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



the Ohio State University. The following notes are taken 

 from his report: 



Hg Br. Metallic mercury unmistakably present. Tiny globules of mercury 

 found under 'scope on surface of lumps of Hg Br. 



Cd I-i. Showed presence of metallic cadmium by reflected light. 



Zn Br-i. Could get no evidence of presence of free zinc. 



Fe-i Cl(,. Under magnifier, presence of numerous green spots showing existence 

 of FeClo, or Fe2Cl4, (the latter formula being now considered 

 preferable by chemists). 



Ca Br-y. All tests showed presence of metallic calcium. 



Hg I>. In this case there was considerable reduction to Hg I. 



The writers are disposed therefore to believe that there is 

 dissociation and ionization of some portions of the original 

 compound into its basic and radical components, but that there 

 is also, under the proper conditions, a dissociation of the 

 vaporized salt into a positively charged sub-compound and 

 negatively charged radical. To illustrate, let us take mercuric 

 iodide (Hg I2). This in part dissociates into Hg+-|- or 2 Hg+ and 

 1= or 2 I_, or positive ions of mercury carrying a single or 

 double charge and negative ions of iodide carrying a single or 

 double charge. In addition there is, as evidenced by the 

 analysis of the residue after heating, a considerable dissociation 

 into Hg 1+ or positively charged mercurous iodide and I_ or 

 negatively charged ions of iodine. It is probable that the 

 banded or continuous spectra arise from the recombination 

 of the oppositely charged mercurous iodide and iodine which 

 are formed from the vaporized, non-dissociated salt present 

 in the capillary portion of the discharge tube, although it is 

 not likely that there is a permanent recombination of mercuric 

 chloride formed. We can say that we have the emission 

 spectrum of the original compound if we admit such a process. 

 That some such process is operative is made plausible by the 

 fact that the banded regions in the case of the iodides have their 

 heads at or near a prominent iodine line. In the case of the 

 chlorides we find extended continuous regions which include 

 the spectrum from the blue to the orange-yellow region. 

 There is no evidence of lines or bands in the red end of the 

 spectrum. The .spectrum of chlorine exhibits a richness of 

 lines in the region 5()()()-410() Angstroms and a dearth of 

 anything in the red region except the line 6095. The same 

 explanation made for the iodides seems satisfactory here. If 

 stannic chloride dissociates in part into ions of tin and chlorine 

 and in part into positively charged molecules of stannous 



