﻿96 Miss Ellen O'Connor on the Spectrum of 



The magnesium electrodes were held in clamps on an 

 adjustable stand, one electrode being connected to the junction 

 of the inductance and capacity in the oscillatory circuit, the 

 other being earth-connected through a hot wire ammeter. 

 The generator having been started, the electrodes were 

 brought into contact and an arc struck. . The length of the 

 arc which can be maintained with magnesium depends upon 

 the conditions in the oscillatory circuit, and in the following 

 experiments was from 1 to 3 mm. With a frequency of 

 10 3 per sec. the current, as measured by the hot wire 

 ammeter, was - 8 amp. Increasing the frequency causes a 

 decrease in the arc-length and current. With a frequency 

 of 6o x 10 5 per sec. the current falls to 1/10 of its original 

 value and an arc can only just be maintained. 



The light from this arc was concentrated by means of a 

 condenser on to the slit of a Hilger fixed-arm spectrograph. 

 As the prism of this spectrograph was of very dense glass it 

 was not possible to photograph the spectrum below a wave- 

 length of 4000. By making use of panchromatic plates * 

 the spectrum could be obtained between the wave-lengths 

 G570 and 4000. 



PI. V. fig. 2 e shows a typical spectrum of this form of 

 high-frequency arc which, lor the purpose of comparison, 

 was photographed on the same plate as the spectrum of the 

 direct arc (tig. 2 b) and that of a ley den-jar discharge 

 through a large inductance (fig. 2 c), the same electrodes 

 being used in each case. It will be seen that there is a 

 marked difference between the spectrum of the undamped 

 high-frequency arc and that of the direct arc and spark. 

 One interesting feature is the intensity of X 4481. If the 

 relative intensity of this line can be taken as a criterion of 

 the temperature in the arc it would appear that this high- 

 frequency arc was intermediate to the ordinary arc and 

 spark |. It is more probable, however, that the low current 

 density is responsible for its intensity. 



The conditions in the oscillary circuit were next varied to 

 see if these differences could be accentuated, and to trace if 

 possible the cause of any changes produced. 

 . A short visual examination of the spectrum showed that 

 variations could easily be obtained. With a frequency of 

 10 5 per sec. the spectrum consists of a few lines on a con- 

 tinuous background, while at the higher frequency of 

 6 5 X 10 5 per sec. it becomes richer in lines, the continuous 

 spectrum has disappeared to give place to a feeble band 

 spectrum (nitrogen). The change in the appearance of the 



* AVratten nnd WaiDwright. 



t Kayser, Asiron. Ndchr. clxii. pp. 277-282 (1903). 



