Feb., 1916] Spectra of Some Halogen Compounds 123 



chloride and negatively charged ions of chlorine we should 

 expect the experimental results detailed above. Furthermore, 

 none of these banded regions appeared to be associated in any 

 manner with the element making up the base of the salt. This 

 is clearly proven, we believe, by the fact that the banded 

 regions were practically identical in any halogen group irrespec- 

 tive of the base present. Possibly the best illustration is 

 afforded by mercuric iodide in which the lines characteristic 

 of mercury coexisted with the broad bands and were prominent 

 and sharp with no tendency towards broadening or diffuseness. 

 These experiments also throw light upon the investigations 

 of Professor O. W. Richardson on the specific charge of the ions. 

 In investigating zinc iodide he found that the negative ions were 

 iodine and the positive ions, during the initial stages of heating, 

 correspond most closely to ions of zinc carr3dng a double charge. 

 We found the spectrum of zinc and iodine present in the early 

 stages of heating, but also found that the zinc lines soon dis- 

 appeared. In the case of manganous chloride, Richardson 

 found initial values which would correspond to manganese, 

 but with further heating a value of "m" equal to 90 indicating 

 possibly the existence of positively charged Mn CI. The 

 results of the experiments made by the writers support such 

 conclusions as these. In the case of ferric chloride no spectral 

 evidence of the presence of iron was found at any period of the 

 experimentation, but after sufficient heating a continuous 

 spectrum within the limits of 5900 and 4800 Angstroms was 

 developed. This may indicate, if we follow out the line of 

 argument given in the preceding paragraph, a dissociation into 

 negative ions of chlorine, (CL) or (2C1_) and positively 

 charged sub-compounds of iron and chlorine (Fe C1++). If such 

 a positively charged molecular structure existed it would have a 

 value of "m" equal to 45, which would be in accord with 

 Richardson's experimental values of 40.2, 39.3 and 39.8. It is 

 to be stated in this connection, however, that the analysis of 

 the residue from Fe CI3 did not show the existence of such a 

 compound as Fe CI, but did show the sub-compound Fe CI2. 

 The chemistry of compounds at high temperatures is still 

 somewhat of an unopened book ; without doubt there are formed 

 and exist at such temperatures molecular structures or sub- 

 compounds or recombinations of the elements of the original 

 substance of which we know little. The postulation of the 



