388 CHEMISTRY: RICHARDS AND BOYER 
remains to be done in order that perfect certainty may be attained, never- 
theless the outcome must be regarded as promising in pointing toward an 
adequate and convenient method of separating gallium from other metals. 
Three fractional samples of the chloride discussed in the previous section 
were analyzed, using the usual methods employed in Harvard University 
in order to make a preliminary determination of the atomic weight. The 
samples were small and for various reasons the results cannot be considered 
as anything more than preliminary, but the outcome is, nevertheless, worth 
recounting. In the last and best determination 0.43947 gram of gallium 
chloride (weighed in vacuo in a sealed glass tube) yielded 1.07087 grams of 
silver chloride having required 0.80587 gram of silver for complete precipita- 
tion. The atomic weight of gallium computed from these two sets of data 
are respectively 70.09 and 70.11 — concordant results indicating a value some- 
what higher than that usually accepted for gallium, but near enough to show 
that the chloride was at least not far from being pure, and that the whole 
proceeding is capable, when employed with larger quantities of material and 
with the experience already gained, of affording an accurate evaluation of 
this atomic weight. Of course such meagre data as those thus far secured 
furnish no worthy evidence concerning it; they are as merely preliminary 
as the data of Lecoq de Boisbandrau. At the conclusion of the war it is 
hoped that the joint investigation (which was stopped by the departure of 
W. M. Craig into War Service) may be continued and completed. 
1 The principle of this method has since been published by Dennis and Bridgman, al- 
though the details o" their treatment differed from ours. Their work was entirely inde- 
pendent of ours (which was brought to a close in March, 1918) and was entirely unknown 
to us. /. Amer. Chem. Soc, Easton, Pa., 40, 1918, (1540). 
THE PURIFICATION OF GALLIUM BY ELECTROLYSIS, AND THE 
COMPRESSIBILITY AND DENSITY OF GALLIUM 
By Theodore W. Richards and Sylvester Boyer 
WOLCOTT GlBBS MEMORIAL LABORATORY, HARVARD UNIVERSITY 
Communicated, October 31, 1918 
The method of separating gallium from indium by means of the different 
solubilities of the hydroxides in caustic alkali, recommended by various au- 
thorities, was tested without success. The separation was found to be so in- 
complete that several per cent of indium remained in the gallium, at least 
under the conditions used in our work, and it seemed clear that this differ- 
ence in solubility is not enough to effect a complete separation. Much more 
promising results were obtained by the electrolytic method. Gallium occu- 
pies a place in the electrolytic series between indium and zinc. It is far 
less easy to deposit than indium, but, on the other hand, much more easy to 
deposit than zinc. By carefully regulating the hydrogen-ion concentration 
