The Botvos Gravity Balance in the Science Museum, South Kensington, is described, and the theory of its operation considered. A full account is given of certain preliminary experiments made on the torsion wires, with a view to ascertaining the daily variation of the equilibrium position, due to strains set up in the wires during manufacture, Until this daily shift has been eliminated the instrument cannot be employed successfully, The experiments show that a preliminary treatment of the wires

more »

... desirable, and a test was also made of the " baking " method recommended by Eotvos for ageing the mires. This method was found to give the desired result, and after this treatment the reading of the instrument remained practically constant from day to day. Tables are given showing the variation in the reading of the instrument when its beam is rotated into different azimuths, and the constancy of these readings when repeated. The differences in the readings as the azimuth varies are sufficiently marked to indicate the changes in the local gravitational field of .the laboratory, and further experiments are being carried out to determine these changes in detail, as a preliminary to field tests ; it is hoped to communicate the results of these experiments at a later date. PREVIOUS to the introduction of this instrument, researches on gravity were confined almost exclusively to investigations with the pendulum and the bubble level, but in 1888 a new line was followed by Eotvos, who endeavoured to measure the variation of the force of gravity in the vicinity of a point, or more exactly, to determine the derivatives of its components, while analogous methods were also employed to determine the derivatives of the magnetic force. As the variations of gravity are extremely small in comparison with the total force, EGtvos concluded that the method to be employed should measure the differences of gravity directly, rather than the force of gravity itself. This method, it is claimed, is extremely sensitive, and is capable of giving results which are unobtainable by pendulum methods. According ta Eotvos, it is possible by this method to estimate the desired magnitudes at any place in the course of one night, with an accuracy of 1 x