144 Professor T. E. Thorpe [June 9, 



torially mounted, so that the plates were kept in a constant position 

 with respect to the sun during the times of exposure. 



Three different methods were employed to obtain photographic 

 records of the spectrum of the corona. Jn the first, which was sug- 

 gested by Professor Norman Lockyer more than twenty years ago, 

 the eclipsed sun was to be photographed through a prism attached 

 to a telescope of 6-inch aperture. In this manner an image of the 

 corona would be obtained corresponding with each kind of light 

 emitted by it. Thus, if the corona consisted entirely of glowing 

 hydrogen, there would be an image in the position occupied by each 

 of the lines in the hydrogen spectrum. If, as may be expected, the 

 materials composing the corona are different in different regions, the 

 images obtained will not exactly resemble each other, but the form 

 of each image will depend upon the distribution of that particular 

 spectral line through the corona. The complete spectrum of every 

 part of the corona which is bright enough to be photographed will, 

 therefore, be obtained with a single exposure. 



The other method of studying the spectrum of the corona is by 

 means of the ordinary slit spectroscoj^es. The arrangement, employed 

 by Captain Hills, consisted of two spectroscopes, each provided with 

 a condensing lens and camera, mounted on an equatorial stand. The 

 spectroscopes were of different dispersive power, one having two 

 prisms, and the other one. The slits were placed parallel to each 

 other, and were so arranged as to cut across opposite limbs of the 

 sun at right angles to the sun's equator. An image of the sun is 

 thrown on the slit by the condensing lens, and the slit is long enough 

 to cover the whole width of the corona. The resulting photographs 

 ought then to show at least three different spectra : a continuous 

 spectrum over the dark body of the moon, on either side of which 

 will appear the prominence spectrum, and outside of which again will 

 be the true corona spectrum, which may or may not be broken up 

 into bands by the occurrence of rifts or dark spaces in the corona. 

 This method has the great advantage of discriminating between the 

 different spectra of every portion of the corona along the line of the 

 slit ; the main difficulty connected with it is the want of light, which 

 makes it almost impossible to give a sufficient exposure unless the 

 slit is opened rather wide. It was decided, therefore, to mako only 

 one exposure with each spectroscope ; this was to last as nearly as 

 possible the whole time of totality, and the most rapid photographic 

 plates procurable were to be used. 



The measurement of the visual brightness of the coronal light' 

 was to be effected by the following arrangement. An image of the 

 corona is accurately focused on a white screen by means of an equa- 

 torial of 6-inch aperture and 78-inch focal length, and the intensity 

 of the light from different portions of the corona at definite distances 

 from the limb is compared with that of a standard glow-lamp by 

 means of an arrangement constructed on the principle of the Bunsen 



