MEMOIR XIII.] EXPERIMENTS MADK IN VIRGINIA. 



199 



being placed as at a, Fig. 25, a narrow streak of light 

 passes through. The trough, Fig. 26, is then placed 

 behind in such a position that half the light coming 

 through the slit may pass through the liquid in the 

 trough, and the other half pass by its side unintercepted. 

 Behind the trough is placed a flint-glass prism, as in Fig. 

 _ } r>, and further still a white pasteboard screen,^ a is 

 the slit plate, b the trough, d the prism. 



The action of this arrangement is as follows: The 

 beam of light cast by the mirror into the room is inter- 

 cepted, except the small portion that passes the slit. Of 

 this a part passes through the trough and a part on one 

 side of it. Two beams of lio-ht, therefore, fall on the 



O ' 



prism, one of which has passed through the trough 

 and suffered absorption, the other has not. There are, 

 therefore, on the pasteboard screen two spectra side 

 by side. 



Suppose, for example, that the trough is filled with 

 distilled water; the two spectra on the screen are alike, 

 as in Fig. 28. If it be filled with a solution of chromate 

 of potassa, in that coming from the light that has passed 

 through the trough the blue, indigo, and violet rays are 

 missing, as in Fig. 29. If the trough be filled with am- 

 monia-sulphate of copper, the red, the orange, the yellow 

 have disappeared or been absorbed, as in Fig. 30. The 



Fig. 28. 



Fig. 20. 



Fig.sa 



reference or undisturbed spectrum enables us to deter- 

 mine what rays have been absorbed with precision. 

 For the investigation of the absorption of heat the 



