19 14.] HALE— SOLAR MAGNETIC PHENOMENA. 253 



vapors within it, thus providing the separation called for in the 

 hypothesis. As for the existence of the vortex, in a form different 

 from that first assumed, it is abundantly confirmed by the discovery 

 of Evershed, and the subsequent observations of Evershed and St. 

 John on the motion of vapors in the solar atmosphere surrounding 

 spots. 



Assume the axis of the vortex to coincide approximately with 

 a solar radius. Then, if the spot were central on the sun, the lines 

 of force at its center would lie in the line of sight. Such an iron 

 line as A 6302.709, which is resolved by a magnetic field into three 

 components, should then appear in the spot as a doublet, the central 

 component being absent when observed along the lines of force. The 

 two outer components should be circularly polarized in opposite 

 directions, and it should be possible to extinguish either one at will 

 with the aid of a Nicol prism and quarter-wave plate. Furthermore, 

 two spot vortices rotating in opposite directions should show the 

 opposite components of the line, with the same adjustment of the 

 polarizing apparatus. 



This test was successfully applied, and has since been repeated 

 on many sun-spots. Under the most favorable conditions, either 

 component can be completely extinguished. In general, however, 

 the observations cannot be made exactly along the lines of force, and 

 under such circumstances the elliptically polarized components are 

 not completely cut off. Moreover, such a line as A 6302.709 usually 

 appears as a triplet, the relative intensities of the central and side 

 components varying, as would be expected from the Zeeman effect, 

 with the angle between the lines of force and the line of sight. 



Speaking generally, this angle should increase as the spot ap- 

 proaches the sun's limb. We should therefore be able in this case to 

 distinguish the phenomena of plane polarization, since in the labora- 

 tory the three components are plane polarized when observed normal 

 to the lines of force. As the central component is polarized in a 

 plane at right angles to the plane of polarization of the side com- 

 ponents, it should be possible to extinguish this line in the spectrum 

 of a spot near the limb by rotating the Nicol prism, used with- 

 out quarter-wave plate. This experiment has been successfully 

 performed. 



