1336. An observation, closely connected with the preceding, namely, that, in clear and still nights, frosts are less severe upon the hills than in the neighbouring plains, has excited more attention, chiefly from its contradicting what is commonly regarded an established fact, that the cold of the atmosphere always increases with the distance from the earth. But, on the contrary, the fact is certain, that, in very clear and still nights, the air near to the earth is colder than that which is more distant from it, to the height of at least 220 feet, this being the greatest to which experiments relate. If, then, a hill be supposed to rise from a plain to the height of 220 feet, having upon its summit a small flat surface covered with grass; and if the atmosphere, during a calm and serene night, be admitted to be 10ﾦ warmer there than it is near the surface of the low grounds, which is a less difference than what sometimes occurs in such circumstances, it is manifest that, should both the grass upon the hill, and that upon the plain, acquire a cold of 10ﾦ by radiation, the former will, notwithstanding, be 10ﾦ warmer than the latter. Hence, also, the tops of trees are sometimes found dry, when the grass on the ground's surface has been found covered with dew.
