TKE SET. 105 
mathematical formula to express the diminution of the 
blue, in arcs drawn east and west from the zenith down- 
ward. This diminution is a natural consequence of the 
predominance of coarser particles in the lower regions of 
the atmosphere. Were the particles which produce the 
purer celestial vault all swept away, we should, unless 
helped by what has been called "cosmic dust," look into 
the blackness of celestial space. And were the whole 
atmosphere abolished along with its suspended matter, we 
should have the " blackness "spangled with steady stars; for 
the twinkling of the stars is caused by our atmosphere. 
Now, the higher we ascend, the more do we leave behind 
us the particles which scatter the light; the nearer, in fact, 
do we approach to that vision of celestial space mentioned 
a moment ago. Viewed, therefore, from the loftiest 
Alpine summits, the firmamental blue is darker than it is 
ever observed to be from the plains. 
It is thus shown that by the scattering action of minute 
particles the blue of the sky can be produced; but there 
is yet more to be said upon the subject. Let the natural 
sky be looked at on a fine day through a piece of trans- 
parent Iceland spar cut into the form known as a Nicol 
prism. It may be well to begin by looking through the 
prism at a snow slope, or a white wall. Turning the prism 
round its axis, the light coming from these objects does 
not undergo any sensible change. But when the prism is 
directed toward the sky the great probability is that, on 
turning it, variations in the amount of light reaching the 
eye will be observed. Testing various portions of the sky 
with due diligence, we at length discover one particular 
direction where the difference of illumination becomes a 
maximum. Here the Nicol, in one position, seems to 
offer no impediment to the passage of the skylight, while, 
when turned through an arc of ninetv degrees from this 
position, the light is almost entirely quenched. We soon 
discern that the particular line of vision in which this 
maximum difference is observed is perpendicular to the 
direction of the solar rays. The Nicol acts thus upon sky- 
light because that light is polarized, while the light from 
the white wall or the white snow, being unpolarized, is not 
affected by the rotation of the prism. 
In the case of our manufactured sky not only is the 
azure of the firmament reproduced, but these phenomena 
