234 ELECTRIC LIGHTING. 



them. The two actions which formerly combined now oppose 

 each other ; so far from lengthening the arc, they shorten it ; 

 instead of lessening its resistance to rupture and diminishing the 

 intensity of the current, they increase both. 



" This arc may be said to be compressed between two opposite 

 actions ; it is shorter, narrower, less expanded, denser, and con- 

 sequently hotter, and the number of the lights may be increased. 

 The Jablochkoff candles, otherwise so well arranged, possess the 

 inconvenience of having their points upwards. The arcs they 

 form naturally tend to curve and rise up, a tendency which is 

 strengthened by the electro-magnetic action of the current 

 ascending in one carbon and descending in the other : an action 

 identical with that in my directive circuit, though smaller in 

 amount. The burners with the points downwards should there- 

 fore surpass these candles, as, indeed, experience proves. With 

 a machine which is sufficient, with difficulty, to light three candles, 

 I easily supply five burners, fitted with much larger carbons, 

 each burner giving twice as much light as a candle ; and as the 

 points are immersed in the mass of the arc, they are more bril- 

 liant and of a much whiter colour. Even six lamps may be 

 lighted, but they give a sum total of light less than five do ; the 

 number may be doubled, but there is a loss in the quantity of 

 light. It is always thus when we try to divide the electric light 

 beyond a certain extent ; the division is purchased by a propor- 

 tional loss. 



" It is interesting to examine the behaviour of these burners- 

 When the points are upwards, the first lighting is very difficult, 

 because the arc is at once briskly thrown upwards by the in- 

 fluence of the directive current, which is proportional to the 

 square of the intensity. When this increases it becomes abso- 

 lutely impossible to light the carbons ; we get nothing but a great 

 flame, which immediately disappears with a snap. If the current 

 is weaker the light remains, but very diffused, very high, and 

 always very noisy, on account of the oscillations which take place 

 on each inversion of the current. Again, the equilibrium is 

 unstable ; if an accidental current of air for an instant raises the 

 height of the arc, it cannot be brought back ; the limit of its 

 elasticity is exceeded, and it immediately breaks. In the burners 

 with the points downwards the lighting is easy and the equili- 

 brium stable, for if a movement of the air, or a failure, causes the 



