ASTRONOMY — HALS. 71 



The identification of the faint lines in the spot spectrum, although it 

 cleared up the uncertainty as to their origin, by no means solved the 

 problem regarding the cause of the characteristic spectral phenomena of 

 spots. It seemed evident that extensive laboratory investigations might 

 be needed for this purpose. The question was, however, to find some 

 logical basis for a plan of attack. I have long been impressed with 

 the following characteristic features of spot spectra: (i) The diverse 

 behavior of different lines of the same element; (2) the fact that all of 

 the strengthened lines lie in the visible spectrum, and that the most 

 conspicuous of them are in the red, yellow, and green ; and (3) the rela- 

 tively great intensity in the less refrangible region of the general back- 

 ground of the spot spectrum. In considering these peculiarities, I could 

 not fail to recall: (i) That in the spectra of the elements some lines 

 increase in intensity, while others decrease, when the temperature falls; 

 (2) that in spectra observed at low temperatures, the most prominent 

 lines are likely to appear in the less refrangible region ; and (3) that in 

 a continuous spectrum (as well as in a spectrum of bright lines) a reduc- 

 tion in temperature involves a shift of the maximum toward the red. 

 The similarity of these two groups of facts led to the belief that the most 

 characteristic phenomena could in all probability be accounted for on 

 the hypothesis that the temperature of the vapors within the spot is 

 below that of the corresponding vapors within the sun's reversing layer. 

 It was on this basis that the laboratory investigations, described on 

 another page of this report, were undertaken. 



The results so far obtained in our study of spot spectra, while strongly 

 confirmatory of the hypothesis outlined above, of course constitute only 

 the first steps in the extended researches required to test it in a com- 

 plete manner. The use of an electric furnace, for the vaporization of 

 metals in a carbon tube, where no electrical phenomena, other than those 

 which result in the production of heat, can influence their radiation, is a 

 most important element in the general investigation. This work, which 

 is now in progress, will include all of the characteristic sun-spot metals 

 that can be volatilized in the furnace. The solar side of the investigation 

 will involve photographic and visual studies of spot spectra, carried on 

 through a period of several years. Further stellar spectroscopic investi- 

 gations must be delayed until the completion of the 60-inch reflector, as 

 the Snow telescope is not well adapted for this phase of the work. 



It must not be supposed that our hypothesis is intended to account for 

 all of the phenomena of spot spectra. The bright lines occasionally 

 observed are probably due to the overlying reversing layer, or to the 

 chromosphere, especially in cases where eruptive action is concerned. 

 The hypothesis will serve a useful purpose if it provides a basis for the 



