The vital task now is to establish a safety culture on the railways

We owe it to the memory of Robert Alcorn, Steve Arthur, Leslie Gray and Peter Monkhouse to learn the right lessons from the train crash at Hatfield in which they died. It has been pointed out that more than twice as many people die virtually unreported on the roads every day, which is true, but irrelevant.

We owe it to the memory of Robert Alcorn, Steve Arthur, Leslie Gray and Peter Monkhouse to learn the right lessons from the train crash at Hatfield in which they died. It has been pointed out that more than twice as many people die virtually unreported on the roads every day, which is true, but irrelevant.

The crucial difference is that when people get into a cars they are, to a large extent, the author of their own destiny. If you drive at 15mph in the gutter all the way from Edinburgh to London you will annoy other road users, but you are most unlikely to die. When, on the other hand, you board a train - or a plane or a boat - you are entrusting yourself to the competence of strangers and the effectiveness of unseen systems of human organisation.

The fact that train travel is relatively far safer than motoring and has been getting safer, even since privatisation, is not only tragically irrelevant to the families of the bereaved but beside the point for the rest of us. No one expects zero risk from travelling at speeds of more than 100mph, nor does anyone expect unlimited spending of ticket-buyer or tax-payer's money on safety, but the level of avoidable risk in the present system is still clearly too high.

So what should be done? At least the Hatfield derailment has diverted attention from the search for expensive systems to prevent train collisions and focused attention where it should be, on the regulation of human organisations. In fact, the response to the accident has been surprisingly measured. Although the poor state of the rails on that high-speed curve was quickly established as the likely cause of the accident, lodging responsibility firmly with Railtrack, not a single newspaper suggested that the resignation of Gerald Corbett, its chief executive, should be accepted.

This may be sensible, because he has improved the safety culture of the company in his three years there. But it cannot be enough to pat him on the back and say, "Carry on with the good work".

Nor is it any use Mr Corbett blaming the "fragmentation" of the rail industry at privatisation for its safety problems. It is true, but again largely irrelevant, that privatisation was botched and that the structure of the rail industry is a mess. There are too many operating companies which are not in meaningful competition with each other, and it is a pity that the current round of franchises is only reducing the number from 25 to 23.

What really matters, however, is the fragmentation within Railtrack itself. One of Mr Corbett's claims to stay in his job rests on his record of tightening up the company's control of its contractors, but the evidence of Railtrack's letter to Balfour Beatty published yesterday suggests that it is still far from adequate.

One of the respects in which the industry has become less fragmented since privatisation, however, is the creation of the Strategic Rail Authority under Sir Alastair Morton. But its legislative powers have still not been enacted, and the Government is waiting for the report of the Cullen inquiry into the Paddington rail disaster before considering other ways of trying to insulate safety at Railtrack from commercial pressures.

But John Prescott, the deputy prime minister, does not need to wait before addressing the question: if Railtrack knew the line at Hatfield was unsafe, was the threat of loss of revenue a factor in keeping it open? He needs to do more than knock heads together: he needs to create institutions and incentives which ensure that the culture of safety in the rail industry is as stringent as it is in aviation.