of Metallic Spectra in the Ultra- Violet. 349 



Effect of Change of Temperature of Source of Light on 

 Constancy of Position of Metallic Lines. 



In the progress of the investigation we were much troubled 

 by a slight shifting in position of the metallic lines upon the 

 photographs. This shifting could be observed when the 

 metallic lines were compared with a solar spectrum taken 

 upon the same plate. The amount of this shifting in no case 

 amounted to more than '1 or - 2 of a wave-length. At first we 

 thought it might be possible that there was a change in 

 lvfnmgibility of the metallic lines due to a difference in 

 temperature of the source of light, and a long study was 

 made of the influence of the temperature of the source of light 

 upon its wave-length. When a metal was burned in the 

 carbon electric light with varying strength of current, no dis- 

 placement could be observed between the lines of the metal 

 photographed beneath each other upon the same sensitive 

 plate. When the electric spark with a large battery of 

 Leyden jars was substituted for the electric arc, and the 

 metallic lines obtained by the light of the spark were com- 

 pared with those from the arc, occasionally a small displace- 

 ment could be observed. This did not seem to arise from a 

 change of position of the source of light, or from the heating 

 of the slit of the spectroscope. A careful study of the iron 

 lines showed us that the wave-lengths of the iron lines in the 

 sun and those obtained from burnino- iron in the electric arc 

 were the same to certainly one hundredth of a wave-length. 

 The displacement we observed was noticed only when the 

 electric spark was employed. This shifting did not arise 

 from a change of position of the spark in our apparatus, for 

 it could not be produced at will by changing the position of 

 the source of light. Moreover, when the arc-light was placed 

 in the same position that the spark occupied, no displacement 

 could be observed in photographs taken by the aid of the arc. 

 We were forced to conclude that through the range of 

 temperature afforded by the electric arc and the electric 

 spark the wave-lengths of the metallic lines were constant. 

 The displacement we observed was therefore referred to a 

 jarring of the apparatus due to the noise of the electric spark. 

 When the camera was at a considerable distance from the slit 

 of the spectroscope, the displacement was diminished and 

 sometimes entirely disappeared. The entire apparatus was 

 vry solid, and the camera was clamped to a massive girder. 

 It was difficult, therefore, to believe that the displacement 

 could arise from the noise of the spark. We believe, how- 

 ever, that it can be ascribed to this cause, and that the 



