Fiske and Collins — Surface Travel on Electrolytes. 61 



4. 



In order to give uniformity to a set of results it was found 

 necessary to scrape carefully the cadmium terminals of the 

 spark gap before taking each photograph. The distance 

 between the terminals was kept constantly at l*5 mm . 



The average number of half oscillations of all the isolated 

 sparks on any neg itive was taken as the representative num- 

 ber. The pilot spark was not counted. The oscillations for 

 Plate III were counted by Mr. Collins, those for the other 

 plates by Mr. Fiske. Plates III and IV were made from the 

 same set of observations. 



The free acid in the solution kept the electrodes constantly 

 clean. 



The wires themselves show no surface travel. This is indi- 

 cated by Rayleigh's formula 



W= R(?rr j/^f) (Phil. Mag. (5) xxi, p. 369, 1886). 



or by Stefan's 

 R'=r/ 



/ Iag 



nr a, ~ + x) (Wied. Ann., xli, p. 421, 189<»). 



Here fi = 1 ; r, the radius, = *0165 cm ; <r, the specific conduc- 

 tivity, = T - 5 ±^; T, the time of oscillation, = -annrWiF sec. 

 approximately Neither of these formulae holds because the 

 factor in the parenthesis is in each case less than 1. 



That the wires exhibit no surface travel was also shown 

 experimentally. Wires of different sizes and materials but of 

 approximately the same resistance, laid parallel to avoid self- 

 induction, were placed in circuit, the sparks photographed, 

 and the number of half-oscillations compared. The following 

 table gives the result 











No 



. of half- 





T o. 



Material. 



Radius. 



Resistance. 



oscillatiODS. 



Remarks. 



1 



German silver 



0"36 mm 



8-07 ohms 





5 





2 



Manganine 



16 



8 "()7 





3 55 



Doubtful 



3 



Iron 



0-127 



7-75 





1-66 



a 



4 



German silver 



0-128 



8-17 





4-73 





5 



a u 



0-16 



8-07 





4-8 





6 



Like No. 3 









1-5 





1 



" " 2 









5-17 





