502 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 54 



preparations by cerium, and losses by spattering. To this criticism 

 Jones ' promptly replied, giving a new series of determinations as follows : 



Mean, 57.549, ± .0010 



Hence La = 138.81. 



The material was spectroscopically pure, and the siilphate was neutral 

 and soluble. The operations were performed in porcelain crucibles, and 

 the oxide was perfectly white. Brauner used platinum crucibles, and 

 Jones found that lanthanum oxide, heated in platinum, became per- 

 ceptibly discolored. Two determinations made in platinum gave the 

 following results : 



Mean, 57.586, ± .0027 



Hence La = 139.06, a value in accord with Brauner's. According to 

 Jones the discoloration and variation in atomic weight suggest the 

 presence of some other oxide than the normal compound in Brauner's 

 preparations. The controversy, however, remains unsettled, and addi- 

 tional investigations are needed to determine the truth. 



The two determinations by Brill " arc of slight value, and hardly worth 

 considering. Small quantities of lanthanum sulphate were calcined to 

 oxide, and the weighings were made with the Nernst microbalance, in 

 order to test its applicability to work of this kind. The percentages of 

 oxide in sulphate are given below, more for the sake of completeness 

 than for any real significance in them : 



57.664 

 57.726 



Mean, 57.695, ± .0207 



Hence La= 139.79., a very high value. 



In 1906 Feit and Przibylla ' determined the atomic weights of several 

 rare earth metals by a special volumetric process, which, however, seems 



iZeitsch. anorg. Chem., 36, 92. 1903. Chem. News, 88, 13. 



2 Zeitsch. anorg. Chem., 47, 464. 1905. 



3 Zeitsch. anorg. Chem., 50, 248. 1906. See also an earlier paper in Vol. 43, p. 213. 1905. 



