156 
ME. W. GEOFFREY DUFFIELD ON THE 
for future work; but the effect of any such interaction would be the loading of the 
absorbing metallic systems, and the consequent increase in the wave-length of the 
light they absorb, whereas the opposite effect has been observed by Anderson, # who 
measured the displacements of some spark lines under pressure in atmospheres of 
carbon dioxide and hydrogen, and obtained larger values when the denser gas was 
used, but it is possible that the effect may be less obvious in the arc than in the spark, 
since in the former the surrounding gas will not presumably influence so strongly the 
nature of the discharge. 
Differences in physical properties (density, temperature, specific inductive capacity, 
&c.) between the absorbing layer and the core of the arc, as well as the possible 
influence of abnormal dispersive properties of the surrounding vapours, have also been 
suggested as causes of this phenomenon. There is also considerable evidence in 
favour of the existence in the gaseous envelopes of a modified system derived from 
the iron atom. 
18. The Order of Reversal. 
In their study of the spectrum obtained from a spark discharge in liquids and gases 
under pressure, Hale and Kent! noticed that the reversals appeared first at the violet 
end of the region observed, 3550 to 4500 A.U., becoming more intense and gradually 
extending towards the red end as the pressure about the spark was increased. 
In the case of the arc spectrum of iron within the range of wave-lengths 4000 to 
4500 A.U., this phenomenon does not make its appearance between pressures of 1 and 
101 atmospheres (absolute), the intensities of the reversals at the more refrangible 
end of the plate bearing very nearly the same ratio to the intensities of the reversals 
at the red end for all pressures. 
A recent research]; affords a possible explanation of the difference between the results 
obtained by Hale and the present writer on the hypothesis that different tempe¬ 
rature and radiation gradients exist in the arc and spark. These are shown to be 
determining factors in the reversals of lines when the same amounts of vapour are 
present—the steep gradient of the electric spark being more favourable than that of 
the arc for the production of this phenomenon. 
19. Lines Enhanced and Weakened under Pressure. 
The majority of lines continue throughout the range of pressure with nearly the 
same relative intensities that they possess at 1 atmosphere, but a few undergo a 
change and are enhanced as the pressure is increased, while others appear as weakened 
lines under high pressures. 
* Anderson, ‘Astrophysical Journal,’ XXIV., 221, 1906. 
f Hale and Kent, ‘Astrophysical Journal,’ XVII., 154, 1903; also Hale, ‘Astrophysical Journal,’ 
XV., 227. 
| A. Schuster, “Radiation through a Foggy Atmosphere,” ‘Astrophysical Journal,’ XXI., b 1905 
