The floodwater that covered New Orleans after Hurricane Katrina was not unusually toxic and was "typical of storm water runoff in the region," according to a study published yesterday.

Most of the gasoline-derived substances in the water evaporated quickly, and the bacteria from sewage also declined over time, the scientist leading the study said. The water's chief hazard was from metals that are potentially toxic to fish. However, no fish kills have been reported in Lake Pontchartrain, where the water that once covered 80 percent of the city was pumped.

"What it most looks like is the storm water that is present in New Orleans every time it rains," said John H. Pardue, an environmental engineer at Louisiana State University, who headed the team whose research was published in the journal Environmental Science and Technology. "We still don't think the floodwaters were safe, but it could have been a lot worse. It was not the chemical catastrophe some had expected."

The study was published as President Bush made his eighth visit to the region since the hurricane.

The president put on a hard hat and helped build a house being constructed in Covington, La., by Habitat for Humanity. He also stopped at Delisle Elementary School, which reopened yesterday in Pass Christian, Miss.

"Out of this rubble is going to come some good; out of the devastation is going to come new cities and new hope," Bush told troops at the U.S. Naval Air Station, Joint Reserve Base, in New Orleans.

Cleanup crews continued to arrive in the flood-ravaged city, but it remained largely vacant, and business owners said they were increasingly worried about attracting workers back. Several directed their ire at the Federal Emergency Management Agency, which they faulted for not setting up more trailers in or near the city to house returning residents.

Residents of most of the lower Ninth Ward will be allowed back on a "look-and-leave" basis beginning today, but the vast majority of houses there remain uninhabitable. While upstate emergency shelters are emptying, many of the evacuees are heading out of state to apartments arranged by church groups, not by FEMA.

At an American Red Cross shelter in Gonzales, La., shelter manager Missy Stehr said the last buses leaving were taking people to Denver, where a church has paid for 18-month stays at local apartments.

"They let people sign up," said Jerry Putnam, assistant manager at the shelter. "We've yet to see an offer of a house or trailer. As I understand it, they don't have any space."

Although some experts and many news stories had warned of long-term contamination from the floodwater, yesterday's report was one of two suggesting that those fears were unfounded.

Tests results released by the National Oceanic and Atmospheric Administration found no human or animal fecal bacteria in fish or shellfish sampled in the Gulf of Mexico a month ago.

Industrial chemicals, such as polychlorinated biphenyls (PCBs) and brominated flame retardants (PBDEs), either were not found in the fish or were in the same concentrations as before the storm. Previous studies had shown no contamination from oil-derived chemicals.

The floodwater samples in the new study were collected on Sept. 3 from the residential Lakeview area and on Sept. 7 from Tulane-Gravier, a downtown neighborhood that includes the Louisiana Superdome and Charity Hospital. The hurricane struck the city on Aug. 29.

Pardue, the LSU researcher, and his colleagues -- traveling in a three-boat flotilla that included armed guards -- sampled surface and bottom water, which was 10 feet deep in some places.

The floodwater averaged about 100,000 coliform bacteria per 100 milliliters of water, far above the government standard of 200 for swimming and other "primary contact." For drinking, the standard is zero. After the storm, Lake Pontchartrain's coliform level was about 1,000 bacteria per 100 milliliters, suggesting that the water the city pumped in was substantially dirtier.

However, that level of contamination "was generally within the range . . . for storm water discharges . . . during normal wet weather flow," the researchers wrote.

Fecal bacteria do not necessarily cause diseases.

The tests showed that the water contained zinc and copper levels 10 times as high as is tolerated by freshwater fish, but well below the safe limits for drinking water.

The water pumped into the lake apparently did not kill fish, probably because it was so depleted of oxygen that fish had moved away, Pardue said.

The dissolved metal will settle to the bottom of the lake, he said. "Will that provide a source of metal that will be released over time? That is an unanswered question," he added.

The concentration of lead averaged 38 micrograms per liter. The limit for drinking water is 15 micrograms per liter. For arsenic, it was 50, with a drinking-water limit of 10. There was little evidence, however, that people were drinking the floodwater, which was a quarter as salty as seawater.

Given the huge volume of water from Katrina, the concentration of metals and arsenic would have been much lower if they had come only from roads, lawns and other surfaces. Pardue believes the other sources were submerged vehicles and houses.

There is copper in electrical wiring and brake pads; arsenic in brake linings; lead in batteries; and zinc in many plated metals.

Despite the visible sheen on the water in many places, the concentration of various gasoline compounds was very low. Dangerous substances such as benzene and toluene evaporate quickly, with "half-lives" of two to 20 hours, and are unlikely to remain in the sludge.

Staff writer Peter Whoriskey in New Orleans contributed to this report.