324 Professor H. E. Roscoe [March I, 



Heated to the point of incandescence in any other manner, the 

 vapours of these metals and their salts give out the same coloured 

 light. Thus, if we burn gun-cotton, or gun-paper, steeped in solu- 

 tions of these various salts, we get the characteristic colours. The 

 well-known coloured fires owe their peculiar effects to the ignition of 

 the vapour of some particular substance. Thus, in red fire we have 

 strontium, in green fire we have barium salts present in the state 

 of luminous vapour. 



These facts have long been known and applied ; but it was reserved 

 for Bunsen and Kirchhoff to place these beautiful phenomena in their 

 true position, to apply to them the modern methods of exact research ; 

 and thus to open out a new and rich field for most important investiga- 

 tions. This they accomplished in a most simple and loeautiful manner, 

 by examining these coloured flames, not by the naked eye, but by means 

 of a prism or an apparatus for separating, decomposing, or splitting up 

 the light produced by the incandescent vapour into its different con- 

 stituent parts. 



If we pass white sun-light through a prism, we get the well-known 

 solar spectrum discovered by Newton. The red, or least refrangible 

 rays appear at one end, and we pass through all gradations of colour — 

 noticing on our way certain dark lines or spaces, showing the absence in 

 solar light of some particular rays, lines with which we shall have 

 much to do — until we arrive at the violet, or most refrangible end of 

 the spectrum. If instead of using white sun-light, we pass the rays 

 from the yellow soda flame through the prism, we get the soda spec- 

 trum ; and we find that instead of a continuous spectrum, all we see is 

 one bright yellow line, showing that every kind of light except that 

 bright yellow ray, is absent in the soda flame ; or that the soda flame 

 gives out only one kind of light. 



And as each metal, sodium, potassium, lithium, calcium, stron- 

 tium, barium, &c., communicates a distinct tint to flame, so each gives 

 a distinct and characteristic spectrum, consisting of certain bright 

 coloured lines, or bands of light of the most peculiar form and tint. 



The actual spectra of these metals can be beautifully seen in the 

 simple apparatus designed by Bunsen and Kirchhoff. 



In each spectrum of these metals, the form, number, position, colour, 

 and tone of the bright lines remain perfectly constant and unvarying, 

 so that from the presence or absence of one of these lines, we may 

 with absolute certainty draw conclusions respecting the presence or 

 absence of the particular metal, as we know of no two substances 

 which produce the same bright lines. None of the bright lines pro- 

 duced by any one metal interfere in the least with those of any other, 

 and in a mixture of all these metallic salts together, each ingredient 

 can thus be easily detected. 



As an example of the exactitude with which a very small quantity 

 of a most complicated mixture can thus be analyzed, the speaker 

 quoted Bunsen's words. " I took," says Bunsen, " a mixture of 

 chloride of sodium, chloride of potassium, chloride of lithium, 



