﻿Ionization by Collision in Helium. 



847 



is necessary in consequence to take some precautions in 

 obtaining the sparking potentials. At the higher pressures 

 the distance from the charged plate to the earth connexion 

 to the ebonite supports may be such that pa' is about 4, and a 

 discharge will pass to the supports before the proper potential 

 is reached for a discharge between the plates. To guard 

 against this in all the above determinations (Tables VIII. and 

 IX.) the supports were put at a potential of 75 volts. 



At the lower pressures the spark would often pass round 

 the back of the upper plate in a curved path, in the same 

 way as in a Hittorf tube, the pa' for the curved path giving 

 a lower value of V than the pa' for the straight path between 

 the plates. This difficulty only occurs for values of pa' 

 considerably less than 4, and rendered it impossible to find 

 the sparking potential for p = l, X = 200, a = 2*37, /3 = *165, 

 for which the calculated pa' is 1*2. 



The comparison of the results so far obtained with those 

 obtained previously with a slightly impure gas are so striking 

 that a set of experiments were made to find the effect upon 

 a and /? of adding known amounts of impurity. The leak of 

 the bell-jar per day afforded a convenient method of adding 

 known amounts of air to the helium, and the following results 

 were obtained for the currents between the plates for different 

 distances, the current for distance d-='2 being always taken 

 as 10. 



Table X. 

 Helium pressure 5 mm. X = 200. 



d 



•2 



•4 -6 1 -7 



05 

 00 



Pure Helium 



10 

 !0 



10 

 10 



31-3 

 28-6 

 29-1 

 29-4 



119 

 97 

 90-5 

 92-7 



372 



228 

 194 

 172 



790 

 612 

 408 



2410 



Helium with 1 in 400 air... 



„ 2 in 400 



„ 3 in 400 



The values of a and ft were calculated from these figures 

 and are given below, it not being thought necessary to include 

 for comparison the currents calculated from the values, 

 although the calculation was actually performed and the 

 agreement found as good as those given in tables I. to V. 



The sparking distances a' and the corresponding sparking 

 potentials V are also given. 



