26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. ']'/ 



that here, too, the change was mainly in the blue, violet, and ultra- 

 violet spectrum. 



There are evidently two kinds of solar change. The long-period 

 swings are related to the total solar activity. Great visible activity 

 in the sun, such as numerous sun spots, faculae, or prominences, 

 like stirring a fire, brings hotter radiating surfaces to the front, and 

 produces higher solar constants. 



But on the other hand, whenever a sun spot crosses the sun's 

 center, it carries along with it a cloudlike effect, not a cloud, of 

 course, but a diminished transparency. When this diminished trans- 

 parency points towards the earth, we have for a few days lower 

 solar constants. 



Both kinds of change affect the short-wave rays of the spectrum 

 more than the long-wave rays. This is, of course, what one would 

 expect. Increased effective solar temperature, attending increased 

 activity, would produce its larger effects at shorter wave lengths, 

 in accord with the Wien-Planck Law of temperature radiation. In- 

 creased opacity of the solar envelope, just like increasing opacity of 

 the earth's atmosphere, would also produce its larger effects at 

 shorter wave lengths, quite in accord with our own observations of 

 atmospheric transmission coefficients. It is yet too early to decide 

 by a comparison of Curve B with Curve C that there is a real differ- 

 ence in the quality of these spectrum changes, depending on the 

 character of the solar change involved. Yet so far as this evidence 

 goes, it indicates a less pronounced contrast between short- and long- 

 wave rays in spectrum change for short-period solar variations than 

 for long-fperiod ones. 



CONCLUSION 

 To sum up : 



1. It is not necessary to wait for perfectly impeccable solar- 

 constant determinations to determine changes of the sun's radiation 

 well enough for useful comparison with meteorological phenomena. 

 Better values, however, will soon be available. 



2. Such comparisons as have been made indicate that a higher 

 accuracy than the present in solar-constant determinations will be 

 needed to yield high correlations in forecasts for individual days, 

 but that where mean values can be used, as in forecasts for weeks 

 or months, present values are fairly satisfactory. 



3. There is no reason to think that the independence of the two 

 solar radiation stations, Montezuma and Harqua Hala, has been lost 

 on account of means used to eliminate systematic errors. 



