Short Spark-Discharges. b*9o 



form V=K+/(r I. where d is the spark-length and A is 



the mean free path, and/' a function which is the same for all 

 gases/ 1 and " K a constant which may vary from one gas to 

 another/' 



While the distance is comparatively long, u e. greater than 

 o x 10~ 4 cm., it seems probable that the carrier of the elec- 

 tricity remains the same even though the spark distance is 

 considerably less than the mean free path of the molecules. 

 It follows, however, from Earhart's* experiments, as well 

 as from the results obtained by G. M. Hobbs f , that when 

 the distance is less than 3 x 10 -4 cm. there is an abrupt diver- 

 gence from the previous relation between spark-potential and 

 length. This probably results from a change in the carrier of 

 the electricity. Experiments with coherers % and other loose 

 contacts have shown that when the break in the metallic 

 circuit is small, only a small difference of potential is needed 

 to cause a discharge, followed usually by a coherence between 

 the terminals. The difference of potential necessary to cause 

 coherence was found to be unaffected by a change in the 

 pressure of the gas. If between and 3x 10~ 4 cm. there is 

 no other abrupt change in the relation between spark-potential 

 and distance; it is probable that there is no other change in 

 the carrier of the electrical discharge. 



Some preliminary experiments made by me during the 

 winter of 1904 to determine the relation between pocential- 

 difference and spark-lengths less than 3 x 10 -4 cm. were not 

 decisive, and so a special instrument was designed and its 

 construction begun in April. Shaw§ soon after published a 

 •• preliminary report," to which further reference will be 

 made upon the same subject; but the question seemed of 

 sufficient importance to warrant a prosecution of the work 

 upon which I was engaged. 



Tin- apparatus (a horizontal section of which is shown in 

 tig. 1 ) was so designed as to have the greatest possible 

 rigidity, since measurements of distances as short as 1/u-//,, 

 /'. r. 1<»- 7 centim.. were desired. The discharge surfaces were 

 iridio-platinum face- soldered on to the bra— rodsPand N. 

 These were truncated cones carefully ground to fit their 



* Earhart, Phil. Mag. [6] vol. i. p. 147 (1901). 



An elaborate investigation carried on in the Ryerson Laboratory 

 between 1902 and 1906, but as yet unpublished, in which Hobbs has 

 extended the work of Earhart. 



X Kinsley, Phys. Rev. vol. xii. p. 177 (1901); Eccles, Electrician, 

 vol. xhii. p. 882(1901). 



§ Shaw, Proc. Roy. Soc. .May 2-. 1904. 



