23 
ascended to the points, where it remained so long as the 
current was transmitted. My first impression of this pecu- 
liar action of the arc was, that it was due to the ascending 
current of hot air by which it was surrounded. This how- 
ever was found not to be the cause, as the arc travelled 
towards the points in whatever position the carbons were 
placed, whether horizontally or vertically in an inverted 
position. Moreover, when a pair of carbons were held in 
the middle by the holders, the arc travelled upwards or 
downwards towards the points, according as the circuit was 
established above or below the holders. The action was in 
fact recognised to be the same as that which determines the 
propagation of an electric current through two rectilinear 
and parallel conductors submerged in contact with the 
terrestrial bed, which was described by me in the Philo- 
sophical Magazine , August, 1868. 
In all the arrangements in general use for regulating the 
electric light, the carbon pencils are placed in the same 
straight line, and end to end. When the light is required, 
the ends are brought into momentary contact, and are then 
separated a short distance to enable the arc to form between 
them. The peculiar behaviour of the electric arc when the 
carbons are placed parallel to each other, suggested to me 
the means of lighting the carbons automatically, notwith- 
standing the fact that they could only be made to approach 
each other by a motion laterally, and to come into contact 
at their adjacent sides. To accomplish this object, one of 
the carbon holders is articulated or hinged to a small base 
plate of cast iron, which is so constructed as to become an 
electro-magnet when coiled with a few turns of insulated 
wire. The carbon holder is made in the form of a riofit- 
O 
