22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 66 



contrast of brightness would be zero, and the higher the temperature, 

 the higher the contrast and the higher the solar constant of radiation. 



Turning now to a consideration of the short-period changes, it was 

 the older view that the difference of brightness between the center 

 and the edge of the solar disk was produced b)' an absorbing atmos- 

 phere or envelope. We prefer, however, to regard the decrease of 

 brightness toward the edge of the sun as mainly a consequence of 

 decreased effective temperature of the radiating surface. We sup- 

 pose that the depth to which one looks at the center of the solar disk 

 is very considerable, and that the limit of it is reached when the 

 molecular scattering cuts off the ray. The same quantity of mole- 

 cular scattering will be found in a very much less radial depth near 

 the edge of the sun, because the line of sight there is oblique, so that a 

 comparatively thin layer viewed obliquely will furnish the necessary 

 number of molecules to cut off the radiation. 



But while holding these views, we admit that the escape of radia- 

 tion depends also on the transparency of the outer solar envelope. 

 If now the transparency of this envelope increases, the solar constant 

 of radiation must increase also; but the percentage increase of the 

 intensity of solar rays will be greatest near the edge of the sun, where 

 the path in these imperfectly transparent layers is longest. Thus it 

 would happen that increased transparency of the outer solar layers 

 would produce at the same time increased values of the solar constant 

 of radiation and decreased contrast of brightness between the center 

 and edge of the sun. 



The two contrary effects we have been discussing may sometimes 

 neutralize each other, but it is not to be expected that they will exactly 

 neutralize each other for all wave lengths. Hence, we may find, 

 with very high values of the solar constant of radiation, a contrast 

 almost exactly, on the whole, equal to that which prevailed in the 

 mean in the year 191 3, but the dift'erent wave lengths may differ 

 slightly in their behavior, some indicating" a greater contrast than 

 the mean of 191 3, others less. 



In our former publications on this subject (see Vols. 2 and 3 of the 

 Annals) we have already considered the possibility of short-period 

 irregular fluctuations of contrast and arrived at opposita-conclusions 

 from the two sets of data then published. Both sets of data, however, 

 are so far inferior in accuracy to those we are now publishing that 

 we withdraw altogether those former conclusions in favor of the 

 ones which we now advance. 



