NATIONAL ACADEMY OF SCIENCES. 



was at this time, too, that Dr. Draper suggested that the different 

 regions of the spectrum should be indicated by their wave-lengths. 

 " Since the deviations of the different fixed lines, B, C, D, in the 

 interference spectrum," he says, " are proportional to the lengths of 

 the undulations which they respectively represent, by designating 

 the different points of the spectrum by their wave-lengths, the sub- 

 division may be carried to any degree of minuteness ; the measures 

 of one author will compare with those of another and the different 

 phenomena of chemical changes occurring through the agency of 

 light become allied at once with a multitude of other optical re- 

 sults." 



It was with this grating that the diffraction spectrum was for the 

 first time photographed. A daguerreotype plate, rendered sensitive 

 by iodine and then by bromine, and exposed for half an hour, gave 

 a maximum sensitiveness at wave-length 0.00001538 Paris inch. 

 A plate prepared by iodine, bromine, and chloride of iodine, and ex- 

 posed for an hour, gave a maximum at the same point, the decom- 

 position extending from wave-length 0.00002007 in the green to 

 0.00001257 in the violet. In these photographs the fixed lines were 

 beautifully distinct. 



Dr. Draper, too, was among the first to point out the unsatisfac- 

 tory character of the measurements which have been made with the 

 prismatic spectrum on the distribution of heat. Since the less re- 

 frangible regions are much compressed and the more refrangible 

 much dilated, the measures obtained by means of a uniform move- 

 ment through the spectrum cannot be accepted as expressing the 

 true distribution. Hence, in 1857, he attempted to determine the 

 curve of distribution in a diffraction spectrum ; but the results, though 

 suggestive, were not conclusive. Subsequently, in 1872, he devised 

 a simple modification of the prismatic spectrum method, which ac- 

 complished practically the result which he had desired to obtain 

 with the interference spectrum. Using Angstrom's values expressed 

 in ten millionths of a millimeter, the wave length of the line A is 

 7,604 and that of H 2 3,933, these lines bounding very nearly the 

 visible spectrum. The middle point is, therefore, at 5,768. If, 

 now, the heat be determined first in the region from 7,604 to 5,768, 

 and then from 5,768 to 3,933, may not the question of its uniform 

 distribution be thus settled ? Upon making the experiment it was 

 found that the two halves of the spectrum gave identically the same 

 amount of heat, and this whatever be the material of the prisms. 



368 



