( 217 ) 



The following general statement is borne out by these experiments. 



If an illuminated absorbing vapour is investigated by means of 

 the spectroheliograph, and the camera-slit of the instrument is set 

 on the edge of a dispersion band , marked irregularities in the 

 brightness of the field will only appear at those places in the image 

 which correspond to regions with large density gradients in ilie 

 vapour. Setting the slit nearer to the middle of the dispersion band, 

 we shall get evidence, in the image, also of regions with smaller 

 gradients, a.s.o. Particulars regarding the distribution of a vapour 

 are thus clearly shown by the spectroheliograph through anomalous 

 refraction, even in cases, where the absorbing or emitting power of 

 that medium would have failed to reveal its structure. 



The bearing of these inferences on astrophysical phenomena has 

 now to be considered a little more closely. 



Suppose we have a large mass of absorbing vapour of such 

 average density, that, if it were uniform, its absorption lines would 

 appear rather narrow; and of such temperature and condition of 

 luminescence that its emission lines are very faint. As soon as the 

 density of this mass becomes irregular, some parts of it may give 

 rise, when traversed by light from another source, to the appearance 

 of dark or bright dispersion bands, greatly exceeding in width and 

 strength its absorption or emission lines. 



It is therefore possible, that anomalous refraction plays a very 

 essential part in the production of those phenomena which the 

 student of astrophysics observes with his spectroscope or spectro- 

 heliograph; we must inquire how far this is also probable. 



One might be inclined to object, for instance, that in our experi- 

 ment the use of a narrow and sharply limited source of light, 

 placed at a fair distance behind the vapour, seemed to be a necessary 

 condition for obtaining any marked dispersion effects, and that in 

 the sun similar circumstances are very unlikely to prevail. Indeed, 

 the bulk of the sun — whatever the nature of the photosphere may 

 be — is a large incandescent mass, closely surrounded by the absorbing- 

 vapours, so that the "source of light", if considered from a point 

 of the chromosphere, subtends a solid angle of nearly 2.t. The 

 reversing layer and the chromosphere have sometimes been compared 

 to a thin, transparent layer of selectively absorbing varnish, covering 

 a luminous (e.g. phosphorescent) globe: the photosphere. It seems 

 very improbable that refraction in density gradients of such a trans- 

 parent envelope should be able to disturb to any perceptible degree 

 the uniform brightness of that globe. 



