96 
LUMMER. 
therefore defines a sharp limit between the region from which 
only light of the white incandescent center ol the sun and 
those regions from which only the light of gases emitting 
line spectra reaches our eyes at 0 (Fig. 6). The sharp edge 
of our sun is, according to Schmidt, no real limit between 
liquid and gaseous parts of the sun, but an optical illusion 
produced by the refraction of light on the sun. IIow large 
the radius, p, of the critical sphere is and many other similar 
questions are problems to be solved by future speculations 
and observations. 
In Schmidt’s theory the dispersion phenomenon plays 
only a secondary role, since in general the small dispersion 
of gases causes the radius, p, of the critical sphere to appear 
imperceptibly different for different colors. But a gas emit- 
ting spectrum lines has, as you know, with reference to the 
emitted light waves, the power of anomalous dispersion. 
Starting from this, II. W. Julius, the well-known Dutch 
physicist, has succeeded in giving a simple explanation of 
spectral phenomena with respect to the sun hitherto not 
plausibly explained. 
Anomalous dispersion was observed first for sodium vapor 
by Le Roux and A. Kijndt by using the well-known method 
of crossed spectra. With respect to the theory of Julius, it 
was of great importance to establish anomalous dispersion 
also for substances, such as calcium and strontium, the vapors 
in which are predominant in the chromosphere. Wilson, at 
Potsdam, contradicted Julius’ hypothesis, inasmuch as he 
could not observe the anomalous dispersion of these sub- 
stances. 
Pringsheim and I succeeded by using as a vapor prism 
tile flame of an arc lamp colored by vapors of Ca, Sr , Na, 
etc,, and transmitting through the prismatic arc flame the 
strong light of an ordinary arc lamp. The transmitted light 
falls upon a spectroscope of high resolving power, whose slit 
and prism edges are perpendicular to the edge of the vapor 
prism (cut out from the colored arc flame), as required by 
Kundt’s method of crossed spectra. Using sodium vapor, we 
observe in the spectroscope the interesting phenomena rep- 
