CHEMISTRY. 281 



gave values from 37.33 to 37^36. (average 37.342) grams of nitrate in 

 100 grams of solution, and nine determinations with uranio-lead nitrate 

 gave values from 37.26 to 37.30 (average 37.280) grams of nitrate in 

 100 grams of solution. These results are precisely proportional (within 

 the limit of error of experiment) to the different molecular weights of 

 the two samples of nitrate. Expressed in other terms, the molal solu- 

 bilities of the two samples per 1,000 grams of water are respectively 

 1.7993 for common lead nitrate and 1.7991 for uranio-lead nitrate. The 

 outcome of this work, then, by finding no appreciable difference in the 

 molal solubiUty of the different samples of lead nitrate containing 

 isotopes of widely different atomic weights, confirms earher work upon 

 other properties of these interesting substances. Evidently weight (or 

 mass) is the prime distinguishing feature of the two kinds of lead here 

 studied, as it was in other cases. This work and that described in the 

 preceding section thus afford further support for the hypothesis due 

 to Russell, Fleck, Soddy, and Fajans. 



3. Purification of Gallium by Distillation of the Chloride. 



With the help of Mr. W. M. Craig, a new method for the purification 

 of gallium salts was worked out in some detail. This rests upon the 

 convenient fact that gallium trichloride subHmes and distils at a low 

 temperature, whereas the other chlorides likely to be associated with it 

 are much less volatile. The method rested, therefore, upon fractional 

 distillation, at first in a stream of chlorine and afterwards in a vacuum, 

 of impure gallium chloride. The apparatus was a complex affair, in 

 which gallium could be first burned in pure dry chlorine and then sub- 

 jected to distillation either in pure chlorine or nitrogen or in a vacuum, 

 the whole being fused together without rubber connections and scrupu- 

 lously dried. The gallium was provided through the great kindness of 

 Mr. McCutcheon, of the Bartlesville Zinc Company, Blackwell, Okla- 

 homa, to w^hom grateful thanks are due. Three successive distilla- 

 tions of the trichloride of gallium were made in chlorine at 220° to 230° 

 C, three more at about 175° (the melting-point of gallium dichloride), 

 three in nitrogen at 90° to 110°, and five in vacuo at 65° to 80° — four- 

 teen in all. The resulting product showed no trace of any other sub- 

 stance in its spark spectrum when examined with great care in a Hilger 

 wave-length spectrometer. Although this research did not reach a 

 final stage, and much more remains to be done in order that perfect 

 certainty may be attained, nevertheless the outcome must be regarded 

 as promising in pointing toward an adequate and convenient method 

 of separating galHum from other metals. The analysis of the result- 

 ing material is discussed in the following section. 



4. Preliminary Determination of the Atomic Weight of Gallium in the Chloride. 



Three fractional samples of the chloride discussed in the previous 

 section were analyzed by Dr. J. Sameshima, using the usual methods 



