224 CARNEGIE INSTITUTION OF WASHINGTON. 



Confining our attention, therefore, to single spots, and assuming, as 

 already stated, that two superimposed vortices are involved in each, 

 we may provisionally represent their structure in three different ways: 



First, as presented by Evershed and St. John, who assume that the 

 spot-vortex extends well into the photosphere, the circulation being 

 spirally upward below the surface and then radially outward from the 

 center of the spot above the penumbra through the lower reversing 

 layer. A secondary vortex, with inward flow at high levels, is set up 

 in the chromosphere by suction resulting from reduced pressure over 

 the spot-vortex. The inward and outward flow of the spot-vapors, 

 observed spectroscopically above and below the level of velocity in- 

 version, do not represent the phenomena corresponding to the upper 

 and lower extremities of a dumb-bell vortex, but do indicate the exist- 

 ence of two superposed vortices, in which the directions of whirl are not 

 necessarily the same. The arguments supporting this hypothesis are 

 fully set forth by St. John in Mount Wilson Contributions Nos. 69 

 and 74. 



In the second hypothesis, the spot-vortex, of the dumb-bell type, lies 

 for the most part above the photosphere. The inflow along its lower 

 face is made visible by the penumbral filaments in the spot, while the 

 outward flow near its upper face is shown by the Evershed effect in the 

 lower reversing layer. The secondary vortex induced in the overlying 

 chromosphere is of the same type as in the first hypothesis. 



In the third hypothesis the spot-vortex lies chiefly below the photo- 

 sphere, and its suctional effect at the surface is rendered visible by the 

 structure of the penumbral filaments. The secondary vortex induced 

 in the overljdng chromosphere is of the dumb-bell type, the inflow 

 in the upper chromosphere sweeping spirally downward and then nearly 

 radially outward through the reversing layer. 



In dealing with all of these hypotheses, both the hydrodynamic and 

 the electrodynamic influence of the spot-vortex on the charged particles 

 in the overlying solar atmosphere must be given adequate considera- 

 tion, and the phenomena of bipolar spots must not be overlooked. 



MOUNT WILSON MAP OF THE SUN-SPOT SPECTRUM. 



In discriminating between rival sun-spot hypotheses, some of the 

 most important criteria are to be found in the phenomena of the mag- 

 netic field. It thus becomes necessary to give special attention to the 

 peculiarities of the Zeeman effect as exhibited by thousands of lines in 

 the sun-spot spectrum. To facilitate a general study of these peculiari- 

 ties, a photographic map of the sun-spot spectrum, on a scale of 1 cm. 

 to the angstrom, has been prepared by Mr. Ellerman from enlarge- 

 ments of negatives made in the second order of the 75-foot spectrograph 

 of the 150-foot tower telescope. This covers the region X3900-X6600, 

 and is on a sufficientl}^ large scale to bring out the details of many trip 



