ASTRONOMY: ST. JOHN AND BABCOCK 
133 
attention was given it both in collecting and discussing our data. The 
equalization of the brightness of the two parts of the source by the ro- 
tating disk referred to above was always nearly complete except when 
the Hght from the pole was purposely made fainter than from the center. 
Under these latter conditions it was found that the displacements persist- 
ed, although if due to unsymmetrical widening they should disappear. 
Furthermore, a set of typical lines showing displacements in both 
directions were observed with two different forms of microphotome- 
ter, both of which confirmed the conclusion that the maximum of the 
line is actually displaced at the pole as compared to its position at the 
center of the arc. With either of these instruments the location of the 
maximum of a hne is independent of any bias on the part of the observer. 
It is true that the shifts are often accompanied by considerable dissym- 
metry, though there are some exceptions. 
An examination of our data brings out a partial correspondence 
between the effects observed at the negative pole and those known to 
be due to increase of pressure around the source. But we find a large 
number of Hnes whose wavelengths are unaffected at the negative 
pole, which should show an easily measurable increase in wavelength 
if there were a general increase in pressure near the pole sufficient to 
account for the observed displacements. This would suggest the prob- 
ability that some other agency than pressure is effective, but the possi- 
bility remains of a local increase in pressure affecting only the inner- 
most portion of the vapor in the vicinity of the pole. To determine 
the role played by pressure a fuller knowledge is required of its effect 
upon the wavelengths of these affected lines, a subject now under in- 
vestigation at this observatory. A more complete discussion of this 
question and of the possible effect of density as distinguished from 
pressure, together with complete lists of affected lines, will appear in 
our more extended paper soon to be published in the Astrophysical 
Journal. 
In addition to the observations discussed above, we have measured 
the relative intensities in different parts of the arc of a few typical lines 
in the green part of the iron spectrum. For this purpose still higher dis- 
persion and resolving power were employed in the spectrograph, and the 
slit was placed parallel to the axis of the arc. As precautions against 
the numerous sources of error in such observations, may be mentioned 
the impression of a photometric scale upon each photographic plate 
used, the avoidance of overexposure, etc. The enlarged image of the 
arc and the long sKt of the spectrograph permit an excellent analysis of 
