i8 ELECTRIC LIGHTING. 



easily oxidizable metals, like potassium, sodium, magnesium, 

 it is more intense than with the inoxidizable metals, such as 

 gold or platinum. The shape of the luminous centre depends 

 on the form of the electrodes and on their polarity, as I 

 have shown in my account of RuhmkorfFs induction ap- 

 paratus. Between a point and a conducting surface it 

 assumes a conical form, and between two carbon points it 

 has the appearance of a globe. The maximum length of the 

 arc especially depends on the tension of the current, and with 

 a strong current may reach to 2 or 3 centimetres when the 

 arc has once been established. According to Despretz this 

 length increases more rapidly than the number of battery 

 cells employed, and the increase is more marked for small 

 arcs than for large ones. It is, therefore, greater with bat- 

 teries arranged for tension than with those arranged for 

 quantity. Again, the voltaic arc is better developed when 

 the positive carbon is above than when it is below ; and 

 when the carbons are horizontal the arc is shorter than when 

 they are placed vertically, for the hot air always tends to 

 ascend. On the other hand, the arrangement of the battery 

 for quantity then becomes more advantageous than the 

 arrangement for tension. 



When voltaic currents are transmitted by carbon elec- 

 trodes, the positive electrode has a much higher temperature 

 than the negative, a circumstance which does not occur when 

 induced currents of high tension are used to produce the 

 light. With a moderate electric intensity, and with very pure 

 carbons, the luminous effect emanates from a- bluish radiant 

 centre only ; but when the intensity is greater, a real flame 

 always surrounds that luminous centre, and the less pure are 

 the carbons the larger is the flame. 



Fig. 2 shows the appearance of the carbons producing the 

 electric light when the voltaic arc is projected on a screen- 

 Under these conditions the brightness of the flame, and that 

 of the arc itself, are so far surpassed by that of the carbons 

 that they can scarcely be distinguished. If the carbons are 



