The former Republican governor of Idaho ushered in 60 hours of gushing meant to rebuild the canyon's beaches and sandbars with sand that tributary streams have dumped below the dam. He said the third such experiment since 1996 will help scientists learn how to balance natural and human demands on scarce Southwestern water. The humpback chub, a Colorado Basin native whose bulbous head makes it resemble a spawning salmon, is most endangered among the species that need more natural flows.

"Today we're here to set the river free once again," Kempthorne told a crowd of land, water and wildlife managers at the dam. "We're the stewards. It's our turn to do our part."

His action sent water from two portals at the dam's base jetting 100 yards or more through the air, whipping up a cool wind in the formerly still canyon and frothing whitewater against the orange-and-black rock walls.

Though hailed as an environmental triumph, there is dissension within Kemp- thorne's own department. The U.S. Bureau of Reclamation runs the dam and has agreed to stabilize flows only in the fall for the next five years - not during the summer months of peak air-conditioning demand. The National Park Service, meanwhile, fears that the five-year plan forecloses the option of further spring flushes. Data from the previous tests indicate that it's best to conduct these flushes anytime storms and flash floods have lined up adequate levels of sand from tributaries below the dam, as they have this year, Grand Canyon National Park Superintendent Steve Martin said.

"We have to provide water and power, but we also have to take some actions to begin to truly protect the resources of the canyon," Martin said. "It's clear that we need to do these flows every time we have the sediment."

In an interview after the dam event, Kempthorne said the environmental review that authorized the flush followed by five years of stable fall flows does not preclude another test if scientists agree it's warranted. Martin said it makes one unlikely, though he'll continue to press for it.

Environmental advocates who agree with Martin insist that the Interior Department is caving to commercial interests who get back control of the river for five years once this experiment ends. They say there's no more need to label it an experiment at all, after $80 million in federal studies have shown how floods can restore a more natural river.

"Really this is a five-year power and water deal," said Daniel Patterson, Southwest director for the Public Employees for Environmental Responsibility.

Any benefits of this flood may wash away within a couple of years if not repeated, scientists believe. And following up with steady flows only during the fall, as the Bureau of Reclamation proposes, doesn't help the fish at their most sensitive times. "The scientists have told us we need them April through October, and actually the most important part is July and August," said Bill Hedden, of the Grand Canyon Trust.

Three native fish species have disappeared from the canyon since the dam, built in the early 1960s, began spilling colder water into the Arizona national park and choking off 98 percent of the sands that through the millennia washed down and built sheltering sandbars and beaches. Biologists feared the chub would be next, though they believe experimental floods from the dam in 1996 and 2004 may have helped stabilize the population.

Routinely flushing water through the canyon and then stabilizing flows during the summer spawning and rearing season could jeopardize water supplies and hydroelectricity that the dam provides for some 400,000 Southwestern homes. But that flow regime is exactly what biologists believe offers young fish crucial breakwaters and a measure of peace. Congress has mandated both the continued provision of those commodities and protection of the Grand Canyon's natural resources.