Solar and Lunar Spectrum, 511 



We may, however, satisfy ourselves by visual observation of 

 the Fraunhofer-liues seen through the whole compound system 

 (entirely of salt) both of the optical perfection of the surfaces 

 of our entire double train, and of the accuracy of its purely 

 automatic action. 



Results of Observation. 



With this and the apparatus already described in previous 

 memoirs we have searched the extreme infra-red solar spec- 

 trum, at first without definite success, later with results which 

 will be better understood by the accompanying drawings. 

 Fio-. 2b shows the newly investigated invisible solar-heat 

 spectrum on the normal scale up to an (estimated) wave- 

 length of 18^. Fig. 3a is an enlarged view of that portion of it 

 extending to 5^, and fig. ob a photographic interpretation of 

 the last, obtained automatically by a special device ; so as to 

 present somewhat the appearance which this heat-region 

 might be expected to show to an eye which could see it. 



Inordinately long as our new chart (fig. 2b) may seem, we 

 see that the scale is, nevertheless, contracted to the last degree, 

 so that the entire visible spectrum is compressed into hardly 

 an inch, seen on the left or violet end. Next we have the 

 already described solar infra-red, already shown to exist to 

 2' x, 8, and which includes its great absorption-bands <E>, NP, O, 

 already investigated with glass prisms. 



The principal lunar heat lies chiefly beyond the great wave- 

 length of O'Ol millim. (10^), and ere we reach it we pass over 

 a region between 5 M and 11 M (many times the length of the 

 entire visible spectrum) , where the solar heat seems to have 

 been, to our present means of recognition, entirely absorbed, 

 probably chiefly by our own atmosphere. 



It will be convenient, however, after noting the extent of 

 the whole region shown in fig. 2a, to commence our descrip- 

 tion with the detailed portion shown in figs. 3a and 36, to 

 which the reader is referred in illustration of what imme- 

 diately follows. 



The lowest bands already described, as seen in the spectrum 

 of a flint-glass prism, are the small ones co l and co 2 near 2*\ 

 Below this point all is believed to be here given for the first 

 time. The bands in this new region are undoubtedly due 

 chiefly, if not wholly, to telluric absorption ; and they are 

 notably variable, depending on the season of the year, and 

 still more on the hour of the day. As the sun sinks, its rays, 

 passing through increasing air-masses, suffer absorptions 

 which singularly change the appearance of the bands, as is 

 shown in tigs. 4 and 5, which, however, are drawn upon the 



