626 Profess^or Gmrfjp. E. Hale [May 14, 



one position of the Nicol one of the components should be photo- 

 graphed alone, and by turning the Nicol \)if this should disappear 

 and the other component come into view. 



When this test was applied with the tower telescope, in June 

 1908, the true character of the spot doublets became apparent 

 (Fig. 6). One or the other component of the doublet could be cut 

 off at will by rotating the Nicol, precisely as Zeeman had done in the 

 laboratory. On account of the unique character of the Zeeman 

 doublets, this test alone was almost sufficient to prove the existence 

 of a magnetic field in sun-spots. But one of the great beauties of 

 the Zeeman effect is its many-sided character, which permitted the 

 test to be multiplied and extended. From Zeeman's first experiments 

 it was known, for example, that if the strength of the magnetic field 

 is insufficient to separate completely the components of a doublet, 

 the edges of the resulting widened line should be circularly polarised 

 in opposite directions. Thus those lines which are widened, but not 

 doubled, in spots might be expected to shift in position when the 

 Nicol is rotated. This was found to be the case. Again, the lines 

 which constitute the flu tings of the spectra of compounds are not, 

 in general, affected by a magnetic field. Hence such lines in the 

 spectrum of a sun-spot should not be shifted when the Nicol is 

 rotated. This, also, was found to be true. But a still more satisfactory 

 test was suggested by another laboratory phenomenon. When a 

 doublet is observed along the lines of force, with one of the com- 

 ponents extinguished by the Nicol, reversal of the current through 

 the magnet should extinguish the visible component and cause the 

 invisible one to appear. In the sun, according to our hypothesis, 

 reversal of the direction of revolution in a vortex should correspond 

 to reversal of the current through the coils of a magnet. Hence the 

 red component of a doublet should appear in the spectrum of a vortex 

 rotating in one direction, the violet component in that of a vortex 

 rotating in the reverse direction. Fortunately, the appearance, on 

 opposite sides of the solar equator, of two spot vortices rotating in 

 opposite directions (Fig. 4), made this test possible. The results were 

 perfectly in accord with the hypothesis. 



So far we have been considering only such phenomena as are 

 observed parallel to the lines of force of a magnetic field. But a 

 spectral line which, under such circumstances, appears as a doublet, 

 is usually transformed into a triplet, when the observation is made 

 at right angles to the lines of force. The circularly polarised side 

 components of the doublet give place to plane polarised components, 

 occupying the same position, while another lixie appears centrally 

 between them. The light of this line is also plane polarised, the 

 direction of the vibrations being parallel to the field, while the 

 vibrations of the side components are in a plane at right angles to 

 the field. Thus when a spot is carried by the solar rotation to a 

 point near the limb, we might expect the double lines in its spectrum 



