it Dielectric in an Klectrostatic Field. 



207 



Relation between strains and thickness of dielectric : — 

 Duter found the volume, at the same potential-differences, 

 to increase inversely as the thickness within a probable error 

 of five per cent. Kighi confirms this from his experiment on 

 glass tubes, but Quincke, using both thermometers and tubes, 

 finds the increase to be inversely proportional to the square of 

 the thickness. This is in accordance with Maxwell's formula 

 for the strain in the medium. In support of this law he gives 

 on page 190 a table for changes of volume, and on page o82 



for chanue^ in length. 



Changes are in millionths of the 



original dimensions. An examination of the table shows 

 that errors in this law are very much greater than the sensi- 

 bility of his apparatus warrants, especially in tubes whose 

 thickness varies to a considerable extent. 









Spark-length in mm 







Therm. 



Thick- 



























ness. 



















1. 



2, 



3. 



4. 



0. 



6. 





mm. 















30 



0-142 



2-883 



10-670 











32 



0-203 



1-756 



7-440 



15-920 









18 



0-258 



1-310 



3-960 



8-071 









23 



0-271 



0'980 



3014 



6110 









22 



0-286 



0-739 



2-662 



7 025 



11-770 







21 



0319 



0-604 



1-971 



4-088 



6 65L 







17 



0346 



0742 



2-042 



3'608 



5-889 



8-592 





61 



0407 



0149 



0-736 



] 658 



2950 



4372 



6016 



60 



0591 



0-058 



0-190 



0554 



0744 



1028 



1-524 



Tube. 



Thickness. 





Spark-length in 



mm. 





1. 



1 

 2, 3. 



4. 



5. 





iniii. 











15 



0092 



0-50 



2-26 ! 5-63 



! 13 68 





J9 



0-150 



0-44 



0-97 ' 2-79 



492 



1019 



28 



0160 





1-54 ' 203 



4-96 



7-07 



27 



0-184 





209 







29 



0194 







1-40 



2-59 





26 



0-230 





0-72 





I I 





To account for some of these discrepancies it is important 

 to compare their apparatus with mine. In the first place, as 

 Righi and Quincke used a system of rigid mechanical levers 

 all lateral distortions, which were so prominently in evidence 

 in my experiments, would show in theirs as a change of 

 length alone, resulting from the bending. This would be 



