One of the few remaining joys inherited from a more leisurely age of travel is to climb into a bunk at Euston or Paddington and wake up the next morning to see the snow-covered Cairngorms rolling past or to hear the call of the seagulls as the train pulls into Penzance. That joy may not last much longer. The sleeper to Cornwall loses £1 million a year and is heavily subsidised by the Strategic Rail Authority. And the SRA, which is inviting train companies to tender for the new franchise for the route, has asked them to cost the sleeper service separately and say how they propose to pay for it. The likely reply from most of the bidders is simple: scrap it.

Clearly, the SRA — itself to be scrapped within months — cannot justify the use of large sums of taxpayers’ money to subsidise a moving hotel for holidaymakers and a few wealthy romantics. And if the Penzance sleeper is under threat, the sleepers to Scotland face an even greater challenge. The figures for the overnight trains operated by ScotRail to Edinburgh, Fort William or Inverness are bleak: some services are operating with no more than two dozen people on board, and the two or three who make their way to the dining car will find service more attentive than any hotel or restaurant.

British Rail also found it hard to make money on sleepers. Indeed, it announced ten years ago that the Highland sleeper to Fort William would be withdrawn, and turned down all appeals by passengers to keep it going. The response, however, was extraordinary. A passenger campaign to use the train took off. Hundreds flocked to buy tickets, and berths were booked for weeks ahead. Campaigners went to court to force BR to add more coaches, and the fight to save one of the most romantic rides in Britain caught the imagination of thousands who had never previously laid head on a BR pillow. After nine months, BR admitted defeat. The service was saved.

There are two obvious lessons from that campaign and from the present SRA ruling. The first is that if passengers want to maintain a loss-making service, they must use it. Half a dozen rural lines that planners would dearly like to axe have been been reprieved only because of lively local campaigns. Indeed, the Ipswich-Lowestoft line, scheduled for closure by Richard Beeching, has fought back so successfully that mainline through services to London have now been reintroduced after 40 years. The second is the need for market realism. Sleepers cannot be justified on the ground of social need; and if they are popular but expensive, the operators should raise the price.

The suspicion is, however, that the SRA wants to abolish all such services for bureaucratic convenience. The trains run slowly and, though they leave late, block the route for other late services. They need many staff, special operating rules and complicated pathing. But any attempt to kill them off by bureaucratic subterfuge is unacceptable. If operators can find a market, they should cost, run and promote them vigorously. They are few joys in modern rail travel. One of the last should be saved.