It seems probable that when the radiation is strong, that soil, especially if it is loose and not in good heat-communication with the ground, will get cooled below the dew-point, and have vapor condensed upon it.  On some occasions the soil certainly got wetter on the surface, but the question still remains, Whence the vapor?  Came it from the air, or from the soil underneath?  The latter seems the more probable source; the vapor rising from the hot soil underneath will be trapped by the cold surface-soil, in the same way as it is trapped by grass over grass-land.  During frost, opportunities are afforded of studying this point in a satisfactory manner, as the trapped vapor keeps its place where it is condensed.  On these occasions the under sides of the clods, at the surface of the soil, are found to be thickly covered with hoar-frost, while there is little on their upper or exposed surfaces, showing that the vapor condensed on the surface-soil has come from below.

The next division of the subject is on dew on roads.  It is generally said that dew forms copiously on grass, while none is deposited on roads, because grass is a good radiator and cools quicker, and cools more, than the surface of a road.  It is shown that the above statement is wrong, and that dew really does form abundantly on roads, and that the reason it has not been observed is that it has not been sought for at the correct place.  We are not entitled to expect to find dew on the surface of roads as on the surface of grass. because stones are good conductors of heat, and, the vapor-tension being higher underneath than above the stones, the result is, the rising vapor gets condensed on the under sides of the stones.  If a road is examined on a dewy night, and the gravel turned up, the under sides of the stones are found to be dripping wet.

Another reason why no dew forms on the surface of roads is that the stones, being fair conductors, and in heat communication with the ground, the temperature of the surface of the road is, from observations taken on several occasions, higher than that of the surface of the grass alongside.  The air in contact with the stones is, therefore, not cooled so much as that in contact with the grass.

For studying the formation of dew on roads, slates were found to be useful.  One slate was placed over a gravelly part of the road, and another over a hard dry part.  Examined on dewy nights, the under sides of these slates were always found to be dripping wet, while their upper surfaces, and the ground all round, were quite dry.

The importance of the heat communicated from the ground is illustrated by a simple experiment with two slates or two iron weights, one of them being placed on the ground, either on grass or on bare soil, and the other elevated a few inches above the surface.  The one resting on the ground, and in heat-communication with it, is found always to keep dry on dewy nights, whereas the elevated one gets dewed all over.