Mr. 0. Heaviside on measuring small Resistances. 245 



spark, only the carbon-lines will be visible. It appears to follow 

 that the specific luminosity of carbon under electric excitement 

 is much higher than that of hydrogen. If the light in a Geiss- 

 ler's tube be really produced exclusively by heat, as is commonly 

 assumed, it follows that carbon at low pressure becomes lumi- 

 nous under the influence of very little heat. It is worth noting 

 that with a broken circuit, such as I have described, the carbon- 

 lines are distinct lines equally sharp on both sides. With a full 

 circuit they are (in the prism) sharp only on one side, the least 

 refrangible ; on the other they show an expansion, a tendency to 

 melt into the obscure continuous spectrum that lies behind them ; 

 on this side they much resemble the hydrogen-lines of the third 

 spectrum, produced by intense heat. In short, the carbon-lines 

 are visible when the hydrogen-lines have not yet appeared ; the 

 carbon-lines have begun to expand while the hydrogen-lines still 

 remain sharp. Both substances pass through the same stages 

 in their passage to the continuous spectrum to which all gases, 

 by heat and pressure, can be ultimately reduced ; but carbon 

 begins the process first, and probably finishes it first ; for it may 

 be inferred that the spectrum of carbon vapour reaches the con- 

 tinuous condition proportionally sooner. I think I have seen it 

 argued that there can be no carbon gases in the sun, because 

 there are no carbon-lines. It has certainly been assumed that 

 the carbon in a candle- flame must be in the solid state, because 

 the spectrum is continuous. The above observations may tend 

 to throw doubt on both these positions. If this carbon-spectrum 

 is visible at a low temperature, it may reach the continuous con- 

 dition, and consequently not be recognizable, at a temperature 

 lower than the flame of a candle. Of course, if all carbon-spectra 

 implied a similar temperature, the existence of the Bessemer- 

 spectrum would negative such an idea. But Dr. Watts has 

 shown that this is far from being the case. 



I do not know whether it would be possible, by slight admix- 

 tures with the mercury of a thermometer, to ascertain the lowest 

 temperature of luminosity in the case of other elements; but 

 the investigation would be worth making if it is practicable. 



XXXI. On an advantageous Method of using the Differential Gal- 

 vanometer for measuring small Resistances. By Oliver 

 Heaviside, Great Northern Telegraph Company, Newcastle- 

 on-Tyne*. 



IN the usual method of measuring resistances with the differen- 

 tial galvanometer, the current from the battery is divided 

 between the two coils, having opposite effects on the needle 



* Communicated by the Author. 



