232 THE CHEMISTRY OF THE SUN. [CH. xvn. 



represented as of minimum intensity in the sun, and of maximum 

 intensity by Thalen and in the photographs. In the sun one of 

 the lines of iron is given as of first, and the other as of third 

 intensity, while in the photograph they are both of the second 

 order. Again, in didymium we get a first order line recorded in 

 the photograph which is absent from the sun altogether, whereas 

 another line of the first, order, near it, is there as a line of feeble 

 intensity ; so also in rubidium, and so we might go on. 



2. Coincident Lines in Different Spectra. 



The final reduction of all the photographs summarised all 

 the observations of metallic spectra compared with the Fraun- 

 hofer lines accumulated during the whole period of observation, 

 for the region lying on both sides of H and K. This reduction 

 taught us that the hypothesis that identical lines in different 

 spectra are due to impurities is not sufficient. 



We found, with the dispersion employed (which had been 

 greater than any employed before), short-line coincidences be- 

 tween many metals the impurities of which had been eliminated, 

 or in which the freedom from mutual impurity had been 

 demonstrated by the absence of the longest lines. 



I will begin by directing special attention to what happened 

 with regard to the spectrum of iron. We first mapped all the 

 lines of iron observed on one of the photographs, including, of 

 course, all impurities ; finally we got rid of the impurities by 

 the process which I have already explained, and at last we got 

 what was called a purified spectrum, in which, along the horizon 

 labelled iron, we had only those lines left which we could not 

 by any application of the principle which has been explained 

 show to be due to the admixture of any other substance 

 whatever. What then was the total result ? The accompanying 

 table will show the sort of corner in which we found ourselves 



