August 4, 1881] 



NATURE 



319 



dispersion which Angstrom had employed ; so that, if impurities 

 had been suggested, there was now a method which has not yet 

 been challenged of getting rid of them. If the dispersion was 

 then insufficient there was nothing to prevent it being made 

 very much mi re considerable, because a perfect photograph will 

 bear a very considerable amount of magnification. 



The diagram (Fig. 31) will show the method of photography 

 that was adopted in this work, and by which the various photo- 

 graphs throw n on the screen were taken. The object was to 



compare the light of the sun with the light of the vapour in the 

 electric arc of any particular substance that we w ished to observe. 

 By means of a heliostat and lens an image of the sun was thrown 

 exactly between the poles of an electric lamp, and the rays 

 diverging from it were collected by a second lens and again 

 brought to a focu=, this time on the slit of the spectroscope. 

 The slit was provided with two slides, by means of which either 

 its upper or lower half could be exposed, while the other half 

 was covered. If we wished to take the solar spectrum first, the 



Lltri-vi. let fluting. 

 Fl". 30.— Carbon fluting-;, contrasted with the line-spectra of calciun 



Ulue fluting, 

 n, aluminium, and other impurities of the poles. 



poles were separated so that they might not obstruct the sunlight ; 

 the image of the sun was allowed to fall on one-half of the slit, 

 and the plate was exposed. That half of the slit was then covered 

 up and the other half opened (the sunlight being cut off), and 

 the substance volatilised in the electric arc so that its image fell 

 on the open part of the slit. The plate was again exposed, and 

 so the two spectra were obtained, one above the other. In this 

 way then we had, first of all, a spectrum of the sun compared 

 w ith the spectrum of the particular substance we wished to map. 



After that we had the long and short lines in the same substance 

 photograj^hed on another plate. After that we had all the sub- 

 stances which might exist as impurities in the first substance — 

 that is to s.ay, all the chemical elements photographed with their 

 lines — their long and short lines, in precisely the same manner ; 

 and finally we had a comparison of the substances we wished to 

 photograph, say iron, with a spectrum of every other substance 

 which might contain these impurities. It will be seen therefore 

 that an enormous number of photographs had to be taken. As 



a matter of fact three or four thousand photographs have been 

 taken, and a very considerable amount of time (about four years) 

 w as consumed in that way. 



But it n^ay be said, "Surely if you are going to limit 

 yourself to jiliotography, you will only be dealing with a very 

 small part of the spectrum.'' My reply to that is that already 

 in the year 1S75, when a part of this work had been carried on, 

 o'.her laboratory work had given us reason to believe that 



what was then being done in photography at the blue end of the 

 spectrum would be done by photography in every other portion, 

 for in fact a spectroscopic study of the behaviour of bodies at 

 low temperature, to which I hope I shall have time to refer, had 

 led several to bel'eve— at all events had led me to believe — that 

 what one got in the text-books about actinism and so on was but 

 a very rough approximation to the truth. We had been taking 

 as the functions of light what were really the functions of the 



