512 PROCEEDINGS OF THE AMERICAN ACADEMY 



continuity of the curve of energy was broken at several places by deep 

 valleys, indicating the presence of cold bands instead of hot ones. 



At this stage of his investigation Professor Langley encountered a 

 most delicate and difficult problem. The prism furnishes the most 

 heat ; the grating gives the wave-length ; but the laws of refraction 

 and dispersion necessarily distort the prismatic spectrum, and lead to 

 false conclusions in regard to the distribution of solar energy. Pro- 

 fessor Langley happily combined the grating and prism so as to be 

 able to translate the prismatic spectrum into the normal spectrum. 

 He had here to deal with three unknown quantities, viz. the pris- 

 matic deviation, the index of refraction, and the wave-length. The 

 relation between this deviation and the index of refraction is well 

 established. The relation between this deviation and the wave-length 

 is obscured by theoretical and mathematical difficulties. This relation 

 has been expressed by various formulas, derived from the undulatory 

 theory of light, and involving preconceived ideas in regard to the 

 atomic constitution of the lumiuiferous ether. A fair criticism can 

 be passed upon these formulas, that they disagree with each other, 

 and, in extreme cases, involve results which are contradicted by obser- 

 vation. All of them are expressed by a descending series of terras, 

 involving many constants which must be determined by observations 

 made within a limited range of the visible spectrum. Professor 

 Langley undertook to solve the j^roblem experimentally by a happy 

 Gomliination of a grating and a prism. 



If the deviation of a homogeneous ray by a grating is observed, 

 the corresponding wave-length is given by a simple rule. Within the 

 limits of visibility there is no difficulty. Outside of these limits, the 

 diffraction deviation must be indicated by heat-energy; and that be- 

 comes insufficient with the Rutherford gratings in the extreme parts 

 of the infra-red. But the concentrated heat from one of Rowland's 

 exquisite gratings affects the delicate strips of the latest bolometers 

 even in the longest wave-lengths. Professor Langley's experimental 

 devices in this case were beautifully complex. Rays were condensed 

 upon a fine slit, and then went to the grating. Those of definite wave- 

 lengths were accumulated upon a second slit, and formed there a sharp 

 spectrum. Invisible spectra were superimposed at this slit upon the 

 visible spectrum ; but their wave-lengths were simple multiples of the 

 wave-lengths of the visible spectrum. This compound spectrum was 

 sent next through a prism, and its components deployed out into sepa- 

 rate columns. The prismatic deviation of the visible spectrum was 

 measured with the eye. By groping with the bolometer in the dark 



