SOLAR VORTICES AND MAGNETISM IN SUN SPOTS ABBOT. 329 



Investigations on tlie spectra of iron, manganese, cliromiuni, titanium, vana- 

 dium, and other metals conspicuous in spots, made witli the arc, sparli, and 

 flame, indicated tliat the cliange of tlie relative intensity of lines observed in 

 passing from the solar spectrum to tlie spot spectrum is due to a reduction of 

 tlie temperature of the spot vapors.*^ Subsequent work vi'ith a new electric 

 furnace by Doctor King,^ tlie details of which have not yet been published, 

 seems to leave little doubt that this explanation is correct. It is supported by 

 the presence in the spot of compounds which appear to be dissociated at the 

 higher temperature outside the spot, and by the resemblance of spot spectra 

 to the spectra of red stars." 



While' our investigations have thus furnished a plausible explanation of S(mie 

 of the characteristic phenomena of sun-spot spectra, the widening of lines and 

 the presence of doublets are among the remaining peculiarities that demand 

 consideration. As we have seen, however, these very peculiarities are precisely 

 what would be expected if a magnetic field were present. Prompted by the 

 theoretical considerations outlined above, and encouraged by their "apparent 

 agreement with the facts of observation, I decided to test the components of 

 the spot doublets for evidences of circular polarization and to seek for other 

 indications of the Zeemau effect. 



METHOD OF OBSERVATION. 



The tower telescope forms an image of the sun, about G.7 inches (17 centi- 

 meters) in diameter, on the slit of a vertical spectrograph of 30 feet focal length. 

 This instrument, to which reference has already been made, stands in a well 

 with concrete walls, the grating being about 2GJ feet (8 nieters) below the 

 surface of the ground. The temperature at the bottom of the well is so con- 

 stant that exposures of any desired length may be given, without danger of a 

 shift of the lines resulting from expansion or contraction of the grating. A 

 Fresnel rhomb and Nicol prism <^ are mounted above the slit, so that the light 

 of the solar image passes through them. If the doublets in sitots are produced 

 by a magnetic field, the light of their components, circularly polarized in 

 opposite directions, should be transformed by the rhomb into two plane polar- 

 ized rays, differing 90° in phase. Thus, in a certain position of the Nicol, the 



« Hale, Adams, and Gale, " Preliminary paper on the cause of the character- 

 istic phenomena of sun-spot spectra," Contributicms from the Mount Wilson 

 Solar Observatory, No. 11; Astrophysical Journal, Vol. XXIV, p. 185, 190G; 

 Hale and Adams, " Second paper on the cause of the characteristic phenomena 

 of sun-spot spectra," Contributions from the Mount Wilson Solar Observatory, 

 No. 15; Astrophysical Journal, Vol. XXV, p. 75, 1907. 



^ King, "An electric furnace for spectroscopic investigations, with results for 

 tlie spectra of titanium and vanadium," Contributions from the Mount Wilson 

 Solar Observatory, No. 28; Astrophysical Journal, Vol. XXVIII, p. 300, 1908. 



'^ Hale and Adams, " Sun-spot lines in the spectra of red stars," Contributions 

 from the Mount Wilson Solar Observatory, No. 8; Astrophysical Journal, "\'ol. 

 XXIII, p. 400, 190G; Adams, "Sun-spot lines in the spectrum of Arcturus," 

 Contributions from the Mount Wilson Solar Observatory, No. 12; Astrophysical 

 Journal, Vol. XXIV, p. G9, 1906. 



<* Obtained for this purpose in 1905, when the idea of searching for the Zee- 

 ni:tn effect in sun spots had already occurred to me. A visual test of the spot 

 lines for plane polarization, made with the 18-foot spectrogi-aph iu 1906, before 

 we had photographed the doublets, gave negative results. 



