262 



Dr. W. J. Russell. 



peating the experiments over and over again, and continuing the 

 heating for a long time, there is always the same result, the return on 

 -cooling to the original spectrum. 



From the change in the fusing point of the mixture, from the per- 

 manence of the cobalt chloride when the mixture is fused in the air, 

 ■and from the change of colour, it seems natural to conclude that 

 chemical combination takes place between the chlorides, and that the 

 new spectrum pertains to the double chloride then formed. Experi- 

 ments, to be described further on, will show, however, that this sup- 

 position is not correct ; for this spectrum (fig. 3) is easily obtainable 

 without the presence of any alkaline chloride, and that it is produced, 

 in fact, in all cases when the cobalt chloride dissolves in any menstruum, 

 without chemically combining with it, to form a new and definite 

 compound. At ordinary temperatures then, at all. events, the two 

 chlorides in the above mixture are not chemically combined, but the 

 cobalt salt is simply dissolved in the potassium or other chloride. 



If these mixtures be dissolved in water, on evaporation the two 

 salts separate out perfectly distinct, the potassium chloride often even 

 without colour. Zinc chloride gives a different reaction from the 

 alkaline chlorides, for when mixed in certain proportions this spectrum 

 is no longer visible. On evaporating an aqueous solution of these 

 two salts, a new compound, a double salt having the composition 

 CoCl 2 ZnCl 2 + 6H 2 0, crystallises out. This double salt produces no 

 visible banded spectrum, consequently as long as sufficient zinc chlo- 

 ride is present to combine with the cobalt chloride, no spectrum is 

 visible ; but when even a trace of cobalt beyond this amount is 

 present the banded spectrum (fig. 3) is visible. 



Some interesting applications of these changes have been made and 

 will form a separate communication. 



The visible spectroscopic changes which occur when these mixed 

 chlorides are heated, are as above described, viz., a new spectrum is 

 produced, as if chemical combination occurred, and on cooling the 

 return of original spectrum. This seems to indicate that the action of 

 heat is to produce change of a physical rather than of a chemical 

 nature ; that a dissociation of a cobalt salt into a simpler molecular 

 form occurred rather than the formation of a new chemical compound, 

 which always decomposed as the temperature fell. With zinc chloride, 

 however, the action of heat is different and easily explainable, for a 

 mixture of cobalt and zinc chlorides, which at ordinary temperature 

 gives no bands, on being heated shows very clearly this same band 

 spectrum of cobalt chloride. In this case heat breaks up the double 

 salts existing at ordinary temperatures, and gives the spectrum of the 

 free chloride. If calcium chloride be used in place of potassium 

 chloride, then a much higher temperature is required to fuse the 

 mixture, and it is far more opaque and difficult to examine. There 



