the Existence of Carbon in the Sun. 311 



of the camera. Preliminary experiments showed us the 

 importance in this work of employing a spectroscope of great 

 dispersion and of fine definition, giving also a normal spec- 

 trum. The use of a prism-spectroscope would undoubtedly 

 have masked the phenomenon we have observed. For our 

 purposes, therefore, Rowland's apparatus was peculiarly 

 advantageous. 



Our experiments led us to conclude that there is positive 

 evidence in the solar spectrum of the existence of carbon in 

 the sun. Before giving an account of our experiments in 

 detail, a few observations may not be considered out of place. 



One who studies the solar spectrum by itself, and who has 

 had no experience in the formation and observation of metallic 

 spectra, is apt to regard the dark lines in the solar spectrum 

 as fixed in character and condition. A line which is seen by 

 one observer, and not by another, is generally regarded as 

 a terrestrial line formed by absorption in the earth's atmo- 

 sphere. Certain lines are well known to be due to the terres- 

 trial absorption, as can be easily proved by their appearance 

 when the sun is observed at sunset, when the rays of light 

 have to penetrate a greater thickness of the earth's atmosphere 

 than at midday. The shifting layers of vapour in the sun's 

 atmosphere also may, in certain cases, obliterate or strengthen 

 certain lines of a metal. To understand this it is only 

 necessary to extend the reasoning of the conservation of 

 energy to the subject. It is a common lecture experiment 

 to reverse the metallic lines by passing the rays of light pro- 

 duced by the vapour of the element through a layer of vapour 

 colder than that of the source of the rays. The energy of 

 the rays is thus absorbed in heating the colder layer. When 

 the temperature of the vapour is increased, and becomes 

 equal to that of the source, no reversal takes place. Thus, 

 on the sun's surface the conditions for a reversal may be 

 wanting at certain times, and faint lines may become bright. 

 Their brightness may not be sufficient to affect the general 

 illumination of the solar spectrum of which they form a part. 

 Conditions may arise, moreover, in which the temperature of 

 the reversing vapour may be called critical — at such a 

 temperature that the faint reversal is sufficient to extinguish 

 the bright line of a metal without producing a well-defined 

 dark line. At certain epochs, also, the temperature of the 

 vapour of any element in the sun may be higher than at 

 other times ; and certain lines may thus appear which are 

 wanting when the temperature falls. One is forced to these 

 conclusions in observing the conditions under which the vary- 

 ing character of metallic spectra are produced. For instance, 



