52 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. ']'] 



and about two days earlier than in summer. Allowing for this lag 

 there is a correlation for the interval plotted of o.63±o.io between 

 winter and spring, a correlation of o.6i±o.ii between spring and 

 summer, and a correlation of 0.43 ± 0.14 between summer and autumn. 

 These correlations are clear proof of the influence of short-period 



Fig. 39.— Mean departures from normal maximum temperature in relation to 

 the position of sun spots. 



solar changes on the earth's weather, entirely independent of measure- 

 ments of solar heat radiation and indirectly are a strong confirmation 

 of the existence of such changes as shown by the measurements of 

 Dr. Abbot and his colleagues. 



RELATION OF SOLAR RADIATION TO FACUL^ 



Faculse are seen as a rule onl}- on the east and west limb of the 

 sun, and are invisible in the center of the sun. 



