68o 



NATURE 



[July 29, 1920 



It is very striking that the solar changes pro- 

 duce such large and prolonged temperature 

 effects. On the tenth and seventeenth days after 

 the event the average temperature following 

 solar constants of 2-00 calories differs by more 

 than 6° F. from that following solar constants 

 of 1-90 calories. 



The temperatures following minimum "solar 

 constant " values are generally lower than the 

 normal from the third to the nineteenth day ; 

 they are above the normal before the third day ; 

 while those following high values are above the 

 normal from the sixth to the nineteenth day, they 

 are below the normal before the sixth day ; and 

 those corresponding to mean " solar constant " 

 values differ by little from the normal through 

 the whole interval. 



The latter state of affairs is probably decidedly 



131 



IG. 2. — Five-day means observed at Calama, Chile, 1919-2J. 



modified at other times of the year, for Mr. 

 Clayton finds the following correlation coefficients 

 connecting the temperature departures at Buenos 

 Aires eight to nine days after the event with the 

 solar radiation variations observed at Calama 

 from August, 1918, to May, 1919. 



Taking these figures with the figures given 

 above, we are to conclude that while on the 

 first three or four days after the event high 

 solar radiation tends to produce high tem- 

 peratures in Buenos Aires from October to 

 February, the opposite tendency governs March 

 to September. 



NO. 2648, VOL. 105] 



These reversals of solar variation effects with 

 the time of the year are paralleled by reversals 

 with geographical position, according to Clayton, 

 who was at first led to regard these geographical 

 reversals as zonal. But it now seems more prob- 

 able to him, I think, as well as to Nansen, that 

 they are associated with the great atmospheric- 

 action centres rather than with the earth's zones. 

 As these action centres change place from time 

 to time, it seems possible that the geographical 

 and secular reversals merge as effects of one 

 general cause. While it may seem extraordinary 

 at first sight that the past winter has been excep- 

 tionally severe (at least in Eastern United States), 

 though solar constant values have been steadily 

 exceptionally high from early in October to Feb- 

 ruary, this may be compared with the known fact 

 that when there are many sun-spots high solar 

 radiation and low temperatures 

 also occur together. Unusual 

 cloudiness or prevailing polar 

 winds may well account for 

 low temperatures associated 

 with high radiation. 



Clayton's studies have led 

 him to a system of forecast- 

 ing in which telegraphic reports 

 of daily solar constant values 

 as obtained by Smithsonian 

 observers at Calama, Chile, 

 take a prominent part. He 

 claims decided and valuable 

 success for both temperature 

 and precipitation forecasts. 



If these pioneer results should 

 be confirmed it seems highly 

 desirable to establish several 

 other solar constant observa- 

 tories in the most cloudless 

 far-separated regions of the 

 earth. By telegraphic commu- 

 nication all their results would 

 be available for daily forecasts 

 all over the world. The cost 

 of ^uch stations fully equipped 

 need not exceed 25,000 dollars for the most inac- 

 cessible. The yearly cost of maintenance need 

 not exceed 10,000 dollars. It is now merely the 

 complete lack of funds for the purpose which 

 withholds the Smithsonian Institution from estab- 

 lishing them. 



Fig. 2 shows the march of " solar constant " 

 results from June i, 1919, to March 24, 1920. In 

 no other period since 1903 has there been ob- 

 served three consecutive months of values so high 

 as those of the winter of 1919-20. The following 

 rapidly falling values accompanied the extra- 

 ordinary solar and terrestrial phenomena of 

 March 22, 1920 : — 



Solar Radiation Values of March, 1920. 



Mean 

 Date II to 17 i3 19 20 21 22 23 



Value 1*968 1-954 1-940 1-931 1-941 1-927 1-866 



1-905 



