NICKEL UNDER MODERATE PRESSURES. 
309 
line (see Plate U fig. 2c). Under the microscope the reversal shows a pronounced 
o 
displacement towards the violet, amounting to 0'023 A.U. at eleven atmospheres and 
o 
0'035 A.U. at six atmospheres. As this was the only violet shift observed, further 
investigation seemed desirable. It was found that the spark line showed a displace¬ 
ment of 0*035 A.U. towards the red with respect to the corresponding line in the arc 
(Plate 7, fig. l). This fact, and the anomalous behaviour under pressure, may be 
co-ordinated if we suppose that in the spark spectrum the line is really a doublet 
consisting of the arc line X 35 14TO together with an enhanced line, the mean wave¬ 
length of the doublet being 3514T0 plus 0*035. Under pressure, we should expect the 
arc line to reverse and the enhanced line to broaden without reversal, a conclusion in 
entire agreement with the appearance actually observed. Moreover if the pressure 
o 
displacement of the arc line was less than 0'035 A.U., the reversal should exhibit a 
violet displacement with respect to the normal spark line. On this theory the arc 
line behaves in a perfectly normal fashion, giving a red displacement of 0*035 — ‘023, 
o 
i.e., 0*012 A.U. at eleven atmospheres. 
10. Abnormal Shift of Line 3609. 
A somewhat similar example is furnished by a line of wave-length 3609*02 in the 
arc (possibly due to iron). This line gives a very narrow reversal under pressure 
o 
which shows an abnormally large shift (0*055 A.U. at eleven atmospheres) towards the 
red. (Plate 7, fig. 4.) 
A comparison of the arc and spark spectra (Plate 7, fig. 3) shows that the spai’k 
o 
line is displaced towards the violet by 0*040 A.U. with respect to the arc line. 
Assuming as before that the spark line is really a doublet, the more refrangible 
component of which is, in this case, an enhanced line we get a true shift of 0*015 for 
the arc line, which is of about the right order. 
11. Qualitative Observations. 
A large number of lines have been investigated qualitatively with a view to 
observing their behaviour under pressure. In some cases, where the lines are very 
close together, such as in the doublet 4732*00 and 4732*66, it is impossible to examine 
the behaviour of the individual lines. Such groups have therefore, been bracketed 
together and treated collectively. 
3 B 
VOL. CCXIV.-A. 
