8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 85 



publish it until the unanimous approval of many competent critics 

 encouraged me. I am further supported in this view by having found 

 a similar opposition of relations prevailing not only at Washington 

 but at Williston, North Dakota, and Yuma, Arizona, in all months 

 of the year. 



By what physical connection are these surprising meteorological 

 results produced by such small solar changes? We must discard at 

 once, I think, the idea that changes of ground temperature, directly 

 produced, communicate the effects to the surface air. For firstly, by 

 Stefan's law, in equilibrium conditions radiation varies as the fourth 

 power of the absolute temperature. Hence a change of i per cent in 

 radiation, if acting directly and in equilibrium conditions, should re- 

 quire but I per cent change in the earth's temperature. Actually the 

 change of temperature observed exceeds i per cent, reckoned from 

 the absolute zero. Secondly, in March and some other months, a tem- 

 perature eft"ect at Washington is found to be nearly simultaneous with 

 the solar change. The solid earth has too large a capacity for heat to 

 follow in temperature thus quickly. Thirdly, large effects occur at 

 Washington lo or 12 days, and sometimes 16 or 17 days, after the 

 solar cause ceases. Not all of these effects can be direct. Fourthly, in 

 September a reversal of sign is observed. 



Admitting that the meteorological effects are produced indirectly, 

 let us recall: Firstly, that from 10 to 25 per cent of the solar radia- 

 tion is primarily absorbed in the atmosphere itself, which has a very 

 small capacity for heat. Secondly, that the atmosphere circulates in 

 great cyclonic whirls. Thirdly, that the temperature of a station 

 depends greatly on the prevailing wind direction. May it not be that 

 the instantaneous changes of heat absorption in the atmosphere tend 

 to displace centers of cyclones, and thereby to alter the wind direction 

 at stations, thus altering their temperatures ? 



How shall we explain deferred effects occurring 10 or even 17 days 

 after the culmination of solar sequences? May they not result from 

 atmospheric waves drifting in a southeasterly direction from distant 

 centers of action where primary eft'ects are produced? If so, we 

 must perceive that the average effects shown in tables i and 2 can 

 form no trustworthy basis for forecasting individual cases. For 

 primary and secondary effects, treading on each other's heels, as it 

 were, must often interfere, and either augment or reduce expected 

 weather changes. 



