( 595 ) 



i'cspondiim- to fhroiiiospliore lines a\ ill (liorofore havo a more or less 

 (lai-keiied hackgroiiiKl in the ordinai-v sohir s|)e('lrnin. The rate of 

 darkeninii' al \arions distances from the cenlre of an absorption line 

 is, of eoni'se, connected with the siiape of the dispersion curve near 

 that Hne: whereas the axerage shading depends l^' on the quantity of 

 matter causing aju>makjus dispersion an<l 2 "'b' on the slo|)es and the 

 directions of the density gradients in (he gases through which the 

 ligiit is transmitted, viz. on the Sun's "activity" ^). 



We distinguish, therefore, a twofold origin of the dark lines in 

 the solar spectrum: real absorption of those waves, exactly cor- 

 responding to the pei'iods of the media, and dispersion of die 

 strongly ileviated neighluniring light *j. 



The dispersion will be especially evident where extraordinary diffe- 

 rences in the density of the medium occur ; in this way the widening 

 of most of the Fraunhofer h'nes in the spectra of spots may be 

 accounted for. 



Dispersed bght has not, of course, vanished; the absence of certain 

 rays in the spectrum of a spot is counterbalanced by the increased 

 intensity of the same raiHations in the light coming from the neigh- 

 bouring faculae. Thus the distribution of the density in the solar 

 gases may locally be such, that a limited part of the disk seems to 

 emit a consideral)le amount of rays with abnormally high or abnor- 

 mally low refractive indices. In the spectrum of sucii parts not only 

 will the Fraunhofer lines show nairower and fainter than usually, but 

 here we may even meet with lines contrasting brightly with their 

 surroundings. These l)right lines will not coincide Avith the c(n're- 

 sponding absorption lines ; their average wave-length will in general 

 be greater or smaller than that of the absorbetl light, for, according 

 to the accidental distribution of the density, xve shall find either the 

 rays with high or (iiose with low retVacliw incHces most prominent 

 in the beam. 



The above considerations suggest an explanation of Hale's abnormal 

 spectrum. 



Iji fact, the lines showing especially faijit in this s[>eclrum were 

 exactly those, causing strong anomalous dispersion — witness the 



1) The possible inlUience of the general or regular ray-curving (after Schmidt's 

 principle) on the feature of the spectral lines has, in the present paj)er, been left 

 out of consideration. If we were able to observe or to calculate the radii of the 

 "critical spheres" for radiations undergoing anomalous refraction, it would be 

 possible to estimate that influence ; but as yet sufficient data are wanting. 



2) Proc. Hoy. Acad. Anist. II. p. 580. 



