466 #. L. Nichols—Character and Intensity of the 
comparative intensities of the various parts of the visible 
spectrum by constant temperature, the latter denotes the 
influence of variations of temperature upon the intensity of 
each individual series of rays. 
ow exceedingly small the intensity of the more refrangible 
rays is, even at 1900°, compared with that of the spectral 
regions between line D and the red end of the spectrum, is 
equally evident in both sets of curves. A striking peculiarity 
of the isothermals is the too large value of the intensity of 
region (10). This unlooked-for feature can scarcely be due to 
a corresponding irregularity in the platinum spectrum, and at 
the same time it is in all probability no accidental error. Such 
an error, in order to produce such an effect in all the isothermic 
curves, must be looked for either in my comparison of the 
sun’s spectrum with that of the platinum wire, or in Laman- 
sky’s measurements of the sun’s heat. As to the former, we 
find by reference to the table, that the result in question is cal- 
culated from the mean of three readings of fair agreement. 
An error of several degrees in these determinations would be 
necessary to bring about any such irregularity as exists in the 
cur amansky’s results are also the mean of numerous 
the surrounding atmosphere contained a trace of sodium, 
appears in the curves on Diagrams A and B. ee 
Kirchhoff, in his treatise ($15), draws the following @ prort 
ion : 
ce _ conclusions from his discussio 
2 “Ifa body, a platinum wire for example, is gradually heated, 
_ emits until attaining a certain temperature, only rays the 
_ Wavelengths of which are greater than those of the visible : 
spectru At a certain temperature, rays of wave lengths cor- 
| * Vierordt, Poggendorff’s Annalen, 151. 
es 
