344 
THE OUTSIDE OF A STAR 
more excited atoms present. Excitation and de-excitation no longer 
balance, because the latter is increased proportionately to the greater 
number of atoms ready to be de-excited. Consequently the electron 
collisions on balance reduce the excited atoms without letting them emit 
their energy as radiation and tend to wash out the emission line. 
If the motion of the electrons is purely a temperature motion they 
cannot assist the formation of emission lines; but if they have extraneous 
velocities greater than the random temperature velocities (about 500 km. per 
sec.) then they may be able to perform more excitation than de-excitation. 
Currents in the stellar atmosphere could scarcely have sufficient speed. 
But a disturbed (cyclonic) state of the atmosphere might establish local 
and temporary electric fields—thunderstorms—under which the electrons 
would acquire high speeds, the free path at the low density being fairly 
long. If there is no other way out we may have to suppose that bright 
line spectra in the stars are produced by electric discharges similar to 
those producing bright line spectra in a vacuum tube, and depend entirely 
on disturbed conditions in the atmospheres of the stars which show them*. 
Explanation (2) seems to postulate a much wider deviation from 
thermodynamical equilibrium than we can admit in the stars. It requires 
that a bright line at v shall be accompanied by a dark line at iq, the latter 
representing the absorption which raises the emitting atom to, or above, 
the required state. But since no such effect occurs in thermodynamical 
equilibrium there is only the difference in strength of the iq radiation at 
temperature T e and T 0 respectively available for this extra excitation. 
Similar considerations apply to explanation (3), the dark line at jq being 
replaced by continuous absorption beyond the head of the principal series. 
Again, since the process of absorbing in certain lines and emitting in 
other lines does not alter the total energy in the lines, a bright line which 
has, say, 6 times the intensity of the continuous spectrum would (roughly 
speaking) have to be compensated by 5 wholly black lines—or more, if 
the dark lines are relegated to the unobservable part of the spectrum where 
the intensity is weak. 
* Another possible source of high-speed electrons is radio-active disintegration. 
In order to account for magnetic storms it seems necessary to postulate streams of 
negative and positive charges issuing from the sun. It is usually supposed that 
these are a and /3 particles emitted with high velocity from uranium, etc., though 
it is difficult to believe that the heavy radio-active substances rise sufficiently high 
in the photosphere to give free exit to the emitted particles. The cause of the emission 
may perhaps be radiation pressure (§ 254), or C. T. R. Wilson’s phenomenon (§ 210). 
The difficulty is to account for the escape of positively charged particles; unless 
charges of both signs are leaving the escape is immediately stopped by an electro 
static field. 
High-speed electrons will also be produced by the penetrating radiation (§ 223) 
which, whatever its origin, must be supposed to occur in stellar atmospheres as it 
does in our own atmosphere.