The Salt River bed stretches across the metropolitan midsection of Phoenix like a snakeskin bleaching in the desert sun, bone-dry and barren, a discarded remnant of the past.

The Salt was a free-flowing river once, but now it sees water only at times of flood, when its channel fills with runoff and overflow from upstream reservoirs. The rest of the time it is a phantom waterway, a river of sand pocked with gravel pits, landfills and auto junkyards.

But there is a move afoot to return water permanently to the Salt, through a breathtakingly ambitious project that promoters call "opportunity on a gigantic scale" and opponents deride as a boondoggle of unmatched size.

The project, dubbed Rio Salado (salty river in Spanish) by its backers and "Rio Scamola" by its critics, crystallizes the water debate in Arizona, as competing interests in the cities gird for the same kind of battle that characterized the agricultural irrigation projects of old.

Conjured up by an Arizona State University architectural class 20 years ago and refined by a Boston design firm, the Rio Salado development project would transform more than 30 miles of parched riverbed into a verdant "waterscape," liberally dotted with resort hotels, office buildings, condominiums, restaurants and golf courses and laced together with 1,525 acres of lakes and 3,000 acres of lawn.

It is a concept designed to bring a glistening to the eye of any city promoter, and so it has -- despite the problems that remain.

First, no one is sure where all the water would come from.

And second, the project, in the bed of a river that has seen three damaging floods in the last decade, depends heavily on construction of a $380 million federal dam that would destroy much of Arizona's remaining natural river lands and could push the state's tiny population of desert-nesting bald eagles into extinction.

According to the Bureau of Reclamation, Cliff Dam, a 338-foot-high earthen plug across the Verde River northeast of Phoenix, would reduce flood water flow in the Salt River to 55,000 cubic feet per second, about one-quarter of the 1980 flow that destroyed flood-plain development along the river bed.

Rio Salado is designed to accommodate that reduced flow in its chain of artificial lakes; any more than that, and the water would threaten the proposed luxury hotels, residential districts and boutiques.

"If Cliff is not built, the size of the flood plain goes from 55,000 cubic feet per second to 200,000 cubic feet per second," said Phoenix real-estate developer Bill Schulz, one of the project's most ardent supporters. "It would have a tremendous financial impact. The land values that would rise as a result of building Cliff would not rise."

But if Rio Salado needs Cliff Dam to become a reality, the Bureau of Reclamation needs Rio Salado no less to build Cliff Dam.

"What flood control really means is real-estate development," said Phoenix physician Robert Witzeman, a leading opponent of the dam. "Sixty percent of the flood-control 'benefits' for that dam are attributable to the development for Rio Salado."

Witzeman is an officer of the Maricopa Audubon Society, and his interest in this fight is the preservation of a seven-mile-stretch of the Verde River, Arizona's only federally designated "wild and scenic" river, and the bald eagle nesting sites that would be flooded by waters behind Cliff Dam.

But Witzeman and other opponents of Rio Salado say that economics has as much to do with their position as environmentalism. Private engineering firms think that flood control can be had much more cheaply by upgrading dams above Phoenix, Witzeman said, and Cliff Dam could cost Arizona billions of gallons of what the state contends is its most precious commodity: water.

According to its master plan, Rio Salado would need nearly 7 billion gallons of water a year to maintain its lakes and lawns, enough to supply a city of 100,000 people and significantly more than the reservoir behind Cliff Dam would collect.

To Witzeman, that is a profligate misuse of water when the state is preaching water conservation and pushing farmers to give up water-thirsty crops like alfalfa and cotton.

Schulz demurs. "It would be a very small amount of water consumption for huge benefits," he said. "We have our heads screwed on wrong to think that urban use of lakes and fountains is bad because it uses water."

Where the water would come from, however, is not entirely clear. The project's promoters have suggested tapping into the state or city's Central Arizona Project allotment or using treated sewage effluent, although a significant part of the effluent has already been committed to the Palo Verde nuclear power plant's cooling towers. The city has plans to use much of the rest of the effluent on its parks and sports fields.

The master plan also suggests irrigating Rio Salado's grass with water pumped from beneath the four dozen or so landfills that dot the river bank.

City health officials are somewhat skeptical of the idea. One landfill has been found dangerous enough to be on the federal Superfund list for toxic-waste cleanup. Others are venting explosive methane gas.

The project's promoters say that a less ambitious development could be undertaken without Cliff Dam and with less landscaping, but it would be "less exciting." Schulz, for one, thinks that the water eventually will be available, as a result of retiring more farm land if nothing else.

But Witzeman thinks that developers are about to be tripped up on their own well-worn argument against farmers.

"Cliff's water simply squanders federal dollars to maintain a lavish, water-profligate life style in the Arizona desert," he told a recent congressional subcommittee hearing. "Do the taxpayers in the rest of the nation believe they should pay so Arizonans can live high on the hog, in denial of the realities of their arid environment?"