BAXTER AND CUAPIN. — ATOMIC WEIGUT OF NEODYMIUM. 243 



The average ratio of silver to silver chloride is within four thou- 

 sandths of a per cent of that obtained by Richards and Wells,^ 

 0.752G34 : 1.000000. 



The results in the preceding tables add further evidence to that al- 

 ready obtained by the spectroscope, that all the samples analyzed were 

 of a high degree of purity. Fraction 2928-4 contained all the rare 

 earth impurities of higher atomic weight than neodymium which ac- 

 cumulated in sixty-seven series of crystallizations, and actually was 

 found to contain a trace of samarium, while fraction 2936-7 contained 

 more praseodymium, probably not more than 0.1 per cent, than any 

 other fraction of the last series which was analyzed. The first impurity 

 raises the atomic weight and the second lowers it ; yet the difference 

 in the average results of these two extreme fractions is only 0.024. 

 The fractions 1302 and '1614 gave results differing from those of frac- 

 tions known to be slightly purer by amounts no larger than the experi- 

 mental error. When corrected for the proportion of the chief impurity, 

 praseodymium, the results agree satisfactorily for all the samples ex- 

 cept fraction 2923-4, which certainly contained an unestimated pro- 

 portion of samarium. 



The results obtained from fractions 1302, 1614, and 2923-4 natur- 

 ally are less reliable than those obtained from the other three fractions! 

 The corrected average for fractions 2926-7-8, 2932-3, 2936-7, is, 

 however, 144.268, which differs from the corrected average for all six 

 fractions, 144.275, by less than 0.01. It is evident that the value 

 144.27 can safely be taken to represent with accuracy the atomic 

 weight of the purest material which we have succeeded in preparing. 

 This value is essentially identical with the one chosen by the Inter- 

 national Committee upon Atomic Weights, 144.3, and lies midway be- 

 tween the results of Auer von Welsbach and Feit and Przibylla on the 

 one hand and that of Holmberg on the other. 



We are indebted particularly to the Carnegie Institution of Washing- 

 ton for pecuniary assistance in carrying out this investigation, and also 

 to the Cyrus M. Warren Fund for Research in Harvard University for 

 indispensable platinum vessels, as well as to the Welsbach Light Com- 

 pany for the neodymium material. 



Summary. 



1. It is emphasized that crystallization in more than one form is ad- 

 visable for the preparation in a pure state of salts of rare earths. 



2. The preparation of a very pure neodymium salt is described. 



1 Pub. Car. Inst., No. 28, (1905); Jour. Amer. Chem. Soc, 28, 456. 



