Although fish are neutrally buoyant, they still have to push water out of the way to move forward, he said. That water raises the surface -- a phenomenon that is often imperceptible as it may be spread across an entire lake, stream or ocean.
"The water can only go up because the bottom and sides of the channel are rigid," Bejan said. "That bulge, however undetectable, is the fish's footprint."
Fish must, therefore, work against gravity to lift an amount of water equal to their own mass for each body length they move forward.
"It puts fish in the same physical realm as runners and fliers."
That's from an article about similarities in locomotor dynamics across swimmers, runners, and flyers. A bunch of engineers showed that common physical principles explain relationships between body mass and speed, stride (or wingbeat) frequency, and muscle force.
On the other hand:
"From simple physics, based only on gravity, density and mass, you can explain within an order of magnitude many features of flying, swimming and running," added James Marden, professor of biology at Penn State. "It doesn't matter whether the animal has eight legs, four legs, two, even if it swims with no legs."
I'm fairly sure that my own running speed is within an order of magnitude of almost everything with legs, regardless of its size. So a lot of biological interest is cruising beneath the radar of physical constraints. But these relations may explain why some mouse-elephant type allometries are relatively similar between fish, mammals, and birds.
UPDATE (1/2/06): A reader e-mails: "Am I missing something? Sure fish have to push water out of the way as they move forward, but that water has a perfectly good place to go: into the space the fish just vacated! There might be some small vertical displacement of water, but it's an entirely local phenomenon that has no effect on the surface."
That seems like a good point to me. You might as well say that cars have to push air uphill in order to move forward (albeit with much less mass). Hmmm....