﻿691 Prof. R.W.Wood on Selectice Reflexion, Scattering 



the earlier experiments T excited the lamp with a high 

 potential current from a transformer, running it at a com- 

 paratively low temperature. It immediately occurred to me 

 that in the present case I was working with the lamp at 

 high temperature, and that the 2536 line was very likely 

 reversed, the wave-length necessary for the excitation of 

 the resonance radiation being removed by absorption. I 

 accordingly allowed the lamp to become quite cold, and 

 made my exposure when the lamp was first lighted, when its 

 light is quite violet in colour. On developing the plate 

 1 found that a five seconds exposure gave me a more intense 

 cone of light than anything that I had ever observed beiore. 

 It was absolutely black on the negative. Moreover, the 

 vapour outside of the cone of focussed rays appeared to be 

 glowing at the end of the tube where the beam passed in. 

 As the fused quartz is somewhat fluorescent under the action 

 of the rays emitted by the lamp, it was necessary to prove 

 that this light did not come from the walls of the tube. 

 That such is not the case I shall show when we come to 

 subsequent experiments. 



A photograph of the resonance radiation in the tube is 

 reproduced on PL XI. fig. 1. 



It was next necessary to observe the spectrum of the light 

 emitted by the mercury vapour, and the slit of a quartz 

 spectrograph was opened wide and brought up close to the 

 silica tube. Two photographs were made — the first of ten 

 seconds exposure immediately on starting the Westinghouse 

 lamp ; the second, also of ten seconds after the lamp had been 

 running half a minute. These photographs are shown in 

 PI. XI. fig. 2. In the second spectrum the broadened image 

 of the slit is shown at the 2536 line (indicated), the other 

 lines appearing narrow. These lines result from the circum- 

 stance that some direct light from the arc illuminated one of 

 the jaws of the slit. The upper spectrum, taken after the 

 lamp had been in action for half a minute, is identical with 

 the second, except that the broad image of the slit is absent, 

 showing that the resonance radiation has disappeared. I 

 then made a series of four spectra taken one after the other. 

 These showed that the resonance radiation was very strong 

 during the first five seconds of the lamp's operation, quite 

 faint during the second five seconds operation, the merest 

 trace during the third five seconds, and completely gone in 

 the next five seconds. In all of the subsequent work I 

 accordingly allowed the lamp to become quite cold, and 

 extinguished it after a run of exactly five seconds. This 

 insured uniformity in the exposures. The slit of the spectro- 



