ON OUR KNOWLEDGE OF SPECTRUM ANALYSIS. 301 



point coincident with D2 of the second order was found on Hj of the third 

 order, and the point 64 of the second order had impressed itself near O of 

 the third order. The ray G of the third order, the wave-length of which 

 is known, was impressed photographically on the collodion plate. 



By a simple calculation it was rendered evident that a given ray in the 

 compound H2 was of wave-length 393"01, and that another near had 

 the wave-length 344-46. 



Dr. Draper ruled a fine scale with a dividing engine, and applied this 

 to his spectrum photographs in order that the wave-length of any line 

 could be read ofi" at once to the 10-millionth of a millimetre. 



The followins: remarks concerning: M. Mascart's measurements are 

 worthy of record. The line L, which he regards as single is in reality 

 triple, and does not correspond to wave-length 381'9 but to 382'1 ; M is 

 correctly designated by 372'8, but it is double ; N is really at 358'3 and 

 not at 358-0. 



The spectrum above H, when compared with the region from G to H, 

 is marked by the presence of bolder groups of lines, the most conspicuous 

 of which are those between 382-0-386-0; 370-5-376-0; 362-0-365-0; 

 356-8-359-0; 349-0-3530. Dr. Draper's fine photographs show how 

 impossible it is to depict the relative intensities of lines in the spectrum 

 by any other means than photography, and how groups of lines even may 

 fail to be resolved ; in his original negative there could be readily counted 

 more than fifty lines in the group H. 



In fact ' The exact composition of even a 'part of the spectrum of a metal 

 will not he known until we have ohtained photographs of it on a large 

 scale.' 



M. Cornu has given a description of the solar spectrum from the line 

 called 7i to the ray O, and has drawn a beautiful map made to the scale 

 of wave-lengths.' This work was intended to be a continuation of the 

 labours of Angstrom. The spectra were observed by photography in a 

 manner similar to that devised by. M. Mascart, but as the optical appai-atus 

 was made of glass, all rays more refrangible than O (wave-length=344-ll) 

 were intercepted. 



In a continuation of his experiments using more perfect lenses of 

 quartz and Iceland spa as well as prisms of these materials, M. Cornu 

 has succeeded in photographing the solar spectrum as far as a line called 

 U (wave-length 294-84). Rays more refrangible than this are absorbed 

 by the earth's atmosphere. As a reflector for the ultra-violet rays, 

 metallic mirrors were found to be useless, therefore in order to bring solar 

 rays into the slit of the collimator, a right-angled prism of quartz was 

 used, the light being totally reflected from one side of the prism. The 

 image was received on a photographic plate, a dark-slide being made to 

 replace the eye-piece of the spectroscope. 



By taking two photographs on the same plate, one below the other, 

 the prism being turned to the right or to the left through a measured 

 arc, the sharpness of the lines and their position near the centre of the 

 field gives us a means of ascertaining the position of the pi"ism corre- 

 sponding to the minimum angle of deviation for any particular ray. On 

 accoimt of want of intensity in the rays, photographs are difficult to obtain 

 from diffraction spectra. DiSraction-gratings on glass yield spectra con- 

 tinuous only as far as R. Photographs of diffraction-spectra have been 



' Annales de VJ^colc Kormale, 1874. 



