418 Messrs. J. J. Thomson and H. F, Newall. [June 16 



Table VII. 



Experiment 39. Paraffin Oil. ? 1 inch up straight. Short Charge. 





T)pfl pp4,i nn<? Tva.tvinsi 



Deflections. Ratios. 



380 X ' 184 

 325 ] -' 169 

 280 1 * 161 

 o?n 1 '166 

 240 - <170 



176 



130 } }£ 

 -..o. 1-150 



q« 1-153 



y8 1-166 



2* 1-150 



CA 1 'I 40 



* 1-172 



Fia. 4. 



Curve plotted for paraffin from Table V. Same scales as in fig. 3. 



Experiments on Carbon Bisulphide. 



When carbon disulphide was used as the leaking dielectric, con- 

 siderable discrepancies appeared ; the ratio curve fell or remained 

 steady in what seemed a capricious manner. Filtering seemed to make 

 no difference, although again and again repeated ; the condenser 

 cylinders were cleaned carefully ; the liquid was distilled several times, 

 and the discrepancies only seemed to be exaggerated, especially in 

 experiments made very soon after distillation. The still was cleaned 

 and again used, but to no effect. The irregularities still remained, the 

 ratio curve in some instances being as steadily horizontal as in the 

 case of benzene and paraffin oil ; at other times the values of the 

 ratios falling as much as 12 or 15 per cent. Fig. 5 gives Table VIII 

 graphically, showing the great fall of the ratio curve. The changes 

 in the zero of the electrometer were such as to lead us to the notion 

 that we had to deal with something analogous to electric absorption. 

 We therefore tried the effect of varying the time of charging the 



