442 Dr. G. Gore on the Influence of Proximity 



rotation upon a vertical axis at its middle part, so that either 

 end of the series of tubes might be presented to the centre 

 of the near face of the cube without being shaken. By 

 omitting one of the pieces of lead a horizontal groove was 

 left in that face of the cube, in order to increase the expected 

 effect by permitting the ends of the tubes to penetrate about 

 an inch within the mass. The electrodes were formed of 

 zinc wire "072 inch diameter, all cut from the same piece, 

 not amalgamated, and each fixed in a paraffined cork and 

 projecting about f of an inch into the tube ; they were all 

 connected in series by means of silk-covered thin copper 

 wires. 



By means of this arrangement, whilst the previous experi- 

 ments with the vertical tabes referred to above (toe. cit.) included 

 the effect of considerable difference of pressure, and almost 

 entirely excluded the influence of difference of terrestrial 

 gravitative attraction, those with the horizontal ones included 

 the effect of difference of gravitative attraction of the lead, 

 and completely eliminated that of difference of pressure* 

 caused by difference of attraction of the earth. 



The electrolyte employed consisted of 38 ounces of 

 thoroughly boiled distilled water and 75 grains of potassium 

 chloride, the mixture being subsequently about j 1 ^ saturated 

 with washed chlorine gas. This liquid was used in all the 

 following experiments with this apparatus, and was at all 

 times carefully screened from the light. The use of the 

 chlorine was to prevent the liberation of hydrogen. 



After excluding or rendering negliglible nearly all the 

 numerous sources of error (which need not be here specified), 

 the only remaining disturbances were due to unequal and 

 variable voltaic action at the electrodes, changes of tempera- 

 ture, periodical movements of the tubes, and slightly to 

 variations of terrestrial magnetism. These influences, how- 

 ever, were sufficiently small to permit the detection of a very 

 minute amount of deflexion of the galvanometer needles due 

 to the presence of the lead, provided a proper method of 

 averaging the magnitudes of the deflexions was employed. 



The usual method of manipulation adopted with this appa- 

 ratus was to place the tubes in a line with the centre of the 

 near face of the cube, close the circuit, allow any voltaic 

 current to subside or become steady (this was often a tedious 

 matter), take periodical observations ever}- hour or less, and 

 reverse the ends of the tubes once a day or more frequently. 

 All these experiments were made in a cool room free from 

 draughts, and the temperature of the cube and of the air of 

 the room were recorded regularly with the deflexions. 



These experiments were continued w r ith but little inter- 



