Detection and Measurement of Inflammable Gas in Mines, 127 



are, however, some drawbacks to it : there is a general feeling 

 that it is not sufficiently sensitive ; and many hold, further, that 

 there is a wide difference in the percentage of gas that may 

 be present before the cap makes its appearance. Whether 

 this is correct or not, I am not prepared to give a decided 

 opinion ; but, from experiments made, I consider 2 per cent, 

 of marsh-gas about the limit detectable with the ordinary 

 Davy flame. A small and clean flame is an essential requi- 

 site in applying this test; any particles of ignited matter 

 on the wick will readily produce a spurious cap when no gas 

 exists. 



Of the modes of rendering the cap more visible none is 

 more efficient than turning the flame low, or hiding the 

 luminous portion with some opaque object. Attempts have 

 been made to increase the sensitiveness of this test by viewing 

 the flame through blue glass, with a hope that, the yellow colour 

 of the flame being subdued, the bluish cap would be the more 

 apparent ; but this has failed practically to increase the sensi- 

 tiveness of the test, and, I believe, for this reason — that as 

 the percentage of gas is diminished the cap decreases more 

 in size than in intensity; for it is only those molecules that 

 come into very close proximity to the flame that obtain suffi- 

 cient auxiliary heat to enable them to reach the ignition-tem- 

 perature. 



Of the instrument of M. Coquillion I have nothing to say; 

 for though it may be a very convenient and neat apparatus 

 for gas-analysis, the fact that it cannot be employed in the pit, 

 and that samples of air have to be brought to it puts it en- 

 tirely out of the present question . 



I now come to my own instrument. The principle on which 

 it is based is as follows : — A mixture of marsh-gas and air in 

 which the marsh-gas forms less than 5 per cent, of the mixture 

 is not explosive or capable of continuing its own combustion 

 (at ordinary temperatures and pressures), simply because the 

 heating-value of the CH 4 is insufficient to raise that large excess 

 of atmospheric air to the necessary ignition-temperature. If 

 however, such a mixture is exposed to some sufficiently heated 

 object, especially if that object is platinum, it will burn in its 

 immediate contact and neighbourhood, and in so doing add 

 materially to the temperature of the object, and the more so the 

 larger the percentage of gas present. To apply this principle to 

 the detection and measurement of gas in the air of mines, I have 

 contrived the following arrangement (shown in Plate II.) : — 

 A and B are two spirals of very fine platinum wire, through 

 which a current from a small magneto-electrical machine 

 is made to circulate. Both wires being in the same circuit, 



