An experiment was performed to test Newton's inverse square of law gravitation. The test compared accurately measured gravity values along a 600-m tower with upward continued gravity estimates calculated from ground measurements using algorithms based on Newtonian potential theory. Gravity measurements with an internal consistency of better than 20 μGal (1 μGal=10-8 m/s2) were made on the tower at intervals of about 90 m using a single LaCoste-Romberg model G gravimeter. Field measurements of gravity (using the same instrument) and corresponding geodetic positions were performed at 77 points within a radius of 5 km from the tower. These and other gravimetric data within a radius of about 220 km were used to perform the upward continuation using two independent computational techniques. One is least squares collocation, which is based on a statistical interpretation of the local gravity field, and the other determines a local harmonic expansion on the field from the given data.

The upward continuation from both methods gave almost identical results and is estimated to have an uncertainty comparable to the in situ measurements. The results of the experiment show a substantial departure from the inverse square law, asymptotically approaching about 500 μGal at the top of the tower, with a standard deviation of better than 40 μGal. This indicates that there is a non-Newtonian attractive force which falls off rapidly with elevation. ¿ American Geophysical Union 1989