66 Prof. W. A. Norton on the Physical Constitution of the Sun. 



perceive that the latter supposition should be true, since the 

 solar repulsion should be in operation above the hydrogen 

 envelope or the chromosphere, diminishing the gravitating ten- 

 dency. 



The metallic vapours set free in the region of dissociation 

 should rise to heights varying with their atomic weights. Some 

 of them, especially the lighter ones (sodium, magnesium, cal- 

 cium, &c), may acquire velocities sufficient to bring them above 

 the chromosphere. In fact, the spectroscope has detected, 

 besides hydrogen, magnesium, sodium, iron, and chromium in 

 the solar protuberances. Lockyer states that he has invariably 

 found that in solar storms the chromospheric layers are thrown 

 up in the order of vapour-density. He regards the chromo- 

 sphere as built up of the following layers, which are in the order 

 of vapour-density in the case of known elements : — a new ele- 

 ment giving the green coronal line in the spectroscope, hydrogen, 

 another new element, magnesium, sodium, barium, iron. He 

 remarks that " all the heavier vapours are at or below the level 

 of the photosphere itself"*, 



The green coronal line was traced in the late eclipse by Pro- 

 fessors Young and Winloek as far as 16', or 425,000 miles from 

 the sun's limb. From our present theoretical stand-point we 

 naturally infer, as Lockyer has already done from his observa- 

 tions, that the element present in the solar corona which gives 

 this line is much lighter than hydrogen. We see also that 

 an element several times lighter than hydrogen might be sub- 

 ject to a solar repulsion that would predominate over the at- 

 traction of gravitation at all distances, and urge the subtle 

 vapour indefinitely away from the sun. Since the same line is 

 seen in the light of terrestrial auroras, we must conclude that 

 the same substance is present in our upper atmosphere, either 

 in a permanent upper layer or derived from the sun (as I have 

 elsewhere maintained). We must infer also that it is magnetic, 

 which apparently cannot be the case unless it takes on the con- 

 dition of compound molecules. Such compound molecules 

 might become dispersed in the upper atmosphere of the earth, 

 or in the photosphere of the sun, by electric discharges or sudden 

 evolutions of heat, and then the separate atoms repelled off, form- 

 ing the streamers of the corona and aurora, illuminated either by 

 electric light or, in the case of the corona, in part also by reflected 

 solar light f. 



o 



* It is admitted by Angstrom and Zollner that the absence of spectro- 

 scopic indications of oxygen and nitrogen in the sun is no sufficient evi- 

 dence that these gases are really wanting in the sun's atmosphere. 



t It is also conceivable that the subtle vapour streaming off in the co- 

 ronal rays has been set free by dissociation, like the hydrogen, from some 



