158 
MR. W. GEOFFREY DUFFIELD ON THE 
Above 25 atmospheres there is little change in the relative intensities, though the 
increasing width of some lines and their diffuseness make them appear weaker. 
The line 4238’95 that has diminished in intensity under pressure is marked by an O 
on Plate 5. fig. 1, and it will be seen only in the spectrum under 1 atmosphere—at 
25 atmospheres it seems to have been obliterated. The spectrum taken at a pressure 
of 1 atmosphere is under-exposed, so the difference in relative intensity is greater than 
is apparent on the plate. The other lines that are weakened are marked W on 
Plate 5, fig. 1, and on Plate 8. The two lines fO and e2 do not appear to contribute 
anything to the wings of the symmetrically reversed lines el and e4. 
A third class of phenomena relating to changes in intensity is exhibited by the line 
which occurs in the spectrum under 20 atmospheres, Set A, as a strongly reversed 
line, whose corresponding emission line in the comparison spectrum cannot with any 
certainty be ascertained* (see Plate 6, fig. 3). 
The wave-lengths assigned to the enhanced lines are taken from Kayser and 
Runge’s tables. If any of them are due to impurities, the same impurity must have 
existed in the poles used by both Kayser and Runge and the writer.! 
20. Series of Lines in the Iron Spectrum. 
From the displacement curves we have overwhelming evidence of the existence in 
the iron spectrum of lines of different types, see Section 12, p. 139. The photograph 
reproduced in Plate 5, fig. 2, in which the lines belonging to Groups I., II., and III. 
are linked together, shows that the lines possessing the same shift have in general the 
same appearance, and that their broadening and manner of reversal are similar. 
T1 le lines of Group I., which are characterised by their strong reversals, are linked 
by a dotted line and are marked Group I.R. The unreversed lines of Group I. are 
marked Group I.U. 
The photograph only includes a portion of the region of the spectrum examined. 
Table IX. 
Group I. 
Group II. 
Group III. 
Displacement per atmosphere r Set A . 
in Angstrom units . . \ „ B. 
Ratio of shift.” p ' 
Broadening under pressure .... 
Type of broadening. 
Nature of reversals. 
0-0019 
0-0015 
1 
1 
Rather broadened 
Nearly symmetrical 
Symmetrica] 
0-0040 
0-0033 
2-1 
2 ■ 2 
Much broadened 
U nsymmetrical 
Unsymmetrical 1 
0-0078 
0-0068 
4-1 
4-5 
Yery broad and diffuse. 
Very unsymmetrical, 
almost a band. 
Very unsymmetrical. 
[* Note added December 1907.—Mr. H. F. Newall suggests that this may be due to the calcium line 
4227 produced by an impurity which has found its way into the outer layers of the arc. This explanation 
must, I think, be accepted.] 
f See note p. 138. 
