JBarss — Ionization by Collision in Gases and Vapors. 233 



through a potentiometer by means of a brass on brass contact 

 key. The plate I being raised to the desired potential, the 

 plate G was insulated and the deflection of the electrometer 

 needle during a certain time was measured. By means of the 

 potentiometer the volt sensitiveness of the electrometer at any 

 time could be determined. Four or five trials were usually 

 made for each observation, the separate trials seldom varying 

 by more than three per cent from the mean. 



Two preliminary experiments were made before the uranium 

 oxide was put in the ionization chamber. The first was to test 

 the electrical capacity of the apparatus (i. e., the parallel plate 

 chamber and the electrometer system). To do this the lower 

 plate was raised to a certain known potential which was kept 

 constant. Electrometer deflections caused by the induced 

 potential on the upper plate were measured for different dis- 

 tances between the plates. It was found that these deflections 

 varied inversely as the distances involved, over the range of 

 distances used in the later experiments (0*5 cm to 2"5 cm ). This 

 means that the capacity of the whole system is approximately 

 constant. Since this is true, the ionization current will vary 

 directly as the rate of electrometer deflection, so that the latter 

 may be taken as a measure of the former. 



The second preliminary experiment was to test for any ioni- 

 zation that might be due to the potential gradient between the 

 plates. The vessel was evacuated to a pressure of one milli- 

 meter and the lower plate raised to high potentials. It was 

 found that the electrometer showed no variation until the 

 sparking potential was reached, for all distances down to 0'5 cm 

 between the plates ; so that there was no ionization due to the 

 electric field alone, before a discharge occurred. 



Experimental Results. 



Gases. 



The first experimental results were obtained in air at pres- 

 sures of 4-l7 mm and 2-09 mm . Three sets of readings were taken 

 for each pressure corresponding to three different distances 

 between the plates. Curves showing the relation between field 

 strength (volts per centimeter) and ionization current are given 

 in fig. 2 for a pressure of 4*17 mm . Values of a were calculated 

 from Townsend's formula, page 230, and are given in Tables I 

 and II. The general shape of the curves as well as the values 

 of a agree very well with Townsend's results. In Tables I and 

 II, X is the field strength, L the distance between the plates. 

 The figures in parentheses are taken from Townsend's results,* 

 *Phil. Mag., vol. i, p. 208, 1901. 



