As a vivid demonstration of nature's power, few things beat the passage fish make up rivers to spawn.
This spellbinding phenomenon, long a staple of wildlife documentaries, is found across the world, including in the Middle East where the Shatt al Arab River in Iraq is known for the Indian shad fish that use it to spawn.
Yet in many of the world's mightiest rivers, new hydroelectric dams, often considered to be an environmentally friendly technology, have reduced the once billions-strong movement of migratory fish upstream to a trickle.
This is despite the fact that many dams are designed to allow the free passage of fish thanks to devices known as ladders and lifts.
Research published this year by an Abu Dhabi-based researcher indicates these features are often ineffective.
The work, led by Dr Jed Brown of the Masdar Institute of Science and Technology, found that many dams give passage to only a small fraction of the fish they are built for.
In his study, published earlier in the journal Conservation Letters, Dr Brown analysed publicly available data concerning fish passage rates on three major rivers in the north-east United States: the Susquehanna, Connecticut and Merrimack rivers.
On the Merrimack, the first dam had been designed to allow 300,000 river herring across each year. But between 2008 and 2011 the number that made it was just 706 each year - less than a quarter of 1 per cent of the intended number.
On the Susquehanna, just 3 per cent of fish - fewer than one in 30 - that passed the Conowingo dam, which lies 10 miles (16 kilometres) from the river mouth, went on to reach a spawning area upstream of the river's fourth dam.
Less than 0.5 per cent - one in 200 - of the hoped-for number of American shad fish were passing through dams on the Susquehanna.
A wide variety of migratory fish are affected by dams, including eels, shad, salmon and sturgeon.
The result of the fish failing to reach their spawning grounds has been devastating. Some populations have fallen to less than 10 per cent of their former levels.
It has not been for lack of trying. Dams are often designed with ladders - a stepped series of pools up which certain species of fish can, in theory, travel - or even with elaborate, and expensive, fish elevators, which consist of hoppers that mechanically lift from the lower level of the dam to the top.
But Dr Brown's research suggests such efforts are often in vain.
"Sturgeons typically won't use lifts and ladders at all. They need really specialised structures," he said. "Eels really won't use fish raisers or fish ladders. They use a specialised structure that looks like Astro Turf on a ramp and crawl up the ramp."
Given these problems, Dr Brown, who has been at the Masdar Institute since 2011 and previously worked for the US fish and wildlife service in New Hampshire, said that in many cases there was only one solution: taking out the dams. This is, he said, "typically the best option".
"This is because there are species of migratory fish that will not use fishways at all," he said.
As well as preventing fish from reaching spawning grounds, dams can also stop the creatures from reaching the sea again once they have spawned.
"Downstream passage is frequently problematic. Structures are needed to guide fish to a sluiceway or bypass around the dam so that they don't get harmed or killed by passing through the turbines of the hydroelectric plant," said Dr Brown.
Previous cases have shown that removing dams can offer huge benefits to spawning fish.
On the Kennebec river in Maine the Edwards hydroelectric dam was removed in 1999 and fish passage improved dramatically. According to state fisheries experts, about two million alewives - a fish that had been severely hampered by the Edwards dam - travelled up the Kennebec in 2009 and, by last year, that had grown to more than three million. The environmental organisation, Maine Rivers, has also reported that many other fish - among them the American shad, the striped bass and the shortnose sturgeon - are now able to travel much more freely up the river.
Not everyone believes the emphasis should be on destroying dams.
Steve Minkkinen, a US fish and wildlife service official, said dam removal was not realistic in many cases. Instead, he suggested efforts should focus on improving fish passage while the dams remained in place. "If we don't keep trying to develop new and better technologies we are never going to get there and we are never going to get better passage," he said.
Concerns about the effects dams have on fish are nothing new. In the US they stretch as far back as the 1700s and, later on, they sometimes caused tensions between states.
"In the late 1800s, there was a lot of back-and-forth between fish commissioners. In the states north of Massachusetts they were expressing concern to the fish commissioner in Massachusetts that the fish weren't reaching their state waters," said Dr Brown.
The problems caused by dams may be becoming more acute as global warming increases the pressures on many species. While some types of fish might be able to extend their range as temperatures change, larger numbers are likely to experience population declines.
This is happening as many new dams are being built on major rivers, among them the Amazon in Brazil and China's Yangtze.
"There are a lot of proposals to dam the Mekong, which is one of the most productive rivers from a fishery perspective on the planet," said Dr Brown.
While hesitant about drawing conclusions about what overseas authorities should do, Dr Brown did say his study demonstrates that engineering projects on rivers do have consequences, despite efforts to the contrary.
"There's a lot of damming going on. What we're trying to say is, once you make these fundamental changes to the river it's hard to mitigate or make compensations for the effects you have on the fish resources," he said.
"When you dam it you make these engineering changes and when you try to compensate you run into problems. It's probably putting a lot of faith in technology-based solutions to fix the problems that were created by the dams."