264 



Dr. W. J. Russell. 



possible that since the amount of the 532 producing body appeared, 

 judging very roughly by the eye to diminish on dissolving the solid 

 chloride in hydrochloric acid, that in an excessively dilute hydro- 

 chloric acid solution this substance might be destroyed and the band 

 have disappeared. Experiments, however, did not confirm this, for 

 when a solution so dilute that it required 6 feet of the liquid to bring 

 the absorption up to 600, the 532 band was just as visible as in a 

 strong solution. To sum up, with regard to this spectrum it is 

 remarkably persistent ; strengthen the solution or increase the length 

 of the column looked through, or raise the temperature of a dilute 

 solution, and the effect is the same ; first, the shaded part up nearly to 

 650 becomes uniformly dark, then the absorption closes in as far as 

 the 610 band, and the absorption creeps on up to 600 ; immediately 

 after reaching this point the 532 band becomes visible. 



Still further increase the absorption and a shadow is seen extending 

 from the 532 band towards the red to 560. And lastly, with the 

 strongest solution that can be looked through, there is still discernible 

 a broad black band from 550 to 520, and the absorption from the red 

 end reaches now to 565, and it is remarkable that in the ultra-red, 

 where at first there was complete absorption, now light is transmitted. 

 At the blue end considerable absorption occurs, no light being visible 

 beyond 480. 



Ethylic, methylic, and amylic alcohols, the saline ethers, and 

 glycerine all dissolve cobalt chloride freely, and on being saturated 

 with anhydrous chloride of cobalt give this same spectrum (fig. 3). 

 These liquids must, of course, be free from water, or the true spectra 

 are complicated by the spectra of the hydrates. 



When ordinary alcohol is the solvent, a very large amount of the 

 chloride is taken up and a mere film of the liquid on a microscopic 

 slide is quite sufficient to give clearly the characteristic spectrum 

 (fig- 3). 



With amylic or methylic alcohol no appreciable alteration in the 

 position of the bands takes place ; but, with glycerine, both the bands 

 have moved slightly towards the red. 



That solvents so different as the above should all give the same 

 spectrum with the cobalt chloride seems only to be explainable on the 

 view that the spectrum so formed is that of the chloride itself, and 

 not that of any new combination, in fact, that the solvent has so far 

 altered the molecular structure of this salt that in place of its giving 

 the spectrum (fig. 1) which it does when alone and solid, it gives this 

 second spectrum (fig. 3) ; and further, it seems, whether the chloride be 

 in the liquid or solid state, whether dissolved in alcohol or in a solid 

 alkaline chloride, the spectrum is practically the same. 



Although not probable, at the same time it seemed possible, that the 

 identity of these different spectra might be due to a trace of water in 



