A few observations of the temperature of the ground near the surface, and of the air over it, first raised doubts as to the correctness of the now generally received opinion that dew is formed of vapor existing at the time in the air.  These observations, made at night, showed the ground at a short distance below the surface to be always hotter than the air over it, and it was thought that so long as this excess is sufficient to keep the temperature of the surface of the ground above the dew point of the air, it will, if moist, give off vapor, and it will be this rising vapor that will condense on the grass and form dew, and not the vapor that was previously present in the air.

The first question to be determined was whether vapor does, or does not, rise from the ground on dewy nights.  One method tried of testing this point was by placing over the grass, in an inverted position, shallow trays made of thin metal and painted.  These trays were put over the ground to be tested after sunset and examined at night, and also next morning.  It was expected that, if vapor was rising from the ground during dewy nights, it would be trapped inside the trays.  The result in all the experiments was that the inside was dewed every night, and the grass inside was wetter than that outside.  On some nights there was no dew outside the trays, and on all nights the inside deposit was heavier than the outside one.

An analysis of the action of these trays is given, and it is concluded that they act very much the same as if the air was quite still.  Under these conditions vapor will rise from the ground so long as the vapor-tension on the surface of the ground is higher than that at the top of the grass, and much of this rising vapor is, under ordinary conditions, carried away by the passing air, and mixed with a large amount of drier air, whereas the vapor rising under the trays is not so diluted; and hence, though only cooled to the same amount as the air outside, it yields a heavier deposit of dew.

Another method of testing this point was employed, which consisted in weighing a small area of the exposed surface of the ground, as it was evident that if the soil gave off vapor during a dewy night, it must lose weight.  A small turf about 6 inches (152 mm.) square was cut out of the lawn, and placed in a small shallow pan of about the same size.  The pan with its turf, after being carefully weighed, was put out on the lawn in the place where the turf had been cut.  It was exposed for some hours while dew was forming, and on these occasions it was always found to lose weight.  It was thus evident that vapor was rising from the ground while dew was forming, and therefore the dew found on the grass was formed of part of the rising vapor, trapped or held back by coming into contact with the cold blades of grass.

The difference between these experiments, in which the exposed bodies lose weight, and the well-known ones in which bodies are exposed to radiation, and the amount of dew formed is estimated by the increase in their weight, is pointed out.  In the former case, the bodies are in good heat-communication with the ground, whereas in the latter little or no heat is received by conduction from the earth.