Last week, GEC-Alsthom, the Anglo-French builder of the TGV (Train a Grande Vitesse) and the Eurostar trains, anounced it was talking to its rival Siemens, which makes Germany's Inter City Express. The idea is to create a partnership that would be ready to storm the growing market for high-speed trains outside Europe.
An American attempt to build a TGV line has collapsed, yet the rail-obsessed Euros believe they can export ever more sophisticated trains. Why? In part, strangely, because of the American reaction to the high-speed network. Though the TGV ultimately failed on funding, it was badly damaged by a massive opposition campaign by SouthWest Airlines.
Herb Kelleher, SouthWest's chief executive, indicated to the Texas legislature that if the rail line was built he would move his head office out of Dallas. He also claimed high-speed rail trains were unsafe - even though not one person has ever been killed on one. There was little doubt that his real concern was that a 200mph line linking Dallas, San Antonio and Houston would do great damage, even to one of the most efficient airlines in the US. Mr Kelleher would, it seems, agree with Bill Vantuono, managing editor of Railway Age in New York, that "given the choice, the travelling public will opt for the railroad".
Does this mean that the railway is about to have a second golden age, and that airlines will be driven to the wall by a revived Santa Fe or Canadian Pacific? Well, up to a point, but high-speed trains have to overcome big handicaps to compete with planes. First, even trains travelling at 200mph cannot compete over long distances with planes moving at three times that. It will never be worth building a high- speed track across North America or Australia, and we will never see a high-speed Trans-Siberian Railway. Second, overall track mileages are continuing to shrink: despite the TGV and ICE, lines are being closed across France and Germany. Britain is not the only country to have decimated its regional railways.
Nevertheless, there is wide agreement that many people who now travel by air or road will in future travel by rail. The big opportunities are where cities can be linked in less than three hours.
After Japan's Shinkansen Bullet Trains were introduced in 1964, several domestic air routes were abandoned. The Tokyo-Osaka route takes about two-and- a-half hours, against one hour's flying time. But many more people use the train rather than the plane, because travel on the ground and check-in time can wipe out the advantage, and the Shinkansens are more frequent.
Most of the high-speed activity in the last decade has been in Europe. Since the French launched their TGVs, they have become the acknowledged technical leaders. In 1990, one of their trains reached a world record 320mph on the Atlantique route south-west from Paris; the standard top speed is 186mph. TGV lines now run in three directions from Paris - to Valence in the south, Tours in the west and Calais in the north - and there are plans both to extend existing lines and to build new ones. The Spanish have also bought TGVs, which they call the the Alta Velocidad Espanola and which runs from Madrid to Seville. The Eurostar to London is a close relative of the TGV, though it has complex extra systems to allow it to run on both sides of the Channel (half of it is built in the UK, while the TGV is a thoroughly French affair). Also, it is certainly not a high-speed train once it arrives in Britain, thanks to the old track.
When the TGV route from Paris to Lyon was opened fully in 1983, travel time was slashed from four to two hours. "It decimated the air service," says Gordon Wiseman of Railway Gazette International. But to run flat out, the TGV needs specially laid tracks and, as the table shows, times slip badly if it has to run over ordinary lines. Although three-quarters of the distance from Paris to Nice is on high-speed lines, the rest is enough to slow the journey to seven hours. The consensus in the industry is that most business travellers will choose to fly if the journey takes much more than three hours.
There is an alternative for rail networks that want to cut travel times but do not want to build new lines. In the early 1980s, British Rail experimented with the Advanced Passenger Train, which could take corners faster because it tilted. It clipped an hour off the London to Glasgow time, but ran into all sorts of embarrassing publicity when the tilt mechanism stuck and the train refused to return to the upright. BR ran out of money before it could iron out these technical bugs. The APT was scrapped 10 years ago and pounds 50m was written off.
The Italians and Swedes continued where the British left off. The tilting Pendolino has cut the travel time from Rome to Milan from six to four hours, and is already cutting its way into airline traffic, while the X2000, built by ABB, is tilting its way through Sweden. Although there has been talk of reviving the tilter for the West Coast line, it would almost certainly have to be bought from abroad. Britain's involvement in high-speed trains is now limited to a half share in the Eurostar. The other train factory in Britain, British Rail Engineering Limited (Brel), was bought by ABB but has failed to win enough orders and is being closed.
Taking a European rather than British view, though, there is plenty of reason for optimism. Although the Japanese invented high-speed train travel, they seem to have lost their technical edge to the Europeans. In South Korea, there was a three-cornered fight between GEC-Alsthom, Siemens and a Japanese consortium to build a high-speed line, complete with trains, between Seoul and Pusan. GEC-Alsthom won; Siemens cried foul over the $2.4bn (pounds 1.5bn) deal, but it must have tipped it towards a decision to link up with GEC-Alsthom rather than an attempt to go it alone.
Other state and national governments have said they would like to speed up their railway systems. The Taiwanese are planning a link between Taipei and Kaohsiung, their two biggest cities. Studies have been made on a Moscow- St Petersburg link. In Australia, a consortium called Speedlink is working on a proposal for a Sydney to Canberra high-speed line, while Countrylink in New South Wales recently conducted a trial of a Swedish X2000 tilt train.
Despite the Texan fiasco, the US could also be a land of opportunity for the train makers. High-speed trains are not unknown there: the New York-Washington Metroliner can touch 160mph, and in September Amtrak will decide which of three consortia wins the $700m contract for 26 150mph tilting trains to link New York and Boston. Each competing consortium consists of one of the big three Europeans - Siemens, GEC-Alsthom and ABB - and a North American company (respectively General Motors, Bombardier and General Electric).
Although the new Newt-dominated Congress is certain to fight the White House's plans for public transport, the federal government has committed $184m over three years to fund rail studies and technological development. The concentration again is on how to link cities about three hours apart. The lines will be built, but only if state governments are prepared to support them, politically and probably financially. California is likely to encourage rail, while Washington in the north-west has already leased a Spanish Talgo train as an experiment. But it seems likely that the first real high-speed rail link will be built in Florida.
The Florida Department of Transportation has declared that no new highways can be built, and the state has committed $70m a year for the next 25 years to the railroads. Unlike the North-Eastern links, which rely on existing lines, new tracks would have to be laid linking Orlando, Miami and Tampa. This would be an opportunity for the French to transfer their TGV technology wholesale - the new generation TGVs, which will reach 225mph, will surely tempt the Americans. "The chances of Florida having a high speed system by the turn of the century are quite good," Mr Vantuono says.
How far trains take over from planes in the US depends as much on psychology as economics, Mr Wiseman believes: "Between 1975 and 1990, people got so thoroughly fixed into using air that they may now hear what people say about rail - but they won't always listen."
If we want to see how train technology will develop in the future, we have to return to Europe. The Siemens/GEC-Als- thom link is likely to produce ever swifter conventional trains, although it will be challenged by at least one quite different principle.
If it is built, the magnetic hovertrain between Berlin and Hamburg would run at 250 mph and would cut travel time from three hours to 53 minutes. The train would be kept above and moved along a monorail by magnetic repulsion, which would mean there was no physical contact between them. Magnetic levitation (Maglev) has been around for many years, but this is the first serious attempt to build a long-range version.
But Mr Wiseman is sceptical that Maglev will come to fruition. "They have been struggling to reinvent the wheel for 25 years," he says.
Oddly enough, one country to gain from a high-speed rail network takes it so much for granted that it is barely aware of it. British trains have been travelling at more than 100 mph for most of this century - and the current InterCity 225s can reach 140mph. But unless GEC-Alsthom can grab a bigger share of the rail business from its French partner, British industry will hardly benefit from the growing interest in high- speed trains. The abandonment of the APT in 1985 could turn out to be one of the more unfortunate decisions made by this country's policymakers.
Train versus plane
From Paris to Grenoble (300 miles straight line):
TGV Air France
Journey time 3 hours 55 minutes
(2hr 30 centre to centre)
Frequency 5 daily Twice daily
Cost pounds 45-57 pounds 122 (one class)
Food "OK" meals Drink only
From Paris to Nice (430 miles straight line)
Journey time 7 hours 1hr 20 mins
(2hr 50 centre to centre)
Frequency 10 daily 12 daily
Cost pounds 57-69 pounds 142 (econ), pounds 191 (bus)
Food "OK" meals Drink only
Both these routes involve a mix of dedicated high-speed and ordinary lines.