[ 277 ] 



XVII. On Spark Potentials in Liquid Dielectrics. 

 ^ Dr. P. E. Shaw* 



IN June of last year I read a paper before the Physical 

 Society of London on the above subject ; it was pub- 

 lished in the Phil. Mag. for October. Since then Mr. R. F. 

 Earhart has produced a paper (Phys. Rev. xxiii. pp. 358-369, 

 Nov. 1906) which covers to some extent the same ground. 

 As his results are somewhat at variance with mine it seems 

 of interest to briefly compare and contrast them. 



The conditions are different in several ways : — (1) He 

 uses a 110-volt 60-cycle circuit stepped up by a transformer 

 to 2200 volts ; I used a 400-volt steady current, a town- 

 lighting circuit. (2) He works up to 1000- volts; I used 

 up to only 400 volts. (3) He uses a plane surface and a 

 spherical one of 25*4 mm. diam., whereas I used a plane and 

 a small sphere of 2 mm. radius. (4) His surfaces are steel, 

 mine of iridio-platinum. (5) He immerses the whole of the 

 contact surfaces in the oil used ; 1 used a drop between the 

 two surfaces, renewed for each experiment. (6) The smallest 

 discharge distance in his work is 3 microns, w T hereas I used 

 distances as small as 0*25 micron. (7) His work is confined 

 to four oils, whereas I worked on fifteen oils and other liquids. 



Seeing how the conditions differ in his apparatus and mine, 

 it is scarcely strange that the results also should differ. To 

 contrast the results : — (1) He finds the potential gradient 

 for small distances greater than for large ones, there being a 

 " knee " for potential 330. In my curves, half of the liquids 

 show such variation in strength, as potential varies, that for 

 them it seems impossible to state a potiential gradient. 

 (2) The potential gradient found by him is in all cases much 

 less than in my experiments, and the curves for liquids do not 

 appear to cross the air-line, as is the case in all my experi- 

 ments. (3) He finds, as I did, that for small distances air 

 insulates better than liquids. (4) The most striking result 

 of his research is the indication of the " knee " in the curves. 

 He finds this to be at 330 volts, which is the same as for 

 sparks in air. 



In view of this last result I have examined my results to 

 see if there is any indication of a " knee " in my curves also. 

 As the accompanying figure shows, there seems to be a " knee " 

 in some cases. It is pronounced for pentane, linseed oil, fusel 

 oil, and rape oil. It is slight for heptane, castor oil, and 

 turpentine ; whereas for olive oil, codliver oil, hexane, and 

 octane there is no clear indication of a change of the kind. 



* Communicated bv the Author. 



