lO SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 128 



days. This differs by only f percent from Dr. Marshall's longest ptdse 

 period. Hence all the pulse periods found are aliquot parts of 22f 

 years. 



The original observation that led me to discover the family of 

 periods in weather was the discovery of more than 20 of them in 

 measures of the solar constant of radiation. Meteorologists and others 

 have expressed doubt that the evidence warranted this result. But if it 

 did, they say, the solar variations claimed, generally of the order of 

 Vlo percent of the solar constant, are too small appreciably to affect 

 weather. 



With much recent experience in the study of this family of periodici- 

 ties, I have revised and hope to publish soon my study of more than 

 30 years of daily measures of the solar constant of radiation. This 

 revision has yielded highly satisfactory results, and I believe the evi- 

 dence of the existence, in these measures, of the family referred to 

 has become more convincing. 



But as regards the adequacy of these solar variations to affect 

 weather, I made a suggestion in my paper "Solar Variation, a Lead- 

 ing Weather Element." ^ It is this : Temperature depends on wind 

 direction. Wind direction depends on orientation with respect to 

 atmospheric cyclones. H. H. Clayton showed some 25 years ago, by 

 extensive tabulations of atmospheric pressure, that the position of 

 cyclones varies with the intensity of solar radiation. 



I have been interested in checking the first step in this sequence. 

 My assistant, Mrs. I. W. Windom, collected records of the prevailing 

 winds and temperature departures from the normal at Washington, 

 D. C, for about 1,600 successive days. Of these the prevailing wind 

 could not be determined on about 100 days. The relation of prevailing 

 wind to temperature departures for the remaining 1,500 days is given 

 by figure 7. The result is quite definite. The temperature is about 

 1 2? 5 F. higher when the wind is south, or southwest, than when it is 

 northwest. That the values in figure 6 are not symmetrical about zero 

 is merely resulting from the characteristics of the years chosen. 



Still it makes no difference for the acceptance of the results of my 

 present paper whether meteorologists concede solar variation to be 

 the cause. The family of periodicities, to a number of nearly 30, has 

 been demonstrated in weather, and at least 24 of them have been used 

 in the several syntheses I have shown in illustrations. 



The periods are not apparent in weather at first sight for several 

 reasons. First, the normal values published with weather records are 



'' Smithsonian Misc. Coll., vol. 122, No. 4, August 1953. 



