16 Prol:'. L. T. More on Electrostriction. 



The rise in temperature of the glass is from two causes ; 

 the ohmic current through the glass from armature to arma- 

 ture, and the discharge through the oil and along the surface 

 of the glass to the top and bottom of the apparatus. Of the 

 two the latter is undoubtedly the greater, as the oil is a better 

 conductor. The liquid is always put in violent commotion 

 })y the charge and, at times, th9 discharge vertically is so 

 great as to cause a glow-discharge along the glass from the 

 surface of the oil to the wire-gauze cage on the top of the 

 tube. The deflexions also indicate the same thing as they lag 

 behind the potential, showing that the passage of heat from the 

 oil to the glass requires a small interval of time. 



Although the rise in temperature of the glass itself cannot 

 be measured, it can for the oil. For this purpose one junction 

 of a platinum-iron thermoelectric element was inserted in 

 the oil and the other in ice. The galvanometer was first 

 calibrated, and gave a permanent deflexion of 2 centimetres 

 for an increase of 6°. The tube was then charged ten times 

 to a potential giving a 12 millimetre spark, when the deflexion 

 was 0*3 centimetre, an increase of 0°'9. On another day, 

 the same number of similar charges indicated a rise of 0°'4. 

 This is much greater than is required to expand the glass by 

 the observed amonnt, and will account for the continued 

 increase in the length of the tube after it has been charged 

 and discharged a number of times. 



J^on-adlierent Armatures. 



To experiment with non-adherent armatures, a new glass 

 tube, number 3, was cemented in place and coaxially with it 

 a larger glass tube, 4*9 centimetres inner diameter, which had 

 been silvered on its inner surface a length of 50 centimetres. 

 The sodium carbonate solution was not used, but, instead, 

 both spaces between the three tubes were filled with lard oil. 

 Electrical connexion was made with the silvered surface by 

 a strip of tinfoil, carried through a hole in the outer tube, 

 vv^hich was then closed by a rubber stopper. The other 

 arrangements were as before. Both mirrors were used, but 

 deflexions have been reduced so that one division equals 

 6 X 10"^ mm. as before. 



The tube was now charged to a potential producing a 12 

 millimetre spark, with one jar in the circuit. The image 

 moved to the left, entirely out of the field of the microscope, 

 and this motion was accompanied by a large apparent 

 elongation of the tube, too irregular and jerky to measure. 



No care that I could employ to make the three tubes 

 rigid and parallel would overcome this difliculty until the 



