148 PROCEEDINGS OF THE AMERICAN ACADEMY 



bon as for one of rock-salt, and that when the calcium light was sub- 

 stituted for the sun they entirely disappeared. These lines, he found, 

 were in groups, each of which was quite broad. The index of refrac- 

 tion of the farthest group he estimated at 1.5274. 



There are two other series of experiments, brief notices of which 

 seem to be necessary for the completion of the history of the subject. 

 That of Tyndall* on the distribution of heat in the spectrum of the 

 electric light ; and that of Draper,! in which he compares the heat of 

 equivalent parts of the spectrum. 



The experiments of Professor Tyndall were made with exception- 

 ally fine apparatus, and undoubtedly represent the distribution from 

 the electric light with considerable accuracy. The curve he obtains is 

 somewhat symmetrical about its point of maximum, and this maxi- 

 mum is found at a distance beyond the red about equal to the dis- 

 tance on the other side of the green. The total length of the invisible 

 spectrum measured was about twice that of the visible. The ordinates 

 near this point of maximum he found to be very much greater, com- 

 pared with those of the luminous spectrum, than those found by 

 other experimenters. He also experimented upon the effect of intro- 

 ducing water and various other media in the path of the rays, with 

 results essentially similar to those we have already seen. 



Professor Draper, in his paper, calls attention to the fact, that in 

 the prismatic spectrum the length of any part is not, as in the diffrac- 

 tion spectrum, proportional to the differences in wave-length of its 

 extremes, but is more and more condensed as we go from the violet to 

 the red. 



He supposes this to be the reason of the apparent increase of ther- 

 mal effect, and shows, by an extended series of experiments, that the 

 amount of heat between the A line, for which A = 7604, and the point 

 for which A = 5768, when converged by a concave mirror upon a 

 thermopile, had nearly the same heating effect as the portion between 

 A = 5768 and the H 2 line, for which A = 3933 ; A = 5768 being the 

 mean of the other two wave-lengths, and called by him the optical 

 centre. 



From this, he says, " it necessarily follows that in the spectrum any 

 two equivalent series of undulations will have the same heating power, 

 no matter what their actual wave-length may be." A conclusion, 

 however, which we shall see is by no means correct. 



* Phil. Trans., 1866. j Am. Jour. Sci., 1872. 



