REPORT OF THE SECRETARY. 63 
simultaneously on Mount Wilson. I give below the results obtained at Mount 
Wilson and Mount Whitney in 1909 and 1910: 
Solar constant: : 
MOWING llSON arn senna cs setoacte nes acmaase ecncoceicceecce 1. 943 | 1, 943 1, 924 1,904 
Taking the mean of the differences between the results obtained simultane- 
ously at the two stations, it appears that the results obtained on Mount Whitney 
average 1.4 per cent higher than those obtained on Mount Wilson. But con- 
sidering that the optical apparatus used on Mount Wilson comprised a silvered 
glass mirror coelostat, an ultra-violet crown glass prism, and two silvered 
glass mirrors, while that on Mount Whitney comprised only a quartz prism 
and two magnalium mirrors, and, furthermore, that the pyrheliometers em- 
ployed at the two stations were read at very different temperatures, it is prob- 
able that the slight difference found between the results may be due wholly 
to experimental differences and implies no discrepancy due to the difference 
of altitude between the two stations. 
This conclusion seems worth emphasizing. We have now made simultane- 
ously solar-constant determinations at sea level (Washington), at over a mile 
altitude (Mount Wilson), and again at Mount Wilson, and at nearly 3 miles 
altitude (Mount Whitney). Although both the quantity and the quality of 
the solar radiation found at these stations differ very much, neither tne “ solar 
constant” nor the distribution of the solar energy in the spectrum outside the 
atmosphere, as fixed by the wholly independent measurements at these three 
stations, differs more than would be expected in view of the unavoidable small 
errors of observation. We seem justified in concluding that we do, in fact, 
eliminate the effects of atmospheric losses and actually determine the true 
quantity and quality of the sun’s radiation outside the atmosphere as we might 
do if we could observe in free space with no atmosphere at all to hinder. 
Expeditions to Mount Wilson have now been made in 1905, 1906, 1908, 1909, 
and 1910. The last, like the others, continued from May until November. In 
the earlier years the observations were not made daily, but in 1908, 1909, and 
1910 daily determinations of the solar constant were made when possible. As 
stated in earlier reports, the results indicate a variability of the sun. In order 
to show the strength of the argument for this conclusion, I give in the accom- 
panying figure a diagram showing all the “solar constant” values obtained in 
the first four years of observation (fig. 1). 
The “solar constant” results lie between 1.80 and 2.00 calories per square 
centimeter per minute. I call particular attention to the two later years. It 
will be noted that successive days’ results march step by step regularly from 
low to high values and the reverse, and that this order of march is not the 
exception, but almost without exception the rule. This seems to render it 
highly improbable that the fluctuations are due to accidental error, for such a_ 
regularity of fluctuation is incompatible with that supposition. As it has now 
been shown that the altitude of the observing station is immaterial, at least 
for altitudes below 38 miles, it seems also reasonable to conclude that the 
fluctuation is not due to faulty estimates of the losses of radiation in the air. 
Hence the most probable conclusion is that the sun actually varies from day to 
