M. R. Buiiscn's Bloiqnpe Experimeufs. 517 



to approximative quantitative determinations. A piece of paper 

 about a centimetre square is coated with red iodide of mercury, 

 and placed upon a small holder d (fig. 1), which is moveable on 

 an arm c, also moveable, attached to the chimney. When the least 

 quantity of soda is introduced into the melting-space of the flame, 

 the red iodide becomes white with a faint tinge of fawn colour! 

 Potass, lithia, and lime do not prevent this reaction. 



If the soda is dissolved in water, the experiment is made as 

 follows. The capillary platinum wire d, fig. 3, is bent to a loop, 

 and this is beaten out so as to form a small ring. If this be 

 dipped in the solution, and evaporated without boiling in the 

 neighbourhood of the flame, as much solid substance remains as is 

 necessary for the reaction. With solutions of various degrees of 

 concentration, the time during which the change of colour is ob- 

 served is proportional to the concentration of the liquids. With 

 the same quantity of chloride of sodium diluted with 1, 3, and 

 4 parts of water, the duration of the change of colour was 1, 2, 

 and 4 seconds. The reaction is extremely delicate. To deter- 

 mine its limit, C-186o grm. of salt were mixed with measured 

 quantities of distilled water, and, each time, 2 milligrammes 

 of the solution evaporated and exposed in the flame. When the 

 quantity of water had reached 1 kilogramme, the reaction was 

 still quite distinct ; ttt^o ir of a milligramme of chloride of sodium 

 may thus be detected without difficulty. 



To detect potass compounds in the presence of soda, Cart- 

 mell's* reaction may be used. As Cartmell has observed, all 

 substances which make flame luminous, and especially all organic 

 substances which burn with separation of carbon, give the same 

 violet colour, and must hence be first removed by heat. 



Cartmell deflected lithia in the presence of soda and potash, 

 by comparing the mixed colour of the flames of those bases with 

 that of the flame of pure potash, when both are viewed through 

 an indigo solution. Bunsen has found that the discrimination 

 of these bases in presence of each other is more easily efl'ected by 

 observing the succession of changes of colour which the mixed 

 flame produced by these substances experiences when the rays 

 reach the eye after passing through gradually thicker layers of 

 an indigo solution. For this purpose a hollow plate-glass prism 

 is used, filled with indigo solution : it is 40 millims. high, and 

 its principal section is a triangle with two sides of 150 millims., 

 and the other of 35 millims. The indigo solution is made by 

 dissolving 1 part of indigo in 8 parts of fuming sulphuric acid, 

 diluting with 1500 to 2000 parts of water, and filtering. 



In the following exi)crimeuts the prism was moved horizon- 

 tally before the eye, so that the rays of the flame always passed 

 * riiil. Mag. vol. xvi. p. '3-2H. 



