In order to maintain a good electrical contact between the fixed conducting portions of the lamp and the sliding carbons, Mr. Hedges fits to each carbon-holder a little contact piece, F F, hinged to its respective trough at its upper end, and carrying at its lower or free end a somewhat heavy little block of brass grooved out to fit the cylindrical side of the carbon, against which it presses with an even pressure. This arrangement offers another advantage, namely, that the length of that portion of the carbon rods which is conveying the current is always the same notwithstanding the shortening of their total length by combustion; the resistance of the carbon electrodes is, therefore, maintained constant, and, for the reason that the contact piece presses against the rods very near their lower ends, that resistance is reduced to a minimum. In this way very long carbons, such, for instance, as will burn for ten or sixteen hours, can be used without introducing any increase of resistance into the circuit. The length of the arc can be determined by the adjustment of the screw, G, by which the amount of movement of the armature is limited. Fig. 2 represents a modified form of Mr. Hedges' lamp designed for installation when it is desirable to burn a number of lamps in series. In this arrangement the carbons are separated by the attractive influence of a solenoid upon an iron plunger, to which is attached (by a non-magnetic connection) the armature of an electro-magnet, the coils (which are of fine wire) forming a shunt circuit between the two terminals of the lamp, and so disposed with respect to the armature as to influence it in an opposite direction to that of the solenoid. When the circuit of the lamp is completed with the electric generator the carbons are drawn apart by the action of the solenoid on the plunger, and the distance to which they are separated is determined by the difference of attractive force exercised upon the armature by the solenoid and the magnet; but as the latter forms a short circuit to that of the arc, it follows that should the resistance of the arc circuit increase either through the arc becoming too long or through imperfection in the carbons or contacts, a greater percentage of current will flow through the magnet coils, and the arc will be shortened, thereby reducing its resistance and regulating it to the strength of the current. In other words, the distance between the carbons, that is to say, the length of the arc, is determined by the position of the armature of the electro-magnet between its magnets and the solenoid, which position is in its turn determined by the difference between the strength of current passing through the coil of the solenoid and that of the magnet.