Witson— On the Temperature of certain Stars. — 185 
be as efficient a radiator as a layer of clouds. It is well to remember that a sun- 
spot is only dark by comparison, and is really quite as bright as the lime-light. 
In the Sun, therefore, we seem to have two quite distinct sources of light which 
give us a continuous spectrum. First, we have the photospheric clouds, and 
secondly, at some depth below them, layers of highly heated gases, under con- 
siderable pressure, which are also able to radiate as if solid. Between these two 
layers lie the vapours of titanium, vanadium, and other elements which, from 
their atomic weights, lie at a greater depth in the solar atmosphere than carbon. 
All the elements which, from their atomic weights, have a place near the photo- 
spheric clouds, must get carried up by convection-currents above them, and thus 
give us the Fraunhofer spectrum. In asun-spot where the photospheric clouds are 
volatilized, we then get the absorption-spectrum of all the elements, like titanium, 
which lie between those two great sources of radiation, greatly strengthened. 
Now, what are the causes which give rise to an absorption-spectrum? In the 
first place, the absorbing layer of gas must be at a lower temperature than the 
source which is giving the continuous spectrum; and, secondly, the absorbing 
layer must be thick enough to be opaque to those particular wave-lengths which 
give us the position of the dark lines. Gases are not all equally opaque. Sodium- 
vapour seems particularly opaque. The smallest amount of its vapour will show 
the D lines dark. Other gases, like hydrogen or helium, would probably require 
some miles of thickness before they would show their dark-line spectrum. 
Gases are also not equally opaque for all their lines, which is shown by the 
gradual increase in the number of lines as the depth of the absorbing layer is 
increased. Therefore, in a sun-spot where the source of the continuous spectrum 
is at a greater depth, we see the lines in the absorbing layer increased in number, 
and both widened and darkened. 
Now suppose the Sun’s temperature were raised enough to volatilize entirely 
the photospheric clouds, his brilliancy would fall off probably fifty per cent., 
and his spectrum would be darkened by the increased lines of titanium and other 
elements which now lie below the photosphere. The bands of carbon would 
be dark; and probably the lines of magnesium and other elements which lie high 
up in his atmosphere would show as bright lines in his otherwise darkened 
spectrum. In fact, the solar spectrum would be almost identical with that of 
1V-type and Wolf-Rayet stars. Brilliancy, therefore, is, per se, no criterion as to 
the relative temperature of stars. 
In Sirius, we have got a white star which Sir Norman Lockyer would class as 
at the highest point of the curve of temperature. Its spectrum is a brilliant, con- 
tinuous one, extending far into the ultra-violet, and only crossed by a few dark 
lines, principally those of hydrogen. This type of spectrum would naturally be 
caused by such a deep layer of carbon clouds that they would be quite opaque to 
