52 
The Light from the Stars [ch. ii 
from 3000° to 15,000°; according to Fowler and Milne the maximum activity 
occurs at 6290°. Other lines in stellar spectra provide the means of fixing 
other ranges of temperature. From a combination of all the evidence, it is 
possible to fix the temperature of the sun’s “ reversing layer ” at about 6000°, 
and to estimate the temperatures of other stars. 
Spectral Type. 
52. The circumstance that the H and K lines of calcium were specially 
prominent in the sun was at one time interpreted to mean that the sun was 
in some way especially rich in calcium. That three different stars shewed 
prominently the lines of, say, hydrogen, calcium, and titanium oxide, was 
supposed to indicate that the composition of a star changed as it aged, so 
that the atoms of which it was made changed, by transmutation of the 
chemical elements, from hydrogen into calcium, from calcium into titanium 
oxide, and so on. But it is now clear that the star’s spectrum reveals its 
surface-temperature rather than its chemical composition. If the sun’s tem 
perature were suddenly doubled, without any change in its chemical com 
position taking place, the calcium lines would almost disappear from its 
spectrum and hydrogen lines would take their place; if the sun’s temperature 
were halved, the calcium lines would again disappear and the solar spectrum 
would be dominated by the bands of titanium oxide. 
Two stars whose outer layers were at the same temperature and in the 
same physical state would shew different spectra if they contained different 
chemical elements, or even if they contained the same chemical elements in 
different proportions. A few cases are known in which the lines of certain 
elements appear in abnormal strength in the spectra of particular stars but 
apart from these, it appears to be a general rule that the elements appear in 
the same relative proportions in all the stars*. As a consequence stellar spectra 
fall approximately into a single continuous sequence, different positions in the 
sequence merely representing different temperatures of the star’s surface. 
It is not yet convenient to express a star’s spectrum as a temperature, 
although possibly this means of identification may come into use in time. At 
present a star’s spectrum is described by reference to arbitrarily selected 
standard points in the continuous series of spectra. As regards the spectra of 
normal stars, six selected points are designated by the letters B, A, F, G, K, 
M, in this order; the statement that a star’s spectrum is B 3 means that it 
is three-tenths of the way on from B to A. A star whose spectrum is exactly 
at B is generally said to be of type B0; a, star whose spectrum is described 
merely as being of type B may be anywhere from B0 to B 9. Type 0 comes 
in front of type B, its various sub-classes being temporarily designated as 
Oa, Ob, Oc, Ucl and Oe, Oe itself being further divided with ten sub-divisions. 
Typical spectra of the various classes are shewn in Plate VI facing p. 49. 
* Miss Paj'ne, Stellar Atmospheres, chap. xm.