NO. I WEATHER DOMINATED BY SOLAR CPIANGES ABBOT J 



parency of our earth's atmosphere. No appreciable 12-month peri- 

 odicity appears in the results. This is a good sign of their indepen- 

 dence of atmospheric influences. Full and dotted curves in Figure i 

 mark all the well-supported sequences of rising and of falling solar 

 radiation. They occur in short intervals, averaging 5 days. All of 

 those selected exceed 0.4 per cent in range, averaging 0.8 per cent. 

 These rising and falling sequences are iii and 106 in number, respec- 

 tively. Many are lost because of unfavorable observing conditions. 



Figure 2 shows average changes in the mean temperature and the 

 barometric pressure at Washington, D. C, associated with these rising 

 and falling sequences of solar radiation, during the months of March, 

 April, September, and October. These meteorological exhibits are 

 average values representing the work of 7 years, and of about 10 cases 

 of each kind in each month. 



The method of computing the curves shown in figure 2 is illustrated 

 in tables i and 2 as regards temperatures of March. The temperatures 

 (which are the mean of maximum and minimum at Washington as 

 published by the U. S. Weather Bureau) are arranged in consecutive 

 series of 25 days each. In each series, the fifth day is that on which 

 the solar change examined reached its culmination. Departures of 

 temperatures are always computed from the first day of the series 

 as the base. The mean values of all the departures occurring in 

 March in the years 1924 to 1930 are given at the foot of the table. 

 They are corrected to eliminate the secular rise of temperature which, 

 of course, occurs during any 25-day interval at that season of the year. 

 The final result is plotted in figure 2. The reader will see that in all 

 cases there is a marked opposition between curves corresponding to 

 rising and falling solar radiation, respectively. 



Eleven physicists to whom I have shown these results unanimously 

 concur in advising me that the constant opposition of the weather 

 effects following opposite solar causes demonstrates a physical con- 

 nection between the weather of Washington and the changes in the 

 solar constant of radiation as observed in Chile. Average changes 

 of mean temperature of 5° Fahrenheit are found corresponding to 

 solar changes averaging only 0.8 per cent. Hence we may suppose 

 that on many occasions temperature effects caused by solar changes 

 may reach 10°, and sometimes 15° or 20°. That is to say, major 

 changes in weather are due to short period clianges in the sun. So 

 revolutionary is this conclusion for meteorology, that I hesitated to 



