NO. 6 



MAGNETIC STORMS ABnOT 



pheric conditions— would have tended to raise, not to depress, the 

 solar-constant values. At that time the solar constant was being deter- 

 mined at Calama, Chile, by the fundamental, or "long," method of 

 Langley. The less transparent the atmosphere the greater would have 

 been our estimate of the losses it produced in the solar ray, and the 

 larger the computed solar-constant value outside the atmosphere. 

 But this tendency is counterbalanced exactly by lower observed 

 pyrheliometer readings when the atmosphere is less clear. The solar- 



Fig. I. — Solar constant March-April 1920, upper curve; atmospheric trans- 

 mission green light, lower curve. 



constant values would be too low only if the transparency of the 

 atmosphere was erroneously observed too high. 



Other critics have suggested to me that the observed change of 

 the solar constant of March 1920 might have been produced errone- 

 ously by a change of the absorption of solar rays by ozone, assuming 

 that the solar outburst of electrified ions produces large changes of 

 the concentration of atmospheric ozone. I reply that all solar-constant 

 values done by the long method, as in 1920, take account of such 

 effects because the atmospheric transmission coefficients are neces- 

 sarily appropriately modified owing to the method of obtaining them. 

 All solar-constant values done by the short method since 1923 are 



