238 THE SCIENTIFIC PAPERS OF 



mode of advancing the carbons, which, instead of being effected by 

 clockwork (as has been the case hitherto in constructing regu- 

 lators), is effected simply by the force of gravity, or by spring 

 power urging the carbons forward towards the point of meeting, 

 in which forward motion each carbon is checked by a metallic 

 abutment, in the form of a point or edge of copper or other metal 

 of high conductivity, the exact position of which can be regulated 

 by a screw. 



This metallic ridge touches the carbons laterally, at a distance 

 of 10 to 15 millimetres from the luminous point, where the 

 temperature of the carbon is sufficient to cause its gradual decom- 

 position in contact with the atmosphere, without being high 

 enough to fuse or injure the metal. 



In the lamp before you, represented in Plate 21, the carbons 

 are contained in horizontal holders, suspended from the lamp frame 

 by means of four suspension rods : a solenoid coil is placed vertically 

 above the point of light, the iron core of which is connected to- 

 the suspension rods on either side by means of rods, whereby 

 horizontal motion is applied to the metallic carbon-holders, tending 

 to separate them when the current flowing through the solenoid 

 coil diminishes, and approaching the carbons when the current 

 passing through the coil increases ; the effect is, that an increased 

 resistance in the electric arc causes a shortening, and a decrease, a 

 lengthening of the arc itself. In order to steady the action of this 

 regulator, the iron core carries a piston, working freely up and 

 down in a vertical cylinder, having a throttled aperture for the 

 ingress and egress of atmospheric air above the piston. The two 

 metallic holders are put into conductive connection with the two 

 wires leading up from the dynamo-electric machine, and the 

 regulating solenoid coil is connected to the two holders respectively,, 

 so that the current active in this coil is always proportionate to 

 the potential between the two electrodes, which by this arrange- 

 ment is made practically constant. If this lamp is worked by 

 alternating currents, the two carbons are made equal in diameter ; 

 but if worked by a continuous current, the carbon connected with 

 the positive pole should be made larger than the one connected with 

 the negative pole, in the ratio of about 3 : 2. 



Instead of the solenoid regulator, the steel tape regulator may 

 be employed, which I described in a paper read before the Eoyal 



