120 The Ohio Journal of Science [Vol. XVI, No. 4, 



DEPOSITION OF METALLIC MIRRORS. 



When a discharge tube was opened after a spectral examina- 

 tion of a salt it was found that there was a thin metallic deposit 

 covering the inside of the lower electrode in the portion near 

 the water cooled end. It cannot be stated that these were 

 present in the case of every substance investigated; but in the 

 case of mercuric iodide, cadmium iodide, zinc bromide and tin 

 chloride metallic mirrors, almost as brilliant as if polished, were 

 noticed. This shows that considerable dissociation and reduc- 

 tion of the compound occurred under heat and the electrical 

 excitation. 



These mirrors, as stated in a preceding paragraph, were 

 also obtained with two compounds which were heated in the 

 apparatus used for investigating absorption spectra. The 

 mirrors obtained from stannic chloride were almost as bright 

 as is the surface of clean mercury. In these experiments no 

 external agent was used except the heat. It is therefore appar- 

 ent that under heat alone some metallic compounds (at least 

 halogens) are dissociated, the metal being set free. There can 

 have been no reducing agent present, such as would have been 

 the case had the salts been introduced into the flame, unless the 

 residual gas present in the tube could have acted as such agent. 

 This seems improbable in view of the low pressures used. 



vSome interesting results on "Flame Reactions" have been 

 published recently by Bancroft and Weiser (Journal of Physical 

 Chemistry, Vol. XVIII, 1914) in which they have obtained 

 metallic deposits on cold porcelain introduced into a Bunsen 

 flame which was fed with chlorides and nitrates of copper, 

 cadmium, tin, mercury and silver. These experimenters con- 

 cluded that "the reducing action of the flame gases is not 

 essential, though at times it may increase the decomposition." 

 With this conclusion the writers of this paper are in accord 

 and feel that the results detailed above establish the legitimacy 

 of this conclusion since the question of flame gases does not enter. 



POSSIBLE NATURE OF THE SYSTEMS PRODUCING THE BAND 



SPECTRA OF THE COMPOUNDS. 



In order to present the experimental facts in concise form 

 and to provide a ready reference table in connection with the 

 ensuing discussion the following brief resume of the data 

 obtained on the emission spectra of the compounds is introduced 



