Analysis by Spectrum-observations. 331 



we proceed to recount the experiments which one of us has 

 conducted for the purpose of establishing the properties of the 

 two new elements, and their more important compounds. 



I. Of the Preparation, Atomic Weight, and occurrence of the 

 Rubidium Compounds. 



The pure chloride of rubidium was procured from the saline 

 residue obtained by fusing a mass of about 150 kilogrammes of 

 Saxony lepidolite, from which the alkaline earths and lithium 

 salts had been removed. The separation of the new element, 

 and the preliminary determination of its atomic weight, were 

 effected as follows : — 



The saline residue was dissolved in water, and treated with 

 about 100 grins, of bichloride of platinum, a quantity, however, 

 quite insufficient to precipitate all the potassium ; the double 

 platinum salt was then boiled out twenty times with a small 

 volume of water, and the boilings added to the original solution 

 of the saline residue, whereby a precipitate again occurred, 

 which was treated exactly as the former. In the course of the 

 process of continued boiling with small quantities of water, the 

 solution, which originally was of a dark yellowish-brown colour, 

 becomes gradually lighter, so that it is easy to see, by the light 

 colour of the precipitate remaining unchanged, the point at 

 which the boiling-out has been continued long enough. The 

 extraction is carried on until the whole of the precipitate formed 

 by the saline residue dissolves on repeated boiling with small 

 quantities of water. The several platinum precipitates, after 

 having been again purified by treating them altogether with 

 boiling water, are dried and reduced in a current of hydrogen 

 gas, by which means a mixture of metallic platinum and im- 

 pure chloride of rubidium is obtained, the latter being extracted 

 by water. This aqueous solution is diluted, and, whilst boiling, 

 again precipitated by chloride of platinum, and the insoluble 

 double salt reduced, as before, in a current of hydrogen. 



Of the chloride of rubidium thus prepared, which we will 

 designate as portion A, 2*2496 grammes gave on precipitation 

 2' 7688 grms. chloride of silver. A portion of this same preparation 

 A was dissolved in about thirty times its weight of water, and 

 precipitated whilst hot with a solution of chloride of platinum 

 so diluted that the precipitate appeared only after the lapse of 

 a few minutes. As the liquid cooled, the precipitate became 

 more dense ; and when the temperature had sunk to about 40° C, 

 it was filtered off, dried, and reduced in hydrogen as described. 

 The chloride thus prepared we will call portion B ; of this prepa- 

 ration 0*9022 grm. gave 1*0712 grm. chloride of silver. A similar 

 mode of separation was adopted in the case of the salt B, and 

 another salt, which we call portion C, obtained; 1*3540 grm. of 



Z2 



