Hidden and Warren — ITttrocrasite. 517 



bustion tube reinforced with heavy platinum foil as recom- 

 mended by Penfield and catching the products in sulphuric 

 and potash bulbs respectively. Several blank determinations 

 were run to test the apparatus previous to the actual determina- 

 tion. The combined weights of the water and carbon dioxide 

 are slightly higher (0*15 per cent) than the loss on ignition of 

 the boat and its contents, due probably to the oxidation 

 of the U0 2 . This close agreement indicates that there can 

 be little if any helium or nitrogen present in the mineral and 

 no evidence of their presence could be obtained. After igni- 

 tion the mineral has a light buff color and goes into solution 

 more difficultly than the unignited material. 



A qualitative test for ferrous iron, made as recommended by 

 Penfield,* proved its entire absence. The oxidation, therefore, 

 of potassium permanganate by a cold sulphuric acid solution of 

 the mineral was taken as an indication that part at least of the 

 uranium present was in the form of the dioxide, and the figure 

 given for this oxide in the analysis was obtained by titration 

 with permanganate. 



For the main portion of the analysis, the mineral was decom- 

 posed with hydrofluoric acid (the J. Lawrence Smith method). 

 The greenish white residue of earth fluorides was filtered off, 

 converted into sulphate (a little lead sulphate came out at this 

 point) and twice precipitated with ammonium hydroxide to sep- 

 arate the earths, etc., from the lime. The earths were sepa- 

 rated from the iron and uranium by precipitation with oxalic 

 acid. Thorium and cerium earths were taken out as described by 

 Hillebrand.f The cerium earths were separated from thorium 

 by means of sodium thiosulphate and ammonium oxalate pre- 

 cipitations.^; The earths were all finally precipitated as oxalates, 

 before ignition and weighing. Great care was taken to insure 

 the complete recovery of earth oxalates from the filtrates. 

 Some were almost invariably recovered, which again empha- 

 sizes the caution, in this regard, given in the article re- 

 ferred to above by Dr. Hillebrand. The yttrium earth oxides 

 were of a light buff color and gave a pink nitrate solution. This 

 tested spectroscopically showed the characteristic absorption 

 spectra of erbium. The molecular weight was found to be 

 268-4 (K 2 3 ). 



The cerium oxides were brown in color and when dissolved in 

 hydrochloric acid gave a yellow solution. Tested spectroscopic- 

 ally the didymium bands were obtained and a faint band in the 

 position of the most characteristic erbium line, indicating a slight 

 contamination. The molecular weight, determined as. 331*0 

 (R 2 3 ), is probably not exact owing to the small total weight 



* Brush and Penfield's Determinative Mineralogy, p. 87. 



f This Journal (4), xiii, p. 148, 1902. 



X See Metzger, Am. Chem. Jour., xxiv, 901, 1901. 



