1456 
an ideal homogeneous one for which d = 2000 kilometers (= ’/,,, 
of the sun’s radius). The corresponding value of AB is about 0,157, 
giving for the ratio 
r 
0 ae. 
Our conclusion therefore is, that less than '/,,,, of the sun’s total 
radiation is emitted or scattered by parts of the celestial body lying 
outside the photospheric surface. 
Even though we are free to admit an uncertainty of several 
hundreds percent in some of the estimates on which the above calcu- 
lation is based, our result yet makes it impossible to maintain the 
current ideas on the nature of the photosphere. 
Most solar theories, indeed, consider the photosphere to bea layer 
of incandescent clouds, whose decrease of luminosity from the centre 
toward the limb of the solar disk would be caused by absorption 
and scattering of light in an enveloping atmosphere (“the dusky veil”). 
According to calculations made by Pickrerinc, Wison, SCHUSTER, 
VOGEL, v. SEELIGER, and others, such an atmosphere should intercept 
an important fraction (*/, to '/,) of the photospheric radiation. The 
atmosphere is of course in a stationary condition ; receipts and expenses 
must balance each other. Now, what would become of that immense 
quantity of absorbed energy, of which only something of the order 
of magnitude ‘/,,,, is emitted and scattered ? So long as we have 
no evidence of any other form of solar output, especially proceeding 
from the atmospheric iayers, and comparable in magnitude with the 
sun’s total radiation, we are forced to reject the cloud-theory of the 
photosphere. 
The radial variation of the brightness of the disk depends on the 
nature of the photosphere itself, not of its envelope. A new inter- 
pretation of the photospkere, agreeing with this result, will be 
proposed in a subsequent paper. 
p="), 10. 
Observations made with the thermopile. 
We now proceed to the discussion of the observations made for 
finding the shape of the entire radiation curve. In this part of the 
work our thermopile arrangement had the advantage of the bolo- 
metric apparatus in point of proportionality, within wide limits, 
between radiation and galvanometer deflection. 
The total resistance of the thermopile circuit had to be varied in 
a few steps from 1300 for full sunshine to 100 for the central 
quarter of an hour, and back again. Table I contains the deflections 
