1909] on Solar Vortices and Magnetic Fields. 625 



breaks up the light into a series of spectra, and the rays are returned 

 through the same objective, which brings the spectra to a focus at a 

 point near the slit. By inclining the grating at a slight angle, the 

 image of the spectrum is made to fall at a point slightly to one side 

 of the slit, and here the photographic plate is placed. Thus a portion 

 of the spectrum 17 inches in length can be photographed in a single 

 operation. In the work on sun-spots, most of the photographs are 

 taken in the third order of the grating, where the dispersion and 

 resolving power are very high. When the spot spectrum is being 

 photographed, only the light from the umbra is admitted to the slit. 

 At the end of the exposure this portion of the slit is covered, and 

 light from the photosphere, at a point removed from the spot, 

 is admitted to the slit on either side. Thus the narrow spot spec- 

 trum is photographed between two strips of solar spectrum, used for 

 comparison. 



The advantages of this combined form of telescope and spectro- 

 graph are considerable. On account of the great thickness (12 inches) 

 of the mirrors, the height of the coelostat above the heated earth, 

 and the use of a vertical beam, the definition of the solar image is 

 always better than with the Snow (horizontal) telescope. Another 

 important advantage is the nearly constant temperature at the bottom 

 of the well, where the grating is placed. This permits long exposures 

 to be given, when necessary, without danger of such displacements 

 of the spectral lines as would be caused by expansion or contraction 

 of the grating. The grating used in this spectrograph is a small one, 

 which I have employed in most of my work since 188U, but the 

 unusual focal length of the spectrograph permits the full visual 

 resolution of the grating to be utilised in photographic observations. 

 Thus it has become possible to photograph the widened lines and 

 doublets, as well as a host of narrow lines, most of them due to 

 chemical compounds, which had not previously been recorded in the 

 spot spectrum. 



Lack of time prevents me from discussing in this lecture the 

 various studies of sun-spot lines carried out with this instrument 

 before the attempt to detect a magnetic field in spots was undertaken. 

 An extensive catalogue of these lines is nearly complete, a preUminary 

 map has been issued and a better one is in preparation, and a series 

 of investigations with the arc and electric furnace has suggested that 

 the strengthening and weakening of certain lines is due to a reduction 

 in the temperature of the spot vapours. At present we are concerned 

 with the cause of the widening and doubling of spot lines, and the 

 method of testing this question must now be described. 



A Nicol prism was mounted above the slit of the spectrograph, 

 and just above this a Fresnel rhomb. If the components of a spot 

 doublet were circularly polarised in opposite directions, passage 

 through the rhomb should give two plane polarised beams, the planes 

 of polarisation making an angle of 90° with each other. Thus in 



