BAXTER AND CHAPIN. — ATOMIC WEIGHT OF NEODYMIUM. 221 



crystallizations the number of fractions in each series was maintained 

 nearly constant between twenty and twenty-five, and the different frac- 

 tions were not allowed to vary much in size. As a rule the solutions 

 were allowed to become supersaturated by cooling without crystalliza- 

 tion, then they were seeded with tiny crystals of the purest neodymium 

 salt available, and allowed to stand undisturbed over night in order 

 that by the formation of large crystals the removal of mother liquor 

 might be facilitated. On account of the extra labor and time in- 

 volved, centrifugal drainage of the crystals is of little advantage where 

 the composition of crystals and mother liquor differs as slightly as in 

 the present case, and therefore was not employed. When the end frac- 

 tions became small they were either temporarily removed and added to 

 similar fractions subsequently obtained, or, if it could be plainly seen 

 by the difference in color that they contained considerable impurity, 

 they were rejected. 



A diagram of the crystallizations as double nitrate is given on page 

 222. In any given series of crystallizations a lower number always indi- 

 cates a less soluble fraction. A line not connecting an end fraction 

 with any fraction in a subsequent series indicates rejection. 



"When eighty-six series of crystallizations had been completed, spec- 

 trograms were made of the absorption spectra of selected fractions in 

 the last series, which contained fractions 1287 to 1309. These photo- 

 graphs were taken on Cramer's Trichromatic plates with a spectrograph 

 employing glass lenses and prism and covered the range A 650 — • A 380. 

 The spectra indicated the presence of samarium in only a few of the 

 more soluble fractions, 1306 to 1309. Praseodymium could be de- 

 tected in all the fractions, although only traces were visible in the more 

 soluble fractions. 



In order to throw more light upon the progress of the purification 

 approximate atomic weight determinations by the "permanganate" 

 method ^ were carried.out with portions of fractions 1290, 1299, 1303, 

 and 1305. The oxalate was precipitated by adding a dilute solution 

 of the double nitrate to a dilute solution of an excess of pure oxalic 

 acid, and the precipitate was thoroughly washed, collected, and dried 

 at 120°-130°. After very complete mixing of the highly crystalline 

 material by grinding in an agate mortar, weighed portions of the oxalate 

 were converted to oxide by ignition in platinum crucibles. Other por- 

 tions, weighed out at the same time, were dissolved in dilute sulphuric 

 acid and titrated with a standard solution of potassium permanganate. 

 Fi'om the ratio M2O3 : 3 C2O4, the atomic weight was calculated. 



^ See page 216. 



