John Michell 87 



strengthened by a copper wire which acted as a tie to pre- 

 vent bending, and was attached to a vertical suspension. 



On being slightly displaced from its position of equili- 

 brium the torsion of the wire by which it was suspended 

 would tend to bring the horizontal beam back and make 

 it oscillate slowly in a horizontal plane. Two larger, leaden 

 spheres eight inches in diameter could be brought near the 

 ends of the beam, so that their gravitational attraction 

 on the spheres attached to the beam would displace it, 

 with the result that it would oscillate about the new posi- 

 tion of equilibrium. By bringing the larger spheres round 

 to the other side of the beam the displacement in the 

 opposite direction could be observed and the gravitational 

 effect measured. Cavendish fully realized the difficulties 

 he would have to encounter in consequence of almost 

 unavoidable air currents. Even when the apparatus was 

 enclosed in a box the slightest difference in temperature 

 would cause convection currents and, consequently, irre- 

 gular movements of the beam. He, therefore, had to plan 

 out a scheme which would allow him to conduct the whole 

 of the experiments without entering the room in which 

 the apparatus was placed. The observations were taken, 

 and the large leaden spheres moved one side of the beam 

 to the other from outside. No more delicate measurement 

 had ever been successfully carried out. From the average 

 of the number of observations, Cavendish deduced the 

 value of 5*48 for the density of the earth, a number in fair 

 agreement with, though slightly larger than, that obtained 

 by Maskelyne and Hutton. The extreme difficulty and 

 great charm of the experiment has still in our times 

 attracted the most skilled physicists, and the introduction 

 of quartz fibres by Mr. Vernon Boys has enabled us to 

 increase its accuracy considerably. The final value for the 

 average density of the earth as determined by Mr. Boys 

 is 5*5270, so that Cavendish was correct to within one per 

 cent. 



John Michell (1724-1793), whose name has been mentioned 

 above as the inventor of that most useful and delicate 

 appliance, the torsion balance, has also in other directions 

 given evidence of great originality of mind. He contributed 



