218 PROCEEDINGS OF THE AMERICAN ACADEMY. 



In 1898 Brauner,^ by converting oxide to sulphate, obtained the re- 

 sult 143.6 for neodymium. In this he included a correction for an im- 

 purity of 2.9 per cent of the corresponding praseodymium compound, 

 estimated by comparison of the intensity of absorption spectra. 



In the same year Boudouard ^ obtained a neodymium preparation by 

 precipitation with potassium sulphate from some impure yttrium ma- 

 terial. Apparently to identify the base rather than to establish its 

 atomic weight, he analyzed the sulphate by converting it to oxide, with 

 the result 143.0. Undoubtedly his material contained praseodymium, 

 as no steps were taken to remove this inevitable impurity. 



In 1898 Jones 2 also investigated both neodymium and praseodym- 

 ium with material purified by crystallization as the double ammonium 

 nitrate. Cerium and lanthanum were eliminated by precipitation as 

 basic nitrates, and a correction for residual praseodymium was obtained 

 by comparing the intensities of its absorption bands in the neodymium 

 preparations with those of the same bands in dilute praseodymium 

 solutions of known strength. The atomic weight was found by con- 

 verting the oxide to sulphate. The average of twelve results, ranging 

 in value from 143.46 to 143.62, was 143.55. Corrected for praseodym- 

 ium content this value becomes 143.60. 



In 1901, after further purification of his original material, Brauner* 

 published another determination, which yielded the value 143.89. As 

 this is the result of but one analysis and includes a correction of ques- 

 tionable accuracy for praseodymium impurity, it can hardly be given 

 much weight. 



Auer von Welsbach^ in 1903 published the results of further study 

 of both neodymium and praseodymium. His source of material was a 

 large quantity of double ammonium nitrates obtained as by-products 

 in the technical extraction of lanthanum. A long series of fractional 

 crystallizations as double ammonium nitrate served to free the neodym- 

 ium from all but a trace of praseodymium. With this material, un- 

 doubtedly of very high purity, three analyses were made by the Bunsen 

 method of converting oxide to sulphate, from which the atomic weight 

 values 144.55, 144.52, and 144.57 respectively were obtained. Un- 

 fortunately no details of procedure are given. 



A point of especial interest in connection with Jones's and Brauner's 

 results is Auer's observation that the praseodymium absorption spectrum 



1 Proc. Chem. Soc, 14, 70 (189S); Chcm. News, 77, 161 (1898). 



2 Compt. Rend., 126, 900 (1898). 



3 Amcr. Chem. Jour., 20, 345 (1898), 

 * Troc. Chem. Soo., 17, G6 (1901). 



» Sitzungsber. Akad. Wiss. Wien., 112, 1037 (1903). 



