NOTE: This article is also posted on the Internet under the title "Japan 'Loses' 206kg of Plutonium." - RBL
Japan on Tuesday admitted that 206kg of its plutonium - enough to make about 25 nuclear bombs - is unaccounted for.
Government scientists said that 6,890kg of plutonium had been extracted since 1977 from spent nuclear fuel at a processing plant about 120km north east of Tokyo. But that is 3 per cent short of the amount the plant was estimated to have produced.
About 5kg to 8kg of plutonium are needed to make a 20-kiloton atomic bomb similar to the one that destroyed Nagasaki in 1945.
Experts said the missing amount was surprisingly large.
There is normally a margin of error of 1 per cent or less when measuring liquid plutonium, which can dissolve into other elements.
Japan's admission comes at a time of acute sensitivity because of the threat of nuclear proliferation in north-east Asia following North Korea's revival of its mothballed nuclear programme.
However, there is no evidence that North Korea was linked to the missing plutonium even though it is known to smuggle goods in and out of Japan.
"This is an unusually large amount of plutonium to be unaccounted for, which makes me uncomfortable, although I think it's highly unlikely that it was stolen," said Tatsujiro Suzuki, senior research scientist at the Central Research Institute of Electric Power Industry.
The science ministry, which reported the discrepancy to the International Atomic Energy Agency (IAEA), dismissed the idea that the plutonium had been stolen. It said about 90kg was probably diluted into waste-water and about 30kg probably dissolved into other elements.
It admitted it was baffled by the remaining 86kg but said initial output projections may have been too high and the plutonium may not have been produced.
Mohamed El-Baradei, director general of the IAEA, said: "The Agency [is] confident that no nuclear material has been diverted from the facility."
The IAEA, the United Nations nuclear watchdog, has urged Japan to strengthen its procedures for measuring nuclear material since it first noted discrepancies in 1998.
Additional reporting by Clive Cookson in London and Andrew Ward in Seoul