522 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 54 



To reduce these ratios we have 8 = 32.0667, ±.00075; C = 12.0038,± 

 .0002; and H = 1.00779, ±.00001. 

 Hence, 



From ratio ?, Pr = 139.623, ± .1200 



" 5 140.545, ± .0189 



" 2 140.628, ± .0035 



" 4 140.954, ± .0160 



" 1 141.090, ± .0236 



General mean, Pr = 140.619, ± .0033 



Brainier's deteniiinations make Pr=141, very nearly, and must be 

 taken into consideration in criticizino- the foregoing combination. His 

 value may be nearer the truth, but the work of Jones and of Scheele 

 cannot yet be. rejected. There is still an uncertainty of half a unit in 

 the atomic weight of praseodymium. The later determinations by 

 Welsbach are in harmony with the general mean of all the other esti- 

 mations. 



NEODYMIUM. 



Our knowledge of the atomic weight of neodymium is almost entirely 

 based upon a study of the sulphate. Welsbach's first determination was 

 cited under praseodymium, and needs no farther consideration. So 

 also Brauner's^ first, preliminary figure, N"d = 143.63, given without 

 analytical details, may be dismissed here. The first important series of 

 determinations is that by Jones,'' published in 1898. The synthesis of 

 the sulphate was effected in the usual way, with the following results: 



Mean, 58.246, ± .0015 



Hence Xd = 143.54. 



^ Proc. Chem. Soc, 14, 72. 1898. 



^ Amcr. (■hem. .loiirn., 20, 345. 1898. See additional note in Zeitsjli. anorer. Chem., 19, .339. 1899. 



