MONTPELIER, Vt. (AP) — When a fish taken from the Connecticut River recently tested positive for radioactive strontium-90, suspicion focused on the nearby Vermont Yankee nuclear plant as the likely source.

Operators of the troubled 38-year-old nuclear plant on the banks of the river, where work is under way to clean up leaking radioactive tritium, revealed this month that it also found soil contaminated with strontium-90, an isotope linked to bone cancer and leukemia.

Three days later, officials said a fish caught four miles upstream from the reactor in February had tested positive for strontium-90 in its bones. State officials say they don't believe the contamination came from Vermont Yankee.

Tritium was reported leaking from the plant in January, and since then has turned up in monitoring wells at levels 100 times the federal Environmental Protection Agency's safety limit for that substance in drinking water. Other radioactive isotopes have been found as well, including cesium-137, zinc-65 and cobalt-60.

Officials have said tritium has been flowing downhill from the plant to the adjacent river, though it is diluted quickly in the fast-flowing stream. Tests on river water have not produced measurable tritium readings. Now the question is whether strontium-90, generally considered a more dangerous isotope than tritium, may also have found its way to the river.

State health officials say Vermont Yankee most likely was not the source of the radioactivity in the fish, a yellow perch. Fish and other living things — including humans — around the world have been absorbing tiny amounts of strontium-90 since the United States, Russia and China tested nuclear weapons in the atmosphere in the 1950s and 1960s. A fresher dose was released by the Chernobyl nuclear disaster in 1986.

"It's clearly consistent with the background levels from Chernobyl and weapons testing that went on until 1965," said Michael Dumond, chief of prevention services, which includes radiological health, for the state of New Hampshire. The river between the states is New Hampshire territory, though Dumond said New Hampshire has largely deferred to Vermont on testing samples from it.

Does that mean strontium-90 is present in fish caught around the world?

"Yes. It's everywhere," said John Till, president of South Carolina-based Risk Assessment Corp. and a consultant for more than three decades in testing for radioactive substances in the environment.

Till said he supports nuclear power but faults the industry for a lack of speed and candor in discussing its risks.

Should people limit fish consumption because strontium and other radioactive substances can collect in their tissue?

"Absolutely not," Till said, adding that the amounts are too tiny to be a concern. (Some states, including Vermont, have urged limits on fish consumption — especially by children and pregnant women — because of mercury contamination.)

The International Agency for Research on Cancer has determined that radioactive strontium is a human carcinogen, but the arm of the U.S. Centers for Disease Control and Prevention that tracks toxic substances says exposures must be at high levels before the risk of cancer is elevated.

David Deen, a Vermont state legislator, Connecticut River Watershed Council river steward and fishing guide, is not mollified.

"As a guide, I'll tell you when the fish you're angling for are identified as having strontium-90 in them, it doesn't do much for the image of pristine fishing," said Deen, chairman of the House Fish and Wildlife Committee.

Some people think Vermont Yankee should not be let off the hook any more easily than was the fish that ended up in a Tennessee lab and tested positive for strontium-90.

William Irwin, radiological health chief for the state of Vermont, acknowledged that it was impossible to establish a baseline for strontium-90 in Connecticut River fish, because the state had not tested for it before this year. For that reason, it can't be determined for certain whether Vermont Yankee has been adding strontium-90 to the river

Irwin said the 59 picocuries per kilogram found in the perch's bones was actually at the low end of measurements taken from fish caught even much farther from nuclear plants.

Still, Irwin's comments troubled Helen Caldicott, a pediatrician by training, an internationally known critic of nuclear power and author most recently of a book debunking nuclear as a solution to global warming, "Nuclear Power is Not the Answer."

"What is the baseline level in fish from (bomb-testing and Chernobyl) fallout?" Caldicott asked in a phone interview from her home country of Australia. "What he's saying is fallacious. He doesn't have a baseline level, so to say it's the same as baseline level is not true."

Irwin said there was strong evidence that the strontium-90 in the fish was not from Vermont Yankee, but added it is impossible to say for sure.

Vermont Yankee spokesman Larry Smith said the only spot on the reactor site where strontium-90 had been found was in the pit plant technicians had dug looking for the source of the tritium leak, in an alley between two plant buildings.

Irwin said strontium-90 appeared not to have migrated from there. "We did not find it in groundwater," he said. "We did not find it in river water." And it was not found in soil samples taken farther from the site of the Vermont Yankee leak.

Irwin said a study last year by the New York state Department of Environmental Conservation found levels of strontium-90 in Hudson River fish at up to three times the level found in the Connecticut River fish. That study looked at fish samples from much farther from the nearest nuclear plant — 80 to 90 miles upriver from Indian Point — and attributed the results to bomb testing and Chernobyl, Irwin said.

Caldicott was not convinced. "Fish can swim 80 miles," she said. "To say that the strontium-90 didn't come from Indian Point, I would be very suspicious."

Irwin said the fish seemed to have caught the public's imagination. Asked what species the fish was, he said he didn't know, but later said it was "a red herring."

Given his doubts that the strontium in the perch came from Vermont Yankee, Irwin said he considered it more important that he and his staff focus on testing roughly 2,000 samples taken since January closer to the reactor than the four-mile distance at which the fish was found.

"We're sampling groundwater, drinking water, river water, river sediments, soil and fish. Starting next week we'll be collecting vegetation," Irwin said. "We're going to be doing all of these analyses of all these samples for a very long time."

"We're in this for the long haul," he said. "We'll get more samples. We'll get more analytical data, and we'll get a better and better picture of what the truth is."