How frost forms (Wilson).

In the day, plants usually receive more heat from the sun than they give off (radiate), and consequently become warmer; but at night the process is reversed, and they radiate more heat than they receive and thus grow colder. When the surface of a plant has lost (radiated) sufficient heat to cause its temperature to fall to 32° or below, frost forms. Any condition that causes increased radiation will increase the liability of frost, and conversely, whatever checks radiation or supplies additional heat to the air will tend to ward off frost.

A clear night is favorable for frost because radiation or loss of heat from the surface of the earth proceeds most rapidly under a clear sky. Clouds act as a blanket. The heat rays do not penetrate them easily, but are reflected back toward the earth, thus checking radiation by confining the heat to the strata of air between the earth and the clouds.

A quiet air is favorable for frost. Radiation proceeds more rapidly from the surface than from the air above the surface. This is shown by the fact that a thermometer placed in the grass on a quiet, clear night will read 10° or even 15° below one suspended three or four feet above the surface. If there is much wind, this difference will not occur, because the wind mixes the colder air at the surface with the warmer air above, thus giving a more uniform temperature.

A moderately dry atmosphere is favorable for frost, because when the air is humid only a slight fall of temperature will occur before the temperature at which dew begins to form (dew-point) is reached, and when the vapor in the air begins to change into water (dew), the heat that was used originally to change the water into vapor is no longer required and is said to be liberated, and tends to raise the temperature of the air, or at least to retard the fall.

The effect of the liberation of heat in the process of the formation of dew may be appreciated when it is said that the heat added to the air in the formation of a pint of dew is sufficient to raise the temperature of more than five pints of water from the freezing to the boiling point.

Under ordinary conditions, when the dew-point is 10° or more above the frost-point, 32°, a frost is not likely to occur, but if the dew-point approaches 32°, frost is likely to occur.

In a cranberry marsh near Mather, Wis., during the season of 1906, Cox found that the minimum temperature averaged 8.2° below the temperature of the dew-point as observed the previous evening, and in extreme cases the difference was as much as 20° and 22°. On a marsh near Berlin, Wis., on the night of September 27, 1906, at 11 p.m. the dew-point was found to be 43°, yet frost began to form in parts of the marsh at 1 p.m. when the temperature had fallen to 28°; frost became general at 2 a.m., and the following morning a minimum temperature of 24.4° was observed.

The dew-point of the previous evening cannot, therefore, be regarded as a safe guide for the minimum temperature of the following night.

The chief value of dew-point observations of the previous evening appears to be in the fact that they indicate the temperature at which the heat from the condensing vapor will begin to be poured into the air, and if this temperature is much above the frost-point, this addition of heat may be reasonably expected to check the fall of temperature and thus ward off a frost.