In spite of the unabated decline, Marsh perseveres. "I've spent a 30-year career watching the animals I love and work on go down the toilet despite my best efforts," he said. "I love these fish. I want my children, my grandchildren, to have an opportunity to love these fish the way I do. They were a remarkable creation of Mother Nature. For us, as a species, to allow these animals to be extinguished is fundamentally wrong. It's a philosophical thing."

It's an outlook shared by Tom Burke, the fishery group manager for the Lower Basin's multispecies conservation program at the Bureau of Reclamation, who has been a linchpin of razorback recovery efforts here for three decades. A large, bearded man famous for breaking out his harmonica at meetings, Burke has a tendency to wax poetic about the river, and he's smitten by the razorback's tenacity. "They evolved in a river of extremes," he said. "These fish are capable of withstanding 200-plus million metric tons of sediment a year coming out of Grand Canyon. That's --" he paused, trying to think of an apt comparison. "Well, it's a bunch."

Marsh and Burke began working on the Lower Colorado the same year, as field biologists. Their mentors were lifelong collaborators on endangered species conservation, and Marsh and Burke are close friends and colleagues. Their relationship is partly responsible for the continuing push to save Mohave's razorbacks. "We've had the same outlook," said Marsh. "We've had a lot of disagreements over the years, but we agree to disagree. His goal is the same as mine: to do the right thing for these critters."

The Colorado River was always, in Burke's words, "a tough neighborhood" for fish, not only because of its famous sediment load, but also because of its wide fluctuations in water volume, temperature and salt content. The razorback's 50-year-lifespan, he explained, gave it a fighting chance of producing at least one surviving offspring. But then humans made the neighborhood even tougher: We crisscrossed it with dams, and stocked it with voracious fish from far-off lands.

In the 1950s and early 1960s, government officials also routinely poisoned waterways, killing native species in order to stock sport fish like rainbow trout. Using money given to the states from taxes on recreational fishing products, government officials poisoned 2,500 miles of streams and 225,000 acres of lakes in 34 states, according to Anders Halversson, whose new book, An Entirely Synthetic Fish, chronicles America's love affair with rainbow trout.

In 1962, the Wyoming and Utah wildlife agencies orchestrated the ne plus ultra of fish genocide, a massive effort to poison native fish using a chemical called rotenone. The idea was to create a clean slate on the Green River -- a main Colorado tributary originating in Wyoming's Wind River Range -- above the soon-to-be-completed Flaming Gorge Dam, eradicating all life in the water and then filling it with millions of rainbow trout. Using a series of "drip stations" strung along 450 miles of river, as well as airboats and helicopters, officials killed off 450 tons of fish in three days. "As the rotenone front progressed downstream," writes Halversson, "it drove large schools of desperate fish in front of it, a sight that deeply impressed itself on those who saw them ‘thrashing about and struggling for air on the surface of the river.' "

Today, the idea of poisoning rivers and lakes is shocking -- unless it's to kill exotic species to protect native fish populations. But the second part of the Flaming Gorge strategy, stocking game fish, remains standard practice up and down the Colorado. In the Lower Basin alone, more than 100 non-native fish species have either become established in reproducing populations or are continually re-stocked. In the eight years following the Green River poisoning, according to Halversson, fish and game officials filled Flaming Gorge Reservoir with upwards of 20 million rainbow trout. The stocking continues today.

At Willow Beach, a short drive south of the Hoover Dam in Arizona, rainbow trout grow in raceways just yards away from the razorback suckers' pens. Every Friday, year-round, 2,000 rainbows are set free in Lake Mohave.

The rainbows are released away from razorback spawning grounds in an effort to limit predation, and at first it seemed that the rainbows wouldn't be too much of a problem -- as long as young razorbacks had a few safe havens in which to grow. "These fish evolved to spawn as the rivers started to rise from melting snow, and then rising waters would distribute the young into these slow eddy backwater areas," explained Burke. There, he said, the sediment would drop out, the sun would shine, and algae would grow, giving young fish a food source in a protected area.

The scientists established a series of backwater ponds, separated from the lake by man-made berms, designed to mimic the oxbow lakes that formed in the river's original floodplain. The fish grew relatively quickly, but they couldn't grow large enough to fend off predators before the ponds either dried out or reconnected to the lake. So the scientists turned to the hatchery. If wild larvae were collected and reared in safety for as long as possible, the thinking went, the young fish would have a better chance of surviving, and overall genetic diversity would be conserved. (In standard hatchery programs, larvae are produced at the hatchery by selectively breeding certain fish, which severely limits the gene pool.) Still, year after year, most of the fish died shortly after being returned to the lake.