an Incandescent Paper-carbon Horseshoe Lamp, 23 



branch passed through a series of adjustable resistances com- 

 posed of German-silver wire, stretched in the free air of the 

 laboratory to avoid heating (careful tests showed that this pre- 

 caution fully accomplished the desired result) ; and the united 

 branches were then carried to the other pole of the battery. 



These arrangements having been made, a certain number of 

 battery-cells were put in circuit and the resistances adjusted 

 until the galvanometer showed no deflection. The condition 

 of the loop was then observed ,in perfect darkness ; and when 

 its light was measurable it was taken by varying the distances 

 of both lamp and candle as circumstances required. 



Thus, for the lowest candle-power taken, the lamp was at 

 15*8 inches from the photometer, and the candle at 50 inches. 

 The results so obtained were as follows : — 



No. of cells Candle- Resistance 



in circuit. power. resistance. 



123-0 ohms. 



5 113-5 „ 



10 dark red 106*0 „ 



20 -1 candle 94'0 „ 



25 -2 „ 89-0 „ 



30 -4 „ 87-0 „ 



35 -9 „ 83-7 „ 



40 1-9 „ 82-0 „ 



45 5-1 candles 79*8 „ 



50 8-4 „ 78-0 „ 



58 18-0 „ 75-0 „ 



These results are also expressed in the curve shown in fig. 1. 



The fact of a decrease of resistance with rise in temperature 

 with carbon was previously noticed by Matthiessen in 1858 

 (see Phil. Mag. vol. xvi. pp. 220-221). This experimenter 

 found the electric conductivity of ordinary gas-coke to rise 

 about 12 per cent, between the common temperature and a 

 light red heat. 



In the case of this delicate thread of impure carbon consti- 

 tuting the loop of the lamp, the rate of increase in conducti- 

 vity or fall in resistance is more rapid. Fig. 1 (p. 24) shows 

 the above observations plotted as a curve, and needs no further 

 explanation. 



In the above discussion we have compared the resistance of 

 the lamps with the luminous emissions only ; but we have 

 once considered it worth while to make an analogous but more 

 extended comparison, namely one between the resistance and 

 the total heat, or total heat and light, generated in the lamp. 

 This enables us to carry the range of comparison below those 

 points at which sensible light is developed. As a matter of 

 course, this relation to total heat is also the relation to energy 



