112 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 30 



From all this I conclude : 



1. An apparent influence of short-period solar variation appears* 

 in the temperature of the United States. 



2. Corresponding to 0.8 per cent change in the sun, there appear 

 to be temperature changes of the order of 5° F. at Washington. 



3. The sign of the correlation changes during the year. 



4. A high negative correlation is found between the temperaturt 

 effects corresponding respectively to rising and to falling solar 

 sequences. 



5. The temperature effects coincide in date with the solar changejj 

 which appear to induce them. 



6. If the connection between solar change and temperature change 

 is a genuine one, it must operate by some indirect atmospheric mech- 

 anism; because if it were a direct effect its sign would not change 

 .during the year. 



7. Although complicated, the relation seems to offer promise for 

 weather forecasting nearly a week in advance. Yet the occasional 

 inversions of effect found inspire caution in a pronouncement of this 

 character. These apparent inversions are, however, doubtless caused 

 frequently by one solar change treading too quickly on the heels of 

 another. Again they may sometimes be caused by delayed receipt 

 from distant centers of action of waves of temperature effect arising 

 from former solar changes. 



The results thus far are tentative. It is proposed to study baro- 

 metric pressures as well as temperatures and to extend the investiga- 

 tion to other parts of the United States and of the world. Prelim- 

 inary studies have been made, too, of 10-day mean values of solar 

 radiation and temperature, and we hope that in this way, if reliable 

 weather- forecasting data are really secured, they may be extended to 

 months and seasons in advance. 



FIELD STATIONS 



Observations of the solar radiation have been continued whenever 

 weather conditions would permit at Table Mountain, Calif., at 

 Mount Montezuma, Chile, and at Mount Brukkaros, South West 

 Africa. All three stations continue to report measurements as made 

 on three-quarters of the days of the year or more. However, not all 

 of these observations prove satisfactory, so that 60 per cent is a bet- 

 ter estimate of available observation days for these selected high- 

 level desert stations. 



A strange and serious accident occurred in December on Mount 

 Bruldvaros. It will be recalled that for tlie sake of uniformity of 



