12 PROCEEDINGS OF THE AMERICAN ACADEMY 



the conditions under which the varying character of metallic spectra 

 are produced. For instance, we have caused the rays from iron vapor 

 to traverse a long and dense layer of iron vapor, and have observed 

 that the strength of the lines and the number of reversals have been 

 largely increased. In another experiment, the lower carbon of the 

 electric lamp we employed occupied the centre of an electro-magnet. 

 This was accomplished by passing the carbon through a hollow iron 

 cone, and surrounding the latter by layers of wire, through which the 

 electrical current employed in generating the light passed. In this 

 case the electric arc was spread out at right angles to the pole of the 

 magnet, into a fan-like, intensely hot flame, which roared loudly, and 

 which rarefied, so to speak, the iron vapor between the carbon termi- 

 nals. The strength of the lines and the number of reversals were 

 diminished under this new condition. 



Another phenomenon may happen. When an excess of the vapor 

 of one metal floats over or is mixed with that of another, the lines of 

 one metal are superimposed upon those of another in the solar spec- 

 trum, and the stronger spectrum of one element may easily obliterate 

 the weaker spectrum of another. Thus we have succeeded in com- 

 pletely obliterating the fluted spectrum of carbon in the green and blue, 

 by photographing upon it the spectrum of iron, of nickel, and of ce- 

 rium. A species of composite photograph was thus obtained. It is 

 possible that in the future Galton's ingenious method of composite 

 photography may be applied to the solar spectrum; and by a judi- 

 cious selection of photographs of the elements, a composite photograph 

 may be obtained which will closely resemble portions of the solar 

 spectrum, and will enable us to judge of the composition of the revers- 

 ing layers of the sun. 



To the varying conditions which we have thus outlined are due, 

 we believe, the disappearance in the sun's spectrum of the marked 

 fluted spectrum of carbon in the green and blue portions. 



A careful examination of the fluted spectrum of carbon, however, 

 with the juxtaposed solar spectrum, discloses a remarkable fact: 

 while traces of obliteration of the evidence of carbon vapor are seen, 

 yet the general character of the lines in the solar spectrum immedi- 

 ately juxtaposed with the fluted spectrum of carbon near H lead us to 

 believe that there is unmistakable evidence of the existence of carbon 

 vapor in the sun. When the arrangement of the fine lines of the 

 spectrum of carbon is plotted as a curve, and that of the dark lines 

 in the solar spectrum immediately above the carbon spectrum is also 

 plotted, the two curves have a remarkable similarity in character, 

 running with a slight convexity toward one axis. 



