EFFECT OF PRESSURE UPON ARC SPECTRA. 
133 
The question of the magnitudes of the displacements are discussed under a separate 
heading. 
3. The Reversal of the Lines .—In the region of the spectrum examined, several 
lines are reversed at ordinary atmospheric pressure, such are 63, 64, cO, dO, d 2, &c., 
but under increased pressure the reversals become stronger and other lines originally 
unreversed exhibit the phenomenon; such are 61, cl,dl,fl. It thus follows that 
the immediate effect of 'pressure is to increase the number and intensity of the 
reversals. But at a pressure of about 25 atmospheres for Set A, the maximum 
tendency to reverse is reached, and above 25 atmospheres the reversals decrease in 
number and generally in width. For Set B, the maximum intensity of the reversals 
occurs at 20 atmospheres; it may also be seen from. Table II. that the reversals are 
more numerous at 10 atmospheres than at 15 atmospheres. In Plate 5, fig. 1, in 
which the spectrum photographed at 100 atmospheres pressure is not exposed as 
much as the other spectra under pressure, the lines 63, e 3, e4, are reversed most 
strongly at 25 atmospheres; a careful examination of the original negatives under a 
high magnifying power is, however, necessary for the detection of the fine reversals 
of 61, cl, &c. A few lines remain strongly reversed even at the highest pressures 
reached, cO, dO, d2, el. 
Some reversals are symmetrical, i.e., the absorption line is in the centre of the 
bright emission line, see lines dO, cl2, Plate 6 , fig. 1 ; others are unsymmetrical, see 
line e3, Plate 6, fig. 2. A separate section, p. 152, is devoted to the discussion of this 
phenomenon. 
9. Measurement of the Photographs. 
The plates were measured on a Kayser measuring machine in which they are 
firmly attached to a table which travels on a fine screw in front of a fixed eye-piece 
containing the cross-wires; depressing a key prints the reading of the scale on a tape 
and enables several measurements to be made without moving the eye from the 
instrument. The shifts were measured by causing the line under pressure and its 
comparison line to travel in succession past the double wire of the eye-piece, the key 
being depressed as the most intense portion of each spectral line came between the 
double threads; the setting on the broadened portion of the line was generally made 
first, because it was then found easier to decide upon its most intense portion. In 
some few cases in which the line was abnormally broad, the film was pricked under a 
low-power magnifying glass which enabled the position of maximum intensity to be 
more easily gauged, and the position of this dot was then measured. Some of the 
plates were taken with the pressure spectrum within the comparison spectrum, but 
most of them in the reverse manner; no important differences were found between 
the readings in the two cases. For pressures above 60 atmospheres the second method 
was alone adopted. The astigmatism of the grating was purposely left uncorrected, 
because it was an advantage to have the extremities of the lines drawn out into 
