314 ASTRONOMY: C. G. ABBOT 
on altitude appeared. In 1914 we sent an automatic recording pyrheliometer 
by sounding balloon to 25,000 meters altitude. The result obtained was 
trustworthy and agreed with what we expected. All these checks confirmed 
the accuracy of our work and strengthened our belief in the solar variability. 
Meanwhile in 1911 and 1912 I had observed in Algeria while my colleagues 
observed in California. Unfortunately for the proof of solar variability 1911 
was cloudy and 1912 was the year the air was charged with volcanic dust by 
the great volcanic eruption of Mount Katmai, June 6, 1912. Nevertheless 
despite this handicap the results left little reasonable doubt that the sun is 
variable. High 'solar constant' values at Bassour, Algeria, coincided with 
high values at Mount Wilson, California, and vice versa, and equal increments 
of radiation were found at the two stations independently, notwithstanding 
that they are separated by one-third the earth's circumference. 
In 1913 and subsequently the proof of solar variability was rendered im- 
pregnable. We investigated daily the distribution of radiation over the solar 
image formed by our tower telescope on Mount Wilson. The sun's image is, 
as you know, unequally bright at centre and edges, so that the curve of its 
intensity along a diameter takes the form of the letter U inverted. The 
steepness of the curves of the U varies with wave-length. But we found also 
that it varies from day to day, and that the variations are such that a greater 
contrast of brightness between center and edge occurs when the solar radia- 
tion as a whole is found to be diminished and vice versa. I suppose the outer 
layers of the sun vary in transparency. When more opaque they diminish 
the 'solar constant' but as the effect is greatest near the limb of the sun where 
the oblique path of the ray in the solar layers is greatest, the result is also to 
increase the contrast. At all events we have found a true variation of the 
sun, independent of the earth's atmosphere, which coincides in time of its 
fluctuations with the observed changes of the 'solar constant.' 
Professor Pickering had kindly undertaken pyrheliometer measurements 
at Arequipa, Peru. These were carried on from 1912 to 1917. They tended 
to confirm in the more outstanding features the variation of the sun observed 
at Mount Wilson. 
Dr. L. A. Bauer, whose remarkable campaign of magnetic observations 
lends lustre to the science of our country, has investigated certain minute 
disturbances of terrestrial magnetism for which no cause had been assigned. 
He finds them to be closely correlated with the variations of solar radiation 
we have observed. 
Dr. H. H. Clayton, formerly of Blue Hill Meteorological Observatory, now 
of the Argentine Meterological Service, has investigated the variations of 
terrestrial atmospheric temperature and pressure for nearly fifty stations 
in all parts of the earth. He finds a close correlation of these meteorological 
variations with the irregular variations we have observed in solar radiation. 
Equatorial stations show a direct correlation in temperature. That is high 
solar radiation is followed by high temperature and vice versa. Many tern- 
