A joint federal and Utah project to kill off one type of fish in order to give endangered and threatened species a better chance got out of hand and may have done extensive damage, an Arizona official says.

Biologist Dean Hendrickson of the Arizona Game and Fish Department said the project may have poisoned fish in as much as 60 miles of the Virgin River in Utah, Nevada and Arizona above Lake Mead.But those who headed the project say it actually may not have been as effective as they had hoped and that in any case they expect to go ahead with the next step late this month.

Hendrickson and officials with the U.S. Fish and Wildlife Service and the Utah Department of Wildlife Resources say a toxic agent, rotenone, was introduced into the river near St. George last week, in an attempt to kill red shiners, a fish species that reached this area from the Midwest.

The intent, according to an environmental assessment, was to treat a 22-mile stretch of the river in three steps in order to expand the area of the river that could serve as home to three other species - the Virgin River spinedace, a threatened fish; the endangered woundfin; and the Virgin River roundtail chub, which has been proposed for the endangered-species list.

The chemical causes asphyxiation by blocking transfer of oxygen from gills to the bloodstream. Arizona and federal agencies say it is not a danger to humans in the quantities used to kill fish.

Don Archer of the federal service, who headed the project with Denise Knight of the Utah agency, said the problem arose when officials overestimated how much water was moving down the river. That led them to put too much rotenone into the river, he said.

A station at which potassium permanganate, a neutralizing agent, was to be introduced into the river wasn't set up before the rotenone was released, and the toxic chemical is said to have passed the proposed location several hours before the station could be established.

That permitted the rotenone to flow on down the river unchecked, which Hendrickson said could result in the death of millions of fish.

Archer, however, said the rotenone should have been absorbed by 20 miles of turbid water before it reached Lake Mead.

He also said the main effect was to kill the red shiners that were the project's target.

"Besides that," he added, "I don't think we're even close to a complete kill (of red shiners) even on the 20-mile stretch we wanted to get. In Arizona, we saw a lot of live fish still, mostly red shiners."

He said the federal and Utah agencies still plan to go ahead with the third step on Oct. 31 on a 22-mile stretch of the river, though they will use less rotenone and will set up the potassium permanganate station earlier.