1900] on Solar Vortices and Magnetic Fields. 627 



to be transformed into triplets, if produced by a magnetic field. 

 The failure of the central line to appear seemed to raise an important 

 argument against the magnetic hypothesis. 



At this point the necessity of conducting laboratory investigations 

 in immediate conjunction with astronomical observations is well 

 illustrated. Fortunately, our laboratory was already well equipped 

 for work of this nature (Fig. 7). In anticipation of the possibility 

 that observations of the Zeeman effect would be needed in the inter- 

 pretation of solar and stellar phenomena, a powerful electro-magnet, 

 with suitable accessory apparatus, had been provided. A brilliant 

 spark produced between metallic electrodes in the field of the magnet 

 furnished the source of hght. As many of the double lines in sun- 

 spot spectra are due to iron, this metal was selected for the first ex- 

 periments. The spectrum was photographed, at various angles with 

 the lines of force, with a powerful spectrograph, like the one used 

 with the tower telescope, similarly mounted in an underground 

 chamber. 



The difficulty of accounting for the behaviour of the iron doub- 

 lets in the sun was removed by these investigations. It appears that 

 these lines do not become triplets, when observed across the lines of 

 force. In reality they are changed to quadruplets, or doublets in 

 which each of the components is a close double line. In the magnetic 

 field of sun-spots, which is much weaker than the field used in the 

 laboratory, the closely adjoining lines which constitute the components 

 of the doublets cannot be separated. Thus these sun-spot lines 

 should appear double at whatever position the spot may occupy on 

 the sun's surface. 



The distance between the components of doublets or triplets sepa- 

 rated in the magnetic field varies greatly for different lines. Some 

 exceptional lines are not affected in the least, others are merely 

 widened, and others are clearly and sometimes greatly separated. It 

 is therefore important to compare the widening and the separation 

 of lines in a sun-spot spectrum with the corresponding phenomena 

 in the magnetic field. With few exceptions, most of which may be 

 accounted for by the presence in the spot spectrum of closely adjoin- 

 ing lines of other elements, the solar and laboratory results were 

 found to be in good agreement. The following table gives a com- 

 parison of certain iron lines in the spot and laboratory. 



Wave-Length AA, Spark AA, Spark aA. Spot 



6213-14 0-703 0-138 0136 -0*002 



6301-72 0-737 0-144 0-138 -0-006 



6302-71 1-230 0-241 0-252 +0-011 



6387-05 0-895 0-175 0-172 -0-003 



The column headed " AS, spark " gives the distance between the 

 components of the lines as observed in the laboratory. As the 

 strength of the magnetic field used in the laboratory was about 5 • 1 



