170 Electric Lighting by Incandescence. [February, 
to mp a tf 1 ’ 233 where m and a depend on the nature of the gas 
at a pressure p , and t, as before, is the excess of temperature. 
The preceding formulas do not, however, give results 
agreeing well with more recent experiments. In the “ Pro- 
ceedings of the Royal Society ” for 1872, Mr. McFarlane ' 
published the results of his experiments on the absolute 
amount of radiation and convention from a polished and from 
a blackened copper ball in air. In addition to the faCt that 
these experiments were made probably under the eye of 
Sir William Thomson, to whom Mr. McFarlane is an 
assistant, and that, therefore, they are probably correCt, I 
have verified them experimentally myself, and extended 
them to other gases than air, and at various pressures.* 
Mr. McFarlane’s experiments show that the measured 
surface emissivity increases less and less rapidly as the dif- 
ference of temperature rises. Probably, therefore, we may 
be safe in saying that for a difference of temperature of 
iooo 0 C. (about the temperature of melting cast-iron), the 
value for platinum and carbon will not be very much higher 
than the highest he obtained, respectively, for polished and 
for blackened copper. 
From experiments made in a vacuum, I find that when 
the difference of temperature is under ioo° C. the value of 
the surface emissivity is about half, and for gases like coal- 
gas about double that obtained by Mr. McFarlane in damp 
air at ordinary pressure. 
In some of the systems of eledtric lighting it has been 
proposed to enclose the incandescent rod in a vacuum, in 
others in some non-decomposable gas, like nitrogen. In 
this latter gas I have not hitherto made absolute radiation 
and convection experiments, but we may fairly take as a 
mean value of this surface emissivity the air value previously 
referred to. 
The specific eleCtric resistance of platinum is a com- 
paratively constant quantity, while that for carbon varies 
within wide limits. It, therefore, becomes an important 
question to determine what kind of carbon, that which 
conducts electricity pretty well, or that which is a much 
worse conductor, is the more suitable to be used for illumi- 
nation by incandescence. Now the temperature equation 
previously referred to, shows that for a given electromotive 
force maintained at the two ends of a carbon rod, the smaller 
be the specific resistance of the carbon employed, the greater 
will be the heat, and, consequently, temperature produced. 
* The results of these experiments of Professor Perry and myself will 
shortly be published. 
