David Williams wrote : "Group B cars were magnificent. They were the most extravagant, outlandish, exciting cars ever seen on public roads - a riotous mix of imagination and science. Often conceived by F1 designers, they made F1 seem drab by comparison, but they soon became lethally dangerous and the toll of dead and injured they left rapidly became Intolerable. In consequence, Group B was banned and rallying's emphasis was switched from endurance to outright speed overnight.

Far from widening the sport's appeal, a host of cars and their makers were excluded in Group B's aftermath. It took a decade to undo the damage of this compelling yet disquieting spectacle. Yet the sense of optimism and excitement when Group B was launched in 1982 was palpable. The pinnacle of a new structure for the sport, it promised radical and exciting cars at lower cost, along with a tidying up of some of the anomalies of the old numerical groups that dated back, in one form or another, to the 1965.

It was what the manufacturers said they wanted and they backed it to the hilt; there was a heady period in 1985 when it seemed that a major new rally car was launched every fortnight. The principal change was to reduce the minimum production requirement from 400 cars per year to 200. The theory was that an increasing number of manufacturers wanted to make special competition cars and the 400 deterred all but a handful.

In essence, Group B would allow other makers to build their equivalents of the Stratos (which was still winning at World Championship level as late as 1981) or the Renualt 5 Turbo. A reduced total would allow car a be sub-contracted to an outside company, just as Ford hired Reliant to RS200s, without disrupting a major car plant at prohibitive cost.

The other significant development was a reinterpretation of the rule. In the past, it had been used to accommodate minor production alterations, say from a Mark 1 Mini to a Mark 2. Once 10% of the production total for the group in question had been made, homologation was granted. It emphatically did not encourage extra performance. But under the new regime, a manufacturer could add a bigger turbo, lightweight or a new gearbox by making just 20 cars.

It could bolt on 100 bhp and shave 100 kilos, thereby overturning the counterbalance to the lower production total, a lower level of modification.

Given the nature of the rules, it might have been predicted that new previously unexploited technology, from torque-splitting centre differentials to electronic clutches, composite materials and water injection, should have flooded into rallying, and that racing car designers, such as Patrick Head from Williams and Tony Southgate at Ford, became seriously involved in unprecedented numbers. Austin Rover even had its own aerodynamicist, Bernie Marcus.

Above: Corrado Provera and Francois Chatriot, 2002

Lowering the production total flirted with danger from the outset. It invited car makers to conjure up a new breed of sports cars with intoxicating levels of performance that bore no relationship to their showroom counterparts, even if their appearance occasionally hinted otherwise. Had the process gone no further than a Landa 037 or a Renault 5 Turbo, rallying might have got away with it, but the seeds of destruction were sown at the very start.

While four-wheel drive and turbocharging already existed, neither had been exploited to anything like their full potential. Neither is intrinsically dangerous - they remain the fundamental requirements for a top rally car at the beginning of the 21st century - but Group B incorporated no mechanism for controlling them, nor for almost any other aspect of performance. The equation was simple; on narrow, twisty, often slippery roads, around 300 bhp had been the effective maximum that could be exploited with two-wheel drive.

Four-wheel drive offered double the traction and in the space of three years, turbocharging provided the power. Moreover, the regulations incorporated none of the safeguards in the use of materials or control of fuel systems that racing demanded. A Delta S4 was built along much the same lines as a 1967 Ferrari P4 Le Mans racer, but with vastly superior performance and very little of the rescue back-up that even a 1960s racing driver would have expected.

It was naive rather than criminal, because in 1982, four-wheel drive was as poorly understood as a distant planet - even if it didn't take long to harness its possibilities. The consequences might easily have been worse. In the chilling words of Tony Pond, Austin Rover's linchpin; "As a group of drivers, we knew it was just a matter of time before someone got killed. Sooner or later, you were going to have the proper Group B accident - an Ari Vatanen job, the proper one - but with a few spectators around; wrap it all up in a big one. I think it would have come sooner or later: a somersaulting Group B car exploding amongst the crowd."