132 
MR. W. GEOFFREY DUFFIELD ON THE 
at pressures of 95 and 100 atmospheres when the arc is of too short duration for the 
formation of the necessary quantity of vapour. 
8. General Features of the Results. 
1. Th.e Broadening of the Lines. —It will he seen from the photographs accompanying 
this paper that the first- obvious effect of an increase of the pressure of the gas 
surrounding the arc is the broadening of the lines. 
Under a pressure of three atmospheres, Plate 4, No. 1, the widths of all the lines 
are appreciably increased, and under higher pressures this becomes still more marked. 
The width is too dependent upon the length of the exposure and the nature of the 
photographic process* for more than a general idea of the broadening to be obtained 
from the photographs, but it will be seen that the lines widen and become more diffuse 
as the pressure is increased, and that the amount of broadening is different for 
different lines; those that reverse strongly are spread out over a large range 
(e.g., lines marked 63, 64, cO, c/0, c/2, &c.), while others remain comparatively sharp 
(6l, 62, cl, c 2, c‘3, c/1, &c.), though these are not nearly as fine under high jiressure 
as they are under one atmosphere. It is in general true that the lines that are 
originally strong are the ones that are most broadened under pressure. 
The broadening of most lines is unsymmetrical, being greater towards the region of 
greater wave-length ; the following lines illustrate this well : 2, 3, c/3, f 1, f 2, Plates 4, 
7 and 8. Those that reverse easily, 63, 64, cO, c/0, c/2, &c., however, appear to broaden 
almost symmetrically, and so do the lines already mentioned as remaining compara¬ 
tively sharp at high pressure. In no instance has a line been found with a greater 
broadening on its more refrangible side. It should, however, be remarked that there 
is invariably some broadening towards the violet, but not always to the same 
extent as towards the red end of the spectrum. The phenomenon of broadening 
may be studied in Plate 5, fig. 1, in which, however, the exposures are not quite 
comparable. 
2. The Displacement of the Lines. —A careful study of the photographs, Plate 4, 
reveals the fact, first discovered by Humphreys and Mohler, that there is a slight- 
displacernent of the lines towards the red end of the spectrum as the pressure of the 
vapour about the arc is increased. At low pressures it is not quite definite that this 
phenomenon is not due to an unsymmetrical widening of the line, but the photographs 
taken at high pressures dispose of this objection, because several lines are then 
displaced so much that they are quite clear of their comparison lines. The lines f 1, 
f 2, /3, at 100 atmospheres, No. 9, Plate 4, show this well, and it is emphasized 
in the enlargement, Plate 8, and in Plate 6, fig. 1, in which the shift of c/3 at 
100 atmospheres is specially remarkable ; the lines 2, 3, al, Plate 7, though not 
displaced quite as much, also testify that the displacement is a real phenomenon. 
* 
The photograph at 10 atmospheres (No. 2 ) is over-exposed in comparison with the other spectra. 
