome catalogue of experiments, we now know how much it takes to injure every major bone and organ in the body. The results show that if humans are packaged carefully they can withstand even severe smashes.

Most of the information has come from subjecting the bodies of the dead to simulated collisions on high speed sleds. Nowadays, crash test dummies are used to fine tune new designs, and to provide repeatable standardised tests for new cars. The dummies are based on real human bodies.

Cadavers were dropped down an empty lift shaft

The work began in the late 1950s, at Wayne State University in Detroit. The first part of the body to be studied was the head. Initially the experiments were fairly primitive. Embalmed corpses from the university medical school were dropped down a disused lift shaft onto a metal plate. It turned out that the head was surprisingly strong - it could take a load of about a ton and a half for a fraction of a second without injury.

Cadaver testing has continued ever since. It's not something we have done in this country - we draw the line at using parts of corpses to study injuries to the legs, for example. But all manufacturers make use of the data from the experiments performed primarily in the USA, France, Germany, and Japan.

Everything from the larynx to the patella has been smashed. Techniques have become more sophisticated in an attempt to make the results more representative of living human beings. The blood vessels are sometimes pumped full of dye under pressure to study injuries to internal organs - the spread of the dye shows where vessels have ruptured.

Experiments on children are rare - and the dummies are less-developed

The cadavers tend to be elderly, since the only major killer of the young in the developed world is the motor car, and corpses which have already been through a crash are no use for this kind of research.

Cadaver testing has always been a delicate issue. Despite the vital information crash injury research has produced, it's never been fully accepted that this is a fitting use for the dead.

The issue isn't about to go away. Despite thirty years of cadaver research there are many questions still unanswered. There's been great progress in preventing fatal injuries to the head and chest, but for every person killed in a crash, a hundred are seriously injured - often with crippling damage to the legs and feet. Designing cars which produce fewer severe injuries is the next challenge.

And there are the children. Child crash protection is hampered by the lack of information about children's tolerance to impact - and the only way to provide that data is to perform experiments on dead children - which at the moment, none of the major car producing nations will permit.