Purification of Gallium. - 395 



washed with distilled water, and tested both chemically and 

 spectroscopically. All the evidence thus obtained favored the 

 belief that the compound is gallium hydroxide. In this con- 

 nection, it may be remarked that an apparently identical mate- 

 rial is produced by bubbling carbon dioxide through a solution 

 of gallium in sodium hydroxide. The formation of the white 

 substance at the anode is doubtless due to the decrease in con- 

 centration of the sodium solvent in this region. This decrease 

 may be brought about both by the absorption of carbon dioxide 

 from the air and by the ionic redistribution involved in the 

 passage of the electric current. Obviously the electrolyte may 

 be cleared up by adding caustic soda. 



Attention may now be directed to figure 4 which not only 

 illustrates the last paragraphs both of the present article and 

 of our preceding paper {loo. ait.) but it also constitutes the only 

 reproduction of the complete arc spectrum of gallium which 

 we have seen. Figure 4a shows the spectrum of the mother 

 liquor of the caesium-gallium alum mentioned below. The 

 purification produced by ten crystallizations of the alum is 

 made evident by figure 4b. The spectrum of a specimen of 

 electrolytic gallium (not heated in hydrogen) is given in figure 

 4c. Spectrogram 4ol pertains to the gallium-indium alloy 

 obtained by a sweating process. The following table of wave- 

 lengths (International System) will assist in the identification 

 of the most prominent lines. The bands and faint lines due to 

 the carbon arc and to slight impurities are not tabulated. The 

 numbers indicating lines belonging to the first order of the con- 

 cave grating have no accents. Lines of the second and third 

 order are designated respectively by single and double accents. 

 H — "head " of gallium band, n = new (see previous paper). 

 R = widely reversed throughout entire length of arc. r = 

 reversed chiefly near positive electrode. Lack of space pre- 

 cluded numbering the strong zinc lines shown in figure 4d 

 between 19' and 21'. Their wave-lengths are 4680-138, 

 4722-164, and 4810-534. In figure 4a the principal caesium 

 lines (AX4555-34 and 4593-21) may be readily seen on the less 

 refrangible side of the very intense indium line, X4511-37. 



As already stated in our previous paper,* a fairly satisfac- 

 tory separation of gallium and indium may be obtained by the 

 action of sodium hydroxide upon the hydroxides of the 

 elements, as recommended by Lecoq de Boisbaudran. Another 

 methodf described by the same investigator, the crystallization 

 of the ammonium alums in 70 per cent ethyl alcohol, proves 

 also good, a very few crystallizations giving a gallium product 

 showing only spectroscopic traces of indium, zinc, copper, and 

 lead. 



*This Journal, vol. xli, p. 351, April, 1916. 

 fCoinpt. rend. (Paris), xcv, 410. 



