238 CARNEGIE INSTITUTION OF WASHINGTON. 



components, respectively, by the Nicol and compound quarter-wave 

 plate. The large percentage of undetermined cases for the Ha whirl 

 arises from the fact that small spots oftentimes show no vortex struc- 

 ture, although the polarity is easily determined. 



It thus appears that while the magnetic polarities of the spot-fields 

 reversed at the minimum, the direction of whirl in the overlying 

 hydrogen vortices did not reverse. This seems to be a fatal objection 

 to the hypothesis that the configuration of the hydrogen vortices is 

 caused by the magnetic field of the spot, which might impel mo\dng 

 ions to arrange themselves along the hnes of force. The vortex 

 experiments undertaken last year also oppose the electromagnetic 

 explanation, and point to the hydrodynamic hypothesis. But this 

 must take into account the fact that the direction of whirl in the 

 low-lying spot-vortex apparently reversed without changing the sign of 

 the secondary vortex above. This would be quite possible if the effect 

 of the spot on the overlying atmosphere were confined to a vertical 

 suction, leaving the sign of the secondary vortex thus induced to be 

 determined by the conditions existing at its higher level. As the 

 rotation law at this level is essentially the same before and after the 

 minimum, the sign of the whirl should remain constant. 



The cause of the reversal of spot polarities at the minimum still 

 remains a mystery, and the possibility that the sign of the charge, 

 rather than the direction of rotation in the vortex, is reversed must not 

 be wholly ignored. We have no evidence pointing to such a con- 

 clusion; indeed, we have as yet no direct evidence of the existence of an 

 electric charge in the spot-vortex. 



THE STARK EFFECT. 



As stated in previous reports, the character of the lines in sun-spot 

 spectra offers no apparent reason to hope for the detection of such phe- 

 nomena as the presence of an electric field might induce. The hydrogen 

 lines, which are enormously widened and split into many components 

 by the electric field within a vacuum tube, are not widened, but actually 

 narrowed, in sun-spot spectra. It is true that they represent a 

 comparatively high level, and that lines originating at greater depths 

 should better serve our purpose. But for most of these the Stark 

 effect, if any exists, is unknown. The spot triplets certainly offer 

 nothing of promise. With a Nicol and quarter-wave plate we can cut 

 off completely their ellipitically polarized n-components, leaving a 

 sharply defined p-component which is narrower than the corresponding 

 line in the solar spectrum beyond the Umits of the spot. Since no 

 circular or elliptical polarization has yet been found for the components 

 of a line split up by an electric field, the narrow p-component must 

 presumably contain any evidence that exists for the Stark effect; 

 and yet we find it narrowed rather than widened. 



