52 Messrs. Boys, Briscoe, and Watson on the 



very great delicacy. To make the resonators aluminium wire 

 was rolled into ribbon of *2 x '001 centim., so that each centi- 

 metre in length weighed only half a milligram. Two pieces 

 30 centim. long were fastened in the cranked form previously 

 described on very light frames made by cementing to almost 

 capillary thin glass tubes cross bars cut from microscope cover- 

 glass, so that the ends of these cross bars held the corners of 

 the cranked pieces of aluminium ribbon. One of these was 

 suspended in the same apparatus by a fine quartz fibre, and 

 all was arranged with the view of obtaining a delicacy 

 comparable with that employed by one of us in the Cavendish 

 experiment. It was now impossible to keep the spot of light 

 at rest, owing mainly to air-currents in the jar set up by 

 differences in temperature of the sides of the jar which could 

 not be avoided in the schools, and also to the perpetual 

 tremor. In the winter, however, the elementary laboratory 

 on the other side of the Exhibition Koad was not being used, 

 and so everything was transferred to a small room there, 

 where a fair degree of quiet and uniformity of temperature 

 could be obtained. The Yoss machine was replaced by a 

 large induction-coil, which gave a greater torrent of sparks. 

 Then, to get more energy in each spark, the primary oscillator, 

 made of 12 millim. tube, was replaced by a pair of spheres 

 100 millim. in diameter, with small spheres on short sliding 

 rods, between which the sparks passed. The scale was placed 

 at a distance of 281 centim. from the mirror. Then, when the 

 wet duster was placed round the jar and the sparks made to 

 pass, deflexions of 5 to 10 millim. were observed. Fearing, 

 however, that even the wet duster failed to completely screen 

 the induction-effects due to the rise and fall of potential of 

 the primary or rather some of the higher harmonics already al- 

 luded to, we soaked it in dilute sulphuric acid instead of water, 

 so that no doubt should exist as to its sufficient conductivity to 

 quench the sham effect while it still remained unable to stop 

 more than a small proportion of the waves of 500,000,000 a 

 second. With this precaution all definite movement of the 

 spot of light ceased. It was seldom that air-currents did not 

 cause a gentle hovering of the index a few millimetres 

 either way, which, however, had no relation to the starting or 

 stopping of the sparks ; on one or two frosty days, however, 

 when the temperature was very uniform, there being no fire, the 

 spot of light was so steady that a movement of even j 1 q millim. 

 on the scale would have been observed, but we have based 

 our calculation on the certainty that the deflexion was not 

 £ millim. 



In order to make sure that this was a true negative result 

 the duster was removed, when at once the index was driven 



