PHYSICS: ST. JOHN AND BABCOCK 
297 
In view of the absence of a general increase in pressure at the nega- 
tive pole, it is probable that some agency other than pressure is in- 
volved, but there is the possibility of a local increase in pressure affect- 
ing only the innermost portion of the vapor in the vicinity of the pole. 
To obtain Hght upon this point, a determination of the pressure shift 
per atmosphere has been made for a large number of the lines imder 
consideration. Their wave-lengths in vacuo and at normal pressure 
have been compared, the light in each case being taken from the cen- 
tral section of the arc. These conditions are well suited to a measure- 
ment of the pressure shift for Hues of this type, as they are of good quality 
in both spectra. In a paper by St. John and Ware^ they said: 
Neither the small pressure-changes of about one-fifth of an atmosphere 
taken advantage of in this investigation, nor the high pressures used by 
Gale and Adams are well adapted to the study of lines of this type, and it 
is purposed to examine in vacuo and under normal pressure the behavior of 
an extended list of lines belonging to groups d and e. 
If the pole effect is due to pressure alone, the connection between it 
and wave-length should be similar to that relating pressure shift and 
wave-length. A brief summary of our results is given in the accom- 
panying table. 
Pole Effect and Pressure Shift Relative to Wave-length 
Group 
No. of Lines 
Mean X 
Pressure Shift 
per Atm. 
Pole Effect 
d 
25 
4085.14 
-F0.0048A 
+0.0096A 
c5 
16 
4766.41 
-f-0.0093 
+0.0119 
d 
12 
5528.44 
+0.0089 
+0.0206 
d 
5 
6350.74 
+0.0160 
+0.0185 
The three sections of group d show pressure displacements which vary 
as the cube of the wave-length, a result in harmony with the observa- 
tions of Gale and Adams. ^ On the other hand there is no relation appar- 
ent between pole effect and wave-length. If we attempt to determine 
the difference in pressure between the center of the arc and a point in the 
core near the negative electrode by comparing the last two columns 
in the table, we find 2.0; 1.3; 2.3; and 1.1 atmospheres, numbers whose 
discrepancies exceed the errors of observation. Such a treatment of 
the data assumes that the pressure in the core of the arc at the negative 
pole is the same for all wave-lengths. The increase of pressure, if it 
exists, must be produced by rapid vaporization and the almost explo- 
sive expulsion of the emission centers, which are not necessarily the 
same for all wave-lengths. Investigations are now to be undertaken 
in order to obtain some measure of the velocities of the emitting particles. 
