and number, as they do near the surface of the earth, or in size only, even at the expense of numbers, as happens in a moist atmosphere, because of their hygroscopic property, light of the shorter wave-lengths becomes more completely absorbed and the sky assumes some longer "wave-length color. Finally, when the particles are large enough to reflect as mirrors the sky becomes whitish. Hence both the morning and the evening twilight sky often shows a series of colors ranging from red, near the horizon, through orange and yellow to a green or even blue-green with increase of elevation and consequent decrease in the number and size of dust particles along the path of light from the sun to that part of the sky in question and thence to the observer.
When the air is filled with fog, or other particles of similar size, the whole sky becomes uniformly gray. This is because the water droplets that together make fog and cloud, though usually so small that it would take from 2,000 to 3,000 of them to make a row an inch long, nevertheless are large enough to reflect, as would little mirrors, and to refract, or transmit in a new direction, light of every color.
It remains now, in preparing the way to an understanding of the weather significance of morning and evening colors, briefly to outline the essential conditions and processes of cloud formation and rain.
Probably that one of these conditions with which the general public is least familiar is the presence, in large numbers, of some sort of nuclei about which water vapor can condense. We can safely assume, too, that in the open atmosphere these nuclei consist only of dust particles, though it is possible in the laboratory, under conditions that rarely, if ever, exist naturally, to obtain condensation without the aid of dust of any kind.
Besides the presence of dust particles, a certain relation between temperature and water content of the atmosphere is also essential to condensation. The warmer the air, so long as the temperature is below the boiling point, the greater, and, for ordinary temperatures, at a rapidly increasing rate, the amount of water vapor it can contain in the form of a transparent gas.
In reality the relation above discussed is between the temperature and amount of moisture per unit volume, a quantity which does not appreciably change with the presence or absence of other gases. But it is allowable, because of this constancy, to use the popular, though unscientific, expression, "water content of the atmosphere," provided one thinks of the atmosphere as a mixture of gases (chiefly nitrogen and oxygen) coexisting with the undisturbed water vapor, and not as a sort of sponge that mechanically holds it in suspension.
If, then, air, which always has dust particles in it, containing all or nearly all the water vapor it can hold, is cooled to a distinctly lower temperature, a corresponding amount of condensation will take place