G50 
Frocecclinys of the Royal Society 
allowed to pass on. The amount to wliicli the rays of the violet 
end are thrown out depends on the number of small particles 
encountered by them. It is generally supposed that water vapour, 
by absorbing the rays of the violet end of the spectrum, has a 
somewhat similar effect on the light passing through it. If so, 
the water vapour in our atmosphere will tend to deepen the red, 
and so intensify the effect of the dust. 
The reason why a great amount of dust in oUr atmosphere 
should give such brilliant sunset effects, is that it causes a more 
perfect sifting action on the sun’s rays in the outer strata of our 
atmosphere, and provides a greater amount of particles in the 
lower strata large enough to reflect the red rays. If there were no 
fine particles in the upper strata, the sunset effect would be whiter, 
and if there were no large particles in either upper or lower strata, 
then no such sunset effect would be possible. If our atmosphere 
were perfectly free from dust, the light would simply pass through 
it into space without revealing itself, and the moment the sun 
dipped below the horizon total darkness would follow. The length 
of our twilight, therefore, depends on the amount of dust in some 
form or other in our atmosphere, and the height to which the dust 
extends. A great amount of dust in the higher regions of our 
atmosphere would fully account for the greatly prolonged twilight 
we enjoyed while these sunset effects were at their brightest. 
While the sun is still high above the horizon, its rays have to 
penetrate but little more than the depth of our atmosphere, and 
they are subjected to but a slight sifting action by the dust, the 
colour of the light being little changed by the small amount of 
blue thrown out of it, before it arrives at the surface of the earth. 
But as the sun approaches the horizon, its light has to pass through 
a rapidly increasing extent of our atmosphere, and its rays are sub- 
jected to a proportionate increase in the amount of the sifting action; 
so that by the time the sun is on the horizon, the sifting action 
is so great that all the rays of short Wate length are stopped and 
scattered, and only those of the red end of the spectrum reach 
the earth ; hence the illumination of any object lighted by the 
direct rays alone is yellow, orange, or red, according to the amount 
of sifting that has taken place. 
After the sun has sunk below the horizon, the amount of air ' 
