NO. 5 SOLAR VARIABILITY ABBOT, FOWLE, ALDRICH 21 



The factors of correlation (omitting A = 0.3737 '^"'^ 1.008) are as 

 follows : 



1913: r = 0.60 1 ±0.067 



1914: r = o.2i3±o.o8o 



The reader will recall that the method of deriving these contrast 

 values is such that algebraically increased contrast numbers corre- 

 spond really with decreased solar contrast. 



Let us now restate these logarithmic results in ordinary terms. 

 Imagine that the spectroscope was dispensed with, and that the con- 

 trast of brightness was determined for the solar radiation as a whole, 

 and not for particular wave lengths. Let us further suppose that all 

 the drift curves so obtained during 191 3 and 191 4 were reduced to 

 unit intensity at the center of the solar disk. We should then find, 

 still confining ourselves to a consideration of the short-period varia- 

 tions, that in 1913 an increase of i per cent in the solar constant of 

 radiation corresponds with a decrease of the ordinate of the drift 

 curve at 92 per cent out on the radius of 0.35 per cent. 



The results of 1914 indicate a larger ratio of change of contrast 

 compared to change of solar constant than those of 1913. This may 

 be possibly a real difference caused in the sun, but it should be noted 

 that the range of variation in 1914 was so small that the error of 

 the determination of this ratio is much larger than it was in 1913. 

 It is probable that the numerous values which will be found from 

 observations of 191 5 will enable us to give a more satisfactory con- 

 clusion on this point. 



We now come to the consideration of the question why the ratio 

 of change of contrast from year to year with respect to change of 

 solar constant differs in its sign from the ratio of change of contrast 

 from day to day with reference to the corresponding change in the 

 solar constant. x\s stated above in years when the solar constant was 

 high the solar contrast was greater than usual, while, as just stated, 

 during 191 3 and 1914 short-period temporary increases of the solar 

 constant of radiation were attended by decrease in the solar contrast. 

 This indicates two causes at work in the sun. We are inclined to 

 suggest the following hypothesis. 



Attending the long-period changes of solar activity indicated by 

 prevalence of sun spots, faculse, prominences, etc., there is, it may be 

 assumed, a change of the effective solar radiating temperature. 

 Higher effective radiating temperatures should prevail at times when 

 increased solar activity brings faster the hot material to the surface 

 to radiate. It is clear that if the solar temperature was zero, the 



