COMPOSITION OP THE SUN — RUSSELL 201 



stein's prediction. Even so, the number of solar lines is so great that 

 a good many coincidences will occur through mere chance. 



The Fraunhof er lines are relatively sparsely strewn in the red ; but, 

 even there, if we write down a fictitious wave length, quite at random, 

 there is 1 chance in 16 that it will agree within 1 part in 500,000 with 

 some solar line. In the green the chance is 1 in 12, in the violet and 

 ultra-violet, 1 in 9. 



In a rich spectrum, with several hundred lines, there should there- 

 fore be dozens of apparent coincidences of wave length with solar 

 lines, many of them to one part in a million or better, by chance 

 alone. We must evidently have some test to distinguish these spur- 

 ious agreements from real ones. The obvious criterion is that, if an 

 element is really present, the relative strength, as well as the absolute 

 position, of its lines will be correctly reproduced in the solar spectrum. 

 Numerical coincidences with faint lines are meaningless unless the 

 strong lines of the element are present, and coincidences even with 

 strong lines are suspicious if other stronger or equally strong lines 

 fail to appear. 



In this respect there are noteworthy differences between different 

 elements. Iron, for example, is very richly represented in the solar 

 spectrum, almost 3,300 of its lines having been identified; but it is 

 also very completely represented, for practically every line that can 

 be produced in arc or spark in the laboratory is to be found in the 

 sun. There are, indeed, many faint iron lines whose positions can 

 be accurately predicted by means of the modern theory of spectral 

 structure. Many of these have been detected in the laboratory by 

 long exposures, but still more appear in the sun in exactly the pre- 

 dicted positions and with the expected intensities. 



Magnesium, though but 25 lines are recorded in the " Kevised Row- 

 land," is almost as completely represented there as iron, for its 

 spectrum is poor in lines, and every line that could be anticipated is 

 present. 



Copper, however, which has a fairly rich spectrum, exhibits only 

 its strongest lines in the sun, while silver reveals its presence only 

 by its two strongest lines, and even these are faint. Gold does not 

 appear at all, nor do mercury and bismuth. What is much more 

 remarkable, there is no evidence of phosphorus or arsenic, or chlorine, 

 bromine, or iodine, or of neon, argon, or their heavier homologues. 



The absence of any evidence of these familiar elements in the sun's 

 atmosphere was for a long time a perplexing problem. It was al- 

 ways suspected that they must really be there, but " concealed " in 

 some fashion or other ; and now this conjecture is changed to a well- 

 founded belief. 



