272 PROCEEDINGS OF THE AMERICAN ACADEMY 



bromhydric acid, and the solutions evaporated to a syrupy con- 

 sistence, I examined two cases. The perfectly colorless and easily 

 soluble crystals obtained from gadolinite earths free from cerite 

 earths by chlorhydric acid after two successive crystallizations were 

 analyzed. 



(111) 0.3248 gr. gave 0.1293 gr. C 2 3 = 39.82 per cent. 



(112) 0.3000 gr. " 0.1450 gr. R 2 3 = 48.34 " 



The atomic mass is 107.10 so that the crystalline chlorides contained 

 the earths in the same proportion in which they were obtained from 

 gadolinite after separation of the cerite earths. 



Precisely the same result was obtained with crystallized bromides 

 prepared by dissolving the crude oxides from the cerite and monazite 

 residues in bromhydric acid and evaporating. The beautiful colorless 

 crystals were not quite free from the mother liquor. Of these 

 crystals, 



(113) 0.3418 gr. gave 0.1064 gr. C 2 3 = 31.11 per cent. 



(114) 0.4266 gr. " 0.1325 gr. " =31.07 " 



(115) 0.6110 gr. " 0.2879 gr. R 2 3 = 47.13 " 



(116) 0.6896 gr. " 0.3293 gr. " =47.02 " 



The atomic mass is 139.55, which is nearly the same as that ob- 

 tained from the oxides directly. 137.25. 



The mother liquor from the crystals was also analyzed. 



(117) 0.3750 gr. gave 0.1171 gr. C 2 3 = 31.25 per cent. 



(118) 0.3237 gr. " 0.1015 gr. " =31.37 " 



(119) 0.6752 gr. « 0.3180 gr. R 2 3 = 47.10 « 



(120) 0.8150 gr. " 0.3837 gr. " =47.08 " 



The atomic mass corresponding is 138.45. From the above it 

 appears that little, if anything, is gained by crystallization of the 

 chlorides and bromides, at least in the cases cited. 



The fact, that potassic and sodic sulphates which do not give pre- 

 cipitates of double, sulphates in cold saturated solutions of certain 

 earths often give crystalline precipitates on boiling, has doubtless been 

 observed. I do not find, however, that such observations have been 

 noted in published papers. The following analyses will serve to 

 show that valuable results may sometimes at least be obtained by 

 this process. 



