67S 



NATURE 



[July 29, 1920 



Solar Variation and the Weather. 

 By Dr. C. G. Abbot, Director, Smithsonian Astrophysical Observatory. 



NEARLY a century ago three pioneers, Sir 

 John Herschel, Pouillet, and Forbes, laid the 

 foundations of the measurement of solar radiation. 

 Each devised an instrument for measuring the 

 heating effect of the solar rays and used it dili- 

 gently. Pouillet and Forbes availed themselves 

 of the law of extinction of light, which had been 

 independently discovered about 1760 by Bouguer 

 and Lambert, to calculate the intensity of the solar 

 rays, as they would be outside our atmosphere. 

 Forbes's researches in the Alps proved that this 

 law is not strictly applicable to the sun's rays as 

 a whole, and he was led to believe that the value 

 of the so-called solar constant of radiation was as 

 high as 2-85 calories per sq. cm. per min. 

 Pouillet 's value, based on the assumed validity of 

 the Bouguer-Lambert law, was 1-76 calories. 



As pointed out by Radau, the problem of esti- 

 mating the intensity of the solar heat outside our 

 atmosphere requires the study of the various spec- 

 trum rays separately, because their transmission 

 through the atmosphere is unequal. Langley in- 

 vented the spectro-bolometer about 1880, and 

 immediately applied it to the problem as analysed 

 by Radau. In the famous Mount Whitney ex- 

 pedition of 1 88 1 Langley carried on spectro-bolo- 

 metric and pyrheliometric measurements simul- 

 taneously at an altitude of 12,000 ft. Misled by 

 certain theoretical considerations, however, these 

 experiments seemed to him to yield the value 307 

 calories per sq. cm. per min. as the most probable 

 value of the solar constant of radiation. A correct 

 reduction, which he also gave in his report, yielded 

 2-22 calories. Later experiments made on Mount 

 Whitney and on Teneriffe indicate that while the 

 spectro-bolometric work was good, Langley 's 

 pyrheliometric determinations were too high, so 

 that the true result should have been 1-9 calories. 



Up to that time no fully satisfactory instrument 

 for measuring the intensity of solar heat at the 

 earth's surface had been perfected. About 1893 

 Prof. Knut Angstrom's highly ingenious electrical 

 compensation pyrheliometer fixed the scale of solar 

 radiation measurements surely within 5 per cent. 

 In recent years the accuracy of the Angstrom pyr- 

 heliometer has reached to 2 per cent, or better. 

 In 1913 three independent series of determinations 

 at the Smithsonian Institution fixed the standard 

 scale of radiation measurements now generally 

 adopted. The Angstrom scale as corrected by 

 A, K. Angstrom lies i-8 per cent, lower. 



At Washington, under Langley 's direction, the 

 spectro-bolometer, which at the time of the Mount 

 Whitney expedition was almost unmanageable, 

 was perfected in the decade 1 890-1900 into a tract- 

 able, trustworthy instrument, and made to trace 

 photographically an autographic solar spectrum 

 energy curve extending from wave-length 0-3 

 micron to 3-0 microns within 10 minutes. 



In the autumn of 1907 experiments were begun 

 NO. 2648, VOL. 105] 



in Washington to fix a standard procedure for 

 solar constant observations. Omitting minor 

 details, the process which resulted is as follows : 

 Beginning when the sun is near 15° above the 

 horizon, about six solar spectrum energy curves, 

 and simultaneously pyrheliometric measurements, 

 are made, ending when the sun's altitude has 

 i reached about 60°. These curves are measured 

 at about forty points corresponding to known 

 wave-lengths from far in the ultra-violet to far in 

 I the infra-red. Taking each wave-length by itself, 

 these intensities on the six separate curves follow 

 j the Bouguer-Lambert law of extinction. Hence 

 j plotting the logarithms of measured intensities 

 ! as ordinates and corresponding values of the 

 I secants of the solar zenith distances as abscissae, 

 each group of six points determines a straight 

 I line. Producing this line to zero of abscissae — 

 ; that is, to the line corresponding to no atmosphere 

 at all — we read there the logarithms of the in- 

 I tensity for the various wave-lengths, as the energy 

 \ curve would be found outside our atmosphere — on 

 i the moon, for instance. The scale of energy in 

 calories per square centimetre per minute comes 

 by comparing the total area included under the 

 spectro-bolometric curves with pyrheliometer read- 

 ings taken simultaneously. Such, in brief, is the 

 process. 



Determinations were begun at Washington in 

 October, 1902. In the springtime of 1903 a large 

 drop amounting to nearly 10 per cent, was found 

 in the values after the end of March. The changed 

 values persisted so steadily that we were led to 

 entertain the hypothesis that the solar radiation 

 had actually diminished. A comparison was made 

 between solar heat and terrestrial temperatures. 

 It actually appeared that just after the apparent 

 drop in solar radiation there occurred a general 

 ' drop in terrestrial temperatures for all available 

 stations of the north temperate zone. This led 

 us to the long campaign of solar radiation observa- 

 ) tions which I shall now describe. 



In 1905 we began sending yearly expeditions 

 to observe solar radiation at Mount Wilson, Cali- 

 fornia, also the seat of the famous Mount Wilson 

 Solar Observatory of the Carnegie Institution. 

 I am happy to acknowledge the great assistance 

 and enthusiastic interest which Dr. Hale and his 

 colleagues have at all times given our work there. 

 From the first the Mount Wilson values, though 

 far more accurate than Washington values, owing 

 to the clearer and more uniform sky conditions of 

 ■ California, showed on their face a variability over 

 an extreme range of 10 per cent, in the emission 

 of solar radiation. The sun appeared to be a 

 variable star having a twofold type of variation : 

 First, a fluctuation with the march of years attend- 

 ing changes in solar activity as indicated by sun- 

 spots, faculae, prominences, etc. ; secondly, a fluc- 

 i tuation running its courses in a few days, weeks, 



