ing a single room, was erected for my benefit upon the wild moorland of Hind Head. From the plateau on which the hut stands there is a free outlook in all directions. Here, amid the heather, I had two stout poles fixed firmly in the ground eight feet asunder, and a stout cord stretched from one to the other. From the center of this cord a thermometer is suspended with its bulb four feet above the ground. On the ground is placed a pad of cotton-wool, and on this cotton-wool a second thermometer, the object of the arrangement being to determine the difference of temperature between the two thermometers, which are only four feet vertically apart.
Permit me at the outset to deal with the subject in a perfectly elementary way. In comparison with the cold of space, the earth must be regarded as a hot body, sending its rays, should nothing intercept them, across the atmosphere into space. The cotton-wool is chosen because it is a powerful, though not the most powerful, radiator. It pours its heat freely into the atmosphere, and by reason of its flocculence, which renders it a non-conductor, it is unable to derive from the earth heat which might atone for its loss. Imagine the cotton wool thus self-chilled. The air in immediate contact with it shares its chill, and the thermometer lying upon it partakes of the refrigeration. In calm weather the chilled air, because of its greater density, remains close to the earth's surface, and in this way we sometimes obtain upon that surface a temperature considerably lower than that of the air a few feet above it. The experiments of Wilson, Six, and Wells have made us familiar with this result. On the other hand, the earth's surface during the day receives from the sun more heat than it loses by its own radiation, so that, when the sun is active, the temperature of the surface exceeds that of the air.
These points will be best illustrated by describing the course of temperature for a day, beginning at sunrise and ending at 10.20 p. m. on the 4th of last March. The observations are recorded in the annexed table, at the head of which are named the place of observation, its elevation above the sea, and the state of the weather. The first column in the table contains the times at which the two thermometers were read. The column under "Air" gives the temperatures of the air, the column under "Wool" gives the temperatures of the wool, while the fourth column gives the differences between the two temperatures. It is seen at a glance that, from sunrise to 9.20 a. m., the cotton-wool is colder than the air; at 9.30 the temperatures are alike. This is the hour of "intersection," which is immediately followed by "inversion." Throughout the day, and up to 4 p. M., the wool is warmer than the air. At 4.5 p. m. the temperatures are again alike; while from that point downward the loss by terrestrial radiation is in excess of the gain derived from all other sources, the refrigeration reaching a maximum at 7.30 p. m., when the difference between the two thermometers amounted to 10° Fahr. When the observations are