﻿710 Prof. R. W.Wood on Selective Reflexion, Scattering 



a drop of mercury, exhausted and sealed, was mounted just 

 above a vertical chimney of sheet iron which served to 

 immerse the bulb in the hot gases rising from a large 

 Bunsen burner. A high temperature quartz thermometer 

 was mounted with its bulb in contact with the quartz bulb, 

 and by regulating the height of the flame of the burner, the 

 bulb could be kept at any desired temperature. A divergent 

 beam of the 2536 light was thrown upon the bulb by the 

 quartz spectrograph, and the bulb photographed with the 

 quartz camera. At room temperature the entire interior of 

 the bulb was shown by the photograph to be filled with a 

 uniform glow. As the temperature rose this glow was confined 

 to the wall which received the direct rays from the spectro- 

 graph, owing to the failure of the radiation to penetrate to 

 the interior. As the density of the vapour increased this 

 glow appeared to draw in or close around the bright point 

 which represented the image of the source reflected from the 

 inner wall of the bulb. Two of these points usually appear 

 in the picture, one reflected from the outer, the other from 

 the inner wall. They are the virtual images of the source in 

 the spherical mirror formed by the wall of the bulb. This 

 contraction of the glow struck me as very peculiar, as it 

 reminded one of the reflexion from a slightly corrugated 

 surface, or a surface w T hich has been very finely ground but 

 not polished. I obtained it with the bulb and with a small 

 flask. PI. XI. figs. 5 and 6 show the phenomenon, the 

 pressure of the mercury vapour being recorded on each 

 photograph. I made a calculation by elementary methods 

 of what was to be expected when the resonators became 

 more closely crowded together, and came to the conclusion 

 that such a drawiug-in or contraction of the glow was not 

 to be expected, but that what would actually occur would be 

 a gradual diminution of the intensity of the glow, which 

 at low pressures is spread uniformly over the illuminated 

 wall, and a simultaneous increase in the intensity of the 

 small image reflected from the inner wall. 1 now 

 prepared a new bulb, which had never been heated or used 

 in any experimental work, and found no trace of the con- 

 traction of the glow. A careful examination of the old bulb 

 and the flask in a strong light revealed the presence of a 

 slight amount of devitrification of the inner surface, the 

 reflected image of the sun appearing surrounded with a 

 fairly bright halo. This circumstance at once explained the 

 contraction of the glow with increasing vapour pressure. 

 As the density of the mercury vapour increased, the reflect- 

 ing power of the inner wall increased, much as if a deposit 



