THE OUTSIDE OF A STAR 
369 
be doubly ionised (by starlight if not by sunlight) ; it will then drop back 
since it has small chance of picking up another electron when outside 
the chromosphere. 
Milne has also pointed out that a calcium atom receding from the sun 
under radiation pressure experiences a rapidly increasing acceleration. 
Owing to the Doppler effect of its growing velocity its individual absorp 
tion shifts farther and farther from the centre of the solar H and K lines. 
At first it is, so to speak, balanced on the summit of the dark lines; but 
it topples off into a clear region and experiences the full force of undimmed 
solar radiation. 
Abundance of the Elements. 
255. In Saha’s early researches the spectral types at which a line first 
appears and finally disappears were connected with the physical conditions 
at which the proportion of atoms in the right stage of ionisation becomes 
appreciable. Owing to the difficulty of assigning numerical significance 
to the word ‘'appreciable,” Fowler and Milne preferred to work with 
maximum intensity rather than marginal appearance. But just because 
the marginal appearance involves extraneous factors and is unsuitable 
for the main purpose, it may yield interesting information as to these 
factors. In particular it involves the abundance of the element since ceteris 
paribus if the atoms are ten times more abundant the required proportion 
in the proper state of ionisation is ten times smaller. 
Although the reservation ceteris paribus covers a multitude of individual 
peculiarities of the elements and their spectra, a first clue to the relative 
abundance may be obtained on the hypothesis that the number of atoms 
required to give a spectrum at the limit of visibility is the same for all 
kinds of atoms. 
This hypothesis is not so wild as we might suppose at first. Observation 
is limited to a rather narrow range of spectrum so that the energy-constants 
are roughly the same for all the lines studied. The absorption coefficient 
is mainly determined by the time taken to relapse from an excited state; 
this in turn is supposed to be connected with the classical radiation in the 
excited orbit which cannot be greatly different for the different orbits 
producing optical spectra. The series of lines in the spectra of the same 
element are of widely different intensities, and in comparing the abundance 
of elements by this method care must be taken to choose comparable lines 
of comparable importance as representative of the spectrum. If the 
element is represented only by weak lines in the visual region we might 
be misled. 
This method of calculating abundance of the elements is due to C. H. 
Payne and the results in Table 47 are taken from her book*. They corre- 
* Stellar Atmospheres, p. 187.