518 ]\I. K. Biuiseu's Blowpipe Eaperiinents. 



through gradually thicker layers of the medium. The alkaline 

 substances, brought singly into the melting-space, exhibited the 

 following changes : — 



1. Chemical/}/ pure chlvride of calcium produces a yellow Hame, 

 which, even with very thin layers of the indigo solution, passed 

 through a tinge of violet into the original lamp tiame. 



2. Chemically pure chloride of sodium, the same. 



3. Chemicallii pure carbonate or chloride of potassium appears 

 of a cserulean blue, then violet, and at last of an intense crimson 

 red, even when seen through the thickest layers of solution. 

 Admixtures of soda or lime do not hinder the reaction. 



4. Chemically pure carbonate or chloride of lithium gives a 

 carmine-red flame, which, with increasing thickness of the 

 medium, becomes gradually feebler, and disapjiears before the 

 thickest layers pass befoi'e the eye. Lime and soda are also 

 without influence ou the reaction. 



As of all lithium compounds the carbonate and chloride give 

 the most intense coloured tlame, it is only necessary to mark on 

 the prism the ])lace at which lithia disappears, to obtain a 

 space above this mark which transmits red potash rays, but 

 never red lithia rays. 



5. Lithia uith an admixture of potash is detected by bring- 

 ing a sample into the melting-space, and comparing its flame 

 with that of a sample of pure potash placed side by side with it 

 in the melting-space. With thin layers of the solution, the 

 lithia flame appears redder than the pure potash flame; with 

 somewhat thicker layers the flames are equally red, if the propor- 

 tion of lithia to potash be very small ; if the lithia be in excess, 

 the intensity of the red lithia flame sensibly diminishes with 

 thicker layers, while the red flame of potash is scarcely weakened. 

 In this manner a few thousandths of lithia may be detected in the 

 presence of potash. Soda has almost no influence on the reaction. 



These reactions are readily applicable with volatile salts, but 

 with silicates it is often difiicult to detect 3 or 4 per cent, of these 

 alkalies. The reactions may, however, be applied to the silicates, 

 by heating them in the melting-space with jnire gypsum, by 

 which there is formed silicate of lime, and a volatile alkaline 

 sulphate which imparts a colour to the flame. If the reaction 

 of a specimen both before and after the addition of gypsum be 

 compared with a standard series of silicates the proportion of 

 alkali in which is known, the various alkaline silicates (for in- 

 stance, the members of the felspar family) may not only be 

 recognized, but the relative quantity of potash, soda, and lithia 

 approximatively determined. 



The process adopted in these cases will be most readily un- 

 derstood by an example. Bunsen selects the discrimination of 



