Failure is not an option for some migratory fishes: if they don't make it upstream from the ocean to their freshwater spawning sites, they don't reproduce. But all is not lost if a fish finds itself struggling in a fast-flowing stream; swimming at a different speed may get the fish out of trouble. Engineers designing structures like fish ladders to help fish pass man-made obstacles while travelling upstream need to know what migratory fish are capable of. To find out just how far upstream fish can get by adjusting their swimming speed, Theodore Castro-Santos presented six fish species with high-velocity water flow challenges(p.421).

Castro-Santos explains that fish have three steady-swimming modes: in sustained mode, fish don't fatigue; prolonged mode characterizes a range of speeds that can be maintained for up to two hours; and in sprint mode, fish fatigue in under 20 seconds. Using the relationship between a fish's swim speed and fatigue time, Castro-Santos produced a model that predicts how far a fish will swim upstream in fast-flowing water. The model predicts that fish can maximise this distance firstly by swimming at a constantgroundspeed, which is the fish's speed relative to the ground, and secondly by selecting a faster groundspeed when they switch from prolonged swimming mode to sprint mode.

To see if fish behave according to the model's predictions, Castro-Santos used tiny tracking devices to record the movements of nearly 1500 individuals of six migratory fish species as they voluntarily swam up a long open flume with water rushing along at speeds between 1.5 and 4.5 m s-1. Sure enough, just as the model predicted, American shad, alewife and blueback herring selected a constant groundspeed against different water speeds, and selected a faster groundspeed when they switched from prolonged to sprint mode. But while the other three species (striped bass, walleye and white sucker) also maintained constant groundspeed appropriate for prolonged mode,they did not switch to the optimum speed for sprint mode. `By sticking to low groundspeeds and failing to switch to the higher optimum, these fish didn't travel as large a distance upstream as they would have if they had made the switch' says Castro-Santos.

So some species don't always select distance-maximising behaviour. Does this mean that the model is wrong? No, says Castro-Santos, it means that we should take a closer look at a species' life history. If a species has to traverse zones of high-velocity flow to reach its spawning habitat, natural selection will ensure that the species selects the appropriate speed to maximise the distance it can travel up fast-flowing zones. But if a species is equally happy laying its eggs above or below a velocity barrier, it shouldn't waste energy trying to get through fast-flowing zones; it should save its energy reserves for finding food and evading predators. And this is precisely what Castro-Santos found: the three species that encounter fast-flowing water in the journey to their freshwater spawning sites shifted to different speeds at points predicted by his model. But the other three species, which do not need to traverse velocity barriers to get to their spawning sites, did not switch to the faster groundspeed as his model predicts.

`This suggests that the relationship between swim speed and fatigue isn't enough on its own to predict the distance that a fish will swim upstream in fast-flowing water' says Castro-Santos, `The diversity of fish behaviour is also important.' Fish may be physiologically capable of achieving the maximum predicted distance but fail to reach this distance simply because they choose to swim at the wrong speed; fish that swim either too slowly or too swiftly will not achieve the maximum distance. Castro-Santos concludes, `When they don't select the right speed, they pay the price.'

Castro-Santos, T. (
). Optimal swim speeds for traversing velocity barriers: an analysis of volitional high-speed swimming behavior of migratory fishes.
J. Exp. Biol.