The new estimate of 9-16 kt with a 50 per cent error was published today in the international journal Science

, and is contrary to Indian claims that the yield from the May 11 tests was 58 kt.

The US scientists say two sub-kiloton tests announced by India on May 13 were not detected by two stations of their network which recorded the May 11 explosions.

Since the May 11 explosions were well recorded at both sites, the seismic network would have detected any nuclear test on May 13, however small, if it was successful.

No signal consistent with a test was found for as much as six hours before and after the announced time in the Hindukush region.

Based on this and analysis of noise levels over a wide range of frequencies, the team has concluded that seismic magnitude of the May 13 event was ``at least 500 times smaller than theMay 11 event''.

The upper limit of the yield of the May 13 test was 30 ton if it was performed under conditions similar to May 11 explosions. If performed in a sand dune, the upper limit could be as large as 300 ton. According to Indian officials, yield from May 13 tests was about 800 ton.

The US scientists said, a blast larger than 30 ton would have been detected by their stations but one ten times larger could have escaped detection if detonated in sand, as reported.

The results are based on analysis of signals recorded at two seismological stations near India and Pakistan, that are part of a global network of 321 stations to record even small underground nuclear tests.

The new estimate for the May 11 blasts has been calculated by a team of 19 academic and US government seismologists.

Accounting for errors associated with ``emplacement conditions and near source effects'', and scaling the analysis to the background noise levels at the stations, the scientists have estimated that the upper limit ofthe May 13 test by India is only 30 ton, if it was performed under conditions performed under conditions similar to May 11 tests.