44 EF. L. Nichols— Optical Method for the 
In point of fact the “‘limzt of the spectrum” admits (see vol. 
xvili, p. 400) of no sharp determination, depending as it does 
upon the constantly varying condition of the observer's eye. 
If, as is commonly supposed, the position in the spectrum of 
the thermal maximum were a function of the temperature, the 
second method would be at best practically applicable to but a 
few of the most intense sources of light. Some recent discov- 
eries of Dr. Jacques in Baltimore,* seem to show that the posi- 
tion of the thermal maximum depends upon the molecular 
weight of the glowing body, and that, for a given source of 
light, its position is in no way affected by a change of tem- 
perature. This newly discovered fact renders Crova’s second 
method useless. I shall show presently that aside from Jacques’ 
experimental evidence there are good reasons for supposing the 
position of the thermal maximum to depend upon the nature of 
the glowing body rather than upon its temperature. 
The third method coincides with that which I have pro- 
posed. The chief difficulty in the development of it lies in 
the varying values of the emissive and absorptive capacity of 
different bodies. 
In Equation (1) the fraction z is to be sure independent of 
the nature of the body in question; not so, however, the quan- 
tities E and A, considered separately. That A, possesses for 
different substances widely different values we know from for- 
mer researches. Hitherto, however, these experiments have 
been confined to ordinary temperatures, and the question of the 
dependence of this quantity upon the temperature has been for 
the most part neglected. . 3 
or the purposes of general discussion it is convenient to 
divide all bodies into four classes. 
odies for which A=Constant, for all wave lengths and 
for all temperatures. 
IL. Bodies for which A varies with the temperature, but has 
the same value for all wave lengths. 
odies for which A varies with the temperature and 
possesses different values for different wavelengths, but for 
which the ratio of these values in any two spectral-regions 4 to 
A+d2 and 2’ to 1’+d2’ is independent of the temperature. 
odies for which A varies with the temperature and 
wave length and for which the above-mentioned ratio is also 4 
function of the temperature. 
Black bodies are, by definition, of the first class. Whether 
bodies exist for which A=Constant <1, can only be deter- 
mined by special experiment. ‘Io the second class belong — 
* Distribution of Heat in the Spectra of various sources of Radiation. Cam- — 
bridge, 1879. ae 
