Prof. L. T. More on Electrostrlction. 3 



has published many experiments on the behaviour o£ glass 

 condensers when charged. For the purpose he used cylin- 

 drical tubes about 70 centimetres long and 7 millimetres in dia- 

 meter, ranging from 0*045 to 0*2 centimetre in thickness. The 

 potentials to charge them occasionally were great enough to 

 give a 10 millimetre spark. The armatures of the condensers 

 were adherent to the glass, and the elongations were measured 

 by the displacement of interference-bands. The greatest dis- 

 placement noted is 2*6 bands, but the great majority were 

 less than one band, and they were measured to one-hundredth. 

 The elongations, reduced to unit dimensions, average 



¥4. =6X10- 



This result is calculated from the deflexions observed when 

 the tubes were charged ; the return, when the tubes were 

 discharged, sometimes equalled the above, and was sometimes 

 less than half. The time of charging also affected the amount 

 of the elongation; but no law could be formulated, and no 

 reasons are given for the time of charging determined on to 

 give the correct value. 



i In spite of the carefulness which Cantone's experiments 

 show, there are evidences that the values given are, at least, 

 partly due to causes other than the one he ascribes them to. 



In the first place, the amount of the elongation increases 

 with the time of charging and to an unknown extent. 



Secondly, they are partly due to heating of the tube. 

 Cantone himself attributes the erratic return of the tubes to 

 the heat evolved ; and this varies between a complete return 

 and a value of only one-half. If so, surely part of the elon- 

 gation in charging must be due to the same cause. 



Thirdly, the armatures touch the glass, and their attraction 

 will elongate the tube, using Poisson^s ratio, approximately 

 one-third the observed amount. 



Fourthly, tubes of the dimensions given are flexible, rarely 

 straight, and do not have a uniform wall-thickness. All of 

 these defects tend to strain the tube when it is charged, as the 

 field is not uniform. The amount of this strain is large, and 

 even with unusually accurate tubes it is difficult to avoid an 

 apparent elongation greater than the supposedly true effect. 



Unless these causes of possible elongation are eliminated, 

 the elongation observed is subject to criticism. 



In answer to a statement made in my paper, that the true 

 effect could be obtained only when free charges were not 

 placed on the dielectric ; or, in other words, when the arma- 

 tures of the condenser were separated from the dielectric and 



B2 



