ASTRONOMY: C E. ST. JOHN 
23 
ments in any single region; but the data are not sufficient to obtain 
the relation between the absolute values for the separate regions to the 
same degree of precision, depending as they do upon varying obser- 
vational conditions. 
To obtain residuals capable of representing any systematic behavior 
of lines within the refractive influence of stronger lines, the standard 
displacements should be derived from the limited homogeneous series 
of observations in which they are to be used and, as far as the data per- 
mit, they should refer to the particular element under consideration. 
Such standard displacements have been obtained by determining the 
mean displacements for the lines of each element of a given intensity for 
each of the homogeneous series. 
With these standards the displacements of 131 Fraunhofer lines within 
0.5 A of stronger ones are compared, 67 on the violet side of stronger 
lines and 64 on the red side. The hypothesis of mutual influence re- 
quires that for the Hnes to the violet the mean residual be negative and 
for the lines to the red, positive. The result of the comparison ap- 
plied to this large number of lines is that 67 lines to the violet give a 
mean residual — 0.0003 A, and 64 lines to the red give a mean residual 
- 0.0004 A. 
Stated in another form, the sum of the favorable residuals is 0.212 A; 
of the unfavorable, 0.218 A. Tested, then, by standards deduced from 
the homogeneous series of observations to which they are to be applied, 
the lines under the influence of stronger ones show no systematic difference 
within the limits of measurement from hnes not so situated. 
Another criterion is whether the effects produced by the influencing 
lines depend upon their power to produce anomalous dispersion in the 
laboratory. A comparison of the residuals given by lines under the 
influence of those having strong anomalous dispersion with the residuals 
given by others under the influence of lines having very weak anomalous 
dispersion shows no greater effect in one case than the other. 
It would seem that the effect of mutual influence must increase with 
nearness of the influenced line to the controlKng line. The change 
should then be especially manifest for those in the near neighborhood 
of such powerful lines as H, K, and 4226.9 of calcium and the strong 
aluminum lines, all of which show large anomalous dispersion in the 
laboratory and are bordered by broad shadings or wings in the solar 
spectrum. From the point of view of the anomalous dispersion theory, 
these wings are considered dispersion bands. In making the original 
series of measures, the lines near these very strong lines were in general 
omitted from the observing list, as they present special difficulties. A 
