One aspect of the railroad industry's proclivity to merge is that it seems to be based on the belief that the size of the railroad revenue pie is finite.

If that is so, it makes the merger movement appear all the more understandable. The only way to get a larger piece of the pie is to take it from one of the other railroads, or to acquire another railroad and its share of the pie. And, since it's really difficult to take market share from a competitor, the choice has been to buy market share.The problem with that thinking, of course, is that it doesn't make any sense for Peter to rob Paul if Paul is going to turn around and rob Peter to get his money back. That makes it a zero-sum game, which is largely what has been playing out for more than 20 years.

You could go back to the round of mergers that began just before deregulation with the 1980 acquisition of the St. Louis-San Francisco by Burlington Northern. That was followed by Union Pacific's roll-up of Missouri Pacific and Western Pacific. UP's argument to regulators was that it simply wanted to be able to compete with the newly enlarged BN.

In the East, Chessie System and Seaboard Coast Line combined to form CSX Corp. in 1980, which was followed by Norfolk & Western and Southern Railway in 1982.

More recently, UP acquired Chicago & North Western in 1995, which was followed almost immediately by the combination of BN and the Atchison, Topeka and Santa Fe. The tit-for-tat continued as UP bought Southern Pacific in 1996, again arguing that it simply wanted to be able to compete with the newly enlarged BN.

The railroad pie was still finite when Conrail tried to expand its franchise by breaking off the Cotton Belt - or SP East, as it called it - from the UP-SP transaction. When that failed, it was only a matter of weeks before Conrail entered merger talks with CSX.

The game played out differently in the East, however, as there were only three players. Norfolk Southern couldn't accept CSX hegemony and, following a bitter war over Conrail, CSX and NS divided the prize.

In their joint merger application to the Surface Transportation Board, CSX and NS said they would increase the size of the pie by creating a north-south intermodal market where trucks controlled 86 percent of the business. Both railroads are still trying to dig out from the congestion and service difficulties that came with their June 1, 1999, division of Conrail, and it remains to be seen if their strategy will work.

Fast-forward to the current attempt to combine Burlington Northern Santa Fe and Canadian National. While they clearly hope to divert existing traffic flows to their combination (taking the other guy's piece of pie), they also plan to grow the pie.

With north-south trade lanes between Canada and Mexico growing much more rapidly than traditional east-west domestic markets in Canada and the United States, BNSF and CN stand to gain some traffic without having to take it from competitors.

But look what happened. The other major railroads, seeing someone eyeing their piece of the pie, ganged up on BNSF and CN. Threatening responsive mergers of their own - and at a time when they really couldn't pull them off - the other rails persuaded the STB to impose a 15-month moratorium on consideration of any combinations of major railroads.

BNSF and CN are seeking appellate-court reversal of the moratorium and the outcome remains to be seen.

Before they were summarily put on hold, BNSF and CN, currently the two most efficient railroads in North America (NS could make that claim prior to June 1), were taking steps designed to enlarge the pie. They had announced that they would guarantee current service levels to all customers and that they would retain all existing gateways to other railroads.

Retaining gateways gets close to removing bottlenecks, long sought by shippers and equally long fought by railroads. Bottlenecks occur when two or more carriers can serve either the origin or destination, but only one serves both.

The STB has ruled, and been upheld by the courts, that railroads do not have to quote bottleneck rates. The difference here is that BNSF and CN were promising to retain what customers already have. Quoting bottleneck rates would give them something they do not now have.

BNSF and CN have made little secret that they don't see competitive access the same way other railroads do. On the theory that the lowest-cost producer is the last one chased out of any market, both carriers are prepared to deal differently with the issue.

Bottleneck rates and other forms of rail-to-rail competition will drive rates lower. But BNSF and CN believe that if they provide consistently superior service, more traffic that now moves by other modes will flow to the railroads.

That way they hope to end up with their share of a larger pie.