OF STOs LIGHT THROUGH THE EARTH’S ATMOSPHERE. 
9 
spheric thicknesses. Now, as the ray is known which, if observed, would give the 
same alteration in luminosity as that of the whole spectrum, it follows that with these 
data the equivalent ray for the photograpliic light can be deduced as shown in § XXY. 
Fig. 4. 
To ensure accuracy several sets of double cells, Co, were prepared, as shown in the 
figure (fig. 4). The top cell, C|, contained turbid water, the turbidity being caused by 
suspended mastic. The mastic was very cautiously precipitated and was thus held in 
suspension. As a matter of fact turbid water of different turbidities was prepared 
nearly two years before it was employed ; and thus any coarse particles had ample 
time to settle; for the vessels containing it were kept undisturbed during that period 
and the liquid was syphoned off as required. The bottom cell, C 2 , contained pure 
w’ater. Behind the cell the exposing box (fig. 1 ), already described, was placed, and 
in front of C 3 sectors which rotated by means of clock work. The sectors could be 
clamped at any desired aperture. The apparatus, as shown, was then placed on a 
stand at such an angle that the sun’s rays fell directly on the platinum paper inside 
the box B, when the exposing apertures were open. The sectors were closed to 
the extent which it was judged should suffice to make the blackness of the two 
exposed squares of platinum paper approximately the same, though it will be seen by 
the experiments to be subsequently described that this was not always attained. 
Exposures for different lengths of time were given in each experiment, and conse¬ 
quently any small error in measurement, or from any other cause, became insignificant 
when the mean was taken. 
XXX .—Examples of the Method. 
The following are examples showing the accord between the different values found 
for 
Experiment I. —Sunlight—length of cell (inside measures), 3'15 inches ; breadth 
MDCCCXCTTT.-A. C 
