366 Mr. W. Galloway on the Influence [Mar. 2, 



1 to 18 — a similar cap 2 inches high, which burned rather more steadily. 



1 to 20 — a cap 1^- inch high, with nearly parallel sides to about 

 two thirds of its height, and drawn out to a point at the top. This cap 

 was perfectly steady, and more distinct than any of the others. 



1 to 25 — a conical cap \ to f inch high. 



I to -30 — a conical cap | inch high. 



1 to 40 — a conical cap ^ to | inch high. 



1 to 50 — an exceedingly faint cap g inch high, the top having the 

 appearance of being broken off. 



1 to 60 — it was hardly possible to distinguish any thing above the 

 small oil-flame. If there was any cap (which the manager of the pit 

 and two of the firemen who were present considered excessively doubt- 

 ful), it was represented by the frustum of a cone, of which the sides 

 only were visible, little more than -fa inch high. 



The firemen, whose principal occupation is to examine the workings 

 daily for any symptoms of firedamp, both asserted positively that they 

 would consider an atmosphere in which the small oil-flame burned as it 

 did in this mixture to be perfectly free from gas. 



. When the height of the oil-flame was increased to J inch it became 

 luminous, and then the cap was invisible in all the mixtures consisting of 

 1 volume of gas to 25 volumes and upwards of air. Similarly, the cap is 

 invisible in all mixtures down to the explosive one with higher oil-flames. 



On comparing these results graphically (fig. 7), the divisions of the 

 base line, measured from left to right, representing the number of volumes 

 of air to one of firedamp, and the length of the ordinates giving the 

 heights of the corresponding caps, measured from the top of the oil- 

 flame, it is seen that the curve drawn through the extremities of the 

 ordinates is roughly approximate to the form of a rectangular hyperbola. 

 I have therefore drawn that curve for comparison through the top of 

 the ordinate which corresponds to 20 volumes of air. 



It is evident that the heights of these caps are in some degree propor 

 tional to the quantity of firedamp consumed above the oil-flame per unit 

 of time ; and, although the methods by which they were obtained are far 

 from perfect, the results seem to point to the existence of a simple rela- 

 tionship between the initial temperature, the heat of combustion, and the 

 specific heat of the various gases contained in the mixtures. 



To recapitulate in a summary manner what has been already said in 

 the preceding pages of this paper, we have now two principal facts before 

 us, which are these : — 



1. A mixture of firedamp and air, in the proportion of 1 volume of the 

 former to 60, or more, volumes of the latter, gives no reliable indication 

 of the presence of the inflammable gas, when tested in the manner usually, 

 if not always, adopted in mines. 



2. A mixture of firedamp and air, in the proportion of 1 volume of 

 the former to 112 of the latter, becomes inflammable at ordinary pressure 



