DIFFUSE MATTER IN SPACE 
379 
un-ionised calcium and presence of un-ionised sodium, since the ionisa 
tion potential for sodium is lower than for calcium and we should naturally 
expect the sodium to become ionised first in accordance with (174-2). But 
the expectation applies only when no second stage ionisation of either 
element is occurring, and owing to the low second potential for calcium 
the reversal might even have been anticipated. The following explanation 
has been given by R. H. Fowler. 
The valency electrons of an element are detached by fairly low ionisa 
tion potentials; after they have been removed there is a sharp rise to the 
next ionisation potential required to detach an electron of the compact 
inner group. We may suppose that the stellar radiation traversing space 
contains constituents of sufficiently high frequency (emanating chiefly 
from the B and 0 stars) to ionise the valency electrons, but practically 
no constituents of the much higher frequency required to break into the 
next group. At the low density electron captures are rare; consequently 
the atoms are usually in the state of having lost all their valency electrons 
but with their inner groups complete. These compact ions give no absorp 
tion in the visible region since their principal lines are far in the ultra 
violet. From time to time one of them will capture an electron and hold 
it for some considerable time*, since the ionising radiation is very weak; 
thus the next most common state is that of an ion with one valency electron. 
Since calcium is divalent and sodium monovalent, this places singly 
ionised calcium and neutral sodium on the same footing as regards chance 
of absorbing. Neutral calcium would only occur through the improbable 
coincidence of two electron captures in quick succession, so that its spectrum 
is not to be expected. 
259. The question now arises whether, granting that there is a cloud 
extending through interstellar space, the fixed lines are produced uni 
formly during the transit of the light from the star to the earth or only 
in the neighbourhood of the star. 
The reason for wishing to restrict the production of the lines to the 
neighbourhood of the star is that, if the star is absolved from all com 
plicity, the phenomenon should be the same for all types of stars, the 
intensity of the fixed lines depending only on the distance traversed by 
the light. Since the phenomenon is detected only in the hottest stars of 
types 0 to B 3 the idea has been put forward that the presence of an 
intensely hot star stimulates the part of the cloud in its neighbourhood 
to perform the absorption. This would explain admirably the fixed calcium 
lines, the calcium being ionised by the high-frequency radiation and so 
brought to a condition to absorb the H and K lines. But the fixed sodium 
* I.e. time enough for it to be excited a number of times, since many quanta 
of excitation frequency will come along before a quantum of the ionisation frequency 
arrives.