332 
ASTRONOMY: ABBOT, FOWLE, AND ALDRICH 
tion are known to occur. Fowle has computed from the transmission 
coefficients that the number of molecules per cubic centimeter of air at 
standard temperature and pressure is (2.70 =t= 0.02) X 10^^ This value 
is very close to Millikan's determination by absolutely independent 
observations and methods, namely (2.705 0.005) X 10^^ 
In the third place simultaneous solar-constant observations at Mount 
Wilson and Bassour, separated by one-third the earth's circumference, 
unite in showing a substantial irregular variability of the sun from day 
to day. This solar variability has been of late independently con- 
firmed by us by examination of the distribution of brightness along the 
diameter of the sun's disk. The latter observations show variations of 
distribution from day to day, and these accompany pretty closely the 
variations of the total solar radiation. It seems to us that, as the fact 
of solar variability is thus independently confirmed as a real phenome- 
non, it speaks favorably for the substantial accuracy of our solar con- 
stant measurements that it was through them that the irregular varia- 
tions of from 1 to 5 or, very rarely, 10% were first discovered. 
Notwithstanding these evidences of the soundness of our solar con- 
stant work, various attacks upon it have been made, tending to show that 
the solar constant is much higher than 1.93 calories, perhaps even 3.5 
to 4.0 calories. A principal argument is that the atmospheric transpar- 
ency continually diminishes as the sun rises within 75° zenith distances, 
so that our values of atmospheric transmission are much too great, 
and have no relation to the transmission of an atmosphere of constant 
transparency. Secondly it is said that measurements of solar radiation 
exceeding 1.93 calories have been made on mountain tops, and from free 
balloons. Various other objections are raised, which we discuss in our 
paper now being published by the Smithsonian Institution. 
On two days, September 20 and 21, 1914, we continued solar con- 
stant observations at Mount Wilson from the instant of sunrise until 
about 10 o'clock. We have reduced the work by the aid of Bemporad's 
air-mass formulae and tables. As these postulated uniform optical 
quality of the atmosphere from bottom to top, it was necessary to apply 
certain corrections to them varying with the wave-length, depending 
upon the extinction by water vapor residing in the lowest atmospheric 
strata. We were enabled to determine these corrections by Fowle's 
studies of the effects of water vapor. We find on both days that the 
atmospheric transparency remained sensibly unaltered from sunrise 
to 10 o'clock. Closely identical values of the solar constant are ob- 
tained, whatever the range of air masses used to determine the atmos- 
pheric transmission. We made three independent estimates for each 
