( 219 ) 



Waves, suffering anomalous refraction, will of course be much 



more scattered by the same medium. Let us consider an absorbing 



substance which, al a certain level, occupies say only 1 percent of 



the solar atmosphere, taken as a perfect mixture, its density gradients 



1 

 will then be only — — of those of the mixture. The refraction constant, 



on the other hand, for waves near one of its absorption lines, may 

 attain values as high as 1000 or 2000. With R = 1600 (as actually 

 observed in sodium vapour, Proc. Roy. Acad. Amst. IX, p. 353), 

 our equation (1) becomes 



1 c/A 1 



100 ds 1600 q 

 In a level where, in our image, the irregular density gradients of 



the envelope were supposed to have an average value — =16, 



ds 



the equation gives 



q = 0,004 cm. 



It is evident that under such circumstances rays may easily deviate 

 90 degrees and more in the thin shell of transparent matter covering 

 our globe, and thus give rise to a very unequal distribution of' the 

 light in photographs of it, secured with the spectroheliograph. 



This conclusion holds just as well with regard to the real sun. 

 It follows directly from our only assumption, that in some level of 

 the sun irregular density gradients exist, comparable in magnitude 

 with the vertical gradient in the earth's atmosphere. At lower levels, 

 greater gradients, at higher levels, smaller gradients may be expected 

 to prevail. As the validity of this assumption can hardly be doubted, 

 we may infer that the existence of some important influence of 

 anomalous dispersion on astrophysical phenomena is not merely 

 possible, but exceedingly probable, in spite of the absence of narrow 

 slits as sources of light. 



Although we are free to admit that the phenomena, observed with 

 the spectroheliograph on the solar disk, are perhaps in part due to 

 selective radiation, dependent on various conditions of temperature 

 or luminescence, we may nevertheless inquire into some consequences 

 to which one is led if only the effects of refraction in a mixture oi 

 vapours are considered. 



The composition of the solar atmosphere cannot be the same at 

 all levels. As we get lower down, the percentage of heavier mole- 

 cules is likely to increase; but we should not presume too much 



15 



Proceedings Royal Acad. Amsterdam. Vol. XI. 



