82 Intelligence and Miscellaneous Articles. 



it, and to diminish at the same time its resistance to breaking and 

 the intensity of the current. It is evident that if they cooperate to 

 fix the light at the summit of the carbons, it is on the condition of 

 lessening the extreme length the arc can attain, or, what is the 

 same thing, the number of foci that can be kept alight with a given 

 machine. 



It is no longer so when the points are turned towards the ground. 

 While the arc tends to ascend along the carbons, the directing cir- 

 cuit drives it back, lowers it, and lodges it between the points, dis- 

 tant from 7 to 8 millims. The two actions, which before were 

 added together, are now subtracted one from the other ; and so far 

 from lengthening the arc, they shorten it ; instead of diminishing 

 its resistance to rupture, and lessening the intensity of the current, 

 they augment both. The arc may be said to be as it were com- 

 pressed between two opposite actions : it is shorter, narrower, less 

 expanded, more dense, and consequently hotter ; and the number 

 of the foci can be augmented. M. Jabloschkoff's candles, in other 

 respects so well combined, have nevertheless the inconvenience that 

 the points are directed upwards. The arc produced by them has 

 a natural tendency to bend and raise itself ; and the same tendency 

 is impressed upon it by the electromagnetic action exerted upon it 

 by the current, which ascends in one carbon and descends in the 

 other — an action identical with that of my directing circuit, though 

 less energetic. The burner with the points below must therefore 

 excel those candles. In fact this is proved by experience. With 

 a machine hardly sufficient to light three of those candles I can 

 easily maintain five burners armed with much bigger carbons, each 

 giving about twice as much light ; and as the points are immersed 

 in the mass of the arc, they acquire a more vivid brightness and an 

 incomparably whiter tint. Six foci, even, can be lighted ; but they 

 give a less sum total of light than five ; if we double the number 

 we lose in quantity. It is always so when the electric light is 

 divided beyond measure : the division must be purchased by a 

 proportional loss. 



The management of these burners requires careful study. When 

 the points are above, the lighting is very difficult, because imme- 

 diately it is produced the force of the directing current projects it 

 strongly upwards, that force being proportional to the square of 

 the intensity. When this increases, it becomes absolutely impos- 

 sible to light the carbons ; one can get only a vast flame which 

 immediately disappears with a noise. If the current is less intense, 

 the light continues, but much spread out, very tall, and always 

 very noisy, on account of the amplitude of the oscillations which 

 take place at each inversion of the current. Finally, the equili- 

 brium is not at all stable. If an accidental current of air comes 

 for a moment to augment the height of the flame, nothing can 

 restore it, the limit of its elasticity is passed, and it soon breaks. 

 In the burners with the points beneath, the lighting is easy, and 

 the equilibrium stable ; for if a movement of the air or a deficiency 

 of the current causes the arc to ascend, it settles between the two 

 carbons at the place where they have not been thinned by combus- 

 tion; it lodges in an interval not exceeding 2 or 3 millims. Far 



