Prof. L. T. More on Electrostriction. 5 



the armatures ; but that it is independent o£ the nature o£ the 

 non-conducting- fluid. 



These experiments are adduced by Ercolini and Sacerdote 

 as proving that the coefficient ki is positive. That is, the 

 dielectric constant increases with a tractile force acting per- 

 pendicularly to the field of force. In a note it is stated that 

 the elongation was too small to be even noted by my apparatus. 

 Possibly that is dismissing the matter too curtly unless the 

 estimate of the sensibility of my apparatus made by Dr. Sa- 

 cerdote is the correct one . In my own opinion, formed after 

 long experience with it, this sensibility is at least 3^ times as 

 great as he assumes it to be ; and I thought such was e\ddent 

 from my explanation published after his first criticism. Even 

 if this criticism were valid in respect to my former apparatus, 

 it certainly is not to my present one. The only apparent 

 advantage of Ercolini's ar ran clement is that he used thinner 

 tubes ; but by doing so he increased the effect of the irregular 

 movements he describes. Also it must be remembered that I 

 used potentials great enough to rupture the tube, and a more 

 intense field could hardly have been obtained. Yet, as will 

 be shown later, no elongation was observed with non-adherent 

 armatures when lateral effects were eliminated. The con- 

 clusions of Ercolini''s work have been questioned by Oorbino, 

 and, I think, rightly. 



The most recent work in this field is by Dr. Shearer"^. 

 The writer experimented on condensers of glass and hard 

 rubber with adherent armatures. The method of measuring 

 the effect was by the displacement of interference-bands. His 

 results on rubber and glass may be sunmiarized as follows, — 

 the supposed elongation being for these substances of the 

 order 10~^ and 10~^^ per unit dimensions respectively. 



Rubber plate condenser 40 cms. long and 3 mm. thick. 

 Difference of potential sufficient to puncture glass 3 mm. 

 thick. A displacement of approximately one-fourth of a band 

 was observed, increasing gradually, and returning to the 

 original position after each discharge. This effect disappeared 

 after the first few charges when there was no spark-gap in 

 the circuit. A very slight twist of the hard rubber, by reason 

 of non-uniformity of the field, might easily cause this effect. 

 When a spark-gap was inserted a considerable displacement 

 of the bands, increasing greatly with the time, occurred. On 

 looking between the condenser-plates the whole region ap- 

 peared filled with a glow-discharge. It may be inferred that 

 this motion was due to a o-radual heatino- of the hard rubber. 



* Shearer, Physical Review, vol. xiv. p. 89 (1902). 



