106 Dr. J. W. Draper on the Distribution of 



refrangible regions is to give an apparent dilution to thenij while 

 the less-refrangible are concentrated. The relative position of 

 the colours must also vary ; the fixed lines must be placed at dis- 

 tances greater than their true distance as the violet end is ap- 

 proached/^ I am quoting from the fifth chapter of a work ' On 

 the Forces which produce the Organization of Plants/ published 

 by me in 1844. In this chapter one of the chief points insisted 

 on is the necessity of using wave-lengths in the measurem^ent 

 and discussion of spectrum results — a suggestion which I believe 

 I v»'as the first to make, and which I renewed in the Philoso- 

 phical i\Iagazine^ June 1845. 



The importance of these remarks respecting the peculiarities 

 of the prismatic or dispersion spectrum may perhaps be most 

 satisfactorily recognized on examining such a spectrum by the 

 side of a diffraction or interference one. By the aid of tig. 1 

 (Plate II.) this may be done. 



Regarding the space between the fixed lines D and E as re- 

 presenting the central region in each^ the fixed lines D and E 

 are made coincident in both. The other lines are laid off in the 

 prismatic spectrum as they appear through the flint-glass prism 

 of the spectroscope; those of the diffraction spectrum are arranged 

 according to their wave-lengths. It thus appears that in the 

 prismatic spectrum from the fixed line D to A the j^ellow, orange, 

 and red regions occupy but little more than half the space they 

 do in the difi'raction spectrum; while the green^ blue^ indigo^ and 

 violet from the fixed line E to H occupy nearly double the space 

 in the prismatic that they do in the diffraction spectrum. The 

 general result is that in the prismatic the less-refrangible regions 

 are much compressed and the more refrangible much dilated. 

 And it is plain that the same will hold good in a still greater 

 degree for any invisible rays that are belovr the red and above 

 the violet respectively. 



XoWj if a tiiermometer of any kind v/ere carried in succession 

 from the greatly dilated more refrangible regions to the greatly 

 condensed less refrangible, could the measures obtained be ac- 

 cepted as expressing the true distribution ? The thermometiic 

 surface being invariable, would it not receive in the less-refran- 

 gible spaces more than its proper amount of heat, and in the 

 more refrangible less than its proper amount ? 



If we were to admit that the distribution of heat in a correctly 

 formed spectrum is uniform, it is plain that measures made by 

 the use of a prism would not substantiate that admission. The 

 concentration to which I have alluded as taking place in the less- 

 refrangible region would give an increased heat for that region ; 

 and, on the contrary, the dilatation of the more refrangible would 

 give an exaggerated diminution of heat for that space. But if 



I 



