514 



M. R. Bunsen's Bloiqnpe Experiments. 



the form represented in vertical section, tig. 3 : aba corresponds 

 to the dark part of the ordinary luminous flame, and contains 

 the unburned gas mixed with air. By arranging the supply of 

 gas so that the point b is exactly level with the upper aperture, 

 a flame is obtained of unvarying dimensions, which does not 

 flicker, and may at all times be obtained, a d a b is the burning 

 cone of the flame, and is so little luminous as to be scarcely per- 

 ceptible. In this cone the external mantle ndac may be di- 

 stinguished from the internal mantle a c a b, by being of a more 

 intense blue. The holder (fig. 3) serves to hold the test speci- 

 Fig. .3. 



mens, which must not be larger than a millet grain, and to 

 expose them to the action of the flame. In the two tubes a a, 

 which are lined with cloth, two glass rods, b b, bent at night 

 angles, move up and down. On the ends of the rods are two 

 glass tubes, c, drawn out to solid ends ; in these are fused 

 platinum wires about 0"145 millim. thick, bent at their extre- 

 mities to fine loops. To allow the glass tubes to be firmly fixed, 

 and yet to move freely, a silk thread, provided with a knot, is 

 introduced, which acts like a spring. An aperture is made 

 in the tube through which the thread passes. 



The temperatui'c of the flame depends mostly on the composi- 

 tion of the gas. This exhibits considerable discrepancy, as is 

 seen from the following analyses of the Heidelberg gas, executed 

 by Professor Landolt at diff"erent times : — 



