1118 
physically significant, and that neighbouring Fraunhofer lines really 
seem to repel each other. 
We have of course also considered the idea, that this phenomenon, 
although undoubtedly involved in the observational results, might 
be caused by systematic errors in the methods, applied in making, 
or judging, or measuring the photographs. 
The chance, however, for such errors to have appreciably affected 
the result, is very small especially in the case of limb-—centre dis- 
placements, because these displacements are usually derived from a 
comparison between photographs, the density and general appearance 
of which have been chosen as similar as possible. If, therefore, in 
estimating the distance between the members of a close pair of 
lines, a systematic error is made owing to their proximity, that 
error will be very nearly the same in the limb spectrum as in the 
centre spectrum, and will thus be eliminated in the limb-centre 
differences. 
As the distance between neighbouring lines is nevertheless found 
to be greater in the limb spectrum than in the centre spectrum, it 
is safe to say that influencing each other is a true property of 
Fraunhofer lines. This property seems only explainable from the 
point of view of the dispersion theory. | 
The skilful collaboration of Dr. van Crrrert and Dr. Minnauer 
in this research is highly appreciated. 
Utrecht, January 1921. Heliophysical Observatory. 
