DECOMPOSITION BY LIGHT lOT 



light continued to act, showing at some places vivid 

 iridescence. 



The lens of the electric lamp was now placed so as 

 to form within the tube a strongly convergent cone of 

 rays. The tube was cleansed and again filled in dark- 

 ness. "When the light was sent through it, the precipi- 

 tation upon the beam was so rapid and intense that the 

 cone, which a moment before was invisible, flashed sud- 

 denly forth like a solid luminous spear. The effect was 

 the same when the air and vapor were allowed to enter 

 the tube in diffuse daylight. The cloud, however, which 

 shone with such extraordinary radiance under the electric 

 beam, was invisible in the ordinary light of the labora- 

 tory. 



The quantity of mixed air and vapor within the experi- 

 mental tube could, of course, be regulated at pleasure. 

 The rapidity of the action diminished with the attenua- 

 tion of the vapor. When, for example, the mercurial col- 

 umn associated with the experimental tube was depressed 

 only ^YQ inches, the action was not nearly so rapid as 

 when the tube was full. In such cases, however, it was 

 exceedingly interesting to observe, after some seconds of 

 waiting, a thin streamer of delicate bluish-white cloud 

 slowly forming along the axis of the tube, and finally 

 swelling so as to fill it. 



When dry oxygen was employed to carry in the vapor, 

 the effect was the same as that obtained with air. 



When dry hydrogen was used as a vehicle, the effect 

 was also the same. 



The effect, therefore, is not due to any interaction 

 between the vapor of the nitrite and its vehicle. 



This was further demonstrated by the deportment of 



