NICKEL UNDER MODERATE PRESSURES. 
367 
A few iron lines occurring as impurity in the nickel have been measured with a 
view to comparing the shifts with the determinations of Gale and Adams, and of 
Duffield. 
The following results (Table III.) have been obtained for six lines particularly 
suitable for measurement. 
Table III. 
Wave-length. 
o 
Mean shift per atmosphere in A.U. 
Gale and Adams 
(A rc). 
Duffield 
(Arc). 
Bilham 
(Spark). 
4045-98 
2-9 
3-0 
3-6 
4063-76 
2-5 
2-5 
3-2 
4071-91 
2-6 
2-6 
2-3 
4308-08 
2-6 
2-1 
4-3 
4325-94 
2-5 
2-3 
3-9 
4383-72 
3-4 
2-3 
4-2 
Means . . 
2-8 
2-5 
3-6 
The mean of all Duffield’s values, set B, # has been taken, assuming proportionality 
of displacement to increase of pressure. The values in the second column have been 
obtained from Gale and Adams’ main series of measurements at nine atmospheres.! 
It will be observed that the values given in the last column are, on the average, 
higher than the other two. This is to be expected from the fact that all the lines 
referred to are reversed in the arc under eleven-atmospheres pressure, but bright in 
the spark under the present experimental conditions. It has been found | that if a 
line is reversed at a given pressure its shift is very much smaller than when it is 
unreversed. 
8. Classification of Shifts. 
As the measured lines belonging to each class are mostly comprised between 
somewhat narrow limits of wave-length, no attempt has been made to investigate the 
law of variation of displacement with wave-length. Consequently the shifts have not 
been reduced to a standard wave-length, which is desirable in discussing their relative 
magnitudes. 
Some light is, however, thrown on this subject by averaging the shifts of lines 
belonging to the same class. The results are given in Table IV. The line 4121'48, 
doubtfully due to nickel, has been omitted in making the calculation. 
* ‘Phil. Trans.,’ A, vol. 208, p. 138. 
t Loc. cit., p. 17. 
t Duffield, loc. cit., p. 155. 
