The properties of flames of various kinds as generating electricity and as serving to conduct a voltaic current, as also their behaviour under the influence of a charge of electricity, have been investigated by several philosophers. But, so far as we know, the effect of flame on the electric spark has not been methodically observed, or at all events has not been methodically measured. This was the problem which at the opening of the College session presented itself to us, and the results now attained by a course of experimenting appear capable of throwing some light on the solution.

We first made a number of tentative experiments, for the purpose of finding out the general aspect of the phenomena, and the conditions most favourable for exact measurement. An ordinary Bunsen burner served to supply a flame luminous or non-luminous. It was always connected with the earth, and the charged body was in general a disc of 4 inches diameter, connected along with a large Leyden jar to the insulated conductor of a Holtz machine. We tried the disc in a position above the flames, but found that the disturbing conditions introduced were, as was to be expected, rather numerous, namely, the heating of the disc and the covering of it with soot, and the melting of the gutta-percha covering of the connecting wire. Hence for a series of measurements we preferred the disc placed in a position at the side, as represented in the accompanying figures. (See Plate IV. a).