Samarium. 3 



. The identity of the absorption-bands leads to the conclusion that 

 decipium, samarium, and Y/Î all three contained the same element. But 

 the atomic weights as determined by Delafontaine and by Marignac 

 differed greatly. 



Later on Delafontaine') admitted that his original decipia could 

 be split u]i into two different oxides, one with an absorption-spectrum 

 and one without. For the element in the former he adopts the name 

 samarium reserving the name decipium for the element in the latter. 

 The respective atomic weights wei-e 101 (or 151 . 5) and 111.3 (or 167), 

 The Ye. of Maeignac he regards as a mixture of terbia and decipia, but 

 as terbia is orange coloured and Y« almost white this supposition 

 seems to be inadmissible. 



SEPARATION. 



There is no complete method of separating samaria from the 

 other earths which accompany it. In order to remove the yttria-earths 

 advantage may be taken of the sparing solubility of the double sul- 

 phate of samarium and potassium in a saturated solution of potassium 

 sulphate, but the most convenient method of separation from didymia 

 is by fractional precipitations of dilute solutions of the nitrates by very 

 dilute and cold ammonia. Samaria is precipitated first but the operations 

 must be repeated a great number of times. In order to obtain a pure 

 oxide it is necessary to combine both methods. The. sulphate of sa- 

 marium is much less soluble than the sulphate of didymium, but as 

 these salts are isomorphous, no good separation can be founded upon 

 this difference. 



MODE OF OCCURRENCE. 



Samarium has been found in Samarskite, in Gadolinite, in Orthite, 

 and in Cerite, but there can be no doubt that this ejement will be found 

 together with didjmiium in all minerals containing the latter metal. Never- 

 theless seems didj'mium always to occur in greater quantity. 



ATOMIC WEIGHT. 



In ni}' paper on samarium which appeared in 1883, I published 

 determinations of the atomic weight of samarium. By combining known 



') Comptes reudus XCIII, 63. 



