The Foucault pendulum, invented in 1851 by French astronomer Jean Bernard Leon Foucault, was the first instrument that could demonstrate Earth's rotation without reference to the stars. The swing direction of the long pendulum remains constant as Earth rotates underneath it, which means its path appears to move, traveling a full circle every 24 hours at the poles and taking longer closer to the equator (32 hours in Paris, for example).

Maurice Allais, an amateur astronomer and 1988 economics Nobelist, claimed to find what he called "remarkable anomalies" in a Foucault pendulum's swing at his Paris laboratory during the total solar eclipses of 1954 and 1959. During one eclipse, he measured a slight deviation --an extra 0.15 degree--in movement of the plane under the pendulum's swing, indicating the pendulum was speeding up slightly. This implied a tiny (3 millionths of 1g) increase in Earth's gravity field. His published report, "Should the Laws of Gravitation Be Reconsidered?" lay in obscurity until recently, when David Noever of NASA's Marshall Space Flight Center in Huntsville, Alabama, was rummaging through the Web for information relating to his work on gravity.

Now Noever and colleague Ron Koczor plan to use a state-of-the-art gravity sensor to test Allais's observation during the upcoming eclipse. The NASA team will compare their results with a similar test by Edcon Inc., a gravitometer manufacturer in Denver, as well as observations at Foucault pendulums in Europe that lie in the path of the eclipse.

The duo doubts they will find the eclipse anomaly. Still, says Noever, "Allais could have stumbled onto something important." Possible explanations are highly speculative, ranging from quantum fluctuations in the vacuum of space to radiation pressure changes from the blockage of sunlight.